diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-27 17:56:50 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-27 17:56:50 +0000 |
commit | 8e9f8f2d0dcdddec35091ddbbdc79650763ec922 (patch) | |
tree | 4f88ed317d44fb4e5d9ab77e565bf1d910a8f560 /src/chrtrans | |
parent | Initial commit. (diff) | |
download | lynx-8e9f8f2d0dcdddec35091ddbbdc79650763ec922.tar.xz lynx-8e9f8f2d0dcdddec35091ddbbdc79650763ec922.zip |
Adding upstream version 2.9.0dev.6.upstream/2.9.0dev.6upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'src/chrtrans')
62 files changed, 20509 insertions, 0 deletions
diff --git a/src/chrtrans/README.format b/src/chrtrans/README.format new file mode 100644 index 0000000..7437b50 --- /dev/null +++ b/src/chrtrans/README.format @@ -0,0 +1,138 @@ +Some notes on the format of table files used here. +(See README.tables for what to do with them.) + +The format is derived from stuff in the console driver of the +Linux kernel (as are the guts of the chartrans machinery). +THAT DOES NOT MEAN that anything here is Linux specific - it isn't. + +[Note that the format may change, this is still somewhat experimental.] + +There are four kinds of lines: + +Summary example: + + # This line is a comment, the next line is a directive + O Brand new Charset! + 0x41 U+0041 U+0391 + U+00cd:I' + +Description: + +a) comment lines start with a '#' character. + (trailing comments are allowed on some of the other lines, if in doubt + check the examples..) + +b) directives: + start with a keyword which may be abbreviated to one letter (first + letter must be capitalized), followed by space and a value. + Currently recognized: + + OptionName + The name under which this should appear on the O)ptions screen + in the list for Display Character Set + MIMEName + The name for this charset in MIME syntax (one word with digits + and some other non-letters allowed, should be IANA registered) + Default + If "Y[es]" or "1", this is the default (fallback) translation table, + it will be used for Unicode -> 8bit (or 7bit) translation if no + translation is found in the specific table. + FallBack + Whether to use the default table if no translation is found in + this table. Normally fallback is used, "FallBack NO" or "FallBack 0" + disables it (actually, other values than "FallBack Y[es]" or + "FallBack 1" disable it). + + RawOrEnc + a number which flags some special property (encoding) for this + charset [see utf8_uni.tbl for example, see UCDefs.h for details]. + + Codepage number (IBM specific) + used by OS/2 font-switching code. + +c) character translation definitions: + they look like + + 0x41 U+0041 U+0391 ... + + and are used for "forward" translation (mapping this charset to Unicode) + AS WELL AS "back" translation (mapping Unicodes to an 8-bit + [incl. 7-bit ASCII] code). + + For the "forward" direction, only the first Unicode is used; for + "back" translation, all listed Unicodes are mapped to the byte (i.e. + code point) on the left. + + The above example line would tell the chartrans mechanism: + "For this charset, code position 65 [hex 0x41] contains Unicode + U+0041 (LATIN CAPITAL LETTER A). For translation of Unicodes to + this charset, use byte value 65 [hex 0x41] for U+0041 (LATIN CAPITAL + LETTER A) as well as for U+0391 (GREEK CAPITAL LETTER ALPHA)." + + [Note that for bytes in the ASCII range 0x00-0x7F, the forward translations + will (probably) not be used by Lynx. It doesn't hurt to list those, + too, for completeness.] + + Some other forms are also accepted: + + * Syntax accepted: + * <fontpos> <unicode> <unicode> ... + * <fontpos> <unicode range> <unicode range> ... + * <fontpos> idem + * <range> idem + * <range> <unicode range> + * + * where <unicode range> ::= <unicode>-<unicode> + * and <unicode> ::= U+<h><h><h><h> + * and <h> ::= <hexadecimal digit> + * + [Note that <fontpos> _without_ targets assumed notdefined, + so tables from ftp.unicode.org need no patching.] + + +d) string replacement definitions: + + They look like + + U+00cd:I' + + which would mean "Replace Unicode U+00cd (LATIN CAPITAL LETTER I WITH + ACUTE" with the string (consisting of two character) I' (if no other + translation is available)." Please note that replacement definitions + in certain charset table will override ones from the Default table. + + Note that everything after the ':' is currently taken VERBATIM, so + careful with trailing blanks etc. Please use <C replace> syntax below + when you need trailing spaces. + + * Syntax accepted: + * <unicode> :<replace> + * <unicode range> :<replace> + * <unicode> "<C replace>" + * <unicode range> "<C replace>" + * + * where <unicode range> ::= <unicode>-<unicode> + * and <unicode> ::= U+<h><h><h><h> + * and <h> ::= <hexadecimal digit> + * and <replace> any string not containing '\n' or '\0', taken verbatim + * and <C replace> any string, with backslash having the usual C meaning. + +Motivation: + +- It is an extension of the format already in use for Linux (kernel, + kbd package), those files can be used with some minimal editing. + +- It is easy to convert Unicode tables for other charsets, as they + are commonly found on ftp sites etc., to this format - the right + sed command should do 99% of the work. + +- The format is independent of details of other parts of the Lynx code, + unlike the "old" LYCharsets.c mechanism. The tables don't have to + be changed in synch when e.g., new entities are added to the entities.h. + + +Note: the Default "7bit approximation" table can be used for +case-insensitive search for non-ascii letters if no upper/lower case +information provided by other means, e.g., locale. It is assumed that +upper/lower case letters have their "7bit approximation" images +in def7_uni.tbl matched case-insensitively. diff --git a/src/chrtrans/README.tables b/src/chrtrans/README.tables new file mode 100644 index 0000000..14431da --- /dev/null +++ b/src/chrtrans/README.tables @@ -0,0 +1,76 @@ +The translation table files in this directory were collected from +several sources (among them ftp://ftp.unicode.org, Linux kbd package, +ftp://dkuug.dk/) and are believed to be correct in their mappings, +but not checked in detail. The Unicode/UCS2 values +for some of the RFC 1345 Mnemonic codes are out of date, +a cleanup and update would be needed for serious use. +[See also http://czyborra.com/charsets/iso8859.html for codepages survey.] + +These changes were made to all of the files used from ftp.unicode.org: + + a) add the MIME name of the charset. + b) add a name for the display charset (used on Options screen) + c) add the codepage number + d) remove lines for control characters 0x00 to 0x1f, 0x7f to 0x9f. + e) comment-out ASCII lines 0x20 to 0x7f + f) use idem to represent the commented-out lines + g) change C-style 0xNNNN constants to Unicode-style U+NNNN. + +Other changes include + + h) add code-points to several lines to provide Unicode equivalents + i) add extra mappings at the end of the files + j) comment-out other one-one mappings in the 0xa0-0xff range. + +More translation files can be easily provided (and new character entities +added to entities.h), this set is just to test whether the system works +in principle (and also how it behaves with incomplete data...) + +See the file README.format for a brief explanation of what's in the +table files. + +The examples have names *_uni or *_suni with a .tbl suffix, but it +doesn't really matter. The auxiliary program makeuctb (MAKE UniCode +TaBle) is used to "compile" them into C header files, which can be +included by UCdomap.c. + +Ideally, this should be taken care of by the Makefiles. On VMS, use +build-chrtrans.com to compile and link makeuctb.exe and create the +set of .h files from the current set of .tbl files. Thereafter, use +build-header.com to update particular .h files. + +To make a new chartrans table available to Lynx (and thereby make a new +charset known to Lynx) you currently have to manually edit UCdomap.c, in +two places: + +a) Near the top, you will find a bunch of lines (some may be commented out) + + #include "<fn>.h" + +Add or comment out as you wish. But it is probably safest to leave the +commonly used ones, referring to "def7_uni.h" and "iso01_uni.h", in place. + +b) At the bottom, you will find a bunch of lines (again, some may be + commented out by default) of the form + + UC_CHARSET_SETUP_<something>; + +which should correspond to the #include lines from a). Again, +add or subtract as you wish (but preferably consistent with what you +did under a)...) [The <something> is derived from the charset's MIME name. +if in doubt, check the last lines of the corresponding ...uni.h file.] + +c) To let make automatically notice when you have changed one of the + table files, and automatically regenerate the *uni.h file(s), +you also have to add any new tables to both src/Makefile *and* +src/chrtrans/Makefile. Or, for auto-config, the equivalent files +named makefile.in before running ./configure, or makefile after running +./configure. (That may be inconvenient, but I didn't want to depend +on features than not all makes may have.) Note that for recompiling +Lynx, a `make clean' should not be necessary if you have *only* made +changes to the files in src/chrtrans. On VMS, add entries for new +tables to build-chrtrans.com, but you can update the particular file +with build-header.com, then use the top directory's build.com and +answer 'n' to it's prompts about whether to update the WWWlibrary +and chrtrans modules. + diff --git a/src/chrtrans/UCkd.h b/src/chrtrans/UCkd.h new file mode 100644 index 0000000..f2de902 --- /dev/null +++ b/src/chrtrans/UCkd.h @@ -0,0 +1,54 @@ +#ifndef _UC_KD_H +#define _UC_KD_H + +/* + * NOTE: THE FOLLOWING #define MAY NEED ADJUSTMENT. + * u16 should be an unsigned type of 16 bit length (two octets). + * u8 should be an unsigned type of 8 bit length (one octet). + */ +#ifndef u16 +#define u16 unsigned short +#endif /* u16 */ + +#ifndef u8 +#define u8 unsigned char +#endif /* u8 */ + +typedef char scrnmap_t; + +#define E_TABSZ 256 + +struct unipair { + u16 unicode; + u16 fontpos; +}; +struct unipair_str { + u16 unicode; + const char *replace_str; +}; +struct unimapdesc { + u16 entry_ct; + struct unipair *entries; +}; +struct unimapdesc_str { + u16 entry_ct; + struct unipair_str *entries; + int isdefault; + int trydefault; +}; + +#define UNI_DIRECT_BASE 0xF000 /* start of Direct Font Region */ +#define UNI_DIRECT_MASK 0x01FF /* Direct Font Region bitmask */ + +#define UC_MAXLEN_ID_APPEND 20 +#define UC_MAXLEN_MIMECSNAME 40 +#define UC_MAXLEN_LYNXCSNAME 40 +#define UC_LEN_LYNXCSNAME 20 + +#undef EX_OK /* may be defined in system headers */ +#define EX_OK 0 /* successful termination */ +#define EX_USAGE 64 /* command line usage error */ +#define EX_DATAERR 65 /* data format error */ +#define EX_NOINPUT 66 /* cannot open input */ + +#endif /* _UC_KD_H */ diff --git a/src/chrtrans/build-chrtrans.com b/src/chrtrans/build-chrtrans.com new file mode 100644 index 0000000..c4f88f5 --- /dev/null +++ b/src/chrtrans/build-chrtrans.com @@ -0,0 +1,141 @@ +$ v0 = 0 +$ v = f$verify(v0) +$! BUILD-CHRTRANS.COM +$! +$! Command file to build MAKEUCTB.EXE on VMS systems +$! and then use it to create the chrtrans header files. +$! +$! 28-Jun-1997 F.Macrides macrides@sci.wfeb.edu +$! Initial version, for Lynx v2.7.1+fotemods +$! +$ ON CONTROL_Y THEN GOTO CLEANUP +$ ON ERROR THEN GOTO CLEANUP +$ CHRproc = f$environment("PROCEDURE") +$ CHRwhere = f$parse(CHRproc,,,"DEVICE") + f$parse(CHRproc,,,"DIRECTORY") +$! +$ if p1 .nes. "" +$ then +$ CHRcc_opts = "/DEBUG/NOOPT" +$ CHRlink_opts = "/DEBUG" +$ else +$ CHRcc_opts = "" +$ CHRlink_opts = "" +$ endif +$! +$ Compile_makeuctb: +$!================ +$ v1 = f$verify(1) +$! +$! Compile the Lynx [.SRC.CHRTRANS]makeuctb module. +$! +$ v1 = f$verify(v0) +$ IF f$getsyi("ARCH_NAME") .eqs. "Alpha" .or. - + f$getsyi("ARCH_NAME") .eqs. "IA64" .or. - + f$trnlnm("VAXCMSG") .eqs. "DECC$MSG" .or. - + f$trnlnm("DECC$CC_DEFAULT") .eqs. "/DECC" .or. - + f$trnlnm("DECC$CC_DEFAULT") .eqs. "/VAXC" +$ THEN +$ CHRcompiler := "DECC" +$ v1 = f$verify(1) +$! DECC: +$ cc := cc/decc/prefix=all /nomember 'CHRcc_opts'- + /INCLUDE=([],[-],[--],[--.WWW.Library.Implementation]) +$ v1 = f$verify(v0) +$ ELSE +$ IF f$search("gnu_cc:[000000]gcclib.olb") .nes. "" +$ THEN +$ CHRcompiler := "GNUC" +$ v1 = f$verify(1) +$! GNUC: +$ cc := gcc 'CHRcc_opts'/INCLUDE=([],[-],[--],[--.WWW.Library.Implementation]) +$ v1 = f$verify(v0) +$ ELSE +$ CHRcompiler := "VAXC" +$ v1 = f$verify(1) +$! VAXC: +$ cc := cc 'CHRcc_opts'/INCLUDE=([],[-],[--],[--.WWW.Library.Implementation]) +$ v1 = f$verify(v0) +$ ENDIF +$ ENDIF +$! +$ v1 = f$verify(1) +$ cc makeuctb +$ v1 = f$verify(v0) +$! +$ Link_makeuctb: +$!============= +$ v1 = f$verify(1) +$! +$! Link the Lynx [.SRC.CHRTRANS]makeuctb module. +$! +$ IF f$getsyi("ARCH_NAME") .eqs. "IA64" +$ THEN +$ optslibs="" +$ ELSE +$ optslibs=", sys$disk:[-]''CHRcompiler'.opt/opt" +$ ENDIF +$ +$ link/exe=makeuctb.exe'CHRlink_opts' makeuctb 'optslibs +$ v1 = f$verify(v0) +$! +$ Create_headers: +$!============== +$ v1 = f$verify(1) +$! +$! Create the Lynx [.SRC.CHRTRANS] header files. +$! +$ makeuctb := $'CHRwhere'makeuctb +$ makeuctb cp1250_uni.tbl +$ makeuctb cp1251_uni.tbl +$ makeuctb cp1252_uni.tbl +$ makeuctb cp1253_uni.tbl +$ makeuctb cp1255_uni.tbl +$ makeuctb cp1256_uni.tbl +$ makeuctb cp1257_uni.tbl +$ makeuctb cp437_uni.tbl +$ makeuctb cp737_uni.tbl +$ makeuctb cp775_uni.tbl +$ makeuctb cp850_uni.tbl +$ makeuctb cp852_uni.tbl +$ makeuctb cp857_uni.tbl +$ makeuctb cp862_uni.tbl +$ makeuctb cp864_uni.tbl +$ makeuctb cp866_uni.tbl +$ makeuctb cp866u_uni.tbl +$ makeuctb cp869_uni.tbl +$ makeuctb def7_uni.tbl +$ makeuctb dmcs_uni.tbl +$ makeuctb hp_uni.tbl +$ makeuctb iso01_uni.tbl +$ makeuctb iso02_uni.tbl +$ makeuctb iso03_uni.tbl +$ makeuctb iso04_uni.tbl +$ makeuctb iso05_uni.tbl +$ makeuctb iso06_uni.tbl +$ makeuctb iso07_uni.tbl +$ makeuctb iso08_uni.tbl +$ makeuctb iso09_uni.tbl +$ makeuctb iso10_uni.tbl +$ makeuctb iso13_uni.tbl +$ makeuctb iso14_uni.tbl +$ makeuctb iso15_uni.tbl +$ makeuctb koi8r_uni.tbl +$ makeuctb koi8u_uni.tbl +$ makeuctb mac_uni.tbl +$ makeuctb mnem_suni.tbl +$ makeuctb mnem2_suni.tbl +$ makeuctb mnem_suni.tbl +$ makeuctb next_uni.tbl +$ makeuctb pt154_uni.tbl +$ makeuctb rfc_suni.tbl +$ makeuctb utf8_uni.tbl +$ makeuctb viscii_uni.tbl +$ v1 = f$verify(v0) +$ exit +$! +$ CLEANUP: +$ v1 = f$verify(0) +$ write sys$output "Default directory:" +$ show default +$ v1 = f$verify(v) +$ exit diff --git a/src/chrtrans/build-header.com b/src/chrtrans/build-header.com new file mode 100644 index 0000000..ff8a6f5 --- /dev/null +++ b/src/chrtrans/build-header.com @@ -0,0 +1,37 @@ +$ v0 = 0 +$ v = f$verify(v0) +$! BUILD-HEADER.COM +$! +$! Command file to use MAKEUCTB.EXE on VMS systems for creating +$! a chrtrans header (foo.h) file from a table (foo.tbl) file. +$! Use the file root as P1, e.g.: +$! +$! $ @build-header iso05_uni +$! +$! will create iso05_uni.h from iso05_uni.tbl. +$! +$! 28-Jun-1997 F.Macrides macrides@sci.wfeb.edu +$! Initial version, for Lynx v2.7.1+fotemods +$! +$ ON CONTROL_Y THEN GOTO CLEANUP +$ ON ERROR THEN GOTO CLEANUP +$ CHRproc = f$environment("PROCEDURE") +$ CHRwhere = f$parse(CHRproc,,,"DEVICE") + f$parse(CHRproc,,,"DIRECTORY") +$! +$ Create_header: +$!============= +$ v1 = f$verify(1) +$! +$! Create a Lynx [.SRC.CHRTRANS] header file. +$! +$ makeuctb := $'CHRwhere'makeuctb +$ makeuctb 'P1'.tbl +$ v1 = f$verify(v0) +$ exit +$! +$ CLEANUP: +$ v1 = f$verify(v0) +$ write sys$output "Default directory:" +$ show default +$ v1 = f$verify(v) +$ exit diff --git a/src/chrtrans/caselower.h b/src/chrtrans/caselower.h new file mode 100644 index 0000000..5894a3f --- /dev/null +++ b/src/chrtrans/caselower.h @@ -0,0 +1,738 @@ +/* + Lynx uses this info for 8bit case-insensitive user search. + + This table is generated from Unicode Character Database, Version 2.1.5 + available from ftp.unicode.org, and looks as natural way to get case mapping + equivalents for unicodes. (well, too much characters the cost of 3 Kb only). + Few words from the original README.txt quoted: + +UNICODE 2.1 CHARACTER DATABASE + +Copyright (c) 1991-1998 Unicode, Inc. +All Rights reserved. + +CASE MAPPINGS + +The case mapping is an informative, default mapping. Certain languages, such +as Turkish, German, French, or Greek may have small deviations from the +default mappings listed in the Unicode Character Database. + + */ + +#include <UCkd.h> /* typedef u16 */ + +typedef struct { + u16 upper; + u16 lower; +} unipair_case; + +static const unipair_case unicode_to_lower_case[] = +/* *INDENT-OFF* */ +{ + {0x0041, 0x0061}, /* LATIN CAPITAL LETTER A */ + {0x0042, 0x0062}, /* LATIN CAPITAL LETTER B */ + {0x0043, 0x0063}, /* LATIN CAPITAL LETTER C */ + {0x0044, 0x0064}, /* LATIN CAPITAL LETTER D */ + {0x0045, 0x0065}, /* LATIN CAPITAL LETTER E */ + {0x0046, 0x0066}, /* LATIN CAPITAL LETTER F */ + {0x0047, 0x0067}, /* LATIN CAPITAL LETTER G */ + {0x0048, 0x0068}, /* LATIN CAPITAL LETTER H */ + {0x0049, 0x0069}, /* LATIN CAPITAL LETTER I */ + {0x004A, 0x006A}, /* LATIN CAPITAL LETTER J */ + {0x004B, 0x006B}, /* LATIN CAPITAL LETTER K */ + {0x004C, 0x006C}, /* LATIN CAPITAL LETTER L */ + {0x004D, 0x006D}, /* LATIN CAPITAL LETTER M */ + {0x004E, 0x006E}, /* LATIN CAPITAL LETTER N */ + {0x004F, 0x006F}, /* LATIN CAPITAL LETTER O */ + {0x0050, 0x0070}, /* LATIN CAPITAL LETTER P */ + {0x0051, 0x0071}, /* LATIN CAPITAL LETTER Q */ + {0x0052, 0x0072}, /* LATIN CAPITAL LETTER R */ + {0x0053, 0x0073}, /* LATIN CAPITAL LETTER S */ + {0x0054, 0x0074}, /* LATIN CAPITAL LETTER T */ + {0x0055, 0x0075}, /* LATIN CAPITAL LETTER U */ + {0x0056, 0x0076}, /* LATIN CAPITAL LETTER V */ + {0x0057, 0x0077}, /* LATIN CAPITAL LETTER W */ + {0x0058, 0x0078}, /* LATIN CAPITAL LETTER X */ + {0x0059, 0x0079}, /* LATIN CAPITAL LETTER Y */ + {0x005A, 0x007A}, /* LATIN CAPITAL LETTER Z */ + {0x00C0, 0x00E0}, /* LATIN CAPITAL LETTER A WITH GRAVE */ + {0x00C1, 0x00E1}, /* LATIN CAPITAL LETTER A WITH ACUTE */ + {0x00C2, 0x00E2}, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX */ + {0x00C3, 0x00E3}, /* LATIN CAPITAL LETTER A WITH TILDE */ + {0x00C4, 0x00E4}, /* LATIN CAPITAL LETTER A WITH DIAERESIS */ + {0x00C5, 0x00E5}, /* LATIN CAPITAL LETTER A WITH RING ABOVE */ + {0x00C6, 0x00E6}, /* LATIN CAPITAL LETTER AE */ + {0x00C7, 0x00E7}, /* LATIN CAPITAL LETTER C WITH CEDILLA */ + {0x00C8, 0x00E8}, /* LATIN CAPITAL LETTER E WITH GRAVE */ + {0x00C9, 0x00E9}, /* LATIN CAPITAL LETTER E WITH ACUTE */ + {0x00CA, 0x00EA}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX */ + {0x00CB, 0x00EB}, /* LATIN CAPITAL LETTER E WITH DIAERESIS */ + {0x00CC, 0x00EC}, /* LATIN CAPITAL LETTER I WITH GRAVE */ + {0x00CD, 0x00ED}, /* LATIN CAPITAL LETTER I WITH ACUTE */ + {0x00CE, 0x00EE}, /* LATIN CAPITAL LETTER I WITH CIRCUMFLEX */ + {0x00CF, 0x00EF}, /* LATIN CAPITAL LETTER I WITH DIAERESIS */ + {0x00D0, 0x00F0}, /* LATIN CAPITAL LETTER ETH */ + {0x00D1, 0x00F1}, /* LATIN CAPITAL LETTER N WITH TILDE */ + {0x00D2, 0x00F2}, /* LATIN CAPITAL LETTER O WITH GRAVE */ + {0x00D3, 0x00F3}, /* LATIN CAPITAL LETTER O WITH ACUTE */ + {0x00D4, 0x00F4}, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX */ + {0x00D5, 0x00F5}, /* LATIN CAPITAL LETTER O WITH TILDE */ + {0x00D6, 0x00F6}, /* LATIN CAPITAL LETTER O WITH DIAERESIS */ + {0x00D8, 0x00F8}, /* LATIN CAPITAL LETTER O WITH STROKE */ + {0x00D9, 0x00F9}, /* LATIN CAPITAL LETTER U WITH GRAVE */ + {0x00DA, 0x00FA}, /* LATIN CAPITAL LETTER U WITH ACUTE */ + {0x00DB, 0x00FB}, /* LATIN CAPITAL LETTER U WITH CIRCUMFLEX */ + {0x00DC, 0x00FC}, /* LATIN CAPITAL LETTER U WITH DIAERESIS */ + {0x00DD, 0x00FD}, /* LATIN CAPITAL LETTER Y WITH ACUTE */ + {0x00DE, 0x00FE}, /* LATIN CAPITAL LETTER THORN */ + {0x0100, 0x0101}, /* LATIN CAPITAL LETTER A WITH MACRON */ + {0x0102, 0x0103}, /* LATIN CAPITAL LETTER A WITH BREVE */ + {0x0104, 0x0105}, /* LATIN CAPITAL LETTER A WITH OGONEK */ + {0x0106, 0x0107}, /* LATIN CAPITAL LETTER C WITH ACUTE */ + {0x0108, 0x0109}, /* LATIN CAPITAL LETTER C WITH CIRCUMFLEX */ + {0x010A, 0x010B}, /* LATIN CAPITAL LETTER C WITH DOT ABOVE */ + {0x010C, 0x010D}, /* LATIN CAPITAL LETTER C WITH CARON */ + {0x010E, 0x010F}, /* LATIN CAPITAL LETTER D WITH CARON */ + {0x0110, 0x0111}, /* LATIN CAPITAL LETTER D WITH STROKE */ + {0x0112, 0x0113}, /* LATIN CAPITAL LETTER E WITH MACRON */ + {0x0114, 0x0115}, /* LATIN CAPITAL LETTER E WITH BREVE */ + {0x0116, 0x0117}, /* LATIN CAPITAL LETTER E WITH DOT ABOVE */ + {0x0118, 0x0119}, /* LATIN CAPITAL LETTER E WITH OGONEK */ + {0x011A, 0x011B}, /* LATIN CAPITAL LETTER E WITH CARON */ + {0x011C, 0x011D}, /* LATIN CAPITAL LETTER G WITH CIRCUMFLEX */ + {0x011E, 0x011F}, /* LATIN CAPITAL LETTER G WITH BREVE */ + {0x0120, 0x0121}, /* LATIN CAPITAL LETTER G WITH DOT ABOVE */ + {0x0122, 0x0123}, /* LATIN CAPITAL LETTER G WITH CEDILLA */ + {0x0124, 0x0125}, /* LATIN CAPITAL LETTER H WITH CIRCUMFLEX */ + {0x0126, 0x0127}, /* LATIN CAPITAL LETTER H WITH STROKE */ + {0x0128, 0x0129}, /* LATIN CAPITAL LETTER I WITH TILDE */ + {0x012A, 0x012B}, /* LATIN CAPITAL LETTER I WITH MACRON */ + {0x012C, 0x012D}, /* LATIN CAPITAL LETTER I WITH BREVE */ + {0x012E, 0x012F}, /* LATIN CAPITAL LETTER I WITH OGONEK */ + {0x0130, 0x0069}, /* LATIN CAPITAL LETTER I WITH DOT ABOVE */ + {0x0132, 0x0133}, /* LATIN CAPITAL LIGATURE IJ */ + {0x0134, 0x0135}, /* LATIN CAPITAL LETTER J WITH CIRCUMFLEX */ + {0x0136, 0x0137}, /* LATIN CAPITAL LETTER K WITH CEDILLA */ + {0x0139, 0x013A}, /* LATIN CAPITAL LETTER L WITH ACUTE */ + {0x013B, 0x013C}, /* LATIN CAPITAL LETTER L WITH CEDILLA */ + {0x013D, 0x013E}, /* LATIN CAPITAL LETTER L WITH CARON */ + {0x013F, 0x0140}, /* LATIN CAPITAL LETTER L WITH MIDDLE DOT */ + {0x0141, 0x0142}, /* LATIN CAPITAL LETTER L WITH STROKE */ + {0x0143, 0x0144}, /* LATIN CAPITAL LETTER N WITH ACUTE */ + {0x0145, 0x0146}, /* LATIN CAPITAL LETTER N WITH CEDILLA */ + {0x0147, 0x0148}, /* LATIN CAPITAL LETTER N WITH CARON */ + {0x014A, 0x014B}, /* LATIN CAPITAL LETTER ENG */ + {0x014C, 0x014D}, /* LATIN CAPITAL LETTER O WITH MACRON */ + {0x014E, 0x014F}, /* LATIN CAPITAL LETTER O WITH BREVE */ + {0x0150, 0x0151}, /* LATIN CAPITAL LETTER O WITH DOUBLE ACUTE */ + {0x0152, 0x0153}, /* LATIN CAPITAL LIGATURE OE */ + {0x0154, 0x0155}, /* LATIN CAPITAL LETTER R WITH ACUTE */ + {0x0156, 0x0157}, /* LATIN CAPITAL LETTER R WITH CEDILLA */ + {0x0158, 0x0159}, /* LATIN CAPITAL LETTER R WITH CARON */ + {0x015A, 0x015B}, /* LATIN CAPITAL LETTER S WITH ACUTE */ + {0x015C, 0x015D}, /* LATIN CAPITAL LETTER S WITH CIRCUMFLEX */ + {0x015E, 0x015F}, /* LATIN CAPITAL LETTER S WITH CEDILLA */ + {0x0160, 0x0161}, /* LATIN CAPITAL LETTER S WITH CARON */ + {0x0162, 0x0163}, /* LATIN CAPITAL LETTER T WITH CEDILLA */ + {0x0164, 0x0165}, /* LATIN CAPITAL LETTER T WITH CARON */ + {0x0166, 0x0167}, /* LATIN CAPITAL LETTER T WITH STROKE */ + {0x0168, 0x0169}, /* LATIN CAPITAL LETTER U WITH TILDE */ + {0x016A, 0x016B}, /* LATIN CAPITAL LETTER U WITH MACRON */ + {0x016C, 0x016D}, /* LATIN CAPITAL LETTER U WITH BREVE */ + {0x016E, 0x016F}, /* LATIN CAPITAL LETTER U WITH RING ABOVE */ + {0x0170, 0x0171}, /* LATIN CAPITAL LETTER U WITH DOUBLE ACUTE */ + {0x0172, 0x0173}, /* LATIN CAPITAL LETTER U WITH OGONEK */ + {0x0174, 0x0175}, /* LATIN CAPITAL LETTER W WITH CIRCUMFLEX */ + {0x0176, 0x0177}, /* LATIN CAPITAL LETTER Y WITH CIRCUMFLEX */ + {0x0178, 0x00FF}, /* LATIN CAPITAL LETTER Y WITH DIAERESIS */ + {0x0179, 0x017A}, /* LATIN CAPITAL LETTER Z WITH ACUTE */ + {0x017B, 0x017C}, /* LATIN CAPITAL LETTER Z WITH DOT ABOVE */ + {0x017D, 0x017E}, /* LATIN CAPITAL LETTER Z WITH CARON */ + {0x0181, 0x0253}, /* LATIN CAPITAL LETTER B WITH HOOK */ + {0x0182, 0x0183}, /* LATIN CAPITAL LETTER B WITH TOPBAR */ + {0x0184, 0x0185}, /* LATIN CAPITAL LETTER TONE SIX */ + {0x0186, 0x0254}, /* LATIN CAPITAL LETTER OPEN O */ + {0x0187, 0x0188}, /* LATIN CAPITAL LETTER C WITH HOOK */ + {0x0189, 0x0256}, /* LATIN CAPITAL LETTER AFRICAN D */ + {0x018A, 0x0257}, /* LATIN CAPITAL LETTER D WITH HOOK */ + {0x018B, 0x018C}, /* LATIN CAPITAL LETTER D WITH TOPBAR */ + {0x018E, 0x01DD}, /* LATIN CAPITAL LETTER REVERSED E */ + {0x018F, 0x0259}, /* LATIN CAPITAL LETTER SCHWA */ + {0x0190, 0x025B}, /* LATIN CAPITAL LETTER OPEN E */ + {0x0191, 0x0192}, /* LATIN CAPITAL LETTER F WITH HOOK */ + {0x0193, 0x0260}, /* LATIN CAPITAL LETTER G WITH HOOK */ + {0x0194, 0x0263}, /* LATIN CAPITAL LETTER GAMMA */ + {0x0196, 0x0269}, /* LATIN CAPITAL LETTER IOTA */ + {0x0197, 0x0268}, /* LATIN CAPITAL LETTER I WITH STROKE */ + {0x0198, 0x0199}, /* LATIN CAPITAL LETTER K WITH HOOK */ + {0x019C, 0x026F}, /* LATIN CAPITAL LETTER TURNED M */ + {0x019D, 0x0272}, /* LATIN CAPITAL LETTER N WITH LEFT HOOK */ + {0x019F, 0x0275}, /* LATIN CAPITAL LETTER O WITH MIDDLE TILDE */ + {0x01A0, 0x01A1}, /* LATIN CAPITAL LETTER O WITH HORN */ + {0x01A2, 0x01A3}, /* LATIN CAPITAL LETTER OI */ + {0x01A4, 0x01A5}, /* LATIN CAPITAL LETTER P WITH HOOK */ + {0x01A7, 0x01A8}, /* LATIN CAPITAL LETTER TONE TWO */ + {0x01A9, 0x0283}, /* LATIN CAPITAL LETTER ESH */ + {0x01AC, 0x01AD}, /* LATIN CAPITAL LETTER T WITH HOOK */ + {0x01AE, 0x0288}, /* LATIN CAPITAL LETTER T WITH RETROFLEX HOOK */ + {0x01AF, 0x01B0}, /* LATIN CAPITAL LETTER U WITH HORN */ + {0x01B1, 0x028A}, /* LATIN CAPITAL LETTER UPSILON */ + {0x01B2, 0x028B}, /* LATIN CAPITAL LETTER V WITH HOOK */ + {0x01B3, 0x01B4}, /* LATIN CAPITAL LETTER Y WITH HOOK */ + {0x01B5, 0x01B6}, /* LATIN CAPITAL LETTER Z WITH STROKE */ + {0x01B7, 0x0292}, /* LATIN CAPITAL LETTER EZH */ + {0x01B8, 0x01B9}, /* LATIN CAPITAL LETTER EZH REVERSED */ + {0x01BC, 0x01BD}, /* LATIN CAPITAL LETTER TONE FIVE */ + {0x01C4, 0x01C6}, /* LATIN CAPITAL LETTER DZ WITH CARON */ + {0x01C5, 0x01C6}, /* LATIN CAPITAL LETTER D WITH SMALL LETTER Z WITH CARON */ + {0x01C7, 0x01C9}, /* LATIN CAPITAL LETTER LJ */ + {0x01C8, 0x01C9}, /* LATIN CAPITAL LETTER L WITH SMALL LETTER J */ + {0x01CA, 0x01CC}, /* LATIN CAPITAL LETTER NJ */ + {0x01CB, 0x01CC}, /* LATIN CAPITAL LETTER N WITH SMALL LETTER J */ + {0x01CD, 0x01CE}, /* LATIN CAPITAL LETTER A WITH CARON */ + {0x01CF, 0x01D0}, /* LATIN CAPITAL LETTER I WITH CARON */ + {0x01D1, 0x01D2}, /* LATIN CAPITAL LETTER O WITH CARON */ + {0x01D3, 0x01D4}, /* LATIN CAPITAL LETTER U WITH CARON */ + {0x01D5, 0x01D6}, /* LATIN CAPITAL LETTER U WITH DIAERESIS AND MACRON */ + {0x01D7, 0x01D8}, /* LATIN CAPITAL LETTER U WITH DIAERESIS AND ACUTE */ + {0x01D9, 0x01DA}, /* LATIN CAPITAL LETTER U WITH DIAERESIS AND CARON */ + {0x01DB, 0x01DC}, /* LATIN CAPITAL LETTER U WITH DIAERESIS AND GRAVE */ + {0x01DE, 0x01DF}, /* LATIN CAPITAL LETTER A WITH DIAERESIS AND MACRON */ + {0x01E0, 0x01E1}, /* LATIN CAPITAL LETTER A WITH DOT ABOVE AND MACRON */ + {0x01E2, 0x01E3}, /* LATIN CAPITAL LETTER AE WITH MACRON */ + {0x01E4, 0x01E5}, /* LATIN CAPITAL LETTER G WITH STROKE */ + {0x01E6, 0x01E7}, /* LATIN CAPITAL LETTER G WITH CARON */ + {0x01E8, 0x01E9}, /* LATIN CAPITAL LETTER K WITH CARON */ + {0x01EA, 0x01EB}, /* LATIN CAPITAL LETTER O WITH OGONEK */ + {0x01EC, 0x01ED}, /* LATIN CAPITAL LETTER O WITH OGONEK AND MACRON */ + {0x01EE, 0x01EF}, /* LATIN CAPITAL LETTER EZH WITH CARON */ + {0x01F1, 0x01F3}, /* LATIN CAPITAL LETTER DZ */ + {0x01F2, 0x01F3}, /* LATIN CAPITAL LETTER D WITH SMALL LETTER Z */ + {0x01F4, 0x01F5}, /* LATIN CAPITAL LETTER G WITH ACUTE */ + {0x01FA, 0x01FB}, /* LATIN CAPITAL LETTER A WITH RING ABOVE AND ACUTE */ + {0x01FC, 0x01FD}, /* LATIN CAPITAL LETTER AE WITH ACUTE */ + {0x01FE, 0x01FF}, /* LATIN CAPITAL LETTER O WITH STROKE AND ACUTE */ + {0x0200, 0x0201}, /* LATIN CAPITAL LETTER A WITH DOUBLE GRAVE */ + {0x0202, 0x0203}, /* LATIN CAPITAL LETTER A WITH INVERTED BREVE */ + {0x0204, 0x0205}, /* LATIN CAPITAL LETTER E WITH DOUBLE GRAVE */ + {0x0206, 0x0207}, /* LATIN CAPITAL LETTER E WITH INVERTED BREVE */ + {0x0208, 0x0209}, /* LATIN CAPITAL LETTER I WITH DOUBLE GRAVE */ + {0x020A, 0x020B}, /* LATIN CAPITAL LETTER I WITH INVERTED BREVE */ + {0x020C, 0x020D}, /* LATIN CAPITAL LETTER O WITH DOUBLE GRAVE */ + {0x020E, 0x020F}, /* LATIN CAPITAL LETTER O WITH INVERTED BREVE */ + {0x0210, 0x0211}, /* LATIN CAPITAL LETTER R WITH DOUBLE GRAVE */ + {0x0212, 0x0213}, /* LATIN CAPITAL LETTER R WITH INVERTED BREVE */ + {0x0214, 0x0215}, /* LATIN CAPITAL LETTER U WITH DOUBLE GRAVE */ + {0x0216, 0x0217}, /* LATIN CAPITAL LETTER U WITH INVERTED BREVE */ + {0x0386, 0x03AC}, /* GREEK CAPITAL LETTER ALPHA WITH TONOS */ + {0x0388, 0x03AD}, /* GREEK CAPITAL LETTER EPSILON WITH TONOS */ + {0x0389, 0x03AE}, /* GREEK CAPITAL LETTER ETA WITH TONOS */ + {0x038A, 0x03AF}, /* GREEK CAPITAL LETTER IOTA WITH TONOS */ + {0x038C, 0x03CC}, /* GREEK CAPITAL LETTER OMICRON WITH TONOS */ + {0x038E, 0x03CD}, /* GREEK CAPITAL LETTER UPSILON WITH TONOS */ + {0x038F, 0x03CE}, /* GREEK CAPITAL LETTER OMEGA WITH TONOS */ + {0x0391, 0x03B1}, /* GREEK CAPITAL LETTER ALPHA */ + {0x0392, 0x03B2}, /* GREEK CAPITAL LETTER BETA */ + {0x0393, 0x03B3}, /* GREEK CAPITAL LETTER GAMMA */ + {0x0394, 0x03B4}, /* GREEK CAPITAL LETTER DELTA */ + {0x0395, 0x03B5}, /* GREEK CAPITAL LETTER EPSILON */ + {0x0396, 0x03B6}, /* GREEK CAPITAL LETTER ZETA */ + {0x0397, 0x03B7}, /* GREEK CAPITAL LETTER ETA */ + {0x0398, 0x03B8}, /* GREEK CAPITAL LETTER THETA */ + {0x0399, 0x03B9}, /* GREEK CAPITAL LETTER IOTA */ + {0x039A, 0x03BA}, /* GREEK CAPITAL LETTER KAPPA */ + {0x039B, 0x03BB}, /* GREEK CAPITAL LETTER LAMDA */ + {0x039C, 0x03BC}, /* GREEK CAPITAL LETTER MU */ + {0x039D, 0x03BD}, /* GREEK CAPITAL LETTER NU */ + {0x039E, 0x03BE}, /* GREEK CAPITAL LETTER XI */ + {0x039F, 0x03BF}, /* GREEK CAPITAL LETTER OMICRON */ + {0x03A0, 0x03C0}, /* GREEK CAPITAL LETTER PI */ + {0x03A1, 0x03C1}, /* GREEK CAPITAL LETTER RHO */ + {0x03A3, 0x03C3}, /* GREEK CAPITAL LETTER SIGMA */ + {0x03A4, 0x03C4}, /* GREEK CAPITAL LETTER TAU */ + {0x03A5, 0x03C5}, /* GREEK CAPITAL LETTER UPSILON */ + {0x03A6, 0x03C6}, /* GREEK CAPITAL LETTER PHI */ + {0x03A7, 0x03C7}, /* GREEK CAPITAL LETTER CHI */ + {0x03A8, 0x03C8}, /* GREEK CAPITAL LETTER PSI */ + {0x03A9, 0x03C9}, /* GREEK CAPITAL LETTER OMEGA */ + {0x03AA, 0x03CA}, /* GREEK CAPITAL LETTER IOTA WITH DIALYTIKA */ + {0x03AB, 0x03CB}, /* GREEK CAPITAL LETTER UPSILON WITH DIALYTIKA */ + {0x03E2, 0x03E3}, /* COPTIC CAPITAL LETTER SHEI */ + {0x03E4, 0x03E5}, /* COPTIC CAPITAL LETTER FEI */ + {0x03E6, 0x03E7}, /* COPTIC CAPITAL LETTER KHEI */ + {0x03E8, 0x03E9}, /* COPTIC CAPITAL LETTER HORI */ + {0x03EA, 0x03EB}, /* COPTIC CAPITAL LETTER GANGIA */ + {0x03EC, 0x03ED}, /* COPTIC CAPITAL LETTER SHIMA */ + {0x03EE, 0x03EF}, /* COPTIC CAPITAL LETTER DEI */ + {0x0401, 0x0451}, /* CYRILLIC CAPITAL LETTER IO */ + {0x0402, 0x0452}, /* CYRILLIC CAPITAL LETTER DJE */ + {0x0403, 0x0453}, /* CYRILLIC CAPITAL LETTER GJE */ + {0x0404, 0x0454}, /* CYRILLIC CAPITAL LETTER UKRAINIAN IE */ + {0x0405, 0x0455}, /* CYRILLIC CAPITAL LETTER DZE */ + {0x0406, 0x0456}, /* CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I */ + {0x0407, 0x0457}, /* CYRILLIC CAPITAL LETTER YI */ + {0x0408, 0x0458}, /* CYRILLIC CAPITAL LETTER JE */ + {0x0409, 0x0459}, /* CYRILLIC CAPITAL LETTER LJE */ + {0x040A, 0x045A}, /* CYRILLIC CAPITAL LETTER NJE */ + {0x040B, 0x045B}, /* CYRILLIC CAPITAL LETTER TSHE */ + {0x040C, 0x045C}, /* CYRILLIC CAPITAL LETTER KJE */ + {0x040E, 0x045E}, /* CYRILLIC CAPITAL LETTER SHORT U */ + {0x040F, 0x045F}, /* CYRILLIC CAPITAL LETTER DZHE */ + {0x0410, 0x0430}, /* CYRILLIC CAPITAL LETTER A */ + {0x0411, 0x0431}, /* CYRILLIC CAPITAL LETTER BE */ + {0x0412, 0x0432}, /* CYRILLIC CAPITAL LETTER VE */ + {0x0413, 0x0433}, /* CYRILLIC CAPITAL LETTER GHE */ + {0x0414, 0x0434}, /* CYRILLIC CAPITAL LETTER DE */ + {0x0415, 0x0435}, /* CYRILLIC CAPITAL LETTER IE */ + {0x0416, 0x0436}, /* CYRILLIC CAPITAL LETTER ZHE */ + {0x0417, 0x0437}, /* CYRILLIC CAPITAL LETTER ZE */ + {0x0418, 0x0438}, /* CYRILLIC CAPITAL LETTER I */ + {0x0419, 0x0439}, /* CYRILLIC CAPITAL LETTER SHORT I */ + {0x041A, 0x043A}, /* CYRILLIC CAPITAL LETTER KA */ + {0x041B, 0x043B}, /* CYRILLIC CAPITAL LETTER EL */ + {0x041C, 0x043C}, /* CYRILLIC CAPITAL LETTER EM */ + {0x041D, 0x043D}, /* CYRILLIC CAPITAL LETTER EN */ + {0x041E, 0x043E}, /* CYRILLIC CAPITAL LETTER O */ + {0x041F, 0x043F}, /* CYRILLIC CAPITAL LETTER PE */ + {0x0420, 0x0440}, /* CYRILLIC CAPITAL LETTER ER */ + {0x0421, 0x0441}, /* CYRILLIC CAPITAL LETTER ES */ + {0x0422, 0x0442}, /* CYRILLIC CAPITAL LETTER TE */ + {0x0423, 0x0443}, /* CYRILLIC CAPITAL LETTER U */ + {0x0424, 0x0444}, /* CYRILLIC CAPITAL LETTER EF */ + {0x0425, 0x0445}, /* CYRILLIC CAPITAL LETTER HA */ + {0x0426, 0x0446}, /* CYRILLIC CAPITAL LETTER TSE */ + {0x0427, 0x0447}, /* CYRILLIC CAPITAL LETTER CHE */ + {0x0428, 0x0448}, /* CYRILLIC CAPITAL LETTER SHA */ + {0x0429, 0x0449}, /* CYRILLIC CAPITAL LETTER SHCHA */ + {0x042A, 0x044A}, /* CYRILLIC CAPITAL LETTER HARD SIGN */ + {0x042B, 0x044B}, /* CYRILLIC CAPITAL LETTER YERU */ + {0x042C, 0x044C}, /* CYRILLIC CAPITAL LETTER SOFT SIGN */ + {0x042D, 0x044D}, /* CYRILLIC CAPITAL LETTER E */ + {0x042E, 0x044E}, /* CYRILLIC CAPITAL LETTER YU */ + {0x042F, 0x044F}, /* CYRILLIC CAPITAL LETTER YA */ + {0x0460, 0x0461}, /* CYRILLIC CAPITAL LETTER OMEGA */ + {0x0462, 0x0463}, /* CYRILLIC CAPITAL LETTER YAT */ + {0x0464, 0x0465}, /* CYRILLIC CAPITAL LETTER IOTIFIED E */ + {0x0466, 0x0467}, /* CYRILLIC CAPITAL LETTER LITTLE YUS */ + {0x0468, 0x0469}, /* CYRILLIC CAPITAL LETTER IOTIFIED LITTLE YUS */ + {0x046A, 0x046B}, /* CYRILLIC CAPITAL LETTER BIG YUS */ + {0x046C, 0x046D}, /* CYRILLIC CAPITAL LETTER IOTIFIED BIG YUS */ + {0x046E, 0x046F}, /* CYRILLIC CAPITAL LETTER KSI */ + {0x0470, 0x0471}, /* CYRILLIC CAPITAL LETTER PSI */ + {0x0472, 0x0473}, /* CYRILLIC CAPITAL LETTER FITA */ + {0x0474, 0x0475}, /* CYRILLIC CAPITAL LETTER IZHITSA */ + {0x0476, 0x0477}, /* CYRILLIC CAPITAL LETTER IZHITSA WITH DOUBLE GRAVE ACCENT */ + {0x0478, 0x0479}, /* CYRILLIC CAPITAL LETTER UK */ + {0x047A, 0x047B}, /* CYRILLIC CAPITAL LETTER ROUND OMEGA */ + {0x047C, 0x047D}, /* CYRILLIC CAPITAL LETTER OMEGA WITH TITLO */ + {0x047E, 0x047F}, /* CYRILLIC CAPITAL LETTER OT */ + {0x0480, 0x0481}, /* CYRILLIC CAPITAL LETTER KOPPA */ + {0x0490, 0x0491}, /* CYRILLIC CAPITAL LETTER GHE WITH UPTURN */ + {0x0492, 0x0493}, /* CYRILLIC CAPITAL LETTER GHE WITH STROKE */ + {0x0494, 0x0495}, /* CYRILLIC CAPITAL LETTER GHE WITH MIDDLE HOOK */ + {0x0496, 0x0497}, /* CYRILLIC CAPITAL LETTER ZHE WITH DESCENDER */ + {0x0498, 0x0499}, /* CYRILLIC CAPITAL LETTER ZE WITH DESCENDER */ + {0x049A, 0x049B}, /* CYRILLIC CAPITAL LETTER KA WITH DESCENDER */ + {0x049C, 0x049D}, /* CYRILLIC CAPITAL LETTER KA WITH VERTICAL STROKE */ + {0x049E, 0x049F}, /* CYRILLIC CAPITAL LETTER KA WITH STROKE */ + {0x04A0, 0x04A1}, /* CYRILLIC CAPITAL LETTER BASHKIR KA */ + {0x04A2, 0x04A3}, /* CYRILLIC CAPITAL LETTER EN WITH DESCENDER */ + {0x04A4, 0x04A5}, /* CYRILLIC CAPITAL LIGATURE EN GHE */ + {0x04A6, 0x04A7}, /* CYRILLIC CAPITAL LETTER PE WITH MIDDLE HOOK */ + {0x04A8, 0x04A9}, /* CYRILLIC CAPITAL LETTER ABKHASIAN HA */ + {0x04AA, 0x04AB}, /* CYRILLIC CAPITAL LETTER ES WITH DESCENDER */ + {0x04AC, 0x04AD}, /* CYRILLIC CAPITAL LETTER TE WITH DESCENDER */ + {0x04AE, 0x04AF}, /* CYRILLIC CAPITAL LETTER STRAIGHT U */ + {0x04B0, 0x04B1}, /* CYRILLIC CAPITAL LETTER STRAIGHT U WITH STROKE */ + {0x04B2, 0x04B3}, /* CYRILLIC CAPITAL LETTER HA WITH DESCENDER */ + {0x04B4, 0x04B5}, /* CYRILLIC CAPITAL LIGATURE TE TSE */ + {0x04B6, 0x04B7}, /* CYRILLIC CAPITAL LETTER CHE WITH DESCENDER */ + {0x04B8, 0x04B9}, /* CYRILLIC CAPITAL LETTER CHE WITH VERTICAL STROKE */ + {0x04BA, 0x04BB}, /* CYRILLIC CAPITAL LETTER SHHA */ + {0x04BC, 0x04BD}, /* CYRILLIC CAPITAL LETTER ABKHASIAN CHE */ + {0x04BE, 0x04BF}, /* CYRILLIC CAPITAL LETTER ABKHASIAN CHE WITH DESCENDER */ + {0x04C1, 0x04C2}, /* CYRILLIC CAPITAL LETTER ZHE WITH BREVE */ + {0x04C3, 0x04C4}, /* CYRILLIC CAPITAL LETTER KA WITH HOOK */ + {0x04C7, 0x04C8}, /* CYRILLIC CAPITAL LETTER EN WITH HOOK */ + {0x04CB, 0x04CC}, /* CYRILLIC CAPITAL LETTER KHAKASSIAN CHE */ + {0x04D0, 0x04D1}, /* CYRILLIC CAPITAL LETTER A WITH BREVE */ + {0x04D2, 0x04D3}, /* CYRILLIC CAPITAL LETTER A WITH DIAERESIS */ + {0x04D4, 0x04D5}, /* CYRILLIC CAPITAL LIGATURE A IE */ + {0x04D6, 0x04D7}, /* CYRILLIC CAPITAL LETTER IE WITH BREVE */ + {0x04D8, 0x04D9}, /* CYRILLIC CAPITAL LETTER SCHWA */ + {0x04DA, 0x04DB}, /* CYRILLIC CAPITAL LETTER SCHWA WITH DIAERESIS */ + {0x04DC, 0x04DD}, /* CYRILLIC CAPITAL LETTER ZHE WITH DIAERESIS */ + {0x04DE, 0x04DF}, /* CYRILLIC CAPITAL LETTER ZE WITH DIAERESIS */ + {0x04E0, 0x04E1}, /* CYRILLIC CAPITAL LETTER ABKHASIAN DZE */ + {0x04E2, 0x04E3}, /* CYRILLIC CAPITAL LETTER I WITH MACRON */ + {0x04E4, 0x04E5}, /* CYRILLIC CAPITAL LETTER I WITH DIAERESIS */ + {0x04E6, 0x04E7}, /* CYRILLIC CAPITAL LETTER O WITH DIAERESIS */ + {0x04E8, 0x04E9}, /* CYRILLIC CAPITAL LETTER BARRED O */ + {0x04EA, 0x04EB}, /* CYRILLIC CAPITAL LETTER BARRED O WITH DIAERESIS */ + {0x04EE, 0x04EF}, /* CYRILLIC CAPITAL LETTER U WITH MACRON */ + {0x04F0, 0x04F1}, /* CYRILLIC CAPITAL LETTER U WITH DIAERESIS */ + {0x04F2, 0x04F3}, /* CYRILLIC CAPITAL LETTER U WITH DOUBLE ACUTE */ + {0x04F4, 0x04F5}, /* CYRILLIC CAPITAL LETTER CHE WITH DIAERESIS */ + {0x04F8, 0x04F9}, /* CYRILLIC CAPITAL LETTER YERU WITH DIAERESIS */ + {0x0531, 0x0561}, /* ARMENIAN CAPITAL LETTER AYB */ + {0x0532, 0x0562}, /* ARMENIAN CAPITAL LETTER BEN */ + {0x0533, 0x0563}, /* ARMENIAN CAPITAL LETTER GIM */ + {0x0534, 0x0564}, /* ARMENIAN CAPITAL LETTER DA */ + {0x0535, 0x0565}, /* ARMENIAN CAPITAL LETTER ECH */ + {0x0536, 0x0566}, /* ARMENIAN CAPITAL LETTER ZA */ + {0x0537, 0x0567}, /* ARMENIAN CAPITAL LETTER EH */ + {0x0538, 0x0568}, /* ARMENIAN CAPITAL LETTER ET */ + {0x0539, 0x0569}, /* ARMENIAN CAPITAL LETTER TO */ + {0x053A, 0x056A}, /* ARMENIAN CAPITAL LETTER ZHE */ + {0x053B, 0x056B}, /* ARMENIAN CAPITAL LETTER INI */ + {0x053C, 0x056C}, /* ARMENIAN CAPITAL LETTER LIWN */ + {0x053D, 0x056D}, /* ARMENIAN CAPITAL LETTER XEH */ + {0x053E, 0x056E}, /* ARMENIAN CAPITAL LETTER CA */ + {0x053F, 0x056F}, /* ARMENIAN CAPITAL LETTER KEN */ + {0x0540, 0x0570}, /* ARMENIAN CAPITAL LETTER HO */ + {0x0541, 0x0571}, /* ARMENIAN CAPITAL LETTER JA */ + {0x0542, 0x0572}, /* ARMENIAN CAPITAL LETTER GHAD */ + {0x0543, 0x0573}, /* ARMENIAN CAPITAL LETTER CHEH */ + {0x0544, 0x0574}, /* ARMENIAN CAPITAL LETTER MEN */ + {0x0545, 0x0575}, /* ARMENIAN CAPITAL LETTER YI */ + {0x0546, 0x0576}, /* ARMENIAN CAPITAL LETTER NOW */ + {0x0547, 0x0577}, /* ARMENIAN CAPITAL LETTER SHA */ + {0x0548, 0x0578}, /* ARMENIAN CAPITAL LETTER VO */ + {0x0549, 0x0579}, /* ARMENIAN CAPITAL LETTER CHA */ + {0x054A, 0x057A}, /* ARMENIAN CAPITAL LETTER PEH */ + {0x054B, 0x057B}, /* ARMENIAN CAPITAL LETTER JHEH */ + {0x054C, 0x057C}, /* ARMENIAN CAPITAL LETTER RA */ + {0x054D, 0x057D}, /* ARMENIAN CAPITAL LETTER SEH */ + {0x054E, 0x057E}, /* ARMENIAN CAPITAL LETTER VEW */ + {0x054F, 0x057F}, /* ARMENIAN CAPITAL LETTER TIWN */ + {0x0550, 0x0580}, /* ARMENIAN CAPITAL LETTER REH */ + {0x0551, 0x0581}, /* ARMENIAN CAPITAL LETTER CO */ + {0x0552, 0x0582}, /* ARMENIAN CAPITAL LETTER YIWN */ + {0x0553, 0x0583}, /* ARMENIAN CAPITAL LETTER PIWR */ + {0x0554, 0x0584}, /* ARMENIAN CAPITAL LETTER KEH */ + {0x0555, 0x0585}, /* ARMENIAN CAPITAL LETTER OH */ + {0x0556, 0x0586}, /* ARMENIAN CAPITAL LETTER FEH */ + {0x10A0, 0x10D0}, /* GEORGIAN CAPITAL LETTER AN */ + {0x10A1, 0x10D1}, /* GEORGIAN CAPITAL LETTER BAN */ + {0x10A2, 0x10D2}, /* GEORGIAN CAPITAL LETTER GAN */ + {0x10A3, 0x10D3}, /* GEORGIAN CAPITAL LETTER DON */ + {0x10A4, 0x10D4}, /* GEORGIAN CAPITAL LETTER EN */ + {0x10A5, 0x10D5}, /* GEORGIAN CAPITAL LETTER VIN */ + {0x10A6, 0x10D6}, /* GEORGIAN CAPITAL LETTER ZEN */ + {0x10A7, 0x10D7}, /* GEORGIAN CAPITAL LETTER TAN */ + {0x10A8, 0x10D8}, /* GEORGIAN CAPITAL LETTER IN */ + {0x10A9, 0x10D9}, /* GEORGIAN CAPITAL LETTER KAN */ + {0x10AA, 0x10DA}, /* GEORGIAN CAPITAL LETTER LAS */ + {0x10AB, 0x10DB}, /* GEORGIAN CAPITAL LETTER MAN */ + {0x10AC, 0x10DC}, /* GEORGIAN CAPITAL LETTER NAR */ + {0x10AD, 0x10DD}, /* GEORGIAN CAPITAL LETTER ON */ + {0x10AE, 0x10DE}, /* GEORGIAN CAPITAL LETTER PAR */ + {0x10AF, 0x10DF}, /* GEORGIAN CAPITAL LETTER ZHAR */ + {0x10B0, 0x10E0}, /* GEORGIAN CAPITAL LETTER RAE */ + {0x10B1, 0x10E1}, /* GEORGIAN CAPITAL LETTER SAN */ + {0x10B2, 0x10E2}, /* GEORGIAN CAPITAL LETTER TAR */ + {0x10B3, 0x10E3}, /* GEORGIAN CAPITAL LETTER UN */ + {0x10B4, 0x10E4}, /* GEORGIAN CAPITAL LETTER PHAR */ + {0x10B5, 0x10E5}, /* GEORGIAN CAPITAL LETTER KHAR */ + {0x10B6, 0x10E6}, /* GEORGIAN CAPITAL LETTER GHAN */ + {0x10B7, 0x10E7}, /* GEORGIAN CAPITAL LETTER QAR */ + {0x10B8, 0x10E8}, /* GEORGIAN CAPITAL LETTER SHIN */ + {0x10B9, 0x10E9}, /* GEORGIAN CAPITAL LETTER CHIN */ + {0x10BA, 0x10EA}, /* GEORGIAN CAPITAL LETTER CAN */ + {0x10BB, 0x10EB}, /* GEORGIAN CAPITAL LETTER JIL */ + {0x10BC, 0x10EC}, /* GEORGIAN CAPITAL LETTER CIL */ + {0x10BD, 0x10ED}, /* GEORGIAN CAPITAL LETTER CHAR */ + {0x10BE, 0x10EE}, /* GEORGIAN CAPITAL LETTER XAN */ + {0x10BF, 0x10EF}, /* GEORGIAN CAPITAL LETTER JHAN */ + {0x10C0, 0x10F0}, /* GEORGIAN CAPITAL LETTER HAE */ + {0x10C1, 0x10F1}, /* GEORGIAN CAPITAL LETTER HE */ + {0x10C2, 0x10F2}, /* GEORGIAN CAPITAL LETTER HIE */ + {0x10C3, 0x10F3}, /* GEORGIAN CAPITAL LETTER WE */ + {0x10C4, 0x10F4}, /* GEORGIAN CAPITAL LETTER HAR */ + {0x10C5, 0x10F5}, /* GEORGIAN CAPITAL LETTER HOE */ + {0x1E00, 0x1E01}, /* LATIN CAPITAL LETTER A WITH RING BELOW */ + {0x1E02, 0x1E03}, /* LATIN CAPITAL LETTER B WITH DOT ABOVE */ + {0x1E04, 0x1E05}, /* LATIN CAPITAL LETTER B WITH DOT BELOW */ + {0x1E06, 0x1E07}, /* LATIN CAPITAL LETTER B WITH LINE BELOW */ + {0x1E08, 0x1E09}, /* LATIN CAPITAL LETTER C WITH CEDILLA AND ACUTE */ + {0x1E0A, 0x1E0B}, /* LATIN CAPITAL LETTER D WITH DOT ABOVE */ + {0x1E0C, 0x1E0D}, /* LATIN CAPITAL LETTER D WITH DOT BELOW */ + {0x1E0E, 0x1E0F}, /* LATIN CAPITAL LETTER D WITH LINE BELOW */ + {0x1E10, 0x1E11}, /* LATIN CAPITAL LETTER D WITH CEDILLA */ + {0x1E12, 0x1E13}, /* LATIN CAPITAL LETTER D WITH CIRCUMFLEX BELOW */ + {0x1E14, 0x1E15}, /* LATIN CAPITAL LETTER E WITH MACRON AND GRAVE */ + {0x1E16, 0x1E17}, /* LATIN CAPITAL LETTER E WITH MACRON AND ACUTE */ + {0x1E18, 0x1E19}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX BELOW */ + {0x1E1A, 0x1E1B}, /* LATIN CAPITAL LETTER E WITH TILDE BELOW */ + {0x1E1C, 0x1E1D}, /* LATIN CAPITAL LETTER E WITH CEDILLA AND BREVE */ + {0x1E1E, 0x1E1F}, /* LATIN CAPITAL LETTER F WITH DOT ABOVE */ + {0x1E20, 0x1E21}, /* LATIN CAPITAL LETTER G WITH MACRON */ + {0x1E22, 0x1E23}, /* LATIN CAPITAL LETTER H WITH DOT ABOVE */ + {0x1E24, 0x1E25}, /* LATIN CAPITAL LETTER H WITH DOT BELOW */ + {0x1E26, 0x1E27}, /* LATIN CAPITAL LETTER H WITH DIAERESIS */ + {0x1E28, 0x1E29}, /* LATIN CAPITAL LETTER H WITH CEDILLA */ + {0x1E2A, 0x1E2B}, /* LATIN CAPITAL LETTER H WITH BREVE BELOW */ + {0x1E2C, 0x1E2D}, /* LATIN CAPITAL LETTER I WITH TILDE BELOW */ + {0x1E2E, 0x1E2F}, /* LATIN CAPITAL LETTER I WITH DIAERESIS AND ACUTE */ + {0x1E30, 0x1E31}, /* LATIN CAPITAL LETTER K WITH ACUTE */ + {0x1E32, 0x1E33}, /* LATIN CAPITAL LETTER K WITH DOT BELOW */ + {0x1E34, 0x1E35}, /* LATIN CAPITAL LETTER K WITH LINE BELOW */ + {0x1E36, 0x1E37}, /* LATIN CAPITAL LETTER L WITH DOT BELOW */ + {0x1E38, 0x1E39}, /* LATIN CAPITAL LETTER L WITH DOT BELOW AND MACRON */ + {0x1E3A, 0x1E3B}, /* LATIN CAPITAL LETTER L WITH LINE BELOW */ + {0x1E3C, 0x1E3D}, /* LATIN CAPITAL LETTER L WITH CIRCUMFLEX BELOW */ + {0x1E3E, 0x1E3F}, /* LATIN CAPITAL LETTER M WITH ACUTE */ + {0x1E40, 0x1E41}, /* LATIN CAPITAL LETTER M WITH DOT ABOVE */ + {0x1E42, 0x1E43}, /* LATIN CAPITAL LETTER M WITH DOT BELOW */ + {0x1E44, 0x1E45}, /* LATIN CAPITAL LETTER N WITH DOT ABOVE */ + {0x1E46, 0x1E47}, /* LATIN CAPITAL LETTER N WITH DOT BELOW */ + {0x1E48, 0x1E49}, /* LATIN CAPITAL LETTER N WITH LINE BELOW */ + {0x1E4A, 0x1E4B}, /* LATIN CAPITAL LETTER N WITH CIRCUMFLEX BELOW */ + {0x1E4C, 0x1E4D}, /* LATIN CAPITAL LETTER O WITH TILDE AND ACUTE */ + {0x1E4E, 0x1E4F}, /* LATIN CAPITAL LETTER O WITH TILDE AND DIAERESIS */ + {0x1E50, 0x1E51}, /* LATIN CAPITAL LETTER O WITH MACRON AND GRAVE */ + {0x1E52, 0x1E53}, /* LATIN CAPITAL LETTER O WITH MACRON AND ACUTE */ + {0x1E54, 0x1E55}, /* LATIN CAPITAL LETTER P WITH ACUTE */ + {0x1E56, 0x1E57}, /* LATIN CAPITAL LETTER P WITH DOT ABOVE */ + {0x1E58, 0x1E59}, /* LATIN CAPITAL LETTER R WITH DOT ABOVE */ + {0x1E5A, 0x1E5B}, /* LATIN CAPITAL LETTER R WITH DOT BELOW */ + {0x1E5C, 0x1E5D}, /* LATIN CAPITAL LETTER R WITH DOT BELOW AND MACRON */ + {0x1E5E, 0x1E5F}, /* LATIN CAPITAL LETTER R WITH LINE BELOW */ + {0x1E60, 0x1E61}, /* LATIN CAPITAL LETTER S WITH DOT ABOVE */ + {0x1E62, 0x1E63}, /* LATIN CAPITAL LETTER S WITH DOT BELOW */ + {0x1E64, 0x1E65}, /* LATIN CAPITAL LETTER S WITH ACUTE AND DOT ABOVE */ + {0x1E66, 0x1E67}, /* LATIN CAPITAL LETTER S WITH CARON AND DOT ABOVE */ + {0x1E68, 0x1E69}, /* LATIN CAPITAL LETTER S WITH DOT BELOW AND DOT ABOVE */ + {0x1E6A, 0x1E6B}, /* LATIN CAPITAL LETTER T WITH DOT ABOVE */ + {0x1E6C, 0x1E6D}, /* LATIN CAPITAL LETTER T WITH DOT BELOW */ + {0x1E6E, 0x1E6F}, /* LATIN CAPITAL LETTER T WITH LINE BELOW */ + {0x1E70, 0x1E71}, /* LATIN CAPITAL LETTER T WITH CIRCUMFLEX BELOW */ + {0x1E72, 0x1E73}, /* LATIN CAPITAL LETTER U WITH DIAERESIS BELOW */ + {0x1E74, 0x1E75}, /* LATIN CAPITAL LETTER U WITH TILDE BELOW */ + {0x1E76, 0x1E77}, /* LATIN CAPITAL LETTER U WITH CIRCUMFLEX BELOW */ + {0x1E78, 0x1E79}, /* LATIN CAPITAL LETTER U WITH TILDE AND ACUTE */ + {0x1E7A, 0x1E7B}, /* LATIN CAPITAL LETTER U WITH MACRON AND DIAERESIS */ + {0x1E7C, 0x1E7D}, /* LATIN CAPITAL LETTER V WITH TILDE */ + {0x1E7E, 0x1E7F}, /* LATIN CAPITAL LETTER V WITH DOT BELOW */ + {0x1E80, 0x1E81}, /* LATIN CAPITAL LETTER W WITH GRAVE */ + {0x1E82, 0x1E83}, /* LATIN CAPITAL LETTER W WITH ACUTE */ + {0x1E84, 0x1E85}, /* LATIN CAPITAL LETTER W WITH DIAERESIS */ + {0x1E86, 0x1E87}, /* LATIN CAPITAL LETTER W WITH DOT ABOVE */ + {0x1E88, 0x1E89}, /* LATIN CAPITAL LETTER W WITH DOT BELOW */ + {0x1E8A, 0x1E8B}, /* LATIN CAPITAL LETTER X WITH DOT ABOVE */ + {0x1E8C, 0x1E8D}, /* LATIN CAPITAL LETTER X WITH DIAERESIS */ + {0x1E8E, 0x1E8F}, /* LATIN CAPITAL LETTER Y WITH DOT ABOVE */ + {0x1E90, 0x1E91}, /* LATIN CAPITAL LETTER Z WITH CIRCUMFLEX */ + {0x1E92, 0x1E93}, /* LATIN CAPITAL LETTER Z WITH DOT BELOW */ + {0x1E94, 0x1E95}, /* LATIN CAPITAL LETTER Z WITH LINE BELOW */ + {0x1EA0, 0x1EA1}, /* LATIN CAPITAL LETTER A WITH DOT BELOW */ + {0x1EA2, 0x1EA3}, /* LATIN CAPITAL LETTER A WITH HOOK ABOVE */ + {0x1EA4, 0x1EA5}, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND ACUTE */ + {0x1EA6, 0x1EA7}, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND GRAVE */ + {0x1EA8, 0x1EA9}, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND HOOK ABOVE */ + {0x1EAA, 0x1EAB}, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND TILDE */ + {0x1EAC, 0x1EAD}, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND DOT BELOW */ + {0x1EAE, 0x1EAF}, /* LATIN CAPITAL LETTER A WITH BREVE AND ACUTE */ + {0x1EB0, 0x1EB1}, /* LATIN CAPITAL LETTER A WITH BREVE AND GRAVE */ + {0x1EB2, 0x1EB3}, /* LATIN CAPITAL LETTER A WITH BREVE AND HOOK ABOVE */ + {0x1EB4, 0x1EB5}, /* LATIN CAPITAL LETTER A WITH BREVE AND TILDE */ + {0x1EB6, 0x1EB7}, /* LATIN CAPITAL LETTER A WITH BREVE AND DOT BELOW */ + {0x1EB8, 0x1EB9}, /* LATIN CAPITAL LETTER E WITH DOT BELOW */ + {0x1EBA, 0x1EBB}, /* LATIN CAPITAL LETTER E WITH HOOK ABOVE */ + {0x1EBC, 0x1EBD}, /* LATIN CAPITAL LETTER E WITH TILDE */ + {0x1EBE, 0x1EBF}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND ACUTE */ + {0x1EC0, 0x1EC1}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND GRAVE */ + {0x1EC2, 0x1EC3}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND HOOK ABOVE */ + {0x1EC4, 0x1EC5}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND TILDE */ + {0x1EC6, 0x1EC7}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND DOT BELOW */ + {0x1EC8, 0x1EC9}, /* LATIN CAPITAL LETTER I WITH HOOK ABOVE */ + {0x1ECA, 0x1ECB}, /* LATIN CAPITAL LETTER I WITH DOT BELOW */ + {0x1ECC, 0x1ECD}, /* LATIN CAPITAL LETTER O WITH DOT BELOW */ + {0x1ECE, 0x1ECF}, /* LATIN CAPITAL LETTER O WITH HOOK ABOVE */ + {0x1ED0, 0x1ED1}, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX AND ACUTE */ + {0x1ED2, 0x1ED3}, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX AND GRAVE */ + {0x1ED4, 0x1ED5}, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX AND HOOK ABOVE */ + {0x1ED6, 0x1ED7}, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX AND TILDE */ + {0x1ED8, 0x1ED9}, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX AND DOT BELOW */ + {0x1EDA, 0x1EDB}, /* LATIN CAPITAL LETTER O WITH HORN AND ACUTE */ + {0x1EDC, 0x1EDD}, /* LATIN CAPITAL LETTER O WITH HORN AND GRAVE */ + {0x1EDE, 0x1EDF}, /* LATIN CAPITAL LETTER O WITH HORN AND HOOK ABOVE */ + {0x1EE0, 0x1EE1}, /* LATIN CAPITAL LETTER O WITH HORN AND TILDE */ + {0x1EE2, 0x1EE3}, /* LATIN CAPITAL LETTER O WITH HORN AND DOT BELOW */ + {0x1EE4, 0x1EE5}, /* LATIN CAPITAL LETTER U WITH DOT BELOW */ + {0x1EE6, 0x1EE7}, /* LATIN CAPITAL LETTER U WITH HOOK ABOVE */ + {0x1EE8, 0x1EE9}, /* LATIN CAPITAL LETTER U WITH HORN AND ACUTE */ + {0x1EEA, 0x1EEB}, /* LATIN CAPITAL LETTER U WITH HORN AND GRAVE */ + {0x1EEC, 0x1EED}, /* LATIN CAPITAL LETTER U WITH HORN AND HOOK ABOVE */ + {0x1EEE, 0x1EEF}, /* LATIN CAPITAL LETTER U WITH HORN AND TILDE */ + {0x1EF0, 0x1EF1}, /* LATIN CAPITAL LETTER U WITH HORN AND DOT BELOW */ + {0x1EF2, 0x1EF3}, /* LATIN CAPITAL LETTER Y WITH GRAVE */ + {0x1EF4, 0x1EF5}, /* LATIN CAPITAL LETTER Y WITH DOT BELOW */ + {0x1EF6, 0x1EF7}, /* LATIN CAPITAL LETTER Y WITH HOOK ABOVE */ + {0x1EF8, 0x1EF9}, /* LATIN CAPITAL LETTER Y WITH TILDE */ + {0x1F08, 0x1F00}, /* GREEK CAPITAL LETTER ALPHA WITH PSILI */ + {0x1F09, 0x1F01}, /* GREEK CAPITAL LETTER ALPHA WITH DASIA */ + {0x1F0A, 0x1F02}, /* GREEK CAPITAL LETTER ALPHA WITH PSILI AND VARIA */ + {0x1F0B, 0x1F03}, /* GREEK CAPITAL LETTER ALPHA WITH DASIA AND VARIA */ + {0x1F0C, 0x1F04}, /* GREEK CAPITAL LETTER ALPHA WITH PSILI AND OXIA */ + {0x1F0D, 0x1F05}, /* GREEK CAPITAL LETTER ALPHA WITH DASIA AND OXIA */ + {0x1F0E, 0x1F06}, /* GREEK CAPITAL LETTER ALPHA WITH PSILI AND PERISPOMENI */ + {0x1F0F, 0x1F07}, /* GREEK CAPITAL LETTER ALPHA WITH DASIA AND PERISPOMENI */ + {0x1F18, 0x1F10}, /* GREEK CAPITAL LETTER EPSILON WITH PSILI */ + {0x1F19, 0x1F11}, /* GREEK CAPITAL LETTER EPSILON WITH DASIA */ + {0x1F1A, 0x1F12}, /* GREEK CAPITAL LETTER EPSILON WITH PSILI AND VARIA */ + {0x1F1B, 0x1F13}, /* GREEK CAPITAL LETTER EPSILON WITH DASIA AND VARIA */ + {0x1F1C, 0x1F14}, /* GREEK CAPITAL LETTER EPSILON WITH PSILI AND OXIA */ + {0x1F1D, 0x1F15}, /* GREEK CAPITAL LETTER EPSILON WITH DASIA AND OXIA */ + {0x1F28, 0x1F20}, /* GREEK CAPITAL LETTER ETA WITH PSILI */ + {0x1F29, 0x1F21}, /* GREEK CAPITAL LETTER ETA WITH DASIA */ + {0x1F2A, 0x1F22}, /* GREEK CAPITAL LETTER ETA WITH PSILI AND VARIA */ + {0x1F2B, 0x1F23}, /* GREEK CAPITAL LETTER ETA WITH DASIA AND VARIA */ + {0x1F2C, 0x1F24}, /* GREEK CAPITAL LETTER ETA WITH PSILI AND OXIA */ + {0x1F2D, 0x1F25}, /* GREEK CAPITAL LETTER ETA WITH DASIA AND OXIA */ + {0x1F2E, 0x1F26}, /* GREEK CAPITAL LETTER ETA WITH PSILI AND PERISPOMENI */ + {0x1F2F, 0x1F27}, /* GREEK CAPITAL LETTER ETA WITH DASIA AND PERISPOMENI */ + {0x1F38, 0x1F30}, /* GREEK CAPITAL LETTER IOTA WITH PSILI */ + {0x1F39, 0x1F31}, /* GREEK CAPITAL LETTER IOTA WITH DASIA */ + {0x1F3A, 0x1F32}, /* GREEK CAPITAL LETTER IOTA WITH PSILI AND VARIA */ + {0x1F3B, 0x1F33}, /* GREEK CAPITAL LETTER IOTA WITH DASIA AND VARIA */ + {0x1F3C, 0x1F34}, /* GREEK CAPITAL LETTER IOTA WITH PSILI AND OXIA */ + {0x1F3D, 0x1F35}, /* GREEK CAPITAL LETTER IOTA WITH DASIA AND OXIA */ + {0x1F3E, 0x1F36}, /* GREEK CAPITAL LETTER IOTA WITH PSILI AND PERISPOMENI */ + {0x1F3F, 0x1F37}, /* GREEK CAPITAL LETTER IOTA WITH DASIA AND PERISPOMENI */ + {0x1F48, 0x1F40}, /* GREEK CAPITAL LETTER OMICRON WITH PSILI */ + {0x1F49, 0x1F41}, /* GREEK CAPITAL LETTER OMICRON WITH DASIA */ + {0x1F4A, 0x1F42}, /* GREEK CAPITAL LETTER OMICRON WITH PSILI AND VARIA */ + {0x1F4B, 0x1F43}, /* GREEK CAPITAL LETTER OMICRON WITH DASIA AND VARIA */ + {0x1F4C, 0x1F44}, /* GREEK CAPITAL LETTER OMICRON WITH PSILI AND OXIA */ + {0x1F4D, 0x1F45}, /* GREEK CAPITAL LETTER OMICRON WITH DASIA AND OXIA */ + {0x1F59, 0x1F51}, /* GREEK CAPITAL LETTER UPSILON WITH DASIA */ + {0x1F5B, 0x1F53}, /* GREEK CAPITAL LETTER UPSILON WITH DASIA AND VARIA */ + {0x1F5D, 0x1F55}, /* GREEK CAPITAL LETTER UPSILON WITH DASIA AND OXIA */ + {0x1F5F, 0x1F57}, /* GREEK CAPITAL LETTER UPSILON WITH DASIA AND PERISPOMENI */ + {0x1F68, 0x1F60}, /* GREEK CAPITAL LETTER OMEGA WITH PSILI */ + {0x1F69, 0x1F61}, /* GREEK CAPITAL LETTER OMEGA WITH DASIA */ + {0x1F6A, 0x1F62}, /* GREEK CAPITAL LETTER OMEGA WITH PSILI AND VARIA */ + {0x1F6B, 0x1F63}, /* GREEK CAPITAL LETTER OMEGA WITH DASIA AND VARIA */ + {0x1F6C, 0x1F64}, /* GREEK CAPITAL LETTER OMEGA WITH PSILI AND OXIA */ + {0x1F6D, 0x1F65}, /* GREEK CAPITAL LETTER OMEGA WITH DASIA AND OXIA */ + {0x1F6E, 0x1F66}, /* GREEK CAPITAL LETTER OMEGA WITH PSILI AND PERISPOMENI */ + {0x1F6F, 0x1F67}, /* GREEK CAPITAL LETTER OMEGA WITH DASIA AND PERISPOMENI */ + {0x1F88, 0x1F80}, /* GREEK CAPITAL LETTER ALPHA WITH PSILI AND PROSGEGRAMMENI */ + {0x1F89, 0x1F81}, /* GREEK CAPITAL LETTER ALPHA WITH DASIA AND PROSGEGRAMMENI */ + {0x1F8A, 0x1F82}, /* GREEK CAPITAL LETTER ALPHA WITH PSILI AND VARIA AND PROSGEGRAMMENI */ + {0x1F8B, 0x1F83}, /* GREEK CAPITAL LETTER ALPHA WITH DASIA AND VARIA AND PROSGEGRAMMENI */ + {0x1F8C, 0x1F84}, /* GREEK CAPITAL LETTER ALPHA WITH PSILI AND OXIA AND PROSGEGRAMMENI */ + {0x1F8D, 0x1F85}, /* GREEK CAPITAL LETTER ALPHA WITH DASIA AND OXIA AND PROSGEGRAMMENI */ + {0x1F8E, 0x1F86}, /* GREEK CAPITAL LETTER ALPHA WITH PSILI AND PERISPOMENI AND PROSGEGRAMMENI */ + {0x1F8F, 0x1F87}, /* GREEK CAPITAL LETTER ALPHA WITH DASIA AND PERISPOMENI AND PROSGEGRAMMENI */ + {0x1F98, 0x1F90}, /* GREEK CAPITAL LETTER ETA WITH PSILI AND PROSGEGRAMMENI */ + {0x1F99, 0x1F91}, /* GREEK CAPITAL LETTER ETA WITH DASIA AND PROSGEGRAMMENI */ + {0x1F9A, 0x1F92}, /* GREEK CAPITAL LETTER ETA WITH PSILI AND VARIA AND PROSGEGRAMMENI */ + {0x1F9B, 0x1F93}, /* GREEK CAPITAL LETTER ETA WITH DASIA AND VARIA AND PROSGEGRAMMENI */ + {0x1F9C, 0x1F94}, /* GREEK CAPITAL LETTER ETA WITH PSILI AND OXIA AND PROSGEGRAMMENI */ + {0x1F9D, 0x1F95}, /* GREEK CAPITAL LETTER ETA WITH DASIA AND OXIA AND PROSGEGRAMMENI */ + {0x1F9E, 0x1F96}, /* GREEK CAPITAL LETTER ETA WITH PSILI AND PERISPOMENI AND PROSGEGRAMMENI */ + {0x1F9F, 0x1F97}, /* GREEK CAPITAL LETTER ETA WITH DASIA AND PERISPOMENI AND PROSGEGRAMMENI */ + {0x1FA8, 0x1FA0}, /* GREEK CAPITAL LETTER OMEGA WITH PSILI AND PROSGEGRAMMENI */ + {0x1FA9, 0x1FA1}, /* GREEK CAPITAL LETTER OMEGA WITH DASIA AND PROSGEGRAMMENI */ + {0x1FAA, 0x1FA2}, /* GREEK CAPITAL LETTER OMEGA WITH PSILI AND VARIA AND PROSGEGRAMMENI */ + {0x1FAB, 0x1FA3}, /* GREEK CAPITAL LETTER OMEGA WITH DASIA AND VARIA AND PROSGEGRAMMENI */ + {0x1FAC, 0x1FA4}, /* GREEK CAPITAL LETTER OMEGA WITH PSILI AND OXIA AND PROSGEGRAMMENI */ + {0x1FAD, 0x1FA5}, /* GREEK CAPITAL LETTER OMEGA WITH DASIA AND OXIA AND PROSGEGRAMMENI */ + {0x1FAE, 0x1FA6}, /* GREEK CAPITAL LETTER OMEGA WITH PSILI AND PERISPOMENI AND PROSGEGRAMMENI */ + {0x1FAF, 0x1FA7}, /* GREEK CAPITAL LETTER OMEGA WITH DASIA AND PERISPOMENI AND PROSGEGRAMMENI */ + {0x1FB8, 0x1FB0}, /* GREEK CAPITAL LETTER ALPHA WITH VRACHY */ + {0x1FB9, 0x1FB1}, /* GREEK CAPITAL LETTER ALPHA WITH MACRON */ + {0x1FBA, 0x1F70}, /* GREEK CAPITAL LETTER ALPHA WITH VARIA */ + {0x1FBB, 0x1F71}, /* GREEK CAPITAL LETTER ALPHA WITH OXIA */ + {0x1FBC, 0x1FB3}, /* GREEK CAPITAL LETTER ALPHA WITH PROSGEGRAMMENI */ + {0x1FC8, 0x1F72}, /* GREEK CAPITAL LETTER EPSILON WITH VARIA */ + {0x1FC9, 0x1F73}, /* GREEK CAPITAL LETTER EPSILON WITH OXIA */ + {0x1FCA, 0x1F74}, /* GREEK CAPITAL LETTER ETA WITH VARIA */ + {0x1FCB, 0x1F75}, /* GREEK CAPITAL LETTER ETA WITH OXIA */ + {0x1FCC, 0x1FC3}, /* GREEK CAPITAL LETTER ETA WITH PROSGEGRAMMENI */ + {0x1FD8, 0x1FD0}, /* GREEK CAPITAL LETTER IOTA WITH VRACHY */ + {0x1FD9, 0x1FD1}, /* GREEK CAPITAL LETTER IOTA WITH MACRON */ + {0x1FDA, 0x1F76}, /* GREEK CAPITAL LETTER IOTA WITH VARIA */ + {0x1FDB, 0x1F77}, /* GREEK CAPITAL LETTER IOTA WITH OXIA */ + {0x1FE8, 0x1FE0}, /* GREEK CAPITAL LETTER UPSILON WITH VRACHY */ + {0x1FE9, 0x1FE1}, /* GREEK CAPITAL LETTER UPSILON WITH MACRON */ + {0x1FEA, 0x1F7A}, /* GREEK CAPITAL LETTER UPSILON WITH VARIA */ + {0x1FEB, 0x1F7B}, /* GREEK CAPITAL LETTER UPSILON WITH OXIA */ + {0x1FEC, 0x1FE5}, /* GREEK CAPITAL LETTER RHO WITH DASIA */ + {0x1FF8, 0x1F78}, /* GREEK CAPITAL LETTER OMICRON WITH VARIA */ + {0x1FF9, 0x1F79}, /* GREEK CAPITAL LETTER OMICRON WITH OXIA */ + {0x1FFA, 0x1F7C}, /* GREEK CAPITAL LETTER OMEGA WITH VARIA */ + {0x1FFB, 0x1F7D}, /* GREEK CAPITAL LETTER OMEGA WITH OXIA */ + {0x1FFC, 0x1FF3}, /* GREEK CAPITAL LETTER OMEGA WITH PROSGEGRAMMENI */ + {0x2160, 0x2170}, /* ROMAN NUMERAL ONE */ + {0x2161, 0x2171}, /* ROMAN NUMERAL TWO */ + {0x2162, 0x2172}, /* ROMAN NUMERAL THREE */ + {0x2163, 0x2173}, /* ROMAN NUMERAL FOUR */ + {0x2164, 0x2174}, /* ROMAN NUMERAL FIVE */ + {0x2165, 0x2175}, /* ROMAN NUMERAL SIX */ + {0x2166, 0x2176}, /* ROMAN NUMERAL SEVEN */ + {0x2167, 0x2177}, /* ROMAN NUMERAL EIGHT */ + {0x2168, 0x2178}, /* ROMAN NUMERAL NINE */ + {0x2169, 0x2179}, /* ROMAN NUMERAL TEN */ + {0x216A, 0x217A}, /* ROMAN NUMERAL ELEVEN */ + {0x216B, 0x217B}, /* ROMAN NUMERAL TWELVE */ + {0x216C, 0x217C}, /* ROMAN NUMERAL FIFTY */ + {0x216D, 0x217D}, /* ROMAN NUMERAL ONE HUNDRED */ + {0x216E, 0x217E}, /* ROMAN NUMERAL FIVE HUNDRED */ + {0x216F, 0x217F}, /* ROMAN NUMERAL ONE THOUSAND */ + {0x24B6, 0x24D0}, /* CIRCLED LATIN CAPITAL LETTER A */ + {0x24B7, 0x24D1}, /* CIRCLED LATIN CAPITAL LETTER B */ + {0x24B8, 0x24D2}, /* CIRCLED LATIN CAPITAL LETTER C */ + {0x24B9, 0x24D3}, /* CIRCLED LATIN CAPITAL LETTER D */ + {0x24BA, 0x24D4}, /* CIRCLED LATIN CAPITAL LETTER E */ + {0x24BB, 0x24D5}, /* CIRCLED LATIN CAPITAL LETTER F */ + {0x24BC, 0x24D6}, /* CIRCLED LATIN CAPITAL LETTER G */ + {0x24BD, 0x24D7}, /* CIRCLED LATIN CAPITAL LETTER H */ + {0x24BE, 0x24D8}, /* CIRCLED LATIN CAPITAL LETTER I */ + {0x24BF, 0x24D9}, /* CIRCLED LATIN CAPITAL LETTER J */ + {0x24C0, 0x24DA}, /* CIRCLED LATIN CAPITAL LETTER K */ + {0x24C1, 0x24DB}, /* CIRCLED LATIN CAPITAL LETTER L */ + {0x24C2, 0x24DC}, /* CIRCLED LATIN CAPITAL LETTER M */ + {0x24C3, 0x24DD}, /* CIRCLED LATIN CAPITAL LETTER N */ + {0x24C4, 0x24DE}, /* CIRCLED LATIN CAPITAL LETTER O */ + {0x24C5, 0x24DF}, /* CIRCLED LATIN CAPITAL LETTER P */ + {0x24C6, 0x24E0}, /* CIRCLED LATIN CAPITAL LETTER Q */ + {0x24C7, 0x24E1}, /* CIRCLED LATIN CAPITAL LETTER R */ + {0x24C8, 0x24E2}, /* CIRCLED LATIN CAPITAL LETTER S */ + {0x24C9, 0x24E3}, /* CIRCLED LATIN CAPITAL LETTER T */ + {0x24CA, 0x24E4}, /* CIRCLED LATIN CAPITAL LETTER U */ + {0x24CB, 0x24E5}, /* CIRCLED LATIN CAPITAL LETTER V */ + {0x24CC, 0x24E6}, /* CIRCLED LATIN CAPITAL LETTER W */ + {0x24CD, 0x24E7}, /* CIRCLED LATIN CAPITAL LETTER X */ + {0x24CE, 0x24E8}, /* CIRCLED LATIN CAPITAL LETTER Y */ + {0x24CF, 0x24E9}, /* CIRCLED LATIN CAPITAL LETTER Z */ + {0xFF21, 0xFF41}, /* FULLWIDTH LATIN CAPITAL LETTER A */ + {0xFF22, 0xFF42}, /* FULLWIDTH LATIN CAPITAL LETTER B */ + {0xFF23, 0xFF43}, /* FULLWIDTH LATIN CAPITAL LETTER C */ + {0xFF24, 0xFF44}, /* FULLWIDTH LATIN CAPITAL LETTER D */ + {0xFF25, 0xFF45}, /* FULLWIDTH LATIN CAPITAL LETTER E */ + {0xFF26, 0xFF46}, /* FULLWIDTH LATIN CAPITAL LETTER F */ + {0xFF27, 0xFF47}, /* FULLWIDTH LATIN CAPITAL LETTER G */ + {0xFF28, 0xFF48}, /* FULLWIDTH LATIN CAPITAL LETTER H */ + {0xFF29, 0xFF49}, /* FULLWIDTH LATIN CAPITAL LETTER I */ + {0xFF2A, 0xFF4A}, /* FULLWIDTH LATIN CAPITAL LETTER J */ + {0xFF2B, 0xFF4B}, /* FULLWIDTH LATIN CAPITAL LETTER K */ + {0xFF2C, 0xFF4C}, /* FULLWIDTH LATIN CAPITAL LETTER L */ + {0xFF2D, 0xFF4D}, /* FULLWIDTH LATIN CAPITAL LETTER M */ + {0xFF2E, 0xFF4E}, /* FULLWIDTH LATIN CAPITAL LETTER N */ + {0xFF2F, 0xFF4F}, /* FULLWIDTH LATIN CAPITAL LETTER O */ + {0xFF30, 0xFF50}, /* FULLWIDTH LATIN CAPITAL LETTER P */ + {0xFF31, 0xFF51}, /* FULLWIDTH LATIN CAPITAL LETTER Q */ + {0xFF32, 0xFF52}, /* FULLWIDTH LATIN CAPITAL LETTER R */ + {0xFF33, 0xFF53}, /* FULLWIDTH LATIN CAPITAL LETTER S */ + {0xFF34, 0xFF54}, /* FULLWIDTH LATIN CAPITAL LETTER T */ + {0xFF35, 0xFF55}, /* FULLWIDTH LATIN CAPITAL LETTER U */ + {0xFF36, 0xFF56}, /* FULLWIDTH LATIN CAPITAL LETTER V */ + {0xFF37, 0xFF57}, /* FULLWIDTH LATIN CAPITAL LETTER W */ + {0xFF38, 0xFF58}, /* FULLWIDTH LATIN CAPITAL LETTER X */ + {0xFF39, 0xFF59}, /* FULLWIDTH LATIN CAPITAL LETTER Y */ + {0xFF3A, 0xFF5A} /* FULLWIDTH LATIN CAPITAL LETTER Z */ +}; +/* *INDENT-ON* */ diff --git a/src/chrtrans/cp1250_uni.tbl b/src/chrtrans/cp1250_uni.tbl new file mode 100644 index 0000000..8a19d55 --- /dev/null +++ b/src/chrtrans/cp1250_uni.tbl @@ -0,0 +1,172 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mwindows-1250 + +#Name as a Display Charset (used on Options screen) +OEastern European (windows-1250) + +#Codepage number +C1250 + +# +# Name: cp1250 to Unicode table +# Unicode version: 2.0 +# Table version: 2.01 +# Table format: Format A +# Date: 04/15/98 +# +# Contact: cpxlate@microsoft.com +# +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp1250 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp1250 order +# +################## +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw + +0x20-0x7e idem +# +0x80 U+20AC #EURO SIGN +0x81 #UNDEFINED +0x82 U+201A #SINGLE LOW-9 QUOTATION MARK +0x83 #UNDEFINED +0x84 U+201E #DOUBLE LOW-9 QUOTATION MARK +0x85 U+2026 #HORIZONTAL ELLIPSIS +0x86 U+2020 #DAGGER +0x87 U+2021 #DOUBLE DAGGER +0x88 #UNDEFINED +0x89 U+2030 #PER MILLE SIGN +0x8A U+0160 U+0428 #LATIN CAPITAL LETTER S WITH CARON +0x8B U+2039 #SINGLE LEFT-POINTING ANGLE QUOTATION MARK +0x8C U+015A #LATIN CAPITAL LETTER S WITH ACUTE +0x8D U+0164 #LATIN CAPITAL LETTER T WITH CARON +0x8E U+017D U+0416 #LATIN CAPITAL LETTER Z WITH CARON +0x8F U+0179 #LATIN CAPITAL LETTER Z WITH ACUTE +0x90 #UNDEFINED +0x91 U+2018 #LEFT SINGLE QUOTATION MARK +0x92 U+2019 #RIGHT SINGLE QUOTATION MARK +0x93 U+201C #LEFT DOUBLE QUOTATION MARK +0x94 U+201D #RIGHT DOUBLE QUOTATION MARK +0x95 U+2022 #BULLET +0x96 U+2013 #EN DASH +0x97 U+2014 #EM DASH +0x98 #UNDEFINED +0x99 U+2122 #TRADE MARK SIGN +0x9A U+0161 U+0448 #LATIN SMALL LETTER S WITH CARON +0x9B U+203A #SINGLE RIGHT-POINTING ANGLE QUOTATION MARK +0x9C U+015B #LATIN SMALL LETTER S WITH ACUTE +0x9D U+0165 #LATIN SMALL LETTER T WITH CARON +0x9E U+017E U+0436 #LATIN SMALL LETTER Z WITH CARON +0x9F U+017A #LATIN SMALL LETTER Z WITH ACUTE +0xA0 U+00A0 #NO-BREAK SPACE +0xA1 U+02C7 U+030c #CARON +0xA2 U+02D8 U+0306 #BREVE +0xA3 U+0141 #LATIN CAPITAL LETTER L WITH STROKE +0xA4 U+00A4 #CURRENCY SIGN +0xA5 U+0104 #LATIN CAPITAL LETTER A WITH OGONEK +0xA6 U+00A6 #BROKEN BAR +0xA7 U+00A7 #SECTION SIGN +0xA8 U+00A8 U+0308 #DIAERESIS +0xA9 U+00A9 #COPYRIGHT SIGN +0xAA U+015E #LATIN CAPITAL LETTER S WITH CEDILLA +0xAB U+00AB #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC #NOT SIGN +0xAD U+00AD #SOFT HYPHEN +0xAE U+00AE #REGISTERED SIGN +0xAF U+017B #LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xB0 U+00B0 U+030a #DEGREE SIGN +0xB1 U+00B1 #PLUS-MINUS SIGN +0xB2 U+02DB U+0328 #OGONEK +0xB3 U+0142 #LATIN SMALL LETTER L WITH STROKE +0xB4 U+00B4 #ACUTE ACCENT +0xB5 U+00B5 U+03bc #MICRO SIGN +0xB6 U+00B6 #PILCROW SIGN +0xB7 U+00B7 #MIDDLE DOT +0xB8 U+00B8 U+0327 #CEDILLA +0xB9 U+0105 #LATIN SMALL LETTER A WITH OGONEK +0xBA U+015F #LATIN SMALL LETTER S WITH CEDILLA +0xBB U+00BB #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+013D #LATIN CAPITAL LETTER L WITH CARON +0xBD U+02DD U+030b #DOUBLE ACUTE ACCENT +0xBE U+013E #LATIN SMALL LETTER L WITH CARON +0xBF U+017C #LATIN SMALL LETTER Z WITH DOT ABOVE +0xC0 U+0154 #LATIN CAPITAL LETTER R WITH ACUTE +0xC1 U+00C1 #LATIN CAPITAL LETTER A WITH ACUTE +0xC2 U+00C2 #LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 U+0102 #LATIN CAPITAL LETTER A WITH BREVE +0xC4 U+00C4 #LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+0139 #LATIN CAPITAL LETTER L WITH ACUTE +0xC6 U+0106 #LATIN CAPITAL LETTER C WITH ACUTE +0xC7 U+00C7 #LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 U+010C U+0427 # LATIN CAPITAL LETTER C WITH CARON +0xC9 U+00C9 #LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+0118 #LATIN CAPITAL LETTER E WITH OGONEK +0xCB U+00CB #LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC U+011A #LATIN CAPITAL LETTER E WITH CARON +0xCD U+00CD #LATIN CAPITAL LETTER I WITH ACUTE +0xCE U+00CE #LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF U+010E #LATIN CAPITAL LETTER D WITH CARON +0xD0 U+0110 #LATIN CAPITAL LETTER D WITH STROKE +0xD1 U+0143 #LATIN CAPITAL LETTER N WITH ACUTE +0xD2 U+0147 #LATIN CAPITAL LETTER N WITH CARON +0xD3 U+00D3 #LATIN CAPITAL LETTER O WITH ACUTE +0xD4 U+00D4 #LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 U+0150 #LATIN CAPITAL LETTER O WITH DOUBLE ACUTE +0xD6 U+00D6 #LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+00D7 #MULTIPLICATION SIGN +0xD8 U+0158 #LATIN CAPITAL LETTER R WITH CARON +0xD9 U+016E #LATIN CAPITAL LETTER U WITH RING ABOVE +0xDA U+00DA #LATIN CAPITAL LETTER U WITH ACUTE +0xDB U+0170 #LATIN CAPITAL LETTER U WITH DOUBLE ACUTE +0xDC U+00DC #LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+00DD #LATIN CAPITAL LETTER Y WITH ACUTE +0xDE U+0162 #LATIN CAPITAL LETTER T WITH CEDILLA +0xDF U+00DF #LATIN SMALL LETTER SHARP S +0xE0 U+0155 #LATIN SMALL LETTER R WITH ACUTE +0xE1 U+00E1 #LATIN SMALL LETTER A WITH ACUTE +0xE2 U+00E2 #LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 U+0103 #LATIN SMALL LETTER A WITH BREVE +0xE4 U+00E4 #LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+013A #LATIN SMALL LETTER L WITH ACUTE +0xE6 U+0107 #LATIN SMALL LETTER C WITH ACUTE +0xE7 U+00E7 #LATIN SMALL LETTER C WITH CEDILLA +0xE8 U+010D U+02a7 U+0447 # LATIN SMALL LETTER C WITH CARON +0xE9 U+00E9 #LATIN SMALL LETTER E WITH ACUTE +0xEA U+0119 #LATIN SMALL LETTER E WITH OGONEK +0xEB U+00EB #LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+011B #LATIN SMALL LETTER E WITH CARON +0xED U+00ED #LATIN SMALL LETTER I WITH ACUTE +0xEE U+00EE #LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF U+010F #LATIN SMALL LETTER D WITH CARON +0xF0 U+0111 #LATIN SMALL LETTER D WITH STROKE +0xF1 U+0144 #LATIN SMALL LETTER N WITH ACUTE +0xF2 U+0148 #LATIN SMALL LETTER N WITH CARON +0xF3 U+00F3 #LATIN SMALL LETTER O WITH ACUTE +0xF4 U+00F4 #LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 U+0151 #LATIN SMALL LETTER O WITH DOUBLE ACUTE +0xF6 U+00F6 #LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+00F7 #DIVISION SIGN +0xF8 U+0159 #LATIN SMALL LETTER R WITH CARON +0xF9 U+016F #LATIN SMALL LETTER U WITH RING ABOVE +0xFA U+00FA #LATIN SMALL LETTER U WITH ACUTE +0xFB U+0171 #LATIN SMALL LETTER U WITH DOUBLE ACUTE +0xFC U+00FC #LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+00FD #LATIN SMALL LETTER Y WITH ACUTE +0xFE U+0163 #LATIN SMALL LETTER T WITH CEDILLA +0xFF U+02D9 U+0307 U+0387 #DOT ABOVE + +U+2218 " \260 " # RING OPERATOR +U+2219 " \225 " # BULLET OPERATOR +U+2297 "(\327)" # CIRCLED TIMES +U+2299 "(\267)" # CIRCLED DOT OPERATOR +U+229A "(\260)" # CIRCLED RING OPERATOR +U+22A0 "[\327]" # SQUARED TIMES +U+22A1 "[\267]" # SQUARED DOT OPERATOR +U+22C5 " \267 " # DOT OPERATOR diff --git a/src/chrtrans/cp1251_uni.tbl b/src/chrtrans/cp1251_uni.tbl new file mode 100644 index 0000000..0d928f9 --- /dev/null +++ b/src/chrtrans/cp1251_uni.tbl @@ -0,0 +1,161 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mwindows-1251 + +#Name as a Display Charset (used on Options screen) +OCyrillic (windows-1251) + +#Codepage number +C1251 + +# +# Name: cp1251 to Unicode table +# Unicode version: 2.0 +# Table version: 2.01 +# Table format: Format A +# Date: 04/15/98 +# +# Contact: cpxlate@microsoft.com +# +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp1251 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp1251 order +# +################## + +0x20-0x7e idem +# +0x80 U+0402 #CYRILLIC CAPITAL LETTER DJE +0x81 U+0403 #CYRILLIC CAPITAL LETTER GJE +0x82 U+201A #SINGLE LOW-9 QUOTATION MARK +0x83 U+0453 #CYRILLIC SMALL LETTER GJE +0x84 U+201E #DOUBLE LOW-9 QUOTATION MARK +0x85 U+2026 #HORIZONTAL ELLIPSIS +0x86 U+2020 #DAGGER +0x87 U+2021 #DOUBLE DAGGER +0x88 U+20AC #EURO SIGN +0x89 U+2030 #PER MILLE SIGN +0x8A U+0409 #CYRILLIC CAPITAL LETTER LJE +0x8B U+2039 #SINGLE LEFT-POINTING ANGLE QUOTATION MARK +0x8C U+040A #CYRILLIC CAPITAL LETTER NJE +0x8D U+040C #CYRILLIC CAPITAL LETTER KJE +0x8E U+040B #CYRILLIC CAPITAL LETTER TSHE +0x8F U+040F #CYRILLIC CAPITAL LETTER DZHE +0x90 U+0452 #CYRILLIC SMALL LETTER DJE +0x91 U+2018 #LEFT SINGLE QUOTATION MARK +0x92 U+2019 #RIGHT SINGLE QUOTATION MARK +0x93 U+201C #LEFT DOUBLE QUOTATION MARK +0x94 U+201D #RIGHT DOUBLE QUOTATION MARK +0x95 U+2022 #BULLET +0x96 U+2013 #EN DASH +0x97 U+2014 #EM DASH +0x98 #UNDEFINED +0x99 U+2122 #TRADE MARK SIGN +0x9A U+0459 #CYRILLIC SMALL LETTER LJE +0x9B U+203A #SINGLE RIGHT-POINTING ANGLE QUOTATION MARK +0x9C U+045A #CYRILLIC SMALL LETTER NJE +0x9D U+045C #CYRILLIC SMALL LETTER KJE +0x9E U+045B #CYRILLIC SMALL LETTER TSHE +0x9F U+045F #CYRILLIC SMALL LETTER DZHE +0xA0 U+00A0 #NO-BREAK SPACE +0xA1 U+040E #CYRILLIC CAPITAL LETTER SHORT U +0xA2 U+045E #CYRILLIC SMALL LETTER SHORT U +0xA3 U+0408 #CYRILLIC CAPITAL LETTER JE +0xA4 U+00A4 #CURRENCY SIGN +0xA5 U+0490 #CYRILLIC CAPITAL LETTER GHE WITH UPTURN +0xA6 U+00A6 #BROKEN BAR +0xA7 U+00A7 #SECTION SIGN +0xA8 U+0401 #CYRILLIC CAPITAL LETTER IO +0xA9 U+00A9 #COPYRIGHT SIGN +0xAA U+0404 #CYRILLIC CAPITAL LETTER UKRAINIAN IE +0xAB U+00AB #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC #NOT SIGN +0xAD U+00AD #SOFT HYPHEN +0xAE U+00AE #REGISTERED SIGN +0xAF U+0407 #CYRILLIC CAPITAL LETTER YI +0xB0 U+00B0 #DEGREE SIGN +0xB1 U+00B1 #PLUS-MINUS SIGN +0xB2 U+0406 #CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I +0xB3 U+0456 #CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I +0xB4 U+0491 #CYRILLIC SMALL LETTER GHE WITH UPTURN +0xB5 U+00B5 #MICRO SIGN +0xB6 U+00B6 #PILCROW SIGN +0xB7 U+00B7 #MIDDLE DOT +0xB8 U+0451 #CYRILLIC SMALL LETTER IO +0xB9 U+2116 #NUMERO SIGN +0xBA U+0454 #CYRILLIC SMALL LETTER UKRAINIAN IE +0xBB U+00BB #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+0458 #CYRILLIC SMALL LETTER JE +0xBD U+0405 #CYRILLIC CAPITAL LETTER DZE +0xBE U+0455 #CYRILLIC SMALL LETTER DZE +0xBF U+0457 #CYRILLIC SMALL LETTER YI +0xC0 U+0410 #CYRILLIC CAPITAL LETTER A +0xC1 U+0411 #CYRILLIC CAPITAL LETTER BE +0xC2 U+0412 #CYRILLIC CAPITAL LETTER VE +0xC3 U+0413 #CYRILLIC CAPITAL LETTER GHE +0xC4 U+0414 #CYRILLIC CAPITAL LETTER DE +0xC5 U+0415 #CYRILLIC CAPITAL LETTER IE +0xC6 U+0416 #CYRILLIC CAPITAL LETTER ZHE +0xC7 U+0417 #CYRILLIC CAPITAL LETTER ZE +0xC8 U+0418 #CYRILLIC CAPITAL LETTER I +0xC9 U+0419 #CYRILLIC CAPITAL LETTER SHORT I +0xCA U+041A #CYRILLIC CAPITAL LETTER KA +0xCB U+041B #CYRILLIC CAPITAL LETTER EL +0xCC U+041C #CYRILLIC CAPITAL LETTER EM +0xCD U+041D #CYRILLIC CAPITAL LETTER EN +0xCE U+041E #CYRILLIC CAPITAL LETTER O +0xCF U+041F #CYRILLIC CAPITAL LETTER PE +0xD0 U+0420 #CYRILLIC CAPITAL LETTER ER +0xD1 U+0421 #CYRILLIC CAPITAL LETTER ES +0xD2 U+0422 #CYRILLIC CAPITAL LETTER TE +0xD3 U+0423 #CYRILLIC CAPITAL LETTER U +0xD4 U+0424 #CYRILLIC CAPITAL LETTER EF +0xD5 U+0425 #CYRILLIC CAPITAL LETTER HA +0xD6 U+0426 #CYRILLIC CAPITAL LETTER TSE +0xD7 U+0427 #CYRILLIC CAPITAL LETTER CHE +0xD8 U+0428 #CYRILLIC CAPITAL LETTER SHA +0xD9 U+0429 #CYRILLIC CAPITAL LETTER SHCHA +0xDA U+042A #CYRILLIC CAPITAL LETTER HARD SIGN +0xDB U+042B #CYRILLIC CAPITAL LETTER YERU +0xDC U+042C #CYRILLIC CAPITAL LETTER SOFT SIGN +0xDD U+042D #CYRILLIC CAPITAL LETTER E +0xDE U+042E #CYRILLIC CAPITAL LETTER YU +0xDF U+042F #CYRILLIC CAPITAL LETTER YA +0xE0 U+0430 #CYRILLIC SMALL LETTER A +0xE1 U+0431 #CYRILLIC SMALL LETTER BE +0xE2 U+0432 #CYRILLIC SMALL LETTER VE +0xE3 U+0433 #CYRILLIC SMALL LETTER GHE +0xE4 U+0434 #CYRILLIC SMALL LETTER DE +0xE5 U+0435 #CYRILLIC SMALL LETTER IE +0xE6 U+0436 #CYRILLIC SMALL LETTER ZHE +0xE7 U+0437 #CYRILLIC SMALL LETTER ZE +0xE8 U+0438 #CYRILLIC SMALL LETTER I +0xE9 U+0439 #CYRILLIC SMALL LETTER SHORT I +0xEA U+043A #CYRILLIC SMALL LETTER KA +0xEB U+043B #CYRILLIC SMALL LETTER EL +0xEC U+043C #CYRILLIC SMALL LETTER EM +0xED U+043D #CYRILLIC SMALL LETTER EN +0xEE U+043E #CYRILLIC SMALL LETTER O +0xEF U+043F #CYRILLIC SMALL LETTER PE +0xF0 U+0440 #CYRILLIC SMALL LETTER ER +0xF1 U+0441 #CYRILLIC SMALL LETTER ES +0xF2 U+0442 #CYRILLIC SMALL LETTER TE +0xF3 U+0443 #CYRILLIC SMALL LETTER U +0xF4 U+0444 #CYRILLIC SMALL LETTER EF +0xF5 U+0445 #CYRILLIC SMALL LETTER HA +0xF6 U+0446 #CYRILLIC SMALL LETTER TSE +0xF7 U+0447 #CYRILLIC SMALL LETTER CHE +0xF8 U+0448 #CYRILLIC SMALL LETTER SHA +0xF9 U+0449 #CYRILLIC SMALL LETTER SHCHA +0xFA U+044A #CYRILLIC SMALL LETTER HARD SIGN +0xFB U+044B #CYRILLIC SMALL LETTER YERU +0xFC U+044C #CYRILLIC SMALL LETTER SOFT SIGN +0xFD U+044D #CYRILLIC SMALL LETTER E +0xFE U+044E #CYRILLIC SMALL LETTER YU +0xFF U+044F #CYRILLIC SMALL LETTER YA diff --git a/src/chrtrans/cp1252_uni.tbl b/src/chrtrans/cp1252_uni.tbl new file mode 100644 index 0000000..50ce967 --- /dev/null +++ b/src/chrtrans/cp1252_uni.tbl @@ -0,0 +1,177 @@ +# This file has been modified for lynx (see README.tables) + +#Shall this become the "default" translation? +#There has to be exactly one table marked as "default". +D0 +# +#The MIME name of this charset. +Mwindows-1252 + +#Name as a Display Charset (used on Options screen) +OWestern (windows-1252) + +#Codepage number +C1252 + +# +# Name: cp1252 to Unicode table +# Unicode version: 2.0 +# Table version: 2.01 +# Table format: Format A +# Date: 04/15/98 +# +# Contact: cpxlate@microsoft.com +# +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp1252 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp1252 order +# +################## +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw + +0x20-0x7e idem +# +0x80 U+20AC #EURO SIGN +0x81 #UNDEFINED +0x82 U+201A #SINGLE LOW-9 QUOTATION MARK +0x83 U+0192 #LATIN SMALL LETTER F WITH HOOK +0x84 U+201E #DOUBLE LOW-9 QUOTATION MARK +0x85 U+2026 #HORIZONTAL ELLIPSIS +0x86 U+2020 #DAGGER +0x87 U+2021 #DOUBLE DAGGER +0x88 U+02C6 U+0302 #MODIFIER LETTER CIRCUMFLEX ACCENT +0x89 U+2030 #PER MILLE SIGN +0x8A U+0160 #LATIN CAPITAL LETTER S WITH CARON +0x8B U+2039 #SINGLE LEFT-POINTING ANGLE QUOTATION MARK +0x8C U+0152 #LATIN CAPITAL LIGATURE OE +0x8D #UNDEFINED +0x8E U+017D #LATIN CAPITAL LETTER Z WITH CARON +0x8F #UNDEFINED +0x90 #UNDEFINED +0x91 U+2018 #LEFT SINGLE QUOTATION MARK +0x92 U+2019 #RIGHT SINGLE QUOTATION MARK +0x93 U+201C #LEFT DOUBLE QUOTATION MARK +0x94 U+201D U+02dd U+030b #RIGHT DOUBLE QUOTATION MARK +0x95 U+2022 #BULLET +0x96 U+2013 #EN DASH +0x97 U+2014 #EM DASH +0x98 U+02DC #SMALL TILDE +0x99 U+2122 #TRADE MARK SIGN +0x9A U+0161 #LATIN SMALL LETTER S WITH CARON +0x9B U+203A #SINGLE RIGHT-POINTING ANGLE QUOTATION MARK +0x9C U+0153 #LATIN SMALL LIGATURE OE +0x9D #UNDEFINED +0x9E U+017E #LATIN SMALL LETTER Z WITH CARON +0x9F U+0178 #LATIN CAPITAL LETTER Y WITH DIAERESIS +0xA0 U+00A0 #NO-BREAK SPACE +0xA1 U+00A1 #INVERTED EXCLAMATION MARK +0xA2 U+00A2 #CENT SIGN +0xA3 U+00A3 #POUND SIGN +0xA4 U+00A4 #CURRENCY SIGN +0xA5 U+00A5 #YEN SIGN +0xA6 U+00A6 #BROKEN BAR +0xA7 U+00A7 #SECTION SIGN +0xA8 U+00A8 U+0308 #DIAERESIS +0xA9 U+00A9 #COPYRIGHT SIGN +0xAA U+00AA #FEMININE ORDINAL INDICATOR +0xAB U+00AB #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC #NOT SIGN +0xAD U+00AD #SOFT HYPHEN +0xAE U+00AE #REGISTERED SIGN +0xAF U+00AF U+0304 #MACRON +0xB0 U+00B0 U+030a #DEGREE SIGN +0xB1 U+00B1 #PLUS-MINUS SIGN +0xB2 U+00B2 #SUPERSCRIPT TWO +0xB3 U+00B3 #SUPERSCRIPT THREE +0xB4 U+00B4 #ACUTE ACCENT +0xB5 U+00B5 U+03bc #MICRO SIGN +0xB6 U+00B6 #PILCROW SIGN +0xB7 U+00B7 U+0307 U+0387 U+2027 #MIDDLE DOT +0xB8 U+00B8 U+0327 #CEDILLA +0xB9 U+00B9 #SUPERSCRIPT ONE +0xBA U+00BA #MASCULINE ORDINAL INDICATOR +0xBB U+00BB #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+00BC #VULGAR FRACTION ONE QUARTER +0xBD U+00BD #VULGAR FRACTION ONE HALF +0xBE U+00BE #VULGAR FRACTION THREE QUARTERS +0xBF U+00BF #INVERTED QUESTION MARK +0xC0 U+00C0 #LATIN CAPITAL LETTER A WITH GRAVE +0xC1 U+00C1 #LATIN CAPITAL LETTER A WITH ACUTE +0xC2 U+00C2 #LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 U+00C3 #LATIN CAPITAL LETTER A WITH TILDE +0xC4 U+00C4 #LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+00C5 #LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 U+00C6 #LATIN CAPITAL LETTER AE +0xC7 U+00C7 #LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 U+00C8 #LATIN CAPITAL LETTER E WITH GRAVE +0xC9 U+00C9 #LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+00CA #LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB U+00CB #LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC U+00CC #LATIN CAPITAL LETTER I WITH GRAVE +0xCD U+00CD #LATIN CAPITAL LETTER I WITH ACUTE +0xCE U+00CE #LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF U+00CF #LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 U+00D0 #LATIN CAPITAL LETTER ETH +0xD1 U+00D1 #LATIN CAPITAL LETTER N WITH TILDE +0xD2 U+00D2 #LATIN CAPITAL LETTER O WITH GRAVE +0xD3 U+00D3 #LATIN CAPITAL LETTER O WITH ACUTE +0xD4 U+00D4 #LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 U+00D5 #LATIN CAPITAL LETTER O WITH TILDE +0xD6 U+00D6 #LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+00D7 #MULTIPLICATION SIGN +0xD8 U+00D8 #LATIN CAPITAL LETTER O WITH STROKE +0xD9 U+00D9 #LATIN CAPITAL LETTER U WITH GRAVE +0xDA U+00DA #LATIN CAPITAL LETTER U WITH ACUTE +0xDB U+00DB #LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC U+00DC #LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+00DD #LATIN CAPITAL LETTER Y WITH ACUTE +0xDE U+00DE #LATIN CAPITAL LETTER THORN +0xDF U+00DF #LATIN SMALL LETTER SHARP S +0xE0 U+00E0 #LATIN SMALL LETTER A WITH GRAVE +0xE1 U+00E1 #LATIN SMALL LETTER A WITH ACUTE +0xE2 U+00E2 #LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 U+00E3 #LATIN SMALL LETTER A WITH TILDE +0xE4 U+00E4 #LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+00E5 #LATIN SMALL LETTER A WITH RING ABOVE +0xE6 U+00E6 #LATIN SMALL LETTER AE +0xE7 U+00E7 #LATIN SMALL LETTER C WITH CEDILLA +0xE8 U+00E8 #LATIN SMALL LETTER E WITH GRAVE +0xE9 U+00E9 #LATIN SMALL LETTER E WITH ACUTE +0xEA U+00EA #LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB U+00EB #LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+00EC #LATIN SMALL LETTER I WITH GRAVE +0xED U+00ED #LATIN SMALL LETTER I WITH ACUTE +0xEE U+00EE #LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF U+00EF #LATIN SMALL LETTER I WITH DIAERESIS +0xF0 U+00F0 #LATIN SMALL LETTER ETH +0xF1 U+00F1 #LATIN SMALL LETTER N WITH TILDE +0xF2 U+00F2 #LATIN SMALL LETTER O WITH GRAVE +0xF3 U+00F3 #LATIN SMALL LETTER O WITH ACUTE +0xF4 U+00F4 #LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 U+00F5 #LATIN SMALL LETTER O WITH TILDE +0xF6 U+00F6 #LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+00F7 #DIVISION SIGN +0xF8 U+00F8 #LATIN SMALL LETTER O WITH STROKE +0xF9 U+00F9 #LATIN SMALL LETTER U WITH GRAVE +0xFA U+00FA #LATIN SMALL LETTER U WITH ACUTE +0xFB U+00FB #LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC U+00FC #LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+00FD #LATIN SMALL LETTER Y WITH ACUTE +0xFE U+00FE #LATIN SMALL LETTER THORN +0xFF U+00FF #LATIN SMALL LETTER Y WITH DIAERESIS + +U+2218 " \260 " # RING OPERATOR +U+2219 " \225 " # BULLET OPERATOR +U+221b " ROOT\263 " +U+2297 "(\327)" # CIRCLED TIMES +U+2299 "(\267)" # CIRCLED DOT OPERATOR +U+229A "(\260)" # CIRCLED RING OPERATOR +U+22A0 "[\327]" # SQUARED TIMES +U+22A1 "[\267]" # SQUARED DOT OPERATOR +U+22C5 " \267 " # DOT OPERATOR diff --git a/src/chrtrans/cp1253_uni.tbl b/src/chrtrans/cp1253_uni.tbl new file mode 100644 index 0000000..49523d4 --- /dev/null +++ b/src/chrtrans/cp1253_uni.tbl @@ -0,0 +1,161 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mwindows-1253 + +#Name as a Display Charset (used on Options screen) +OGreek (windows-1253) + +#Codepage number +C1253 + +# +# Name: cp1253 to Unicode table +# Unicode version: 2.0 +# Table version: 2.01 +# Table format: Format A +# Date: 04/15/98 +# +# Contact: cpxlate@microsoft.com +# +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp1253 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp1253 order +# +################## + +0x20-0x7e idem +# +0x80 U+20AC #EURO SIGN +0x81 #UNDEFINED +0x82 U+201A #SINGLE LOW-9 QUOTATION MARK +0x83 U+0192 #LATIN SMALL LETTER F WITH HOOK +0x84 U+201E #DOUBLE LOW-9 QUOTATION MARK +0x85 U+2026 #HORIZONTAL ELLIPSIS +0x86 U+2020 #DAGGER +0x87 U+2021 #DOUBLE DAGGER +0x88 #UNDEFINED +0x89 U+2030 #PER MILLE SIGN +0x8A #UNDEFINED +0x8B U+2039 #SINGLE LEFT-POINTING ANGLE QUOTATION MARK +0x8C #UNDEFINED +0x8D #UNDEFINED +0x8E #UNDEFINED +0x8F #UNDEFINED +0x90 #UNDEFINED +0x91 U+2018 U+02bd #LEFT SINGLE QUOTATION MARK +0x92 U+2019 U+02bc #RIGHT SINGLE QUOTATION MARK +0x93 U+201C #LEFT DOUBLE QUOTATION MARK +0x94 U+201D #RIGHT DOUBLE QUOTATION MARK +0x95 U+2022 #BULLET +0x96 U+2013 #EN DASH +0x97 U+2014 #EM DASH +0x98 #UNDEFINED +0x99 U+2122 #TRADE MARK SIGN +0x9A #UNDEFINED +0x9B U+203A #SINGLE RIGHT-POINTING ANGLE QUOTATION MARK +0x9C #UNDEFINED +0x9D #UNDEFINED +0x9E #UNDEFINED +0x9F #UNDEFINED +0xA0 U+00A0 #NO-BREAK SPACE +0xA1 U+0385 #GREEK DIALYTIKA TONOS +0xA2 U+0386 #GREEK CAPITAL LETTER ALPHA WITH TONOS +0xA3 U+00A3 #POUND SIGN +0xA4 U+00A4 #CURRENCY SIGN +0xA5 U+00A5 #YEN SIGN +0xA6 U+00A6 #BROKEN BAR +0xA7 U+00A7 #SECTION SIGN +0xA8 U+00A8 #DIAERESIS +0xA9 U+00A9 #COPYRIGHT SIGN +0xAA #UNDEFINED +0xAB U+00AB #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC #NOT SIGN +0xAD U+00AD #SOFT HYPHEN +0xAE U+00AE #REGISTERED SIGN +0xAF U+2015 #HORIZONTAL BAR +0xB0 U+00B0 #DEGREE SIGN +0xB1 U+00B1 #PLUS-MINUS SIGN +0xB2 U+00B2 #SUPERSCRIPT TWO +0xB3 U+00B3 #SUPERSCRIPT THREE +0xB4 U+0384 #GREEK TONOS +0xB5 U+00B5 #MICRO SIGN +0xB6 U+00B6 #PILCROW SIGN +0xB7 U+00B7 #MIDDLE DOT +0xB8 U+0388 #GREEK CAPITAL LETTER EPSILON WITH TONOS +0xB9 U+0389 #GREEK CAPITAL LETTER ETA WITH TONOS +0xBA U+038A #GREEK CAPITAL LETTER IOTA WITH TONOS +0xBB U+00BB #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+038C #GREEK CAPITAL LETTER OMICRON WITH TONOS +0xBD U+00BD #VULGAR FRACTION ONE HALF +0xBE U+038E #GREEK CAPITAL LETTER UPSILON WITH TONOS +0xBF U+038F #GREEK CAPITAL LETTER OMEGA WITH TONOS +0xC0 U+0390 #GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS +0xC1 U+0391 #GREEK CAPITAL LETTER ALPHA +0xC2 U+0392 #GREEK CAPITAL LETTER BETA +0xC3 U+0393 #GREEK CAPITAL LETTER GAMMA +0xC4 U+0394 #GREEK CAPITAL LETTER DELTA +0xC5 U+0395 #GREEK CAPITAL LETTER EPSILON +0xC6 U+0396 #GREEK CAPITAL LETTER ZETA +0xC7 U+0397 #GREEK CAPITAL LETTER ETA +0xC8 U+0398 #GREEK CAPITAL LETTER THETA +0xC9 U+0399 #GREEK CAPITAL LETTER IOTA +0xCA U+039A #GREEK CAPITAL LETTER KAPPA +0xCB U+039B #GREEK CAPITAL LETTER LAMDA +0xCC U+039C #GREEK CAPITAL LETTER MU +0xCD U+039D #GREEK CAPITAL LETTER NU +0xCE U+039E #GREEK CAPITAL LETTER XI +0xCF U+039F #GREEK CAPITAL LETTER OMICRON +0xD0 U+03A0 #GREEK CAPITAL LETTER PI +0xD1 U+03A1 #GREEK CAPITAL LETTER RHO +0xD2 #UNDEFINED +0xD3 U+03A3 #GREEK CAPITAL LETTER SIGMA +0xD4 U+03A4 #GREEK CAPITAL LETTER TAU +0xD5 U+03A5 #GREEK CAPITAL LETTER UPSILON +0xD6 U+03A6 #GREEK CAPITAL LETTER PHI +0xD7 U+03A7 #GREEK CAPITAL LETTER CHI +0xD8 U+03A8 #GREEK CAPITAL LETTER PSI +0xD9 U+03A9 #GREEK CAPITAL LETTER OMEGA +0xDA U+03AA #GREEK CAPITAL LETTER IOTA WITH DIALYTIKA +0xDB U+03AB #GREEK CAPITAL LETTER UPSILON WITH DIALYTIKA +0xDC U+03AC #GREEK SMALL LETTER ALPHA WITH TONOS +0xDD U+03AD #GREEK SMALL LETTER EPSILON WITH TONOS +0xDE U+03AE #GREEK SMALL LETTER ETA WITH TONOS +0xDF U+03AF #GREEK SMALL LETTER IOTA WITH TONOS +0xE0 U+03B0 #GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS +0xE1 U+03B1 #GREEK SMALL LETTER ALPHA +0xE2 U+03B2 #GREEK SMALL LETTER BETA +0xE3 U+03B3 #GREEK SMALL LETTER GAMMA +0xE4 U+03B4 #GREEK SMALL LETTER DELTA +0xE5 U+03B5 #GREEK SMALL LETTER EPSILON +0xE6 U+03B6 #GREEK SMALL LETTER ZETA +0xE7 U+03B7 #GREEK SMALL LETTER ETA +0xE8 U+03B8 #GREEK SMALL LETTER THETA +0xE9 U+03B9 #GREEK SMALL LETTER IOTA +0xEA U+03BA #GREEK SMALL LETTER KAPPA +0xEB U+03BB #GREEK SMALL LETTER LAMDA +0xEC U+03BC #GREEK SMALL LETTER MU +0xED U+03BD #GREEK SMALL LETTER NU +0xEE U+03BE #GREEK SMALL LETTER XI +0xEF U+03BF #GREEK SMALL LETTER OMICRON +0xF0 U+03C0 #GREEK SMALL LETTER PI +0xF1 U+03C1 #GREEK SMALL LETTER RHO +0xF2 U+03C2 #GREEK SMALL LETTER FINAL SIGMA +0xF3 U+03C3 #GREEK SMALL LETTER SIGMA +0xF4 U+03C4 #GREEK SMALL LETTER TAU +0xF5 U+03C5 #GREEK SMALL LETTER UPSILON +0xF6 U+03C6 #GREEK SMALL LETTER PHI +0xF7 U+03C7 #GREEK SMALL LETTER CHI +0xF8 U+03C8 #GREEK SMALL LETTER PSI +0xF9 U+03C9 #GREEK SMALL LETTER OMEGA +0xFA U+03CA #GREEK SMALL LETTER IOTA WITH DIALYTIKA +0xFB U+03CB #GREEK SMALL LETTER UPSILON WITH DIALYTIKA +0xFC U+03CC #GREEK SMALL LETTER OMICRON WITH TONOS +0xFD U+03CD #GREEK SMALL LETTER UPSILON WITH TONOS +0xFE U+03CE #GREEK SMALL LETTER OMEGA WITH TONOS +0xFF #UNDEFINED diff --git a/src/chrtrans/cp1255_uni.tbl b/src/chrtrans/cp1255_uni.tbl new file mode 100644 index 0000000..3f0af9e --- /dev/null +++ b/src/chrtrans/cp1255_uni.tbl @@ -0,0 +1,161 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mwindows-1255 + +#Name as a Display Charset (used on Options screen). +OHebrew (windows-1255) + +#Codepage number +C1255 + +# +# Name: cp1255 to Unicode table +# Unicode version: 2.0 +# Table version: 2.01 +# Table format: Format A +# Date: 04/15/98 +# +# Contact: cpxlate@microsoft.com +# +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp1255 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp1255 order +# +################## + +0x20-0x7e idem +# +0x80 U+20AC #EURO SIGN +0x81 #UNDEFINED +0x82 U+201A #SINGLE LOW-9 QUOTATION MARK +0x83 U+0192 #LATIN SMALL LETTER F WITH HOOK +0x84 U+201E #DOUBLE LOW-9 QUOTATION MARK +0x85 U+2026 #HORIZONTAL ELLIPSIS +0x86 U+2020 #DAGGER +0x87 U+2021 #DOUBLE DAGGER +0x88 U+02C6 #MODIFIER LETTER CIRCUMFLEX ACCENT +0x89 U+2030 #PER MILLE SIGN +0x8A #UNDEFINED +0x8B U+2039 #SINGLE LEFT-POINTING ANGLE QUOTATION MARK +0x8C #UNDEFINED +0x8D #UNDEFINED +0x8E #UNDEFINED +0x8F #UNDEFINED +0x90 #UNDEFINED +0x91 U+2018 #LEFT SINGLE QUOTATION MARK +0x92 U+2019 #RIGHT SINGLE QUOTATION MARK +0x93 U+201C #LEFT DOUBLE QUOTATION MARK +0x94 U+201D #RIGHT DOUBLE QUOTATION MARK +0x95 U+2022 #BULLET +0x96 U+2013 #EN DASH +0x97 U+2014 #EM DASH +0x98 U+02DC #SMALL TILDE +0x99 U+2122 #TRADE MARK SIGN +0x9A #UNDEFINED +0x9B U+203A #SINGLE RIGHT-POINTING ANGLE QUOTATION MARK +0x9C #UNDEFINED +0x9D #UNDEFINED +0x9E #UNDEFINED +0x9F #UNDEFINED +0xA0 U+00A0 #NO-BREAK SPACE +0xA1 U+00A1 #INVERTED EXCLAMATION MARK +0xA2 U+00A2 #CENT SIGN +0xA3 U+00A3 #POUND SIGN +0xA4 U+20AA #NEW SHEQEL SIGN +0xA5 U+00A5 #YEN SIGN +0xA6 U+00A6 #BROKEN BAR +0xA7 U+00A7 #SECTION SIGN +0xA8 U+00A8 #DIAERESIS +0xA9 U+00A9 #COPYRIGHT SIGN +0xAA U+00D7 #MULTIPLICATION SIGN +0xAB U+00AB #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC #NOT SIGN +0xAD U+00AD #SOFT HYPHEN +0xAE U+00AE #REGISTERED SIGN +0xAF U+00AF #MACRON +0xB0 U+00B0 #DEGREE SIGN +0xB1 U+00B1 #PLUS-MINUS SIGN +0xB2 U+00B2 #SUPERSCRIPT TWO +0xB3 U+00B3 #SUPERSCRIPT THREE +0xB4 U+00B4 #ACUTE ACCENT +0xB5 U+00B5 #MICRO SIGN +0xB6 U+00B6 #PILCROW SIGN +0xB7 U+00B7 #MIDDLE DOT +0xB8 U+00B8 #CEDILLA +0xB9 U+00B9 #SUPERSCRIPT ONE +0xBA U+00F7 #DIVISION SIGN +0xBB U+00BB #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+00BC #VULGAR FRACTION ONE QUARTER +0xBD U+00BD #VULGAR FRACTION ONE HALF +0xBE U+00BE #VULGAR FRACTION THREE QUARTERS +0xBF U+00BF #INVERTED QUESTION MARK +0xC0 U+05B0 #HEBREW POINT SHEVA +0xC1 U+05B1 #HEBREW POINT HATAF SEGOL +0xC2 U+05B2 #HEBREW POINT HATAF PATAH +0xC3 U+05B3 #HEBREW POINT HATAF QAMATS +0xC4 U+05B4 #HEBREW POINT HIRIQ +0xC5 U+05B5 #HEBREW POINT TSERE +0xC6 U+05B6 #HEBREW POINT SEGOL +0xC7 U+05B7 #HEBREW POINT PATAH +0xC8 U+05B8 #HEBREW POINT QAMATS +0xC9 U+05B9 #HEBREW POINT HOLAM +0xCA #UNDEFINED +0xCB U+05BB #HEBREW POINT QUBUTS +0xCC U+05BC #HEBREW POINT DAGESH OR MAPIQ +0xCD U+05BD #HEBREW POINT METEG +0xCE U+05BE #HEBREW PUNCTUATION MAQAF +0xCF U+05BF #HEBREW POINT RAFE +0xD0 U+05C0 #HEBREW PUNCTUATION PASEQ +0xD1 U+05C1 #HEBREW POINT SHIN DOT +0xD2 U+05C2 #HEBREW POINT SIN DOT +0xD3 U+05C3 #HEBREW PUNCTUATION SOF PASUQ +0xD4 U+05F0 #HEBREW LIGATURE YIDDISH DOUBLE VAV +0xD5 U+05F1 #HEBREW LIGATURE YIDDISH VAV YOD +0xD6 U+05F2 #HEBREW LIGATURE YIDDISH DOUBLE YOD +0xD7 U+05F3 #HEBREW PUNCTUATION GERESH +0xD8 U+05F4 #HEBREW PUNCTUATION GERSHAYIM +0xD9 #UNDEFINED +0xDA #UNDEFINED +0xDB #UNDEFINED +0xDC #UNDEFINED +0xDD #UNDEFINED +0xDE #UNDEFINED +0xDF #UNDEFINED +0xE0 U+05D0 #HEBREW LETTER ALEF +0xE1 U+05D1 #HEBREW LETTER BET +0xE2 U+05D2 #HEBREW LETTER GIMEL +0xE3 U+05D3 #HEBREW LETTER DALET +0xE4 U+05D4 #HEBREW LETTER HE +0xE5 U+05D5 #HEBREW LETTER VAV +0xE6 U+05D6 #HEBREW LETTER ZAYIN +0xE7 U+05D7 #HEBREW LETTER HET +0xE8 U+05D8 #HEBREW LETTER TET +0xE9 U+05D9 #HEBREW LETTER YOD +0xEA U+05DA #HEBREW LETTER FINAL KAF +0xEB U+05DB #HEBREW LETTER KAF +0xEC U+05DC #HEBREW LETTER LAMED +0xED U+05DD #HEBREW LETTER FINAL MEM +0xEE U+05DE #HEBREW LETTER MEM +0xEF U+05DF #HEBREW LETTER FINAL NUN +0xF0 U+05E0 #HEBREW LETTER NUN +0xF1 U+05E1 #HEBREW LETTER SAMEKH +0xF2 U+05E2 #HEBREW LETTER AYIN +0xF3 U+05E3 #HEBREW LETTER FINAL PE +0xF4 U+05E4 #HEBREW LETTER PE +0xF5 U+05E5 #HEBREW LETTER FINAL TSADI +0xF6 U+05E6 #HEBREW LETTER TSADI +0xF7 U+05E7 #HEBREW LETTER QOF +0xF8 U+05E8 #HEBREW LETTER RESH +0xF9 U+05E9 #HEBREW LETTER SHIN +0xFA U+05EA #HEBREW LETTER TAV +0xFB #UNDEFINED +0xFC #UNDEFINED +0xFD U+200E #LEFT-TO-RIGHT MARK +0xFE U+200F #RIGHT-TO-LEFT MARK +0xFF #UNDEFINED diff --git a/src/chrtrans/cp1256_uni.tbl b/src/chrtrans/cp1256_uni.tbl new file mode 100644 index 0000000..1ab99c0 --- /dev/null +++ b/src/chrtrans/cp1256_uni.tbl @@ -0,0 +1,161 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mwindows-1256 + +#Name as a Display Charset (used on Options screen). +OArabic (windows-1256) + +#Codepage number +C1256 + +# +# Name: cp1256 to Unicode table +# Unicode version: 2.1 +# Table version: 2.01 +# Table format: Format A +# Date: 01/5/99 +# +# Contact: cpxlate@microsoft.com +# +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp1256 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp1256 order +# +################## + +0x20-0x7e idem +# +0x80 U+20AC #EURO SIGN +0x81 U+067E #ARABIC LETTER PEH +0x82 U+201A #SINGLE LOW-9 QUOTATION MARK +0x83 U+0192 #LATIN SMALL LETTER F WITH HOOK +0x84 U+201E #DOUBLE LOW-9 QUOTATION MARK +0x85 U+2026 #HORIZONTAL ELLIPSIS +0x86 U+2020 #DAGGER +0x87 U+2021 #DOUBLE DAGGER +0x88 U+02C6 #MODIFIER LETTER CIRCUMFLEX ACCENT +0x89 U+2030 #PER MILLE SIGN +0x8A U+0679 #ARABIC LETTER TTEH +0x8B U+2039 #SINGLE LEFT-POINTING ANGLE QUOTATION MARK +0x8C U+0152 #LATIN CAPITAL LIGATURE OE +0x8D U+0686 #ARABIC LETTER TCHEH +0x8E U+0698 #ARABIC LETTER JEH +0x8F U+0688 #ARABIC LETTER DDAL +0x90 U+06AF #ARABIC LETTER GAF +0x91 U+2018 #LEFT SINGLE QUOTATION MARK +0x92 U+2019 #RIGHT SINGLE QUOTATION MARK +0x93 U+201C #LEFT DOUBLE QUOTATION MARK +0x94 U+201D #RIGHT DOUBLE QUOTATION MARK +0x95 U+2022 #BULLET +0x96 U+2013 #EN DASH +0x97 U+2014 #EM DASH +0x98 U+06A9 #ARABIC LETTER KEHEH +0x99 U+2122 #TRADE MARK SIGN +0x9A U+0691 #ARABIC LETTER RREH +0x9B U+203A #SINGLE RIGHT-POINTING ANGLE QUOTATION MARK +0x9C U+0153 #LATIN SMALL LIGATURE OE +0x9D U+200C #ZERO WIDTH NON-JOINER +0x9E U+200D #ZERO WIDTH JOINER +0x9F U+06BA #ARABIC LETTER NOON GHUNNA +0xA0 U+00A0 #NO-BREAK SPACE +0xA1 U+060C #ARABIC COMMA +0xA2 U+00A2 #CENT SIGN +0xA3 U+00A3 #POUND SIGN +0xA4 U+00A4 #CURRENCY SIGN +0xA5 U+00A5 #YEN SIGN +0xA6 U+00A6 #BROKEN BAR +0xA7 U+00A7 #SECTION SIGN +0xA8 U+00A8 #DIAERESIS +0xA9 U+00A9 #COPYRIGHT SIGN +0xAA U+06BE #ARABIC LETTER HEH DOACHASHMEE +0xAB U+00AB #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC #NOT SIGN +0xAD U+00AD #SOFT HYPHEN +0xAE U+00AE #REGISTERED SIGN +0xAF U+00AF #MACRON +0xB0 U+00B0 #DEGREE SIGN +0xB1 U+00B1 #PLUS-MINUS SIGN +0xB2 U+00B2 #SUPERSCRIPT TWO +0xB3 U+00B3 #SUPERSCRIPT THREE +0xB4 U+00B4 #ACUTE ACCENT +0xB5 U+00B5 #MICRO SIGN +0xB6 U+00B6 #PILCROW SIGN +0xB7 U+00B7 #MIDDLE DOT +0xB8 U+00B8 #CEDILLA +0xB9 U+00B9 #SUPERSCRIPT ONE +0xBA U+061B #ARABIC SEMICOLON +0xBB U+00BB #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+00BC #VULGAR FRACTION ONE QUARTER +0xBD U+00BD #VULGAR FRACTION ONE HALF +0xBE U+00BE #VULGAR FRACTION THREE QUARTERS +0xBF U+061F #ARABIC QUESTION MARK +0xC0 U+06C1 #ARABIC LETTER HEH GOAL +0xC1 U+0621 #ARABIC LETTER HAMZA +0xC2 U+0622 #ARABIC LETTER ALEF WITH MADDA ABOVE +0xC3 U+0623 #ARABIC LETTER ALEF WITH HAMZA ABOVE +0xC4 U+0624 #ARABIC LETTER WAW WITH HAMZA ABOVE +0xC5 U+0625 #ARABIC LETTER ALEF WITH HAMZA BELOW +0xC6 U+0626 #ARABIC LETTER YEH WITH HAMZA ABOVE +0xC7 U+0627 #ARABIC LETTER ALEF +0xC8 U+0628 #ARABIC LETTER BEH +0xC9 U+0629 #ARABIC LETTER TEH MARBUTA +0xCA U+062A #ARABIC LETTER TEH +0xCB U+062B #ARABIC LETTER THEH +0xCC U+062C #ARABIC LETTER JEEM +0xCD U+062D #ARABIC LETTER HAH +0xCE U+062E #ARABIC LETTER KHAH +0xCF U+062F #ARABIC LETTER DAL +0xD0 U+0630 #ARABIC LETTER THAL +0xD1 U+0631 #ARABIC LETTER REH +0xD2 U+0632 #ARABIC LETTER ZAIN +0xD3 U+0633 #ARABIC LETTER SEEN +0xD4 U+0634 #ARABIC LETTER SHEEN +0xD5 U+0635 #ARABIC LETTER SAD +0xD6 U+0636 #ARABIC LETTER DAD +0xD7 U+00D7 #MULTIPLICATION SIGN +0xD8 U+0637 #ARABIC LETTER TAH +0xD9 U+0638 #ARABIC LETTER ZAH +0xDA U+0639 #ARABIC LETTER AIN +0xDB U+063A #ARABIC LETTER GHAIN +0xDC U+0640 #ARABIC TATWEEL +0xDD U+0641 #ARABIC LETTER FEH +0xDE U+0642 #ARABIC LETTER QAF +0xDF U+0643 #ARABIC LETTER KAF +0xE0 U+00E0 #LATIN SMALL LETTER A WITH GRAVE +0xE1 U+0644 #ARABIC LETTER LAM +0xE2 U+00E2 #LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 U+0645 #ARABIC LETTER MEEM +0xE4 U+0646 #ARABIC LETTER NOON +0xE5 U+0647 #ARABIC LETTER HEH +0xE6 U+0648 #ARABIC LETTER WAW +0xE7 U+00E7 #LATIN SMALL LETTER C WITH CEDILLA +0xE8 U+00E8 #LATIN SMALL LETTER E WITH GRAVE +0xE9 U+00E9 #LATIN SMALL LETTER E WITH ACUTE +0xEA U+00EA #LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB U+00EB #LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+0649 #ARABIC LETTER ALEF MAKSURA +0xED U+064A #ARABIC LETTER YEH +0xEE U+00EE #LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF U+00EF #LATIN SMALL LETTER I WITH DIAERESIS +0xF0 U+064B #ARABIC FATHATAN +0xF1 U+064C #ARABIC DAMMATAN +0xF2 U+064D #ARABIC KASRATAN +0xF3 U+064E #ARABIC FATHA +0xF4 U+00F4 #LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 U+064F #ARABIC DAMMA +0xF6 U+0650 #ARABIC KASRA +0xF7 U+00F7 #DIVISION SIGN +0xF8 U+0651 #ARABIC SHADDA +0xF9 U+00F9 #LATIN SMALL LETTER U WITH GRAVE +0xFA U+0652 #ARABIC SUKUN +0xFB U+00FB #LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC U+00FC #LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+200E #LEFT-TO-RIGHT MARK +0xFE U+200F #RIGHT-TO-LEFT MARK +0xFF U+06D2 #ARABIC LETTER YEH BARREE diff --git a/src/chrtrans/cp1257_uni.tbl b/src/chrtrans/cp1257_uni.tbl new file mode 100644 index 0000000..6cd0081 --- /dev/null +++ b/src/chrtrans/cp1257_uni.tbl @@ -0,0 +1,162 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mwindows-1257 + +#Name as a Display Charset (used on Options screen) +OBaltic Rim (windows-1257) + +#Codepage number +C1257 + +# +# Name: cp1257 to Unicode table +# Unicode version: 2.0 +# Table version: 2.01 +# Table format: Format A +# Date: 04/15/98 +# +# Contact: cpxlate@microsoft.com +# +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp1257 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp1257 order +# +################## + +0x20-0x7e idem +# +0x80 U+20AC #EURO SIGN +0x81 #UNDEFINED +0x82 U+201A #SINGLE LOW-9 QUOTATION MARK +0x83 #UNDEFINED +0x84 U+201E #DOUBLE LOW-9 QUOTATION MARK +0x85 U+2026 #HORIZONTAL ELLIPSIS +0x86 U+2020 #DAGGER +0x87 U+2021 #DOUBLE DAGGER +0x88 #UNDEFINED +0x89 U+2030 #PER MILLE SIGN +0x8A #UNDEFINED +0x8B U+2039 #SINGLE LEFT-POINTING ANGLE QUOTATION MARK +0x8C #UNDEFINED +0x8D U+00A8 #DIAERESIS +0x8E U+02C7 #CARON +0x8F U+00B8 #CEDILLA +0x90 #UNDEFINED +0x91 U+2018 #LEFT SINGLE QUOTATION MARK +0x92 U+2019 #RIGHT SINGLE QUOTATION MARK +0x93 U+201C #LEFT DOUBLE QUOTATION MARK +0x94 U+201D #RIGHT DOUBLE QUOTATION MARK +0x95 U+2022 #BULLET +0x96 U+2013 #EN DASH +0x97 U+2014 #EM DASH +0x98 #UNDEFINED +0x99 U+2122 #TRADE MARK SIGN +0x9A #UNDEFINED +0x9B U+203A #SINGLE RIGHT-POINTING ANGLE QUOTATION MARK +0x9C #UNDEFINED +0x9D U+00AF #MACRON +0x9E U+02DB #OGONEK +0x9F #UNDEFINED +0xA0 U+00A0 #NO-BREAK SPACE +0xA1 #UNDEFINED +0xA2 U+00A2 #CENT SIGN +0xA3 U+00A3 #POUND SIGN +0xA4 U+00A4 #CURRENCY SIGN +0xA5 #UNDEFINED +0xA6 U+00A6 #BROKEN BAR +0xA7 U+00A7 #SECTION SIGN +0xA8 U+00D8 #LATIN CAPITAL LETTER O WITH STROKE +0xA9 U+00A9 #COPYRIGHT SIGN +0xAA U+0156 #LATIN CAPITAL LETTER R WITH CEDILLA +0xAB U+00AB #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC #NOT SIGN +0xAD U+00AD #SOFT HYPHEN +0xAE U+00AE #REGISTERED SIGN +0xAF U+00C6 #LATIN CAPITAL LETTER AE +0xB0 U+00B0 #DEGREE SIGN +0xB1 U+00B1 #PLUS-MINUS SIGN +0xB2 U+00B2 #SUPERSCRIPT TWO +0xB3 U+00B3 #SUPERSCRIPT THREE +0xB4 U+00B4 #ACUTE ACCENT +0xB5 U+00B5 #MICRO SIGN +0xB6 U+00B6 #PILCROW SIGN +0xB7 U+00B7 #MIDDLE DOT +0xB8 U+00F8 #LATIN SMALL LETTER O WITH STROKE +0xB9 U+00B9 #SUPERSCRIPT ONE +0xBA U+0157 #LATIN SMALL LETTER R WITH CEDILLA +0xBB U+00BB #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+00BC #VULGAR FRACTION ONE QUARTER +0xBD U+00BD #VULGAR FRACTION ONE HALF +0xBE U+00BE #VULGAR FRACTION THREE QUARTERS +0xBF U+00E6 #LATIN SMALL LETTER AE +0xC0 U+0104 #LATIN CAPITAL LETTER A WITH OGONEK +0xC1 U+012E #LATIN CAPITAL LETTER I WITH OGONEK +0xC2 U+0100 #LATIN CAPITAL LETTER A WITH MACRON +0xC3 U+0106 #LATIN CAPITAL LETTER C WITH ACUTE +0xC4 U+00C4 #LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+00C5 #LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 U+0118 #LATIN CAPITAL LETTER E WITH OGONEK +0xC7 U+0112 #LATIN CAPITAL LETTER E WITH MACRON +0xC8 U+010C #LATIN CAPITAL LETTER C WITH CARON +0xC9 U+00C9 #LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+0179 #LATIN CAPITAL LETTER Z WITH ACUTE +0xCB U+0116 #LATIN CAPITAL LETTER E WITH DOT ABOVE +0xCC U+0122 #LATIN CAPITAL LETTER G WITH CEDILLA +0xCD U+0136 #LATIN CAPITAL LETTER K WITH CEDILLA +0xCE U+012A #LATIN CAPITAL LETTER I WITH MACRON +0xCF U+013B #LATIN CAPITAL LETTER L WITH CEDILLA +0xD0 U+0160 #LATIN CAPITAL LETTER S WITH CARON +0xD1 U+0143 #LATIN CAPITAL LETTER N WITH ACUTE +0xD2 U+0145 #LATIN CAPITAL LETTER N WITH CEDILLA +0xD3 U+00D3 #LATIN CAPITAL LETTER O WITH ACUTE +0xD4 U+014C #LATIN CAPITAL LETTER O WITH MACRON +0xD5 U+00D5 #LATIN CAPITAL LETTER O WITH TILDE +0xD6 U+00D6 #LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+00D7 #MULTIPLICATION SIGN +0xD8 U+0172 #LATIN CAPITAL LETTER U WITH OGONEK +0xD9 U+0141 #LATIN CAPITAL LETTER L WITH STROKE +0xDA U+015A #LATIN CAPITAL LETTER S WITH ACUTE +0xDB U+016A #LATIN CAPITAL LETTER U WITH MACRON +0xDC U+00DC #LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+017B #LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xDE U+017D #LATIN CAPITAL LETTER Z WITH CARON +0xDF U+00DF #LATIN SMALL LETTER SHARP S +0xE0 U+0105 #LATIN SMALL LETTER A WITH OGONEK +0xE1 U+012F #LATIN SMALL LETTER I WITH OGONEK +0xE2 U+0101 #LATIN SMALL LETTER A WITH MACRON +0xE3 U+0107 #LATIN SMALL LETTER C WITH ACUTE +0xE4 U+00E4 #LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+00E5 #LATIN SMALL LETTER A WITH RING ABOVE +0xE6 U+0119 #LATIN SMALL LETTER E WITH OGONEK +0xE7 U+0113 #LATIN SMALL LETTER E WITH MACRON +0xE8 U+010D #LATIN SMALL LETTER C WITH CARON +0xE9 U+00E9 #LATIN SMALL LETTER E WITH ACUTE +0xEA U+017A #LATIN SMALL LETTER Z WITH ACUTE +0xEB U+0117 #LATIN SMALL LETTER E WITH DOT ABOVE +0xEC U+0123 #LATIN SMALL LETTER G WITH CEDILLA +0xED U+0137 #LATIN SMALL LETTER K WITH CEDILLA +0xEE U+012B #LATIN SMALL LETTER I WITH MACRON +0xEF U+013C #LATIN SMALL LETTER L WITH CEDILLA +0xF0 U+0161 #LATIN SMALL LETTER S WITH CARON +0xF1 U+0144 #LATIN SMALL LETTER N WITH ACUTE +0xF2 U+0146 #LATIN SMALL LETTER N WITH CEDILLA +0xF3 U+00F3 #LATIN SMALL LETTER O WITH ACUTE +0xF4 U+014D #LATIN SMALL LETTER O WITH MACRON +0xF5 U+00F5 #LATIN SMALL LETTER O WITH TILDE +0xF6 U+00F6 #LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+00F7 #DIVISION SIGN +0xF8 U+0173 #LATIN SMALL LETTER U WITH OGONEK +0xF9 U+0142 #LATIN SMALL LETTER L WITH STROKE +0xFA U+015B #LATIN SMALL LETTER S WITH ACUTE +0xFB U+016B #LATIN SMALL LETTER U WITH MACRON +0xFC U+00FC #LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+017C #LATIN SMALL LETTER Z WITH DOT ABOVE +0xFE U+017E #LATIN SMALL LETTER Z WITH CARON +0xFF U+02D9 #DOT ABOVE + diff --git a/src/chrtrans/cp437_uni.tbl b/src/chrtrans/cp437_uni.tbl new file mode 100644 index 0000000..4f45ce7 --- /dev/null +++ b/src/chrtrans/cp437_uni.tbl @@ -0,0 +1,181 @@ +# This file has been modified for lynx (see README.tables) + +#Shall this become the "default" translation? +#There has to be exactly one table marked as "default". +D0 +# +#The MIME name of this charset. +Mcp437 + +#Name as a Display Charset (used on Options screen) +OIBM PC US codepage (cp437) + +#Codepage number +C437 + +# +# Name: cp437_DOSLatinUS to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp437_DOSLatinUS code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp437_DosLatinUS order +# +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw +# +####################################### + +0x20-0x7f idem +# +0x80 U+00c7 #LATIN CAPITAL LETTER C WITH CEDILLA +0x81 U+00fc U+03cb #LATIN SMALL LETTER U WITH DIAERESIS +0x82 U+00e9 #LATIN SMALL LETTER E WITH ACUTE +0x83 U+00e2 #LATIN SMALL LETTER A WITH CIRCUMFLEX +0x84 U+00e4 #LATIN SMALL LETTER A WITH DIAERESIS +0x85 U+00e0 #LATIN SMALL LETTER A WITH GRAVE +0x86 U+00e5 #LATIN SMALL LETTER A WITH RING ABOVE +0x87 U+00e7 #LATIN SMALL LETTER C WITH CEDILLA +0x88 U+00ea #LATIN SMALL LETTER E WITH CIRCUMFLEX +0x89 U+00eb #LATIN SMALL LETTER E WITH DIAERESIS +0x8a U+00e8 #LATIN SMALL LETTER E WITH GRAVE +0x8b U+00ef U+03ca #LATIN SMALL LETTER I WITH DIAERESIS +0x8c U+00ee #LATIN SMALL LETTER I WITH CIRCUMFLEX +0x8d U+00ec #LATIN SMALL LETTER I WITH GRAVE +0x8e U+00c4 #LATIN CAPITAL LETTER A WITH DIAERESIS +0x8f U+00c5 #LATIN CAPITAL LETTER A WITH RING ABOVE +0x90 U+00c9 U+0388 #LATIN CAPITAL LETTER E WITH ACUTE +0x91 U+00e6 #LATIN SMALL LIGATURE AE +0x92 U+00c6 #LATIN CAPITAL LIGATURE AE +0x93 U+00f4 #LATIN SMALL LETTER O WITH CIRCUMFLEX +0x94 U+00f6 #LATIN SMALL LETTER O WITH DIAERESIS +0x95 U+00f2 #LATIN SMALL LETTER O WITH GRAVE +0x96 U+00fb #LATIN SMALL LETTER U WITH CIRCUMFLEX +0x97 U+00f9 #LATIN SMALL LETTER U WITH GRAVE +0x98 U+00ff #LATIN SMALL LETTER Y WITH DIAERESIS +0x99 U+00d6 #LATIN CAPITAL LETTER O WITH DIAERESIS +0x9a U+00dc U+03ab #LATIN CAPITAL LETTER U WITH DIAERESIS +0x9b U+00a2 #CENT SIGN +0x9c U+00a3 #POUND SIGN +0x9d U+00a5 #YEN SIGN +0x9e U+20a7 #PESETA SIGN +0x9f U+0192 #LATIN SMALL LETTER F WITH HOOK +0xa0 U+00e1 U+03ac #LATIN SMALL LETTER A WITH ACUTE +0xa1 U+00ed U+03af #LATIN SMALL LETTER I WITH ACUTE +0xa2 U+00f3 U+03cc #LATIN SMALL LETTER O WITH ACUTE +0xa3 U+00fa U+03cd #LATIN SMALL LETTER U WITH ACUTE +0xa4 U+00f1 #LATIN SMALL LETTER N WITH TILDE +0xa5 U+00d1 #LATIN CAPITAL LETTER N WITH TILDE +0xa6 U+00aa #FEMININE ORDINAL INDICATOR +0xa7 U+00ba #MASCULINE ORDINAL INDICATOR +0xa8 U+00bf #INVERTED QUESTION MARK +0xa9 U+2310 #REVERSED NOT SIGN +0xaa U+00ac #NOT SIGN +0xab U+00bd #VULGAR FRACTION ONE HALF +0xac U+00bc #VULGAR FRACTION ONE QUARTER +0xad U+00a1 #INVERTED EXCLAMATION MARK +0xae U+00ab #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xaf U+00bb #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+2561 #BOX DRAWINGS VERTICAL SINGLE AND LEFT DOUBLE +0xb6 U+2562 #BOX DRAWINGS VERTICAL DOUBLE AND LEFT SINGLE +0xb7 U+2556 #BOX DRAWINGS DOWN DOUBLE AND LEFT SINGLE +0xb8 U+2555 #BOX DRAWINGS DOWN SINGLE AND LEFT DOUBLE +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+255c #BOX DRAWINGS UP DOUBLE AND LEFT SINGLE +0xbe U+255b #BOX DRAWINGS UP SINGLE AND LEFT DOUBLE +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+255e #BOX DRAWINGS VERTICAL SINGLE AND RIGHT DOUBLE +0xc7 U+255f #BOX DRAWINGS VERTICAL DOUBLE AND RIGHT SINGLE +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+2567 #BOX DRAWINGS UP SINGLE AND HORIZONTAL DOUBLE +0xd0 U+2568 #BOX DRAWINGS UP DOUBLE AND HORIZONTAL SINGLE +0xd1 U+2564 #BOX DRAWINGS DOWN SINGLE AND HORIZONTAL DOUBLE +0xd2 U+2565 #BOX DRAWINGS DOWN DOUBLE AND HORIZONTAL SINGLE +0xd3 U+2559 #BOX DRAWINGS UP DOUBLE AND RIGHT SINGLE +0xd4 U+2558 #BOX DRAWINGS UP SINGLE AND RIGHT DOUBLE +0xd5 U+2552 #BOX DRAWINGS DOWN SINGLE AND RIGHT DOUBLE +0xd6 U+2553 #BOX DRAWINGS DOWN DOUBLE AND RIGHT SINGLE +0xd7 U+256b #BOX DRAWINGS VERTICAL DOUBLE AND HORIZONTAL SINGLE +0xd8 U+256a #BOX DRAWINGS VERTICAL SINGLE AND HORIZONTAL DOUBLE +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+258c #LEFT HALF BLOCK +0xde U+2590 #RIGHT HALF BLOCK +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+03b1 #GREEK SMALL LETTER ALPHA +0xe1 U+00df U+03b2 #LATIN SMALL LETTER SHARP S +0xe2 U+0393 #GREEK CAPITAL LETTER GAMMA +0xe3 U+03c0 #GREEK SMALL LETTER PI +0xe4 U+03a3 U+2211 #GREEK CAPITAL LETTER SIGMA +0xe5 U+03c3 #GREEK SMALL LETTER SIGMA +0xe6 U+00b5 U+03bc #MICRO SIGN +0xe7 U+03c4 #GREEK SMALL LETTER TAU +0xe8 U+03a6 #GREEK CAPITAL LETTER PHI +0xe9 U+0398 U+03b8 #GREEK CAPITAL LETTER THETA +0xea U+03a9 U+2126 #GREEK CAPITAL LETTER OMEGA +0xeb U+03b4 #GREEK SMALL LETTER DELTA +0xec U+221e #INFINITY +0xed U+03c6 U+00f8 #GREEK SMALL LETTER PHI +0xee U+03b5 U+2208 U+220a #GREEK SMALL LETTER EPSILON +0xef U+2229 #INTERSECTION +0xf0 U+2261 #IDENTICAL TO +0xf1 U+00b1 #PLUS-MINUS SIGN +0xf2 U+2265 U+2267 #GREATER-THAN OR EQUAL TO +0xf3 U+2264 U+2266 #LESS-THAN OR EQUAL TO +0xf4 U+2320 U+0283 #TOP HALF INTEGRAL +0xf5 U+2321 #BOTTOM HALF INTEGRAL +0xf6 U+00f7 #DIVISION SIGN +0xf7 U+2248 #ALMOST EQUAL TO +0xf8 U+00b0 U+030a #DEGREE SIGN +0xf9 U+2219 U+0307 U+0387 #BULLET OPERATOR +0xfa U+00b7 U+2027 #MIDDLE DOT +0xfb U+221a #SQUARE ROOT +0xfc U+207f #SUPERSCRIPT LATIN SMALL LETTER N +0xfd U+00b2 #SUPERSCRIPT TWO +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE + +U+03ad "\356'" #:î' +U+03ae:h' +U+03cd:u' +U+03ce:w' + +U+2209 " !\356 " +U+221b " 3\373" +U+221c " 4\373" +U+2262 " !\360" +U+2299 "(\372)" +U+229a "(\370)" +U+22a1 "[\372]" +U+02a7 "t\364" diff --git a/src/chrtrans/cp737_uni.tbl b/src/chrtrans/cp737_uni.tbl new file mode 100644 index 0000000..e57e261 --- /dev/null +++ b/src/chrtrans/cp737_uni.tbl @@ -0,0 +1,172 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mcp737 + +#Name as a Display Charset (used on Options screen) +OGreek (cp737) + +#Codepage number +C737 + +# +# Name: cp737_DOSGreek to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp737_DOSGreek code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp737_DOSGreek order +# +################## +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw + +0x20-0x7f idem +# +0x80 U+0391 #GREEK CAPITAL LETTER ALPHA +0x81 U+0392 #GREEK CAPITAL LETTER BETA +0x82 U+0393 #GREEK CAPITAL LETTER GAMMA +0x83 U+0394 #GREEK CAPITAL LETTER DELTA +0x84 U+0395 #GREEK CAPITAL LETTER EPSILON +0x85 U+0396 #GREEK CAPITAL LETTER ZETA +0x86 U+0397 #GREEK CAPITAL LETTER ETA +0x87 U+0398 #GREEK CAPITAL LETTER THETA +0x88 U+0399 #GREEK CAPITAL LETTER IOTA +0x89 U+039a #GREEK CAPITAL LETTER KAPPA +0x8a U+039b #GREEK CAPITAL LETTER LAMDA +0x8b U+039c #GREEK CAPITAL LETTER MU +0x8c U+039d #GREEK CAPITAL LETTER NU +0x8d U+039e #GREEK CAPITAL LETTER XI +0x8e U+039f #GREEK CAPITAL LETTER OMICRON +0x8f U+03a0 #GREEK CAPITAL LETTER PI +0x90 U+03a1 #GREEK CAPITAL LETTER RHO +0x91 U+03a3 #GREEK CAPITAL LETTER SIGMA +0x92 U+03a4 #GREEK CAPITAL LETTER TAU +0x93 U+03a5 #GREEK CAPITAL LETTER UPSILON +0x94 U+03a6 #GREEK CAPITAL LETTER PHI +0x95 U+03a7 #GREEK CAPITAL LETTER CHI +0x96 U+03a8 #GREEK CAPITAL LETTER PSI +0x97 U+03a9 #GREEK CAPITAL LETTER OMEGA +0x98 U+03b1 #GREEK SMALL LETTER ALPHA +0x99 U+03b2 #GREEK SMALL LETTER BETA +0x9a U+03b3 U+0263 #GREEK SMALL LETTER GAMMA +0x9b U+03b4 #GREEK SMALL LETTER DELTA +0x9c U+03b5 #GREEK SMALL LETTER EPSILON +0x9d U+03b6 #GREEK SMALL LETTER ZETA +0x9e U+03b7 #GREEK SMALL LETTER ETA +0x9f U+03b8 #GREEK SMALL LETTER THETA +0xa0 U+03b9 U+0131 #GREEK SMALL LETTER IOTA +0xa1 U+03ba #GREEK SMALL LETTER KAPPA +0xa2 U+03bb #GREEK SMALL LETTER LAMDA +0xa3 U+03bc U+00b5 #GREEK SMALL LETTER MU +0xa4 U+03bd #GREEK SMALL LETTER NU +0xa5 U+03be #GREEK SMALL LETTER XI +0xa6 U+03bf #GREEK SMALL LETTER OMICRON +0xa7 U+03c0 #GREEK SMALL LETTER PI +0xa8 U+03c1 #GREEK SMALL LETTER RHO +0xa9 U+03c3 #GREEK SMALL LETTER SIGMA +0xaa U+03c2 #GREEK SMALL LETTER FINAL SIGMA +0xab U+03c4 #GREEK SMALL LETTER TAU +0xac U+03c5 U+028a #GREEK SMALL LETTER UPSILON +0xad U+03c6 #GREEK SMALL LETTER PHI +0xae U+03c7 #GREEK SMALL LETTER CHI +0xaf U+03c8 #GREEK SMALL LETTER PSI +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+2561 #BOX DRAWINGS VERTICAL SINGLE AND LEFT DOUBLE +0xb6 U+2562 #BOX DRAWINGS VERTICAL DOUBLE AND LEFT SINGLE +0xb7 U+2556 #BOX DRAWINGS DOWN DOUBLE AND LEFT SINGLE +0xb8 U+2555 #BOX DRAWINGS DOWN SINGLE AND LEFT DOUBLE +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+255c #BOX DRAWINGS UP DOUBLE AND LEFT SINGLE +0xbe U+255b #BOX DRAWINGS UP SINGLE AND LEFT DOUBLE +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+255e #BOX DRAWINGS VERTICAL SINGLE AND RIGHT DOUBLE +0xc7 U+255f #BOX DRAWINGS VERTICAL DOUBLE AND RIGHT SINGLE +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+2567 #BOX DRAWINGS UP SINGLE AND HORIZONTAL DOUBLE +0xd0 U+2568 #BOX DRAWINGS UP DOUBLE AND HORIZONTAL SINGLE +0xd1 U+2564 #BOX DRAWINGS DOWN SINGLE AND HORIZONTAL DOUBLE +0xd2 U+2565 #BOX DRAWINGS DOWN DOUBLE AND HORIZONTAL SINGLE +0xd3 U+2559 #BOX DRAWINGS UP DOUBLE AND RIGHT SINGLE +0xd4 U+2558 #BOX DRAWINGS UP SINGLE AND RIGHT DOUBLE +0xd5 U+2552 #BOX DRAWINGS DOWN SINGLE AND RIGHT DOUBLE +0xd6 U+2553 #BOX DRAWINGS DOWN DOUBLE AND RIGHT SINGLE +0xd7 U+256b #BOX DRAWINGS VERTICAL DOUBLE AND HORIZONTAL SINGLE +0xd8 U+256a #BOX DRAWINGS VERTICAL SINGLE AND HORIZONTAL DOUBLE +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+258c #LEFT HALF BLOCK +0xde U+2590 #RIGHT HALF BLOCK +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+03c9 #GREEK SMALL LETTER OMEGA +0xe1 U+03ac #GREEK SMALL LETTER ALPHA WITH TONOS +0xe2 U+03ad #GREEK SMALL LETTER EPSILON WITH TONOS +0xe3 U+03ae #GREEK SMALL LETTER ETA WITH TONOS +0xe4 U+03ca #GREEK SMALL LETTER IOTA WITH DIALYTIKA +0xe5 U+03af #GREEK SMALL LETTER IOTA WITH TONOS +0xe6 U+03cc #GREEK SMALL LETTER OMICRON WITH TONOS +0xe7 U+03cd #GREEK SMALL LETTER UPSILON WITH TONOS +0xe8 U+03cb U+00fc #GREEK SMALL LETTER UPSILON WITH DIALYTIKA +0xe9 U+03ce #GREEK SMALL LETTER OMEGA WITH TONOS +0xea U+0386 #GREEK CAPITAL LETTER ALPHA WITH TONOS +0xeb U+0388 #GREEK CAPITAL LETTER EPSILON WITH TONOS +0xec U+0389 #GREEK CAPITAL LETTER ETA WITH TONOS +0xed U+038a #GREEK CAPITAL LETTER IOTA WITH TONOS +0xee U+038c #GREEK CAPITAL LETTER OMICRON WITH TONOS +0xef U+038e #GREEK CAPITAL LETTER UPSILON WITH TONOS +0xf0 U+038f #GREEK CAPITAL LETTER OMEGA WITH TONOS +0xf1 U+00b1 #PLUS-MINUS SIGN +0xf2 U+2265 #GREATER-THAN OR EQUAL TO +0xf3 U+2264 #LESS-THAN OR EQUAL TO +0xf4 U+03aa #GREEK CAPITAL LETTER IOTA WITH DIALYTIKA +0xf5 U+03ab #GREEK CAPITAL LETTER UPSILON WITH DIALYTIKA +0xf6 U+00f7 #DIVISION SIGN +0xf7 U+2248 #ALMOST EQUAL TO +0xf8 U+00b0 #DEGREE SIGN +0xf9 U+2219 U+0307 U+0387 #BULLET OPERATOR +0xfa U+00b7 #MIDDLE DOT +0xfb U+221a #SQUARE ROOT +0xfc U+207f #SUPERSCRIPT LATIN SMALL LETTER N +0xfd U+00b2 #SUPERSCRIPT TWO +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE + +U+2209 " !\234 " +U+2218 " \370 " # RING OPERATOR +U+221b " 3\373" +U+221c " 4\373" +U+2299 "(\372)" +U+229a "(\370)" +U+22a1 "[\372]" +U+02a4 "d\235" +U+2249 "!\367" diff --git a/src/chrtrans/cp775_uni.tbl b/src/chrtrans/cp775_uni.tbl new file mode 100644 index 0000000..26a3ff7 --- /dev/null +++ b/src/chrtrans/cp775_uni.tbl @@ -0,0 +1,159 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mcp775 + +#Name as a Display Charset (used on Options screen) +OBaltic Rim (cp775) + +#Codepage number +C775 + +# Name: cp775_DOSBaltRim to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp775_DOSBaltRim code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp775_DOSBaltRim order +# +################## + +0x20-0x7e idem +# +0x80 U+0106 #LATIN CAPITAL LETTER C WITH ACUTE +0x81 U+00fc #LATIN SMALL LETTER U WITH DIAERESIS +0x82 U+00e9 #LATIN SMALL LETTER E WITH ACUTE +0x83 U+0101 #LATIN SMALL LETTER A WITH MACRON +0x84 U+00e4 #LATIN SMALL LETTER A WITH DIAERESIS +0x85 U+0123 #LATIN SMALL LETTER G WITH CEDILLA +0x86 U+00e5 #LATIN SMALL LETTER A WITH RING ABOVE +0x87 U+0107 #LATIN SMALL LETTER C WITH ACUTE +0x88 U+0142 #LATIN SMALL LETTER L WITH STROKE +0x89 U+0113 #LATIN SMALL LETTER E WITH MACRON +0x8a U+0156 #LATIN CAPITAL LETTER R WITH CEDILLA +0x8b U+0157 #LATIN SMALL LETTER R WITH CEDILLA +0x8c U+012b #LATIN SMALL LETTER I WITH MACRON +0x8d U+0179 #LATIN CAPITAL LETTER Z WITH ACUTE +0x8e U+00c4 #LATIN CAPITAL LETTER A WITH DIAERESIS +0x8f U+00c5 #LATIN CAPITAL LETTER A WITH RING ABOVE +0x90 U+00c9 #LATIN CAPITAL LETTER E WITH ACUTE +0x91 U+00e6 #LATIN SMALL LIGATURE AE +0x92 U+00c6 #LATIN CAPITAL LIGATURE AE +0x93 U+014d #LATIN SMALL LETTER O WITH MACRON +0x94 U+00f6 #LATIN SMALL LETTER O WITH DIAERESIS +0x95 U+0122 #LATIN CAPITAL LETTER G WITH CEDILLA +0x96 U+00a2 #CENT SIGN +0x97 U+015a #LATIN CAPITAL LETTER S WITH ACUTE +0x98 U+015b #LATIN SMALL LETTER S WITH ACUTE +0x99 U+00d6 #LATIN CAPITAL LETTER O WITH DIAERESIS +0x9a U+00dc #LATIN CAPITAL LETTER U WITH DIAERESIS +0x9b U+00f8 #LATIN SMALL LETTER O WITH STROKE +0x9c U+00a3 #POUND SIGN +0x9d U+00d8 #LATIN CAPITAL LETTER O WITH STROKE +0x9e U+00d7 #MULTIPLICATION SIGN +0x9f U+00a4 #CURRENCY SIGN +0xa0 U+0100 #LATIN CAPITAL LETTER A WITH MACRON +0xa1 U+012a #LATIN CAPITAL LETTER I WITH MACRON +0xa2 U+00f3 #LATIN SMALL LETTER O WITH ACUTE +0xa3 U+017b #LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xa4 U+017c #LATIN SMALL LETTER Z WITH DOT ABOVE +0xa5 U+017a #LATIN SMALL LETTER Z WITH ACUTE +0xa6 U+201d #RIGHT DOUBLE QUOTATION MARK +0xa7 U+00a6 #BROKEN BAR +0xa8 U+00a9 #COPYRIGHT SIGN +0xa9 U+00ae #REGISTERED SIGN +0xaa U+00ac #NOT SIGN +0xab U+00bd #VULGAR FRACTION ONE HALF +0xac U+00bc #VULGAR FRACTION ONE QUARTER +0xad U+0141 #LATIN CAPITAL LETTER L WITH STROKE +0xae U+00ab #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xaf U+00bb #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+0104 #LATIN CAPITAL LETTER A WITH OGONEK +0xb6 U+010c #LATIN CAPITAL LETTER C WITH CARON +0xb7 U+0118 #LATIN CAPITAL LETTER E WITH OGONEK +0xb8 U+0116 #LATIN CAPITAL LETTER E WITH DOT ABOVE +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+012e #LATIN CAPITAL LETTER I WITH OGONEK +0xbe U+0160 #LATIN CAPITAL LETTER S WITH CARON +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+0172 #LATIN CAPITAL LETTER U WITH OGONEK +0xc7 U+016a #LATIN CAPITAL LETTER U WITH MACRON +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+017d #LATIN CAPITAL LETTER Z WITH CARON +0xd0 U+0105 #LATIN SMALL LETTER A WITH OGONEK +0xd1 U+010d #LATIN SMALL LETTER C WITH CARON +0xd2 U+0119 #LATIN SMALL LETTER E WITH OGONEK +0xd3 U+0117 #LATIN SMALL LETTER E WITH DOT ABOVE +0xd4 U+012f #LATIN SMALL LETTER I WITH OGONEK +0xd5 U+0161 #LATIN SMALL LETTER S WITH CARON +0xd6 U+0173 #LATIN SMALL LETTER U WITH OGONEK +0xd7 U+016b #LATIN SMALL LETTER U WITH MACRON +0xd8 U+017e #LATIN SMALL LETTER Z WITH CARON +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+258c #LEFT HALF BLOCK +0xde U+2590 #RIGHT HALF BLOCK +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+00d3 #LATIN CAPITAL LETTER O WITH ACUTE +0xe1 U+00df #LATIN SMALL LETTER SHARP S (GERMAN) +0xe2 U+014c #LATIN CAPITAL LETTER O WITH MACRON +0xe3 U+0143 #LATIN CAPITAL LETTER N WITH ACUTE +0xe4 U+00f5 #LATIN SMALL LETTER O WITH TILDE +0xe5 U+00d5 #LATIN CAPITAL LETTER O WITH TILDE +0xe6 U+00b5 #MICRO SIGN +0xe7 U+0144 #LATIN SMALL LETTER N WITH ACUTE +0xe8 U+0136 #LATIN CAPITAL LETTER K WITH CEDILLA +0xe9 U+0137 #LATIN SMALL LETTER K WITH CEDILLA +0xea U+013b #LATIN CAPITAL LETTER L WITH CEDILLA +0xeb U+013c #LATIN SMALL LETTER L WITH CEDILLA +0xec U+0146 #LATIN SMALL LETTER N WITH CEDILLA +0xed U+0112 #LATIN CAPITAL LETTER E WITH MACRON +0xee U+0145 #LATIN CAPITAL LETTER N WITH CEDILLA +0xef U+2019 #RIGHT SINGLE QUOTATION MARK +0xf0 U+00ad #SOFT HYPHEN +0xf1 U+00b1 #PLUS-MINUS SIGN +0xf2 U+201c #LEFT DOUBLE QUOTATION MARK +0xf3 U+00be #VULGAR FRACTION THREE QUARTERS +0xf4 U+00b6 #PILCROW SIGN +0xf5 U+00a7 #SECTION SIGN +0xf6 U+00f7 #DIVISION SIGN +0xf7 U+201e #DOUBLE LOW-9 QUOTATION MARK +0xf8 U+00b0 #DEGREE SIGN +0xf9 U+2219 #BULLET OPERATOR +0xfa U+00b7 #MIDDLE DOT +0xfb U+00b9 #SUPERSCRIPT ONE +0xfc U+00b3 #SUPERSCRIPT THREE +0xfd U+00b2 #SUPERSCRIPT TWO +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE diff --git a/src/chrtrans/cp850_uni.tbl b/src/chrtrans/cp850_uni.tbl new file mode 100644 index 0000000..9d05af2 --- /dev/null +++ b/src/chrtrans/cp850_uni.tbl @@ -0,0 +1,177 @@ +# This file has been modified for lynx (see README.tables) + +#Shall this become the "default" translation? +#Meaning of that is currently unclear... It's different +#from the default input or default output charset... +#but there has to be exactly one table marked as "default". +D0 +# +#The MIME name of this charset. +Mcp850 + +#Name as a Display Charset (used on Options screen) +OWestern (cp850) + +#Codepage number +C850 + +# +# Name: cp850_DOSLatin1 to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp850_DOSLatin1 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp850_DOSLatin1 order +# +################## +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw + +0x20-0x7e idem +# +0x80 U+00c7 #LATIN CAPITAL LETTER C WITH CEDILLA +0x81 U+00fc U+03cb #LATIN SMALL LETTER U WITH DIAERESIS +0x82 U+00e9 U+03ad #LATIN SMALL LETTER E WITH ACUTE +0x83 U+00e2 #LATIN SMALL LETTER A WITH CIRCUMFLEX +0x84 U+00e4 #LATIN SMALL LETTER A WITH DIAERESIS +0x85 U+00e0 #LATIN SMALL LETTER A WITH GRAVE +0x86 U+00e5 #LATIN SMALL LETTER A WITH RING ABOVE +0x87 U+00e7 #LATIN SMALL LETTER C WITH CEDILLA +0x88 U+00ea #LATIN SMALL LETTER E WITH CIRCUMFLEX +0x89 U+00eb #LATIN SMALL LETTER E WITH DIAERESIS +0x8a U+00e8 #LATIN SMALL LETTER E WITH GRAVE +0x8b U+00ef U+03ca #LATIN SMALL LETTER I WITH DIAERESIS +0x8c U+00ee #LATIN SMALL LETTER I WITH CIRCUMFLEX +0x8d U+00ec #LATIN SMALL LETTER I WITH GRAVE +0x8e U+00c4 #LATIN CAPITAL LETTER A WITH DIAERESIS +0x8f U+00c5 #LATIN CAPITAL LETTER A WITH RING ABOVE +0x90 U+00c9 U+0388 #LATIN CAPITAL LETTER E WITH ACUTE +0x91 U+00e6 #LATIN SMALL LIGATURE AE +0x92 U+00c6 #LATIN CAPITAL LIGATURE AE +0x93 U+00f4 #LATIN SMALL LETTER O WITH CIRCUMFLEX +0x94 U+00f6 #LATIN SMALL LETTER O WITH DIAERESIS +0x95 U+00f2 #LATIN SMALL LETTER O WITH GRAVE +0x96 U+00fb #LATIN SMALL LETTER U WITH CIRCUMFLEX +0x97 U+00f9 #LATIN SMALL LETTER U WITH GRAVE +0x98 U+00ff #LATIN SMALL LETTER Y WITH DIAERESIS +0x99 U+00d6 #LATIN CAPITAL LETTER O WITH DIAERESIS +0x9a U+00dc U+03ab #LATIN CAPITAL LETTER U WITH DIAERESIS +0x9b U+00f8 #LATIN SMALL LETTER O WITH STROKE +0x9c U+00a3 #POUND SIGN +0x9d U+00d8 #LATIN CAPITAL LETTER O WITH STROKE +0x9e U+00d7 #MULTIPLICATION SIGN +0x9f U+0192 #LATIN SMALL LETTER F WITH HOOK +0xa0 U+00e1 U+03ac #LATIN SMALL LETTER A WITH ACUTE +0xa1 U+00ed U+03af #LATIN SMALL LETTER I WITH ACUTE +0xa2 U+00f3 U+03cc #LATIN SMALL LETTER O WITH ACUTE +0xa3 U+00fa U+03cd #LATIN SMALL LETTER U WITH ACUTE +0xa4 U+00f1 #LATIN SMALL LETTER N WITH TILDE +0xa5 U+00d1 #LATIN CAPITAL LETTER N WITH TILDE +0xa6 U+00aa #FEMININE ORDINAL INDICATOR +0xa7 U+00ba #MASCULINE ORDINAL INDICATOR +0xa8 U+00bf #INVERTED QUESTION MARK +0xa9 U+00ae #REGISTERED SIGN +0xaa U+00ac #NOT SIGN +0xab U+00bd #VULGAR FRACTION ONE HALF +0xac U+00bc #VULGAR FRACTION ONE QUARTER +0xad U+00a1 #INVERTED EXCLAMATION MARK +0xae U+00ab #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xaf U+00bb #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+00c1 U+0386 #LATIN CAPITAL LETTER A WITH ACUTE +0xb6 U+00c2 #LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xb7 U+00c0 #LATIN CAPITAL LETTER A WITH GRAVE +0xb8 U+00a9 #COPYRIGHT SIGN +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+00a2 #CENT SIGN +0xbe U+00a5 #YEN SIGN +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+00e3 #LATIN SMALL LETTER A WITH TILDE +0xc7 U+00c3 #LATIN CAPITAL LETTER A WITH TILDE +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+00a4 #CURRENCY SIGN +0xd0 U+00f0 #LATIN SMALL LETTER ETH +0xd1 U+00d0 #LATIN CAPITAL LETTER ETH +0xd2 U+00ca #LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xd3 U+00cb #LATIN CAPITAL LETTER E WITH DIAERESIS +0xd4 U+00c8 #LATIN CAPITAL LETTER E WITH GRAVE +0xd5 U+0131 U+03b9 #LATIN SMALL LETTER DOTLESS I +0xd6 U+00cd U+038a #LATIN CAPITAL LETTER I WITH ACUTE +0xd7 U+00ce #LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xd8 U+00cf U+03aa #LATIN CAPITAL LETTER I WITH DIAERESIS +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+00a6 #BROKEN BAR +0xde U+00cc #LATIN CAPITAL LETTER I WITH GRAVE +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+00d3 U+038c #LATIN CAPITAL LETTER O WITH ACUTE +0xe1 U+00df #LATIN SMALL LETTER SHARP S +0xe2 U+00d4 #LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xe3 U+00d2 #LATIN CAPITAL LETTER O WITH GRAVE +0xe4 U+00f5 #LATIN SMALL LETTER O WITH TILDE +0xe5 U+00d5 #LATIN CAPITAL LETTER O WITH TILDE +0xe6 U+00b5 U+03bc #MICRO SIGN +0xe7 U+00fe #LATIN SMALL LETTER THORN +0xe8 U+00de #LATIN CAPITAL LETTER THORN +0xe9 U+00da #LATIN CAPITAL LETTER U WITH ACUTE +0xea U+00db #LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xeb U+00d9 #LATIN CAPITAL LETTER U WITH GRAVE +0xec U+00fd #LATIN SMALL LETTER Y WITH ACUTE +0xed U+00dd #LATIN CAPITAL LETTER Y WITH ACUTE +0xee U+00af U+0304 #MACRON +0xef U+00b4 U+0301 #ACUTE ACCENT +0xf0 U+00ad #SOFT HYPHEN +0xf1 U+00b1 #PLUS-MINUS SIGN +0xf2 U+2017 U+0333 #DOUBLE LOW LINE +0xf3 U+00be #VULGAR FRACTION THREE QUARTERS +0xf4 U+00b6 #PILCROW SIGN +0xf5 U+00a7 #SECTION SIGN +0xf6 U+00f7 #DIVISION SIGN +0xf7 U+00b8 U+0327 #CEDILLA +0xf8 U+00b0 U+030a #DEGREE SIGN +0xf9 U+00a8 U+0308 #DIAERESIS +0xfa U+00b7 U+0307 U+0387 U+2027 #MIDDLE DOT +0xfb U+00b9 #SUPERSCRIPT ONE +0xfc U+00b3 #SUPERSCRIPT THREE +0xfd U+00b2 #SUPERSCRIPT TWO +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE + +U+2218 " \370 " # RING OPERATOR +U+221b " ROOT\374 " +U+2297 "(\236)" # CIRCLED TIMES +U+2299 "(\372)" # CIRCLED DOT OPERATOR +U+229A "(\370)" # CIRCLED RING OPERATOR +U+22A0 "[\236]" # SQUARED TIMES +U+22A1 "[\372]" # SQUARED DOT OPERATOR +U+22C5 " \372 " # DOT OPERATOR diff --git a/src/chrtrans/cp852_uni.tbl b/src/chrtrans/cp852_uni.tbl new file mode 100644 index 0000000..978cca4 --- /dev/null +++ b/src/chrtrans/cp852_uni.tbl @@ -0,0 +1,170 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mcp852 + +#Name as a Display Charset (used on Options screen) +OEastern European (cp852) + +#Codepage number +C852 + +# +# Name: cp852_DOSLatin2 to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp852_DOSLatin2 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp852_DOSLatin2 order +# +################## +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw + +0x20-0x7e idem +# +0x80 U+00c7 #LATIN CAPITAL LETTER C WITH CEDILLA +0x81 U+00fc U+03cb #LATIN SMALL LETTER U WITH DIAERESIS +0x82 U+00e9 U+03ad #LATIN SMALL LETTER E WITH ACUTE +0x83 U+00e2 #LATIN SMALL LETTER A WITH CIRCUMFLEX +0x84 U+00e4 #LATIN SMALL LETTER A WITH DIAERESIS +0x85 U+016f #LATIN SMALL LETTER U WITH RING ABOVE +0x86 U+0107 #LATIN SMALL LETTER C WITH ACUTE +0x87 U+00e7 #LATIN SMALL LETTER C WITH CEDILLA +0x88 U+0142 #LATIN SMALL LETTER L WITH STROKE +0x89 U+00eb #LATIN SMALL LETTER E WITH DIAERESIS +0x8a U+0150 #LATIN CAPITAL LETTER O WITH DOUBLE ACUTE +0x8b U+0151 #LATIN SMALL LETTER O WITH DOUBLE ACUTE +0x8c U+00ee #LATIN SMALL LETTER I WITH CIRCUMFLEX +0x8d U+0179 #LATIN CAPITAL LETTER Z WITH ACUTE +0x8e U+00c4 #LATIN CAPITAL LETTER A WITH DIAERESIS +0x8f U+0106 #LATIN CAPITAL LETTER C WITH ACUTE +0x90 U+00c9 U+0388 #LATIN CAPITAL LETTER E WITH ACUTE +0x91 U+0139 #LATIN CAPITAL LETTER L WITH ACUTE +0x92 U+013a #LATIN SMALL LETTER L WITH ACUTE +0x93 U+00f4 #LATIN SMALL LETTER O WITH CIRCUMFLEX +0x94 U+00f6 #LATIN SMALL LETTER O WITH DIAERESIS +0x95 U+013d #LATIN CAPITAL LETTER L WITH CARON +0x96 U+013e #LATIN SMALL LETTER L WITH CARON +0x97 U+015a #LATIN CAPITAL LETTER S WITH ACUTE +0x98 U+015b #LATIN SMALL LETTER S WITH ACUTE +0x99 U+00d6 #LATIN CAPITAL LETTER O WITH DIAERESIS +0x9a U+00dc U+03ab #LATIN CAPITAL LETTER U WITH DIAERESIS +0x9b U+0164 #LATIN CAPITAL LETTER T WITH CARON +0x9c U+0165 #LATIN SMALL LETTER T WITH CARON +0x9d U+0141 #LATIN CAPITAL LETTER L WITH STROKE +0x9e U+00d7 #MULTIPLICATION SIGN +0x9f U+010d U+02a7 U+0447 #LATIN SMALL LETTER C WITH CARON +0xa0 U+00e1 U+03ac #LATIN SMALL LETTER A WITH ACUTE +0xa1 U+00ed U+03af #LATIN SMALL LETTER I WITH ACUTE +0xa2 U+00f3 U+03cc #LATIN SMALL LETTER O WITH ACUTE +0xa3 U+00fa U+03cd #LATIN SMALL LETTER U WITH ACUTE +0xa4 U+0104 #LATIN CAPITAL LETTER A WITH OGONEK +0xa5 U+0105 #LATIN SMALL LETTER A WITH OGONEK +0xa6 U+017d U+0416 #LATIN CAPITAL LETTER Z WITH CARON +0xa7 U+017e U+0436 #LATIN SMALL LETTER Z WITH CARON +0xa8 U+0118 #LATIN CAPITAL LETTER E WITH OGONEK +0xa9 U+0119 #LATIN SMALL LETTER E WITH OGONEK +0xaa U+00ac #NOT SIGN +0xab U+017a #LATIN SMALL LETTER Z WITH ACUTE +0xac U+010c U+0427 #LATIN CAPITAL LETTER C WITH CARON +0xad U+015f #LATIN SMALL LETTER S WITH CEDILLA +0xae U+00ab #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xaf U+00bb #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+00c1 U+0386 #LATIN CAPITAL LETTER A WITH ACUTE +0xb6 U+00c2 #LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xb7 U+011a #LATIN CAPITAL LETTER E WITH CARON +0xb8 U+015e #LATIN CAPITAL LETTER S WITH CEDILLA +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+017b #LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xbe U+017c #LATIN SMALL LETTER Z WITH DOT ABOVE +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+0102 #LATIN CAPITAL LETTER A WITH BREVE +0xc7 U+0103 #LATIN SMALL LETTER A WITH BREVE +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+00a4 #CURRENCY SIGN +0xd0 U+0111 #LATIN SMALL LETTER D WITH STROKE +0xd1 U+0110 #LATIN CAPITAL LETTER D WITH STROKE +0xd2 U+010e #LATIN CAPITAL LETTER D WITH CARON +0xd3 U+00cb #LATIN CAPITAL LETTER E WITH DIAERESIS +0xd4 U+010f #LATIN SMALL LETTER D WITH CARON +0xd5 U+0147 #LATIN CAPITAL LETTER N WITH CARON +0xd6 U+00cd U+038a #LATIN CAPITAL LETTER I WITH ACUTE +0xd7 U+00ce #LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xd8 U+011b #LATIN SMALL LETTER E WITH CARON +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+0162 #LATIN CAPITAL LETTER T WITH CEDILLA +0xde U+016e #LATIN CAPITAL LETTER U WITH RING ABOVE +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+00d3 U+038c #LATIN CAPITAL LETTER O WITH ACUTE +0xe1 U+00df #LATIN SMALL LETTER SHARP S +0xe2 U+00d4 #LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xe3 U+0143 #LATIN CAPITAL LETTER N WITH ACUTE +0xe4 U+0144 #LATIN SMALL LETTER N WITH ACUTE +0xe5 U+0148 #LATIN SMALL LETTER N WITH CARON +0xe6 U+0160 U+0428 #LATIN CAPITAL LETTER S WITH CARON +0xe7 U+0161 U+0448 #LATIN SMALL LETTER S WITH CARON +0xe8 U+0154 #LATIN CAPITAL LETTER R WITH ACUTE +0xe9 U+00da #LATIN CAPITAL LETTER U WITH ACUTE +0xea U+0155 #LATIN SMALL LETTER R WITH ACUTE +0xeb U+0170 #LATIN CAPITAL LETTER U WITH DOUBLE ACUTE +0xec U+00fd #LATIN SMALL LETTER Y WITH ACUTE +0xed U+00dd #LATIN CAPITAL LETTER Y WITH ACUTE +0xee U+0163 #LATIN SMALL LETTER T WITH CEDILLA +0xef U+00b4 U+0301 #ACUTE ACCENT +0xf0 U+00ad #SOFT HYPHEN +0xf1 U+02dd U+030b #DOUBLE ACUTE ACCENT +0xf2 U+02db U+0328 #OGONEK +0xf3 U+02c7 U+030c #CARON +0xf4 U+02d8 U+0306 #BREVE +0xf5 U+00a7 #SECTION SIGN +0xf6 U+00f7 #DIVISION SIGN +0xf7 U+00b8 U+0327 #CEDILLA +0xf8 U+00b0 U+030a #DEGREE SIGN +0xf9 U+00a8 U+0308 #DIAERESIS +0xfa U+02d9 U+0307 U+0387 #DOT ABOVE +0xfb U+0171 #LATIN SMALL LETTER U WITH DOUBLE ACUTE +0xfc U+0158 #LATIN CAPITAL LETTER R WITH CARON +0xfd U+0159 #LATIN SMALL LETTER R WITH CARON +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE + +U+2218 " \370 " # RING OPERATOR +U+2297 "(\236)" # CIRCLED TIMES +U+2299 "(\372)" # CIRCLED DOT OPERATOR +U+229A "(\370)" # CIRCLED RING OPERATOR +U+22A0 "[\236]" # SQUARED TIMES +U+22A1 "[\372]" # SQUARED DOT OPERATOR +U+22C5 " \372 " # DOT OPERATOR diff --git a/src/chrtrans/cp857_uni.tbl b/src/chrtrans/cp857_uni.tbl new file mode 100644 index 0000000..bd30996 --- /dev/null +++ b/src/chrtrans/cp857_uni.tbl @@ -0,0 +1,159 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mcp857 + +#Name as a Display Charset (used on Options screen). +OTurkish (cp857) + +#Codepage number +C857 + +# Name: cp857_DOSTurkish to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp857_DOSTurkish code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp857_DOSTurkish order +# +################## + +0x20-0x7f idem + +0x80 U+00c7 #LATIN CAPITAL LETTER C WITH CEDILLA +0x81 U+00fc #LATIN SMALL LETTER U WITH DIAERESIS +0x82 U+00e9 #LATIN SMALL LETTER E WITH ACUTE +0x83 U+00e2 #LATIN SMALL LETTER A WITH CIRCUMFLEX +0x84 U+00e4 #LATIN SMALL LETTER A WITH DIAERESIS +0x85 U+00e0 #LATIN SMALL LETTER A WITH GRAVE +0x86 U+00e5 #LATIN SMALL LETTER A WITH RING ABOVE +0x87 U+00e7 #LATIN SMALL LETTER C WITH CEDILLA +0x88 U+00ea #LATIN SMALL LETTER E WITH CIRCUMFLEX +0x89 U+00eb #LATIN SMALL LETTER E WITH DIAERESIS +0x8a U+00e8 #LATIN SMALL LETTER E WITH GRAVE +0x8b U+00ef #LATIN SMALL LETTER I WITH DIAERESIS +0x8c U+00ee #LATIN SMALL LETTER I WITH CIRCUMFLEX +0x8d U+0131 #LATIN SMALL LETTER DOTLESS I +0x8e U+00c4 #LATIN CAPITAL LETTER A WITH DIAERESIS +0x8f U+00c5 #LATIN CAPITAL LETTER A WITH RING ABOVE +0x90 U+00c9 #LATIN CAPITAL LETTER E WITH ACUTE +0x91 U+00e6 #LATIN SMALL LIGATURE AE +0x92 U+00c6 #LATIN CAPITAL LIGATURE AE +0x93 U+00f4 #LATIN SMALL LETTER O WITH CIRCUMFLEX +0x94 U+00f6 #LATIN SMALL LETTER O WITH DIAERESIS +0x95 U+00f2 #LATIN SMALL LETTER O WITH GRAVE +0x96 U+00fb #LATIN SMALL LETTER U WITH CIRCUMFLEX +0x97 U+00f9 #LATIN SMALL LETTER U WITH GRAVE +0x98 U+0130 #LATIN CAPITAL LETTER I WITH DOT ABOVE +0x99 U+00d6 #LATIN CAPITAL LETTER O WITH DIAERESIS +0x9a U+00dc #LATIN CAPITAL LETTER U WITH DIAERESIS +0x9b U+00f8 #LATIN SMALL LETTER O WITH STROKE +0x9c U+00a3 #POUND SIGN +0x9d U+00d8 #LATIN CAPITAL LETTER O WITH STROKE +0x9e U+015e #LATIN CAPITAL LETTER S WITH CEDILLA +0x9f U+015f #LATIN SMALL LETTER S WITH CEDILLA +0xa0 U+00e1 #LATIN SMALL LETTER A WITH ACUTE +0xa1 U+00ed #LATIN SMALL LETTER I WITH ACUTE +0xa2 U+00f3 #LATIN SMALL LETTER O WITH ACUTE +0xa3 U+00fa #LATIN SMALL LETTER U WITH ACUTE +0xa4 U+00f1 #LATIN SMALL LETTER N WITH TILDE +0xa5 U+00d1 #LATIN CAPITAL LETTER N WITH TILDE +0xa6 U+011e #LATIN CAPITAL LETTER G WITH BREVE +0xa7 U+011f #LATIN SMALL LETTER G WITH BREVE +0xa8 U+00bf #INVERTED QUESTION MARK +0xa9 U+00ae #REGISTERED SIGN +0xaa U+00ac #NOT SIGN +0xab U+00bd #VULGAR FRACTION ONE HALF +0xac U+00bc #VULGAR FRACTION ONE QUARTER +0xad U+00a1 #INVERTED EXCLAMATION MARK +0xae U+00ab #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xaf U+00bb #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+00c1 #LATIN CAPITAL LETTER A WITH ACUTE +0xb6 U+00c2 #LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xb7 U+00c0 #LATIN CAPITAL LETTER A WITH GRAVE +0xb8 U+00a9 #COPYRIGHT SIGN +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+00a2 #CENT SIGN +0xbe U+00a5 #YEN SIGN +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+00e3 #LATIN SMALL LETTER A WITH TILDE +0xc7 U+00c3 #LATIN CAPITAL LETTER A WITH TILDE +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+00a4 #CURRENCY SIGN +0xd0 U+00ba #MASCULINE ORDINAL INDICATOR +0xd1 U+00aa #FEMININE ORDINAL INDICATOR +0xd2 U+00ca #LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xd3 U+00cb #LATIN CAPITAL LETTER E WITH DIAERESIS +0xd4 U+00c8 #LATIN CAPITAL LETTER E WITH GRAVE +0xd5 #UNDEFINED +0xd6 U+00cd #LATIN CAPITAL LETTER I WITH ACUTE +0xd7 U+00ce #LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xd8 U+00cf #LATIN CAPITAL LETTER I WITH DIAERESIS +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+00a6 #BROKEN BAR +0xde U+00cc #LATIN CAPITAL LETTER I WITH GRAVE +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+00d3 #LATIN CAPITAL LETTER O WITH ACUTE +0xe1 U+00df #LATIN SMALL LETTER SHARP S +0xe2 U+00d4 #LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xe3 U+00d2 #LATIN CAPITAL LETTER O WITH GRAVE +0xe4 U+00f5 #LATIN SMALL LETTER O WITH TILDE +0xe5 U+00d5 #LATIN CAPITAL LETTER O WITH TILDE +0xe6 U+00b5 #MICRO SIGN +0xe7 #UNDEFINED +0xe8 U+00d7 #MULTIPLICATION SIGN +0xe9 U+00da #LATIN CAPITAL LETTER U WITH ACUTE +0xea U+00db #LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xeb U+00d9 #LATIN CAPITAL LETTER U WITH GRAVE +0xec U+00ec #LATIN SMALL LETTER I WITH GRAVE +0xed U+00ff #LATIN SMALL LETTER Y WITH DIAERESIS +0xee U+00af #MACRON +0xef U+00b4 #ACUTE ACCENT +0xf0 U+00ad #SOFT HYPHEN +0xf1 U+00b1 #PLUS-MINUS SIGN +0xf2 #UNDEFINED +0xf3 U+00be #VULGAR FRACTION THREE QUARTERS +0xf4 U+00b6 #PILCROW SIGN +0xf5 U+00a7 #SECTION SIGN +0xf6 U+00f7 #DIVISION SIGN +0xf7 U+00b8 #CEDILLA +0xf8 U+00b0 #DEGREE SIGN +0xf9 U+00a8 #DIAERESIS +0xfa U+00b7 #MIDDLE DOT +0xfb U+00b9 #SUPERSCRIPT ONE +0xfc U+00b3 #SUPERSCRIPT THREE +0xfd U+00b2 #SUPERSCRIPT TWO +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE diff --git a/src/chrtrans/cp862_uni.tbl b/src/chrtrans/cp862_uni.tbl new file mode 100644 index 0000000..ebf1222 --- /dev/null +++ b/src/chrtrans/cp862_uni.tbl @@ -0,0 +1,160 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mcp862 + +#Name as a Display Charset (used on Options screen). +OHebrew (cp862) + +#Codepage number +C862 + +# Name: cp862_DOSHebrew to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp862_DOSHebrew code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp862_DOSHebrew order +# +################## + +0x20-0x7f idem +# +0x80 U+05d0 #HEBREW LETTER ALEF +0x81 U+05d1 #HEBREW LETTER BET +0x82 U+05d2 #HEBREW LETTER GIMEL +0x83 U+05d3 #HEBREW LETTER DALET +0x84 U+05d4 #HEBREW LETTER HE +0x85 U+05d5 #HEBREW LETTER VAV +0x86 U+05d6 #HEBREW LETTER ZAYIN +0x87 U+05d7 #HEBREW LETTER HET +0x88 U+05d8 #HEBREW LETTER TET +0x89 U+05d9 #HEBREW LETTER YOD +0x8a U+05da #HEBREW LETTER FINAL KAF +0x8b U+05db #HEBREW LETTER KAF +0x8c U+05dc #HEBREW LETTER LAMED +0x8d U+05dd #HEBREW LETTER FINAL MEM +0x8e U+05de #HEBREW LETTER MEM +0x8f U+05df #HEBREW LETTER FINAL NUN +0x90 U+05e0 #HEBREW LETTER NUN +0x91 U+05e1 #HEBREW LETTER SAMEKH +0x92 U+05e2 #HEBREW LETTER AYIN +0x93 U+05e3 #HEBREW LETTER FINAL PE +0x94 U+05e4 #HEBREW LETTER PE +0x95 U+05e5 #HEBREW LETTER FINAL TSADI +0x96 U+05e6 #HEBREW LETTER TSADI +0x97 U+05e7 #HEBREW LETTER QOF +0x98 U+05e8 #HEBREW LETTER RESH +0x99 U+05e9 #HEBREW LETTER SHIN +0x9a U+05ea #HEBREW LETTER TAV +0x9b U+00a2 #CENT SIGN +0x9c U+00a3 #POUND SIGN +0x9d U+00a5 #YEN SIGN +0x9e U+20a7 #PESETA SIGN +0x9f U+0192 #LATIN SMALL LETTER F WITH HOOK +0xa0 U+00e1 #LATIN SMALL LETTER A WITH ACUTE +0xa1 U+00ed #LATIN SMALL LETTER I WITH ACUTE +0xa2 U+00f3 #LATIN SMALL LETTER O WITH ACUTE +0xa3 U+00fa #LATIN SMALL LETTER U WITH ACUTE +0xa4 U+00f1 #LATIN SMALL LETTER N WITH TILDE +0xa5 U+00d1 #LATIN CAPITAL LETTER N WITH TILDE +0xa6 U+00aa #FEMININE ORDINAL INDICATOR +0xa7 U+00ba #MASCULINE ORDINAL INDICATOR +0xa8 U+00bf #INVERTED QUESTION MARK +0xa9 U+2310 #REVERSED NOT SIGN +0xaa U+00ac #NOT SIGN +0xab U+00bd #VULGAR FRACTION ONE HALF +0xac U+00bc #VULGAR FRACTION ONE QUARTER +0xad U+00a1 #INVERTED EXCLAMATION MARK +0xae U+00ab #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xaf U+00bb #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+2561 #BOX DRAWINGS VERTICAL SINGLE AND LEFT DOUBLE +0xb6 U+2562 #BOX DRAWINGS VERTICAL DOUBLE AND LEFT SINGLE +0xb7 U+2556 #BOX DRAWINGS DOWN DOUBLE AND LEFT SINGLE +0xb8 U+2555 #BOX DRAWINGS DOWN SINGLE AND LEFT DOUBLE +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+255c #BOX DRAWINGS UP DOUBLE AND LEFT SINGLE +0xbe U+255b #BOX DRAWINGS UP SINGLE AND LEFT DOUBLE +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+255e #BOX DRAWINGS VERTICAL SINGLE AND RIGHT DOUBLE +0xc7 U+255f #BOX DRAWINGS VERTICAL DOUBLE AND RIGHT SINGLE +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+2567 #BOX DRAWINGS UP SINGLE AND HORIZONTAL DOUBLE +0xd0 U+2568 #BOX DRAWINGS UP DOUBLE AND HORIZONTAL SINGLE +0xd1 U+2564 #BOX DRAWINGS DOWN SINGLE AND HORIZONTAL DOUBLE +0xd2 U+2565 #BOX DRAWINGS DOWN DOUBLE AND HORIZONTAL SINGLE +0xd3 U+2559 #BOX DRAWINGS UP DOUBLE AND RIGHT SINGLE +0xd4 U+2558 #BOX DRAWINGS UP SINGLE AND RIGHT DOUBLE +0xd5 U+2552 #BOX DRAWINGS DOWN SINGLE AND RIGHT DOUBLE +0xd6 U+2553 #BOX DRAWINGS DOWN DOUBLE AND RIGHT SINGLE +0xd7 U+256b #BOX DRAWINGS VERTICAL DOUBLE AND HORIZONTAL SINGLE +0xd8 U+256a #BOX DRAWINGS VERTICAL SINGLE AND HORIZONTAL DOUBLE +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+258c #LEFT HALF BLOCK +0xde U+2590 #RIGHT HALF BLOCK +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+03b1 #GREEK SMALL LETTER ALPHA +0xe1 U+00df #LATIN SMALL LETTER SHARP S (GERMAN) +0xe2 U+0393 #GREEK CAPITAL LETTER GAMMA +0xe3 U+03c0 #GREEK SMALL LETTER PI +0xe4 U+03a3 #GREEK CAPITAL LETTER SIGMA +0xe5 U+03c3 #GREEK SMALL LETTER SIGMA +0xe6 U+00b5 #MICRO SIGN +0xe7 U+03c4 #GREEK SMALL LETTER TAU +0xe8 U+03a6 #GREEK CAPITAL LETTER PHI +0xe9 U+0398 #GREEK CAPITAL LETTER THETA +0xea U+03a9 #GREEK CAPITAL LETTER OMEGA +0xeb U+03b4 #GREEK SMALL LETTER DELTA +0xec U+221e #INFINITY +0xed U+03c6 #GREEK SMALL LETTER PHI +0xee U+03b5 #GREEK SMALL LETTER EPSILON +0xef U+2229 #INTERSECTION +0xf0 U+2261 #IDENTICAL TO +0xf1 U+00b1 #PLUS-MINUS SIGN +0xf2 U+2265 #GREATER-THAN OR EQUAL TO +0xf3 U+2264 #LESS-THAN OR EQUAL TO +0xf4 U+2320 #TOP HALF INTEGRAL +0xf5 U+2321 #BOTTOM HALF INTEGRAL +0xf6 U+00f7 #DIVISION SIGN +0xf7 U+2248 #ALMOST EQUAL TO +0xf8 U+00b0 #DEGREE SIGN +0xf9 U+2219 #BULLET OPERATOR +0xfa U+00b7 #MIDDLE DOT +0xfb U+221a #SQUARE ROOT +0xfc U+207f #SUPERSCRIPT LATIN SMALL LETTER N +0xfd U+00b2 #SUPERSCRIPT TWO +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE + diff --git a/src/chrtrans/cp864_uni.tbl b/src/chrtrans/cp864_uni.tbl new file mode 100644 index 0000000..6568577 --- /dev/null +++ b/src/chrtrans/cp864_uni.tbl @@ -0,0 +1,160 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mcp864 + +#Name as a Display Charset (used on Options screen). +OArabic (cp864) + +#Codepage number +C864 + +# Name: cp864_DOSArabic to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp864_DOSArabic code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp864_DOSArabic order +# +################## + +0x20-0x7f idem +# +0x80 U+00b0 #DEGREE SIGN +0x81 U+00b7 #MIDDLE DOT +0x82 U+2219 #BULLET OPERATOR +0x83 U+221a #SQUARE ROOT +0x84 U+2592 #MEDIUM SHADE +0x85 U+2500 #FORMS LIGHT HORIZONTAL +0x86 U+2502 #FORMS LIGHT VERTICAL +0x87 U+253c #FORMS LIGHT VERTICAL AND HORIZONTAL +0x88 U+2524 #FORMS LIGHT VERTICAL AND LEFT +0x89 U+252c #FORMS LIGHT DOWN AND HORIZONTAL +0x8a U+251c #FORMS LIGHT VERTICAL AND RIGHT +0x8b U+2534 #FORMS LIGHT UP AND HORIZONTAL +0x8c U+2510 #FORMS LIGHT DOWN AND LEFT +0x8d U+250c #FORMS LIGHT DOWN AND RIGHT +0x8e U+2514 #FORMS LIGHT UP AND RIGHT +0x8f U+2518 #FORMS LIGHT UP AND LEFT +0x90 U+03b2 #GREEK SMALL BETA +0x91 U+221e #INFINITY +0x92 U+03c6 #GREEK SMALL PHI +0x93 U+00b1 #PLUS-OR-MINUS SIGN +0x94 U+00bd #FRACTION 1/2 +0x95 U+00bc #FRACTION 1/4 +0x96 U+2248 #ALMOST EQUAL TO +0x97 U+00ab #LEFT POINTING GUILLEMET +0x98 U+00bb #RIGHT POINTING GUILLEMET +0x99 U+fef7 #ARABIC LIGATURE LAM WITH ALEF WITH HAMZA ABOVE ISOLATED FORM +0x9a U+fef8 #ARABIC LIGATURE LAM WITH ALEF WITH HAMZA ABOVE FINAL FORM +#0x9b #UNDEFINED +#0x9c #UNDEFINED +0x9d U+fefb #ARABIC LIGATURE LAM WITH ALEF ISOLATED FORM +0x9e U+fefc #ARABIC LIGATURE LAM WITH ALEF FINAL FORM +#0x9f #UNDEFINED +0xa0 U+00a0 #NON-BREAKING SPACE +0xa1 U+00ad #SOFT HYPHEN +0xa2 U+fe82 #ARABIC LETTER ALEF WITH MADDA ABOVE FINAL FORM +0xa3 U+00a3 #POUND SIGN +0xa4 U+00a4 #CURRENCY SIGN +0xa5 U+fe84 #ARABIC LETTER ALEF WITH HAMZA ABOVE FINAL FORM +#0xa6 #UNDEFINED +#0xa7 #UNDEFINED +0xa8 U+fe8e #ARABIC LETTER ALEF FINAL FORM +0xa9 U+fe8f #ARABIC LETTER BEH ISOLATED FORM +0xaa U+fe95 #ARABIC LETTER TEH ISOLATED FORM +0xab U+fe99 #ARABIC LETTER THEH ISOLATED FORM +0xac U+060c #ARABIC COMMA +0xad U+fe9d #ARABIC LETTER JEEM ISOLATED FORM +0xae U+fea1 #ARABIC LETTER HAH ISOLATED FORM +0xaf U+fea5 #ARABIC LETTER KHAH ISOLATED FORM +0xb0 U+0660 #ARABIC-INDIC DIGIT ZERO +0xb1 U+0661 #ARABIC-INDIC DIGIT ONE +0xb2 U+0662 #ARABIC-INDIC DIGIT TWO +0xb3 U+0663 #ARABIC-INDIC DIGIT THREE +0xb4 U+0664 #ARABIC-INDIC DIGIT FOUR +0xb5 U+0665 #ARABIC-INDIC DIGIT FIVE +0xb6 U+0666 #ARABIC-INDIC DIGIT SIX +0xb7 U+0667 #ARABIC-INDIC DIGIT SEVEN +0xb8 U+0668 #ARABIC-INDIC DIGIT EIGHT +0xb9 U+0669 #ARABIC-INDIC DIGIT NINE +0xba U+fed1 #ARABIC LETTER FEH ISOLATED FORM +0xbb U+061b #ARABIC SEMICOLON +0xbc U+feb1 #ARABIC LETTER SEEN ISOLATED FORM +0xbd U+feb5 #ARABIC LETTER SHEEN ISOLATED FORM +0xbe U+feb9 #ARABIC LETTER SAD ISOLATED FORM +0xbf U+061f #ARABIC QUESTION MARK +0xc0 U+00a2 #CENT SIGN +0xc1 U+fe80 #ARABIC LETTER HAMZA ISOLATED FORM +0xc2 U+fe81 #ARABIC LETTER ALEF WITH MADDA ABOVE ISOLATED FORM +0xc3 U+fe83 #ARABIC LETTER ALEF WITH HAMZA ABOVE ISOLATED FORM +0xc4 U+fe85 #ARABIC LETTER WAW WITH HAMZA ABOVE ISOLATED FORM +0xc5 U+feca #ARABIC LETTER AIN FINAL FORM +0xc6 U+fe8b #ARABIC LETTER YEH WITH HAMZA ABOVE INITIAL FORM +0xc7 U+fe8d #ARABIC LETTER ALEF ISOLATED FORM +0xc8 U+fe91 #ARABIC LETTER BEH INITIAL FORM +0xc9 U+fe93 #ARABIC LETTER TEH MARBUTA ISOLATED FORM +0xca U+fe97 #ARABIC LETTER TEH INITIAL FORM +0xcb U+fe9b #ARABIC LETTER THEH INITIAL FORM +0xcc U+fe9f #ARABIC LETTER JEEM INITIAL FORM +0xcd U+fea3 #ARABIC LETTER HAH INITIAL FORM +0xce U+fea7 #ARABIC LETTER KHAH INITIAL FORM +0xcf U+fea9 #ARABIC LETTER DAL ISOLATED FORM +0xd0 U+feab #ARABIC LETTER THAL ISOLATED FORM +0xd1 U+fead #ARABIC LETTER REH ISOLATED FORM +0xd2 U+feaf #ARABIC LETTER ZAIN ISOLATED FORM +0xd3 U+feb3 #ARABIC LETTER SEEN INITIAL FORM +0xd4 U+feb7 #ARABIC LETTER SHEEN INITIAL FORM +0xd5 U+febb #ARABIC LETTER SAD INITIAL FORM +0xd6 U+febf #ARABIC LETTER DAD INITIAL FORM +0xd7 U+fec1 #ARABIC LETTER TAH ISOLATED FORM +0xd8 U+fec5 #ARABIC LETTER ZAH ISOLATED FORM +0xd9 U+fecb #ARABIC LETTER AIN INITIAL FORM +0xda U+fecf #ARABIC LETTER GHAIN INITIAL FORM +0xdb U+00a6 #BROKEN VERTICAL BAR +0xdc U+00ac #NOT SIGN +0xdd U+00f7 #DIVISION SIGN +0xde U+00d7 #MULTIPLICATION SIGN +0xdf U+fec9 #ARABIC LETTER AIN ISOLATED FORM +0xe0 U+0640 #ARABIC TATWEEL +0xe1 U+fed3 #ARABIC LETTER FEH INITIAL FORM +0xe2 U+fed7 #ARABIC LETTER QAF INITIAL FORM +0xe3 U+fedb #ARABIC LETTER KAF INITIAL FORM +0xe4 U+fedf #ARABIC LETTER LAM INITIAL FORM +0xe5 U+fee3 #ARABIC LETTER MEEM INITIAL FORM +0xe6 U+fee7 #ARABIC LETTER NOON INITIAL FORM +0xe7 U+feeb #ARABIC LETTER HEH INITIAL FORM +0xe8 U+feed #ARABIC LETTER WAW ISOLATED FORM +0xe9 U+feef #ARABIC LETTER ALEF MAKSURA ISOLATED FORM +0xea U+fef3 #ARABIC LETTER YEH INITIAL FORM +0xeb U+febd #ARABIC LETTER DAD ISOLATED FORM +0xec U+fecc #ARABIC LETTER AIN MEDIAL FORM +0xed U+fece #ARABIC LETTER GHAIN FINAL FORM +0xee U+fecd #ARABIC LETTER GHAIN ISOLATED FORM +0xef U+fee1 #ARABIC LETTER MEEM ISOLATED FORM +0xf0 U+fe7d #ARABIC SHADDA MEDIAL FORM +0xf1 U+0651 #ARABIC SHADDAH +0xf2 U+fee5 #ARABIC LETTER NOON ISOLATED FORM +0xf3 U+fee9 #ARABIC LETTER HEH ISOLATED FORM +0xf4 U+feec #ARABIC LETTER HEH MEDIAL FORM +0xf5 U+fef0 #ARABIC LETTER ALEF MAKSURA FINAL FORM +0xf6 U+fef2 #ARABIC LETTER YEH FINAL FORM +0xf7 U+fed0 #ARABIC LETTER GHAIN MEDIAL FORM +0xf8 U+fed5 #ARABIC LETTER QAF ISOLATED FORM +0xf9 U+fef5 #ARABIC LIGATURE LAM WITH ALEF WITH MADDA ABOVE ISOLATED FORM +0xfa U+fef6 #ARABIC LIGATURE LAM WITH ALEF WITH MADDA ABOVE FINAL FORM +0xfb U+fedd #ARABIC LETTER LAM ISOLATED FORM +0xfc U+fed9 #ARABIC LETTER KAF ISOLATED FORM +0xfd U+fef1 #ARABIC LETTER YEH ISOLATED FORM +0xfe U+25a0 #BLACK SQUARE +#0xff #UNDEFINED + diff --git a/src/chrtrans/cp866_uni.tbl b/src/chrtrans/cp866_uni.tbl new file mode 100644 index 0000000..029e025 --- /dev/null +++ b/src/chrtrans/cp866_uni.tbl @@ -0,0 +1,159 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mcp866 + +#Name as a Display Charset (used on Options screen) +OCyrillic (cp866) + +#Codepage number +C866 + +# +# Name: cp866_DOSCyrillicRussian to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp866_DOSCyrillicRussian code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp866_DOSCyrillicRussian order +# +0x20-0x7f idem +# +0x80 U+0410 #CYRILLIC CAPITAL LETTER A +0x81 U+0411 #CYRILLIC CAPITAL LETTER BE +0x82 U+0412 #CYRILLIC CAPITAL LETTER VE +0x83 U+0413 #CYRILLIC CAPITAL LETTER GHE +0x84 U+0414 #CYRILLIC CAPITAL LETTER DE +0x85 U+0415 #CYRILLIC CAPITAL LETTER IE +0x86 U+0416 #CYRILLIC CAPITAL LETTER ZHE +0x87 U+0417 #CYRILLIC CAPITAL LETTER ZE +0x88 U+0418 #CYRILLIC CAPITAL LETTER I +0x89 U+0419 #CYRILLIC CAPITAL LETTER SHORT I +0x8a U+041a #CYRILLIC CAPITAL LETTER KA +0x8b U+041b #CYRILLIC CAPITAL LETTER EL +0x8c U+041c #CYRILLIC CAPITAL LETTER EM +0x8d U+041d #CYRILLIC CAPITAL LETTER EN +0x8e U+041e #CYRILLIC CAPITAL LETTER O +0x8f U+041f #CYRILLIC CAPITAL LETTER PE +0x90 U+0420 #CYRILLIC CAPITAL LETTER ER +0x91 U+0421 #CYRILLIC CAPITAL LETTER ES +0x92 U+0422 #CYRILLIC CAPITAL LETTER TE +0x93 U+0423 #CYRILLIC CAPITAL LETTER U +0x94 U+0424 #CYRILLIC CAPITAL LETTER EF +0x95 U+0425 #CYRILLIC CAPITAL LETTER HA +0x96 U+0426 #CYRILLIC CAPITAL LETTER TSE +0x97 U+0427 #CYRILLIC CAPITAL LETTER CHE +0x98 U+0428 #CYRILLIC CAPITAL LETTER SHA +0x99 U+0429 #CYRILLIC CAPITAL LETTER SHCHA +0x9a U+042a #CYRILLIC CAPITAL LETTER HARD SIGN +0x9b U+042b #CYRILLIC CAPITAL LETTER YERU +0x9c U+042c #CYRILLIC CAPITAL LETTER SOFT SIGN +0x9d U+042d #CYRILLIC CAPITAL LETTER E +0x9e U+042e #CYRILLIC CAPITAL LETTER YU +0x9f U+042f #CYRILLIC CAPITAL LETTER YA +0xa0 U+0430 #CYRILLIC SMALL LETTER A +0xa1 U+0431 #CYRILLIC SMALL LETTER BE +0xa2 U+0432 #CYRILLIC SMALL LETTER VE +0xa3 U+0433 #CYRILLIC SMALL LETTER GHE +0xa4 U+0434 #CYRILLIC SMALL LETTER DE +0xa5 U+0435 #CYRILLIC SMALL LETTER IE +0xa6 U+0436 #CYRILLIC SMALL LETTER ZHE +0xa7 U+0437 #CYRILLIC SMALL LETTER ZE +0xa8 U+0438 #CYRILLIC SMALL LETTER I +0xa9 U+0439 #CYRILLIC SMALL LETTER SHORT I +0xaa U+043a #CYRILLIC SMALL LETTER KA +0xab U+043b #CYRILLIC SMALL LETTER EL +0xac U+043c #CYRILLIC SMALL LETTER EM +0xad U+043d #CYRILLIC SMALL LETTER EN +0xae U+043e #CYRILLIC SMALL LETTER O +0xaf U+043f #CYRILLIC SMALL LETTER PE +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+2561 #BOX DRAWINGS VERTICAL SINGLE AND LEFT DOUBLE +0xb6 U+2562 #BOX DRAWINGS VERTICAL DOUBLE AND LEFT SINGLE +0xb7 U+2556 #BOX DRAWINGS DOWN DOUBLE AND LEFT SINGLE +0xb8 U+2555 #BOX DRAWINGS DOWN SINGLE AND LEFT DOUBLE +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+255c #BOX DRAWINGS UP DOUBLE AND LEFT SINGLE +0xbe U+255b #BOX DRAWINGS UP SINGLE AND LEFT DOUBLE +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+255e #BOX DRAWINGS VERTICAL SINGLE AND RIGHT DOUBLE +0xc7 U+255f #BOX DRAWINGS VERTICAL DOUBLE AND RIGHT SINGLE +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+2567 #BOX DRAWINGS UP SINGLE AND HORIZONTAL DOUBLE +0xd0 U+2568 #BOX DRAWINGS UP DOUBLE AND HORIZONTAL SINGLE +0xd1 U+2564 #BOX DRAWINGS DOWN SINGLE AND HORIZONTAL DOUBLE +0xd2 U+2565 #BOX DRAWINGS DOWN DOUBLE AND HORIZONTAL SINGLE +0xd3 U+2559 #BOX DRAWINGS UP DOUBLE AND RIGHT SINGLE +0xd4 U+2558 #BOX DRAWINGS UP SINGLE AND RIGHT DOUBLE +0xd5 U+2552 #BOX DRAWINGS DOWN SINGLE AND RIGHT DOUBLE +0xd6 U+2553 #BOX DRAWINGS DOWN DOUBLE AND RIGHT SINGLE +0xd7 U+256b #BOX DRAWINGS VERTICAL DOUBLE AND HORIZONTAL SINGLE +0xd8 U+256a #BOX DRAWINGS VERTICAL SINGLE AND HORIZONTAL DOUBLE +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+258c #LEFT HALF BLOCK +0xde U+2590 #RIGHT HALF BLOCK +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+0440 #CYRILLIC SMALL LETTER ER +0xe1 U+0441 #CYRILLIC SMALL LETTER ES +0xe2 U+0442 #CYRILLIC SMALL LETTER TE +0xe3 U+0443 #CYRILLIC SMALL LETTER U +0xe4 U+0444 #CYRILLIC SMALL LETTER EF +0xe5 U+0445 #CYRILLIC SMALL LETTER HA +0xe6 U+0446 #CYRILLIC SMALL LETTER TSE +0xe7 U+0447 #CYRILLIC SMALL LETTER CHE +0xe8 U+0448 #CYRILLIC SMALL LETTER SHA +0xe9 U+0449 #CYRILLIC SMALL LETTER SHCHA +0xea U+044a #CYRILLIC SMALL LETTER HARD SIGN +0xeb U+044b #CYRILLIC SMALL LETTER YERU +0xec U+044c #CYRILLIC SMALL LETTER SOFT SIGN +0xed U+044d #CYRILLIC SMALL LETTER E +0xee U+044e #CYRILLIC SMALL LETTER YU +0xef U+044f #CYRILLIC SMALL LETTER YA +0xf0 U+0401 #CYRILLIC CAPITAL LETTER IO +0xf1 U+0451 #CYRILLIC SMALL LETTER IO +0xf2 U+0404 #CYRILLIC CAPITAL LETTER UKRAINIAN IE +0xf3 U+0454 #CYRILLIC SMALL LETTER UKRAINIAN IE +0xf4 U+0407 #CYRILLIC CAPITAL LETTER YI +0xf5 U+0457 #CYRILLIC SMALL LETTER YI +0xf6 U+040e #CYRILLIC CAPITAL LETTER SHORT U +0xf7 U+045e #CYRILLIC SMALL LETTER SHORT U +0xf8 U+00b0 #DEGREE SIGN +0xf9 U+2219 #BULLET OPERATOR +0xfa U+00b7 #MIDDLE DOT +0xfb U+221a #SQUARE ROOT +0xfc U+2116 #NUMERO SIGN +0xfd U+00a4 #CURRENCY SIGN +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE + diff --git a/src/chrtrans/cp866u_uni.tbl b/src/chrtrans/cp866u_uni.tbl new file mode 100644 index 0000000..8d2dee0 --- /dev/null +++ b/src/chrtrans/cp866u_uni.tbl @@ -0,0 +1,157 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mcp866u + +#Name as a Display Charset (used on Options screen) +OUkrainian Cyrillic (cp866u) + +#Codepage number +#? + +# +# Name: cp866_DOSCyrillicUkrainian to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# General notes: based on Cyrillic (cp866) table, +# have different mapping in 0xF2-0xF9 region. +# +# Format: Three tab-separated columns +# Column #1 is the cp866_DOSCyrillicUkrainian code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp866_DOSCyrillicUkrainian order +# +0x20-0x7f idem +# +0x80 U+0410 #CYRILLIC CAPITAL LETTER A +0x81 U+0411 #CYRILLIC CAPITAL LETTER BE +0x82 U+0412 #CYRILLIC CAPITAL LETTER VE +0x83 U+0413 #CYRILLIC CAPITAL LETTER GHE +0x84 U+0414 #CYRILLIC CAPITAL LETTER DE +0x85 U+0415 #CYRILLIC CAPITAL LETTER IE +0x86 U+0416 #CYRILLIC CAPITAL LETTER ZHE +0x87 U+0417 #CYRILLIC CAPITAL LETTER ZE +0x88 U+0418 #CYRILLIC CAPITAL LETTER I +0x89 U+0419 #CYRILLIC CAPITAL LETTER SHORT I +0x8a U+041a #CYRILLIC CAPITAL LETTER KA +0x8b U+041b #CYRILLIC CAPITAL LETTER EL +0x8c U+041c #CYRILLIC CAPITAL LETTER EM +0x8d U+041d #CYRILLIC CAPITAL LETTER EN +0x8e U+041e #CYRILLIC CAPITAL LETTER O +0x8f U+041f #CYRILLIC CAPITAL LETTER PE +0x90 U+0420 #CYRILLIC CAPITAL LETTER ER +0x91 U+0421 #CYRILLIC CAPITAL LETTER ES +0x92 U+0422 #CYRILLIC CAPITAL LETTER TE +0x93 U+0423 #CYRILLIC CAPITAL LETTER U +0x94 U+0424 #CYRILLIC CAPITAL LETTER EF +0x95 U+0425 #CYRILLIC CAPITAL LETTER HA +0x96 U+0426 #CYRILLIC CAPITAL LETTER TSE +0x97 U+0427 #CYRILLIC CAPITAL LETTER CHE +0x98 U+0428 #CYRILLIC CAPITAL LETTER SHA +0x99 U+0429 #CYRILLIC CAPITAL LETTER SHCHA +0x9a U+042a #CYRILLIC CAPITAL LETTER HARD SIGN +0x9b U+042b #CYRILLIC CAPITAL LETTER YERU +0x9c U+042c #CYRILLIC CAPITAL LETTER SOFT SIGN +0x9d U+042d #CYRILLIC CAPITAL LETTER E +0x9e U+042e #CYRILLIC CAPITAL LETTER YU +0x9f U+042f #CYRILLIC CAPITAL LETTER YA +0xa0 U+0430 #CYRILLIC SMALL LETTER A +0xa1 U+0431 #CYRILLIC SMALL LETTER BE +0xa2 U+0432 #CYRILLIC SMALL LETTER VE +0xa3 U+0433 #CYRILLIC SMALL LETTER GHE +0xa4 U+0434 #CYRILLIC SMALL LETTER DE +0xa5 U+0435 #CYRILLIC SMALL LETTER IE +0xa6 U+0436 #CYRILLIC SMALL LETTER ZHE +0xa7 U+0437 #CYRILLIC SMALL LETTER ZE +0xa8 U+0438 #CYRILLIC SMALL LETTER I +0xa9 U+0439 #CYRILLIC SMALL LETTER SHORT I +0xaa U+043a #CYRILLIC SMALL LETTER KA +0xab U+043b #CYRILLIC SMALL LETTER EL +0xac U+043c #CYRILLIC SMALL LETTER EM +0xad U+043d #CYRILLIC SMALL LETTER EN +0xae U+043e #CYRILLIC SMALL LETTER O +0xaf U+043f #CYRILLIC SMALL LETTER PE +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+2561 #BOX DRAWINGS VERTICAL SINGLE AND LEFT DOUBLE +0xb6 U+2562 #BOX DRAWINGS VERTICAL DOUBLE AND LEFT SINGLE +0xb7 U+2556 #BOX DRAWINGS DOWN DOUBLE AND LEFT SINGLE +0xb8 U+2555 #BOX DRAWINGS DOWN SINGLE AND LEFT DOUBLE +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+255c #BOX DRAWINGS UP DOUBLE AND LEFT SINGLE +0xbe U+255b #BOX DRAWINGS UP SINGLE AND LEFT DOUBLE +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+255e #BOX DRAWINGS VERTICAL SINGLE AND RIGHT DOUBLE +0xc7 U+255f #BOX DRAWINGS VERTICAL DOUBLE AND RIGHT SINGLE +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+2567 #BOX DRAWINGS UP SINGLE AND HORIZONTAL DOUBLE +0xd0 U+2568 #BOX DRAWINGS UP DOUBLE AND HORIZONTAL SINGLE +0xd1 U+2564 #BOX DRAWINGS DOWN SINGLE AND HORIZONTAL DOUBLE +0xd2 U+2565 #BOX DRAWINGS DOWN DOUBLE AND HORIZONTAL SINGLE +0xd3 U+2559 #BOX DRAWINGS UP DOUBLE AND RIGHT SINGLE +0xd4 U+2558 #BOX DRAWINGS UP SINGLE AND RIGHT DOUBLE +0xd5 U+2552 #BOX DRAWINGS DOWN SINGLE AND RIGHT DOUBLE +0xd6 U+2553 #BOX DRAWINGS DOWN DOUBLE AND RIGHT SINGLE +0xd7 U+256b #BOX DRAWINGS VERTICAL DOUBLE AND HORIZONTAL SINGLE +0xd8 U+256a #BOX DRAWINGS VERTICAL SINGLE AND HORIZONTAL DOUBLE +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+258c #LEFT HALF BLOCK +0xde U+2590 #RIGHT HALF BLOCK +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+0440 #CYRILLIC SMALL LETTER ER +0xe1 U+0441 #CYRILLIC SMALL LETTER ES +0xe2 U+0442 #CYRILLIC SMALL LETTER TE +0xe3 U+0443 #CYRILLIC SMALL LETTER U +0xe4 U+0444 #CYRILLIC SMALL LETTER EF +0xe5 U+0445 #CYRILLIC SMALL LETTER HA +0xe6 U+0446 #CYRILLIC SMALL LETTER TSE +0xe7 U+0447 #CYRILLIC SMALL LETTER CHE +0xe8 U+0448 #CYRILLIC SMALL LETTER SHA +0xe9 U+0449 #CYRILLIC SMALL LETTER SHCHA +0xea U+044a #CYRILLIC SMALL LETTER HARD SIGN +0xeb U+044b #CYRILLIC SMALL LETTER YERU +0xec U+044c #CYRILLIC SMALL LETTER SOFT SIGN +0xed U+044d #CYRILLIC SMALL LETTER E +0xee U+044e #CYRILLIC SMALL LETTER YU +0xef U+044f #CYRILLIC SMALL LETTER YA +0xf0 U+0401 #CYRILLIC CAPITAL LETTER IO +0xf1 U+0451 #CYRILLIC SMALL LETTER IO +0xf2 U+0490 #CYRILLIC CAPITAL LETTER GHE WITH UPTURN +0xf3 U+0491 #CYRILLIC SMALL LETTER GHE WITH UPTURN +0xf4 U+0404 #CYRILLIC CAPITAL LETTER UKRAINIAN IE +0xf5 U+0454 #CYRILLIC SMALL LETTER UKRAINIAN IE +0xf6 U+0406 #CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I +0xf7 U+0456 #CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I +0xf8 U+0407 #CYRILLIC CAPITAL LETTER YI +0xf9 U+0457 #CYRILLIC SMALL LETTER YI +0xfa U+00b7 #MIDDLE DOT +0xfb U+221a #SQUARE ROOT +0xfc U+2116 #NUMERO SIGN +0xfd U+00a4 #CURRENCY SIGN +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE + diff --git a/src/chrtrans/cp869_uni.tbl b/src/chrtrans/cp869_uni.tbl new file mode 100644 index 0000000..d662b83 --- /dev/null +++ b/src/chrtrans/cp869_uni.tbl @@ -0,0 +1,160 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mcp869 + +#Name as a Display Charset (used on Options screen) +OGreek2 (cp869) + +#Codepage number +C869 + +# Name: cp869_DOSGreek2 to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp869_DOSGreek2 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp869_DOSGreek2 order +# +################## + +0x20-0x7f idem +# +#0x80 #UNDEFINED +#0x81 #UNDEFINED +#0x82 #UNDEFINED +#0x83 #UNDEFINED +#0x84 #UNDEFINED +#0x85 #UNDEFINED +0x86 U+0386 #GREEK CAPITAL LETTER ALPHA WITH TONOS +#0x87 #UNDEFINED +0x88 U+00b7 #MIDDLE DOT +0x89 U+00ac #NOT SIGN +0x8a U+00a6 #BROKEN BAR +0x8b U+2018 U+02bd #LEFT SINGLE QUOTATION MARK +0x8c U+2019 U+02bc #RIGHT SINGLE QUOTATION MARK +0x8d U+0388 #GREEK CAPITAL LETTER EPSILON WITH TONOS +0x8e U+2015 #HORIZONTAL BAR +0x8f U+0389 #GREEK CAPITAL LETTER ETA WITH TONOS +0x90 U+038a #GREEK CAPITAL LETTER IOTA WITH TONOS +0x91 U+03aa #GREEK CAPITAL LETTER IOTA WITH DIALYTIKA +0x92 U+038c #GREEK CAPITAL LETTER OMICRON WITH TONOS +#0x93 #UNDEFINED +#0x94 #UNDEFINED +0x95 U+038e #GREEK CAPITAL LETTER UPSILON WITH TONOS +0x96 U+03ab #GREEK CAPITAL LETTER UPSILON WITH DIALYTIKA +0x97 U+00a9 #COPYRIGHT SIGN +0x98 U+038f #GREEK CAPITAL LETTER OMEGA WITH TONOS +0x99 U+00b2 #SUPERSCRIPT TWO +0x9a U+00b3 #SUPERSCRIPT THREE +0x9b U+03ac #GREEK SMALL LETTER ALPHA WITH TONOS +0x9c U+00a3 #POUND SIGN +0x9d U+03ad #GREEK SMALL LETTER EPSILON WITH TONOS +0x9e U+03ae #GREEK SMALL LETTER ETA WITH TONOS +0x9f U+03af #GREEK SMALL LETTER IOTA WITH TONOS +0xa0 U+03ca #GREEK SMALL LETTER IOTA WITH DIALYTIKA +0xa1 U+0390 #GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS +0xa2 U+03cc #GREEK SMALL LETTER OMICRON WITH TONOS +0xa3 U+03cd #GREEK SMALL LETTER UPSILON WITH TONOS +0xa4 U+0391 #GREEK CAPITAL LETTER ALPHA +0xa5 U+0392 #GREEK CAPITAL LETTER BETA +0xa6 U+0393 #GREEK CAPITAL LETTER GAMMA +0xa7 U+0394 #GREEK CAPITAL LETTER DELTA +0xa8 U+0395 #GREEK CAPITAL LETTER EPSILON +0xa9 U+0396 #GREEK CAPITAL LETTER ZETA +0xaa U+0397 #GREEK CAPITAL LETTER ETA +0xab U+00bd #VULGAR FRACTION ONE HALF +0xac U+0398 #GREEK CAPITAL LETTER THETA +0xad U+0399 #GREEK CAPITAL LETTER IOTA +0xae U+00ab #LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xaf U+00bb #RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xb0 U+2591 #LIGHT SHADE +0xb1 U+2592 #MEDIUM SHADE +0xb2 U+2593 #DARK SHADE +0xb3 U+2502 #BOX DRAWINGS LIGHT VERTICAL +0xb4 U+2524 #BOX DRAWINGS LIGHT VERTICAL AND LEFT +0xb5 U+039a #GREEK CAPITAL LETTER KAPPA +0xb6 U+039b #GREEK CAPITAL LETTER LAMDA +0xb7 U+039c #GREEK CAPITAL LETTER MU +0xb8 U+039d #GREEK CAPITAL LETTER NU +0xb9 U+2563 #BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xba U+2551 #BOX DRAWINGS DOUBLE VERTICAL +0xbb U+2557 #BOX DRAWINGS DOUBLE DOWN AND LEFT +0xbc U+255d #BOX DRAWINGS DOUBLE UP AND LEFT +0xbd U+039e #GREEK CAPITAL LETTER XI +0xbe U+039f #GREEK CAPITAL LETTER OMICRON +0xbf U+2510 #BOX DRAWINGS LIGHT DOWN AND LEFT +0xc0 U+2514 #BOX DRAWINGS LIGHT UP AND RIGHT +0xc1 U+2534 #BOX DRAWINGS LIGHT UP AND HORIZONTAL +0xc2 U+252c #BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0xc3 U+251c #BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0xc4 U+2500 #BOX DRAWINGS LIGHT HORIZONTAL +0xc5 U+253c #BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0xc6 U+03a0 #GREEK CAPITAL LETTER PI +0xc7 U+03a1 #GREEK CAPITAL LETTER RHO +0xc8 U+255a #BOX DRAWINGS DOUBLE UP AND RIGHT +0xc9 U+2554 #BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xca U+2569 #BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xcb U+2566 #BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xcc U+2560 #BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xcd U+2550 #BOX DRAWINGS DOUBLE HORIZONTAL +0xce U+256c #BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xcf U+03a3 #GREEK CAPITAL LETTER SIGMA +0xd0 U+03a4 #GREEK CAPITAL LETTER TAU +0xd1 U+03a5 #GREEK CAPITAL LETTER UPSILON +0xd2 U+03a6 #GREEK CAPITAL LETTER PHI +0xd3 U+03a7 #GREEK CAPITAL LETTER CHI +0xd4 U+03a8 #GREEK CAPITAL LETTER PSI +0xd5 U+03a9 #GREEK CAPITAL LETTER OMEGA +0xd6 U+03b1 #GREEK SMALL LETTER ALPHA +0xd7 U+03b2 #GREEK SMALL LETTER BETA +0xd8 U+03b3 #GREEK SMALL LETTER GAMMA +0xd9 U+2518 #BOX DRAWINGS LIGHT UP AND LEFT +0xda U+250c #BOX DRAWINGS LIGHT DOWN AND RIGHT +0xdb U+2588 #FULL BLOCK +0xdc U+2584 #LOWER HALF BLOCK +0xdd U+03b4 #GREEK SMALL LETTER DELTA +0xde U+03b5 #GREEK SMALL LETTER EPSILON +0xdf U+2580 #UPPER HALF BLOCK +0xe0 U+03b6 #GREEK SMALL LETTER ZETA +0xe1 U+03b7 #GREEK SMALL LETTER ETA +0xe2 U+03b8 #GREEK SMALL LETTER THETA +0xe3 U+03b9 #GREEK SMALL LETTER IOTA +0xe4 U+03ba #GREEK SMALL LETTER KAPPA +0xe5 U+03bb #GREEK SMALL LETTER LAMDA +0xe6 U+03bc #GREEK SMALL LETTER MU +0xe7 U+03bd #GREEK SMALL LETTER NU +0xe8 U+03be #GREEK SMALL LETTER XI +0xe9 U+03bf #GREEK SMALL LETTER OMICRON +0xea U+03c0 #GREEK SMALL LETTER PI +0xeb U+03c1 #GREEK SMALL LETTER RHO +0xec U+03c3 #GREEK SMALL LETTER SIGMA +0xed U+03c2 #GREEK SMALL LETTER FINAL SIGMA +0xee U+03c4 #GREEK SMALL LETTER TAU +0xef U+0384 #GREEK TONOS +0xf0 U+00ad #SOFT HYPHEN +0xf1 U+00b1 #PLUS-MINUS SIGN +0xf2 U+03c5 #GREEK SMALL LETTER UPSILON +0xf3 U+03c6 #GREEK SMALL LETTER PHI +0xf4 U+03c7 #GREEK SMALL LETTER CHI +0xf5 U+00a7 #SECTION SIGN +0xf6 U+03c8 #GREEK SMALL LETTER PSI +0xf7 U+0385 #GREEK DIALYTIKA TONOS +0xf8 U+00b0 #DEGREE SIGN +0xf9 U+00a8 #DIAERESIS +0xfa U+03c9 #GREEK SMALL LETTER OMEGA +0xfb U+03cb #GREEK SMALL LETTER UPSILON WITH DIALYTIKA +0xfc U+03b0 #GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS +0xfd U+03ce #GREEK SMALL LETTER OMEGA WITH TONOS +0xfe U+25a0 #BLACK SQUARE +0xff U+00a0 #NO-BREAK SPACE + diff --git a/src/chrtrans/def7_uni.tbl b/src/chrtrans/def7_uni.tbl new file mode 100644 index 0000000..14b7ac4 --- /dev/null +++ b/src/chrtrans/def7_uni.tbl @@ -0,0 +1,2951 @@ +# $LynxId: def7_uni.tbl,v 1.33 2020/01/21 22:21:19 tom Exp $ +# Default 7bit replacements. +# +# This table is very important and should not be excluded from the distribution +# since this is a default fallback for any 8bit user's "display character set" +# which (nearly) of 256 chars and could not map a rich Unicode repertoire. +# +# Note: there are a few rare replacement strings with trailing spaces +# which should be enclosed as C strings like "... " to make things obvious +# (and doing that we should escape \ as \134 and escape " as \" or \042 +# but this is really rare). +# + +#The MIME name of this charset. +Mus-ascii + +# Like any other charset this may be selected as "display character set": +#Name as a Display Charset (used on Options screen) +O7 bit approximations (US-ASCII) + +# Shall this become the "default" translation table? YES! +# There has to be exactly one table marked as "default". +D1 + + +# us-ascii characters should not normally pass here, +# they are always processed directly but let declare them here: +0x20-0x7e idem + + +# NO-BREAK SPACE, +# should not happen (processed in the code): +#U+00a0:NS +U+00a1:! +U+00a2:-c- +U+00a3:-L- +U+00a4:CUR +U+00a5:YEN +U+00a6:| +U+00a7:S: +U+00a8:" +# COPYRIGHT SIGN: +U+00a9:(c) +U+00aa:-a +U+00ab:<< +U+00ac:NOT +# SOFT HYPHEN, +# should not happen (processed in the code): +#U+00ad:- +# REGISTERED SIGN: +U+00ae:(R) +U+00af:- +U+00b0:DEG +U+00b1:+- +U+00b2:^2 +U+00b3:^3 +U+00b4:' +# My -> u +U+00b5:u +U+00b6:P: +U+00b7:. +U+00b8:, +U+00b9:^1 +U+00ba:-o +U+00bb:>> +U+00bc: 1/4 +U+00bd: 1/2 +U+00be: 3/4 +U+00bf:? +0x41 U+00c0-U+00c3 +U+00c4 "Ae" # Ä, not the best choice for some languages. +U+00c5:AA +U+00c6:AE +U+00c7:C, +0x45 U+00c8-U+00cb +0x49 U+00cc-U+00cf +U+00d0:D- +0x4e U+00d1 +0x4f U+00d2-U+00d5 +U+00d6 "Oe" # Ö, not the best choice for some languages. +U+00d7: * +U+00d8:O/ +0x55 U+00d9-U+00db +U+00dc "Ue" # Ü, not the best choice for some languages. +0x59 U+00dd +U+00de:TH +U+00df:ss +U+00e0:`a +0x61 U+00e1-U+00e3 +U+00e4 "ae" # ä, not the best choice for some languages. +U+00e5:aa +U+00e6:ae +U+00e7:c, +0x65 U+00e8-U+00eb +U+00ec:`i +0x69 U+00ed-U+00ef +U+00f0:d- +0x6e U+00f1 +0x6f U+00f2-U+00f5 +U+00f6 "oe" # ö, not the best choice for some languages. +U+00f7:-: +U+00f8:o/ +0x75 U+00f9-U+00fb +U+00fc "ue" # ü, not the best choice for some languages. +0x79 U+00fd +U+00fe:th +0x79 U+00ff +# end of latin-1 repertoire +0x41 U+0100 U+0102 U+0104 # A +0x61 U+0101 U+0103 U+0105 # a +0x43 U+0106 U+010a U+010c # C +U+0108 "Ch" +# The following line is an example for mapping several accented versions +# of small letter 'c' to 'c': +0x63 U+0107 U+010b U+010d # c +U+0109 "ch" +0x44 U+010e +0x64 U+010f +U+0110:D/ +U+0111:d/ +0x45 U+0112 U+0114 U+0116 U+0118 U+011a # E +0x65 U+0113 U+0115 U+0117 U+0119 U+011b # e +0x47 U+011e U+0120 U+0122 # G +0x67 U+011f U+0121 U+0123 # g +U+011C "Gh" +U+011D "gh" +U+0124 "Hh" +U+0125" "hh" +U+0126:H/ +0x48 U+0127 # LATIN SMALL LETTER H BAR -> H +0x49 U+0128 U+012a U+012c U+012e U+0130 # I +0x69 U+0129 U+012b U+012d U+012f U+0131 # i +U+0132:IJ +U+0133:ij +U+0134 "Jh" +U+0135 "jh" +0x4b U+0136 +0x6b U+0137 +U+0138:kk +0x4c U+0139 U+013b U+013d # L +0x6c U+013a U+013c U+013e # l +U+013f:L. +U+0140:l. +U+0141:L/ +U+0142:l/ +0x4e U+0143 U+0145 U+0147 # N +0x6e U+0144 U+0146 U+0148 # n +U+0149:'n +U+014a:NG +0x4e U+014B # LATIN SMALL LETTER ENG -> N +0x4f U+014c U+014e # O +0x6f U+014d U+014f # o +U+0150:O" +U+0151:o" +U+0152:OE +U+0153:oe +0x52 U+0154 U+0156 U+0158 # R +0x72 U+0155 U+0157 U+0159 # r +0x53 U+015a U+015e U+0160 U+0218 # S +0x73 U+015b U+015f U+0161 U+0219 # s +U+015C "Sh" +U+015d "sh" +0x54 U+0162 U+0164 U+021a # T +0x74 U+0163 U+0165 U+021b # t +U+0166:T/ +U+0167:t/ +0x55 U+0168 U+016a U+016c U+016e U+0172 # U +0x75 U+0169 U+016b U+016d U+016f U+0173 # u +U+0170:U" +U+0171:u" +0x57 U+0174 +0x77 U+0175 +0x59 U+0176 U+0178 +0x79 U+0177 +0x5a U+0179 U+017b U+017d U+021d +0x7a U+017a U+017c U+017e +U+017f:s1 +U+0187:C2 +U+0188:c2 +U+0191:F2 +U+0192: f +U+0198:K2 +U+0199:k2 +U+01a0:O9 +U+01a1:o9 +U+01a2:OI +U+01a3:oi +U+01a6:yr +U+01af:U9 +U+01b0:u9 +U+01b5:Z/ +U+01b6:z/ +U+01b7:ED +0x41 U+01cd +0x61 U+01ce +0x49 U+01cf +0x69 U+01d0 +0x4f U+01d1 +0x6f U+01d2 +0x55 U+01d3 +0x75 U+01d4 +U+01d5:U:- +U+01d6:u:- +U+01d7:U:' +U+01d8:u:' +U+01d9:U:< +U+01da:u:< +U+01db:U:! +U+01dc:u:! +U+01de:A1 +U+01df:a1 +U+01e0:A7 +U+01e1:a7 +U+01e2:A3 +U+01e3:a3 +U+01e4:G/ +U+01e5:g/ +0x47 U+01e6 +0x67 U+01e7 +0x4b U+01e8 +0x6b U+01e9 +0x4f U+01ea +0x6f U+01eb +U+01ec:O1 +U+01ed:o1 +U+01ee:EZ +U+01ef:ez +0x6a U+01f0 +0x47 U+01f4 +0x67 U+01f5 +U+01fa:AA' +U+01fb:aa' +U+01fc:AE' +U+01fd:ae' +U+01fe:O/' +U+01ff:o/' +U+0200:A!! +U+0201:a!! +U+0202:A) +U+0203:a) +U+0204:E!! +U+0205:e!! +U+0206:E) +U+0207:e) +U+0208:I!! +U+0209:i!! +U+020a:I) +U+020b:i) +U+020c:O!! +U+020d:o!! +U+020e:O) +U+020f:o) +U+0210:R!! +U+0211:r!! +U+0212:R) +U+0213:r) +U+0214:U!! +U+0215:u!! +U+0216:U) +U+0217:u) + +# IPA symbols, from +# Linkname: FAQ: Representing IPA Phonetics in ASCII +# URL: http://www.hpl.hp.com/personal/Evan_Kirshenbaum/IPA/faq.html +# (corrected in Russian Cyrillic area). +# (corrected in Greek area). +# +0x41 U+0251 # LATIN SMALL LETTER SCRIPT A -> A +U+0252:A. +U+0253:b` +0x4f U+0254 # LATIN SMALL LETTER OPEN O -> O +U+0256:d. +U+0257:d` +U+0258:@<umd> +0x40 U+0259 # LATIN SMALL LETTER SCHWA -> @ +0x52 U+025A # LATIN SMALL LETTER SCHWA HOOK -> R +0x45 U+025B # LATIN SMALL LETTER EPSILON -> E +U+025c:V" +U+025d:R<umd> +U+025e:O" +0x4a U+025F # LATIN SMALL LETTER DOTLESS J BAR -> J +U+0260:g` +0x67 U+0261 # LATIN SMALL LETTER SCRIPT G +0x47 U+0262 # LATIN LETTER SMALL CAPITAL G +0x51 U+0263 # LATIN SMALL LETTER GAMMA -> Q +U+0264:o- +U+0265:j<rnd> +U+0266:h<?> +U+0268:i" +0x49 U+026A U+0269 # LATIN LETTER SMALL CAPITAL I, LATIN SMALL LETTER IOTA +0x4c U+026B # LATIN SMALL LETTER L WITH MIDDLE TILDE +0x4c U+026C # LATIN SMALL LETTER L BELT +U+026d:l. +U+026e:z<lat> +U+026f:u- +U+0270:j<vel> +0x4d U+0271 # LATIN SMALL LETTER M HOOK +U+0273:n. +U+0274:n" +U+0275:@. +U+0276:&. +0x55 U+0277 # LATIN SMALL LETTER CLOSED OMEGA -> U +0x72 U+0279 # LATIN SMALL LETTER TURNED R -> r +U+027a:*<lat> +U+027b:r. +U+027d:*. +0x2a U+027E # LATIN SMALL LETTER FISHHOOK R -> * +U+0280:r" +0x52 U+0280 # LATIN LETTER SMALL CAPITAL R -> R +U+0281:g" +U+0282:s. +0x53 U+0283 # LATIN SMALL LETTER ESH -> S +U+0284:J` +U+0287:t! +U+0288:t. +U+0289:u" +0x55 U+028A # LATIN SMALL LETTER UPSILON -> U +U+028b:r<lbd> +0x56 U+028C # LATIN SMALL LETTER TURNED V -> V +U+028d:w<vls> +U+028e:l^ +U+028f:I. +U+0290:z. +U+0292:Z +0x3f U+0294 # LATIN SMALL LETTER GLOTTAL STOP -> ? +U+0295:H<vcd> +U+0296:l! +U+0297:c! +U+0298:p! +U+0299:b<trl> +U+029b:G` +0x6a U+029d # LATIN SMALL LETTER CROSSED-TAIL J +U+029e:k! +0x4c U+029F # LATIN LETTER SMALL CAPITAL L +U+02a0:q` +U+02a4:d3 +U+02a6:ts +U+02a7:tS +U+02b0:<h> +U+02b1:<?> +0x3b U+02b2 U+0321 +U+02b3:<r> +U+02b7:<w> +U+02bb:;S +0x27 U+02bc +0x27 U+02bd +U+02c6:^ +U+02c7:'< +U+02c8:| +U+02c9:1- +U+02cb:1! +0x3a U+02d0 +U+02d1 ":\\" +0x2b U+02d6 +0x2d U+02d7 +U+02d8:'( +U+02d9:'. +U+02da:'0 +U+02db:'; +U+02dc:~ +U+02dd:'" +U+02e5:_T +U+02e6:_H +U+02e7:_M +U+02e8:_L +U+02e9:_B +U+02ec:_v +U+02ee:'' +0x60 U+0300 +0x27 U+0301 +0x5e U+0302 +0x7e U+0303 U+0334 +U+030b:'' +0x7c U+030d +U+030e:|| +U+030f:`` +0x2e U+0322 U+0323 +U+0324:<?> +U+0325:<o> +0x2c U+0326 U+0327 +0x2d U+0329 +0x5b U+032a +U+032b:<w> +U+0334:<H> +0x2f U+0337 U+0338 +U+0340:` +U+0341:' +U+0342:~ +U+0344:'% +U+0345:j3 +U+0347:= +U+0360:~~ +U+0374:' +U+0375:, +U+037a:j3 +U+037e:?% +U+0384:'* +U+0385:'% +# Greek letters +U+0386:A' +U+0387:.* +U+0388:E' +U+0389:Y% +U+038a:I' +U+038c:O' +U+038e:U% +U+038f:W% +U+0390:i3 +U+0391:A +U+0392:B +U+0393:G +U+0394:D +U+0395:E +U+0396:Z +U+0397:Y +U+0398:TH +U+0399:I +U+039a:K +U+039b:L +U+039c:M +U+039d:N +U+039e:C +U+039f:O +U+03a0:P +U+03a1:R +U+03a3:S +U+03a4:T +U+03a5:U +U+03a6:F +U+03a7:X +U+03a8:Q +U+03a9:W* +U+03aa:J +U+03ab:V* +U+03ac:a' +U+03ad:e' +U+03ae:y% +U+03af:i' +U+03b0:u3 +U+03b1:a +U+03b2:b +U+03b3:g +U+03b4:d +U+03b5:e +U+03b6:z +U+03b7:y +U+03b8:th +U+03b9:i +U+03ba:k +U+03bb:l +U+03bc:m +U+03bd:n +U+03be:c +U+03bf:o +U+03c0:p +U+03c1:r +U+03c2:*s +U+03c3:s +U+03c4:t +U+03c5:u +U+03c6:f +U+03c7:x +U+03c8:q +U+03c9:w +U+03ca:j +U+03cb:v* +U+03cc:o' +U+03cd:u% +U+03ce:w% +# Greek symbols +U+03d0 "beta " +U+03d1 "theta " +U+03d2 "upsi " +U+03d5 "phi " +U+03d6 "pi " +U+03d7:k. +U+03da:T3 +U+03db:t3 +U+03dc:M3 +U+03dd:m3 +U+03de:K3 +U+03df:k3 +U+03e0:P3 +U+03e1:p3 +U+03f0 "kappa " +U+03f1 "rho " +U+03f3:J +U+03f4:'% +U+03f5:j3 +# Cyrillic capital letters +U+0402:D% +U+0403:G% +U+0404:IE +U+0405:DS +U+0406:II +U+0407:YI +U+0408:J% +U+0409:LJ +U+040a:NJ +U+040b:Ts +U+040c:KJ +U+040e:V% +U+040f:DZ +# Russian Cyrillic letters, transliterated +U+0401:IO +U+0410:A +U+0411:B +U+0412:V +U+0413:G +U+0414:D +U+0415:E +U+0416:ZH +U+0417:Z +U+0418:I +U+0419:J +U+041a:K +U+041b:L +U+041c:M +U+041d:N +U+041e:O +U+041f:P +U+0420:R +U+0421:S +U+0422:T +U+0423:U +U+0424:F +U+0425:H +U+0426:C +U+0427:CH +U+0428:SH +U+0429:SCH +U+042a:" +U+042b:Y +U+042c:' +U+042d:`E +U+042e:YU +U+042f:YA +U+0430:a +U+0431:b +U+0432:v +U+0433:g +U+0434:d +U+0435:e +U+0436:zh +U+0437:z +U+0438:i +U+0439:j +U+043a:k +U+043b:l +U+043c:m +U+043d:n +U+043e:o +U+043f:p +U+0440:r +U+0441:s +U+0442:t +U+0443:u +U+0444:f +U+0445:h +U+0446:c +U+0447:ch +U+0448:sh +U+0449:sch +U+044a:" +U+044b:y +U+044c:' +U+044d:`e +U+044e:yu +U+044f:ya +U+0451:io +# end of Russian Cyrillic letters. +# Cyrillic small letters (and some archaic) +U+0452:d% +U+0453:g% +U+0454:ie +U+0455:ds +U+0456:ii +U+0457:yi +U+0458:j% +U+0459:lj +U+045a:nj +U+045b:ts +U+045c:kj +U+045e:v% +U+045f:dz +U+0462:Y3 +U+0463:y3 +U+046a:O3 +U+046b:o3 +U+0472:F3 +U+0473:f3 +U+0474:V3 +U+0475:v3 +U+0480:C3 +U+0481:c3 +U+0490:G3 +U+0491:g3 +U+04d4:AE +U+04d5:ae +# These may make Yiddish slightly more readable, until we have +# something better. + +0x69 U+05b4 # i +0x61 U+05b7 # a +0x6f U+05b8 # o +0x75 U+05bc # u +0x68 U+05bf # h +0x3a U+05c2 # : + +0x76 U+05f0 # v +U+05f1:oy +U+05f2:ey + +# U+05d0:A+ +0x23 U+05d0 # '#' + +U+05d1:B+ +U+05d2:G+ +U+05d3:D+ +U+05d4:H+ +U+05d5:W+ +U+05d6:Z+ +U+05d7:X+ +U+05d8:Tj +U+05d9:J+ +U+05da:K% +U+05db:K+ +U+05dc:L+ +U+05dd:M% +U+05de:M+ +U+05df:N% +U+05e0:N+ +U+05e1:S+ +U+05e2:E+ +U+05e3:P% +U+05e4:P+ +U+05e5:Zj +U+05e6:ZJ +U+05e7:Q+ +U+05e8:R+ +U+05e9:Sh +U+05ea:T+ + +U+060c:,+ +U+061b:;+ +U+061f:?+ +U+0621:H' +U+0622:aM +U+0623:aH +U+0624:wH +U+0625:ah +U+0626:yH +U+0627:a+ +U+0628:b+ +U+0629:tm +U+062a:t+ +U+062b:tk +U+062c:g+ +U+062d:hk +U+062e:x+ +U+062f:d+ +U+0630:dk +U+0631:r+ +U+0632:z+ +U+0633:s+ +U+0634:sn +U+0635:c+ +U+0636:dd +U+0637:tj +U+0638:zH +U+0639:e+ +U+063a:i+ +U+0640:++ +U+0641:f+ +U+0642:q+ +U+0643:k+ +U+0644:l+ +U+0645:m+ +U+0646:n+ +U+0647:h+ +U+0648:w+ +U+0649:j+ +U+064a:y+ +U+064b::+ +U+064c:"+ +U+064d:=+ +U+064e:/+ +U+064f:'+ +U+0650:1+ +U+0651:3+ +U+0652:0+ +U+0660:0a +U+0661:1a +U+0662:2a +U+0663:3a +U+0664:4a +U+0665:5a +U+0666:6a +U+0667:7a +U+0668:8a +U+0669:9a +U+0670:aS +U+067e:p+ +U+0681:hH +U+0686:tc +U+0698:zj +U+06a4:v+ +U+06af:gf +U+06f0:0a +U+06f1:1a +U+06f2:2a +U+06f3:3a +U+06f4:4a +U+06f5:5a +U+06f6:6a +U+06f7:7a +U+06f8:8a +U+06f9:9a + +# Replacement strings for Ethiopic characters +U+1200:he +U+1201:hu +U+1202:hi +U+1203:ha +U+1204:hE +0x68 U+1205 #:h +U+1206:ho +U+1208:le +U+1209:lu +U+120A:li +U+120B:la +U+120C:lE +0x6c U+120D #:l +U+120E:lo +U+120F:lWa +U+1210:He +U+1211:Hu +U+1212:Hi +U+1213:Ha +U+1214:HE +0x48 U+1215 #:H +U+1216:Ho +U+1217:HWa +U+1218:me +U+1219:mu +U+121A:mi +U+121B:ma +U+121C:mE +0x6d U+121D #:m +U+121E:mo +U+121F:mWa +U+1220:`se +U+1221:`su +U+1222:`si +U+1223:`sa +U+1224:`sE +U+1225:`s +U+1226:`so +U+1227:`sWa +U+1228:re +U+1229:ru +U+122A:ri +U+122B:ra +U+122C:rE +0x72 U+122D #:r +U+122E:ro +U+122F:rWa +U+1230:se +U+1231:su +U+1232:si +U+1233:sa +U+1234:sE +0x73 U+1235 #:s +U+1236:so +U+1237:sWa +U+1238:xe +U+1239:xu +U+123A:xi +U+123B:xa +U+123C:xE +U+123D:xa +U+123E:xo +U+123F:xWa +U+1240:qe +U+1241:qu +U+1242:qi +U+1243:qa +U+1244:qE +0x71 U+1245 #:q +U+1246:qo +U+1248:qWe +U+124A:qWi +U+124B:qWa +U+124C:qWE +U+124D:qW +U+1250:Qe +U+1251:Qu +U+1252:Qi +U+1253:Qa +U+1254:QE +0x51 U+1255 #:Q +U+1256:Qo +U+1258:QWe +U+125A:QWi +U+125B:QWa +U+125C:QWE +U+125D:QW +U+1260:be +U+1261:bu +U+1262:bi +U+1263:ba +U+1264:bE +0x62 U+1265 #:b +U+1266:bo +U+1267:bWa +U+1268:ve +U+1269:vu +U+126A:vi +U+126B:va +U+126C:vE +0x76 U+126D #:v +U+126E:vo +U+126F:vWa +U+1270:te +U+1271:tu +U+1272:ti +U+1273:ta +U+1274:tE +0x74 U+1275 #:t +U+1276:to +U+1277:tWa +U+1278:ce +U+1279:cu +U+127A:ci +U+127B:ca +U+127C:cE +0x63 U+127D #:c +U+127E:co +U+127F:cWa +U+1280:`he +U+1281:`hu +U+1282:`hi +U+1283:`ha +U+1284:`hE +U+1285:`h +U+1286:`ho +U+1288:hWe +U+128A:hWi +U+128B:hWa +U+128C:hWE +U+128D:hW +U+1290:na +U+1291:nu +U+1292:ni +U+1293:na +U+1294:nE +0x6e U+1295 #:n +U+1296:no +U+1297:nWa +U+1298:Ne +U+1299:Nu +U+129A:Ni +U+129B:Na +U+129C:NE +0x4e U+129D #:N +U+129E:No +U+129F:NWa +0x65 U+12A0 #:e +0x75 U+12A1 #:u +0x69 U+12A2 #:i +0x61 U+12A3 #:a +0x45 U+12A4 #:E +0x49 U+12A5 #:I +0x6f U+12A6 #:o +U+12A7:e3 +U+12A8:ke +U+12A9:ku +U+12AA:ki +U+12AB:ka +U+12AC:kE +0x6b U+12AD #:k +U+12AE:ko +U+12B0:kWe +U+12B2:kWi +U+12B3:kWa +U+12B4:kWE +U+12B5:kW +U+12B8:Ke +U+12B9:Ku +U+12BA:Ki +U+12BB:Ka +U+12BC:KE +0x4b U+12BD #:K +U+12BE:Ko +U+12C0:KWe +U+12C2:KWi +U+12C3:KWa +U+12C4:KWE +U+12C5:KW +U+12C8:we +U+12C9:wu +U+12CA:wi +U+12CB:wa +U+12CC:wE +0x77 U+12CD #:w +U+12CE:wo +U+12D0:`e +U+12D1:`u +U+12D2:`i +U+12D3:`a +U+12D4:`E +U+12D5:`I +U+12D6:`o +U+12D8:ze +U+12D9:zu +U+12DA:zi +U+12DB:za +U+12DC:zE +0x7a U+12DD #:z +U+12DE:zo +U+12DF:zWa +U+12E0:Ze +U+12E1:Zu +U+12E2:Zi +U+12E3:Za +U+12E4:ZE +0x5a U+12E5 #:Z +U+12E6:Zo +U+12E7:ZWa +U+12E8:ye +U+12E9:yu +U+12EA:yi +U+12EB:ya +U+12EC:yE +0x79 U+12ED #:y +U+12EE:yo +U+12EF:yWa +U+12F0:de +U+12F1:du +U+12F2:di +U+12F3:da +U+12F4:dE +0x64 U+12F5 #:d +U+12F6:do +U+12F7:dWa +U+12F8:De +U+12F9:Du +U+12FA:Di +U+12FB:Da +U+12FC:DE +0x44 U+12FD #:D +U+12FE:Do +U+12FF:DWa +U+1300:je +U+1301:ju +U+1302:ji +U+1303:ja +U+1304:jE +0x6a U+1305 #:j +U+1306:jo +U+1307:jWa +U+1308:ga +U+1309:gu +U+130A:gi +U+130B:ga +U+130C:gE +0x67 U+130D #:g +U+130E:go +U+1310:gWu +U+1312:gWi +U+1313:gWa +U+1314:gWE +U+1315:gW +U+1318:Ge +U+1319:Gu +U+131A:Gi +U+131B:Ga +U+131C:GE +0x47 U+131D #:G +U+131E:Go +U+131F:GWa +U+1320:Te +U+1321:Tu +U+1322:Ti +U+1323:Ta +U+1324:TE +0x54 U+1325 #:T +U+1326:To +U+1327:TWa +U+1328:Ce +U+1329:Ca +U+132A:Cu +U+132B:Ca +U+132C:CE +0x43 U+132D #:C +U+132E:Co +U+132F:CWa +U+1330:Pe +U+1331:Pu +U+1332:Pi +U+1333:Pa +U+1334:PE +0x50 U+1335 #:P +U+1336:Po +U+1337:PWa +U+1338:SWe +U+1339:SWu +U+133A:SWi +U+133B:SWa +U+133C:SWE +U+133D:SW +U+133E:SWo +U+133F:SWa +U+1340:`Sa +U+1341:`Su +U+1342:`Si +U+1343:`Sa +U+1344:`SE +U+1345:`S +U+1346:`So +U+1348:fa +U+1349:fu +U+134A:fi +U+134B:fa +U+134C:fE +0x6f U+134D #:f +U+134E:fo +U+134F:fWa +U+1350:pe +U+1351:pu +U+1352:pi +U+1353:pa +U+1354:pE +0x70 U+1355 #:p +U+1356:po +U+1357:pWa +U+1358:mYa +U+1359:rYa +U+135A:fYa +# ETHIOPIC SPACE U+1360 mapped to ASCII space +0x20 U+1360 +0x3a U+1361 #:: +U+1362::: +0x2c U+1363 #:, +U+1364:; +U+1365:-: +U+1366::- +U+1367:`? +U+1368::|: +U+1369:`1 +U+136A:`2 +U+136B:`3 +U+136C:`4 +U+136D:`5 +U+136E:`6 +U+136F:`7 +U+1370:`8 +U+1371:`9 +U+1372:`10 +U+1373:`20 +U+1374:`30 +U+1375:`40 +U+1376:`50 +U+1377:`60 +U+1378:`70 +U+1379:`80 +U+137A:`90 +U+137B:`100 +U+137C:`10000 + + +U+1e00:A-0 +U+1e01:a-0 +U+1e02:B. +U+1e03:b. +U+1e04:B-. +U+1e05:b-. +U+1e06:B_ +U+1e07:b_ +U+1e08:C,' +U+1e09:c,' +U+1e0a:D. +U+1e0b:d. +U+1e0c:D-. +U+1e0d:d-. +U+1e0e:D_ +U+1e0f:d_ +U+1e10:D, +U+1e11:d, +U+1e12:D-> +U+1e13:d-> +U+1e14:E-! +U+1e15:e-! +U+1e16:E-' +U+1e17:e-' +U+1e18:E-> +U+1e19:e-> +U+1e1a:E-? +U+1e1b:e-? +U+1e1c:E,( +U+1e1d:e,( +U+1e1e:F. +U+1e1f:f. +U+1e20:G- +U+1e21:g- +U+1e22:H. +U+1e23:h. +U+1e24:H-. +U+1e25:h-. +U+1e26:H: +U+1e27:h: +U+1e28:H, +U+1e29:h, +U+1e2a:H-( +U+1e2b:h-( +U+1e2c:I-? +U+1e2d:i-? +U+1e2e:I:' +U+1e2f:i:' +U+1e30:K' +U+1e31:k' +U+1e32:K-. +U+1e33:k-. +U+1e34:K_ +U+1e35:k_ +U+1e36:L-. +U+1e37:l-. +U+1e38:L--. +U+1e39:l--. +U+1e3a:L_ +U+1e3b:l_ +U+1e3c:L-> +U+1e3d:l-> +U+1e3e:M' +U+1e3f:m' +U+1e40:M. +U+1e41:m. +U+1e42:M-. +U+1e43:m-. +U+1e44:N. +U+1e45:n. +U+1e46:N-. +U+1e47:n-. +U+1e48:N_ +U+1e49:n_ +U+1e4a:N-> +U+1e4b:n-> +U+1e4c:O?' +U+1e4d:o?' +U+1e4e:O?: +U+1e4f:o?: +U+1e50:O-! +U+1e51:o-! +U+1e52:O-' +U+1e53:o-' +U+1e54:P' +U+1e55:p' +U+1e56:P. +U+1e57:p. +U+1e58:R. +U+1e59:r. +U+1e5a:R-. +U+1e5b:r-. +U+1e5c:R--. +U+1e5d:r--. +U+1e5e:R_ +U+1e5f:r_ +U+1e60:S. +U+1e61:s. +U+1e62:S-. +U+1e63:s-. +U+1e64:S'. +U+1e65:s'. +U+1e66:S<. +U+1e67:s<. +U+1e68:S.-. +U+1e69:s.-. +U+1e6a:T. +U+1e6b:t. +U+1e6c:T-. +U+1e6d:t-. +U+1e6e:T_ +U+1e6f:t_ +U+1e70:T-> +U+1e71:t-> +U+1e72:U--: +U+1e73:u--: +U+1e74:U-? +U+1e75:u-? +U+1e76:U-> +U+1e77:u-> +U+1e78:U?' +U+1e79:u?' +U+1e7a:U-: +U+1e7b:u-: +U+1e7c:V? +U+1e7d:v? +U+1e7e:V-. +U+1e7f:v-. +U+1e80:W! +U+1e81:w! +U+1e82:W' +U+1e83:w' +U+1e84:W: +U+1e85:w: +U+1e86:W. +U+1e87:w. +U+1e88:W-. +U+1e89:w-. +U+1e8a:X. +U+1e8b:x. +U+1e8c:X: +U+1e8d:x: +U+1e8e:Y. +U+1e8f:y. +U+1e90:Z> +U+1e91:z> +U+1e92:Z-. +U+1e93:z-. +U+1e94:Z_ +U+1e95:z_ +U+1e96:h_ +U+1e97:t: +U+1e98:w0 +U+1e99:y0 +U+1ea0:A-. +U+1ea1:a-. +U+1ea2:A2 +U+1ea3:a2 +U+1ea4:A>' +U+1ea5:a>' +U+1ea6:A>! +U+1ea7:a>! +U+1ea8:A>2 +U+1ea9:a>2 +U+1eaa:A>? +U+1eab:a>? +U+1eac:A>-. +U+1ead:a>-. +U+1eae:A(' +U+1eaf:a(' +U+1eb0:A(! +U+1eb1:a(! +U+1eb2:A(2 +U+1eb3:a(2 +U+1eb4:A(? +U+1eb5:a(? +U+1eb6:A(-. +U+1eb7:a(-. +U+1eb8:E-. +U+1eb9:e-. +U+1eba:E2 +U+1ebb:e2 +U+1ebc:E? +U+1ebd:e? +U+1ebe:E>' +U+1ebf:e>' +U+1ec0:E>! +U+1ec1:e>! +U+1ec2:E>2 +U+1ec3:e>2 +U+1ec4:E>? +U+1ec5:e>? +U+1ec6:E>-. +U+1ec7:e>-. +U+1ec8:I2 +U+1ec9:i2 +U+1eca:I-. +U+1ecb:i-. +U+1ecc:O-. +U+1ecd:o-. +U+1ece:O2 +U+1ecf:o2 +U+1ed0:O>' +U+1ed1:o>' +U+1ed2:O>! +U+1ed3:o>! +U+1ed4:O>2 +U+1ed5:o>2 +U+1ed6:O>? +U+1ed7:o>? +U+1ed8:O>-. +U+1ed9:o>-. +U+1eda:O9' +U+1edb:o9' +U+1edc:O9! +U+1edd:o9! +U+1ede:O92 +U+1edf:o92 +U+1ee0:O9? +U+1ee1:o9? +U+1ee2:O9-. +U+1ee3:o9-. +U+1ee4:U-. +U+1ee5:u-. +U+1ee6:U2 +U+1ee7:u2 +U+1ee8:U9' +U+1ee9:u9' +U+1eea:U9! +U+1eeb:u9! +U+1eec:U92 +U+1eed:u92 +U+1eee:U9? +U+1eef:u9? +U+1ef0:U9-. +U+1ef1:u9-. +U+1ef2:Y! +U+1ef3:y! +U+1ef4:Y-. +U+1ef5:y-. +U+1ef6:Y2 +U+1ef7:y2 +U+1ef8:Y? +U+1ef9:y? +0x61 U+1f00 +U+1f01:ha +U+1f02:`a +U+1f03:h`a +U+1f04:a' +U+1f05:ha' +U+1f06:a~ +U+1f07:ha~ +0x41 U+1f08 +U+1f09:hA +U+1f0a:`A +U+1f0b:h`A +U+1f0c:A' +U+1f0d:hA' +U+1f0e:A~ +U+1f0f:hA~ +U+1f11:he +U+1f19:hE +U+1f31:hi +U+1f39:hI +U+1f41:ho +U+1f49:hO +U+1f51:hu +U+1f59:hU +U+1fbf:,, +U+1fc0:?* +U+1fc1:?: +U+1fcd:,! +U+1fce:,' +U+1fcf:?, +U+1fdd:;! +U+1fde:;' +U+1fdf:?; +U+1fe5:rh +U+1fec:Rh +U+1fed:!: +U+1fef:!* +U+1ffe:;; +# General punctuation: +0x20 U+2000 U+2002 U+2004-U+200A U+205F # spaces +U+2001 " " +U+2003 " " +U+200e:(->) +U+200f:(<-) +U+200a: +0x2d U+2010 U+2011 U+2013 U+2015 # hyphen-like +U+2014 "--" +U+2016:|| +U+2017:=2 +0x60 U+2018 # left single quotation mark <`> +0x27 U+2019-U+201b # various single quotation marks <'> +0x22 U+201c-U+201f # various double quotation marks <"> +U+2020:/- +U+2021:/= +U+2022 " o " +U+2023 " > " +0x2e U+2024 +U+2025:.. +U+2026:... +U+2027:. +U+2028 "\015" +U+2029 "\015\012" + +# Don't want to see these: +# POP DIRECTIONAL FORMATTING 202C +U+202c: +# LEFT-TO-RIGHT OVERRIDE 202D +U+202d: + +U+202b "" +U+202f "" +U+2030: 0/00 +U+2031: 0/000 +U+2032:' +U+2033:'' +U+2034:''' +U+2035:` +U+2036:`` +U+2037:``` +U+2038:^ +U+2039:< +U+203a:> +U+203b::X +U+203c:!! +U+203d:?! +U+203e:'- +U+2042:*** +U+2043 " - " +U+2044:/ +U+2045:[- +U+2046:-] +U+2047:?? +U+2048:?! +U+2049:!? +U+204b:|P +U+204e:* +U+2051:** +U+2052:./. +U+2053:~ +U+2056 " .: " +U+2057:'''' +U+2058 " .:. " +U+2059 " :.: " +U+205a " : " +U+205b " .:. " +U+205c ":+:" +U+2044:/ +U+2047:?? +U+2048:?! +U+2049:!? +# end of General punctuation. +U+2070:^0 +U+2074:^4 +U+2075:^5 +U+2076:^6 +U+2077:^7 +U+2078:^8 +U+2079:^9 +U+207a:^+ +U+207b:^- +U+207c:^= +U+207d:^( +U+207e:^) +U+207f:^n +U+2080:_0 +U+2081:_1 +U+2082:_2 +U+2083:_3 +U+2084:_4 +U+2085:_5 +U+2086:_6 +U+2087:_7 +U+2088:_8 +U+2089:_9 +U+208a:_+ +U+208b:_- +U+208c:_= +U+208d:( +U+208e:) +# Old euro currency sign glyph: +#U+20A0:CE +U+20a1:C// +U+20a2:Cr +U+20a3:Ff +U+20a4:Li +U+20a5:m/ +U+20a6:N= +U+20a7:Pt +U+20a8:Rs +U+20a9:W= +U+20aa:rJ +U+20ab:d_ +# New euro currency sign glyph: +U+20AC:EUR +U+20ad:K- +U+20ae:T// +U+20af:Dp +U+20b1:P= +U+20b2:G| +U+20b3:A= +U+20b5:C| +U+2100:a/c +U+2101:a/s +U+2103:oC +U+2104:CL +U+2105:c/o +U+2106:c/u +U+2109:oF +0x67 U+210a +0x68 U+210e +U+210f "\134hbar " +U+2111:Im +U+2113:l +U+2116:No. +U+2117:(P) +U+2118:P +U+211C:Re +U+211e:Rx +U+2120:(SM) +U+2121:TEL +# TRADE MARK SIGN: +U+2122:(TM) +U+2125:oz. +U+2126:Ohm +0x4b U+212A # Kelvin sign - K +U+212b:Ang. +U+212E:est. +0x6f U+2134 +U+2135 "Aleph " +U+2136 "Bet " +U+2137 "Gimel " +U+2138 "Dalet " +U+213B: FAX +U+2153: 1/3 +U+2154: 2/3 +U+2155: 1/5 +U+2156: 2/5 +U+2157: 3/5 +U+2158: 4/5 +U+2159: 1/6 +U+215a: 5/6 +U+215b: 1/8 +U+215c: 3/8 +U+215d: 5/8 +U+215e: 7/8 +U+215f: 1/ +U+2160:I +U+2161:II +U+2162:III +U+2163:IV +U+2164:V +U+2165:VI +U+2166:VII +U+2167:VIII +U+2168:IX +U+2169:X +U+216a:XI +U+216b:XII +U+216c:L +U+216d:C +U+216e:D +U+216f:M +U+2170:i +U+2171:ii +U+2172:iii +U+2173:iv +U+2174:v +U+2175:vi +U+2176:vii +U+2177:viii +U+2178:ix +U+2179:x +U+217a:xi +U+217b:xii +U+217c:l +U+217d:c +U+217e:d +U+217f:m +U+2180:1000RCD +U+2181:5000R +U+2182:10000R +# Arrows +U+2190:<- +U+2191 "^|" # upwards arrow "-^" +U+2192:-> +U+2193 "|v" # downwards arrow "-v" +U+2194:<-> +U+2195 "^|v" # up down arrow "UD" +U+2196:^\ +U+2197:/^ +U+2198:\v +U+2199:v/ +U+219a:</- +U+219b:-/> +U+219c:<~ +U+219d:~> +U+219e:<<- +U+219f:^^| +U+21a0:->> +U+21a1:|vv +U+21a2:<-< +U+21a3:>-> +U+21a4:<-| +U+21a5:^|_ +U+21a6:|-> +U+21a8 "^|v_" # up down arrow with base "UD-" +U+21ad:<~> +U+21ae:<-/-> +U+21af:Nv +U+21b0:<^| +U+21b1:|^> +U+21b2:<v| +U+21b3:|v> +U+21b4:-v +U+21B5:RET +U+21ba:u< +U+21bb:>u +U+21bc:<- +U+21bd:<- +U+21be:^| +U+21bf:^| +U+21c0:-> +U+21c1:-> +U+21c2:|v +U+21c3:|v +U+21c4:<=> +U+21c5:^||v +U+21c6:<=> +U+21c7:<<= +U+21c8:^|^| +U+21c9:=>> +U+21ca:|v|v +U+21cb:<=> +U+21cc:<=> +U+21cd:<=/= +U+21ce:<=/=> +U+21cf:=/=> +U+21d0:<= +U+21d1 "^||" # upwards double arrow "^^" +U+21d2:=> +U+21d3 "||v" # downwards double arrow "vv" +U+21d4:<=> +U+21d5:^||v +U+21d6:^\\ +U+21d7://^ +U+21d8:\\v +U+21d9:v// +U+21da:<-= +U+21db:=-> +U+21dc:<~ +U+21dd:~> +U+21de:^|=| +U+21df:|=|v +U+21e0:<- +U+21e1:^: +U+21e2:-> +U+21e3::v +U+21e4:|<- +U+21e5:->| +U+21e6:<- +U+21e7:^| +U+21e8:-> +U+21e9:|v +U+21ea:^! +U+21eb:^I +U+21ec:^-I +U+21ed:^|I +U+21ee:^^| +U+21ef:^^I +U+21f0:|-> +U+21f2:\v_| +U+21f3:^|v +U+21f4:-o> +U+21f5:|v^| +U+21f6:=->>> +U+21f7:<-|- +U+21f8:-|-> +U+21f9:<-|-> +U+21fa:<-||- +U+21fb:-||-> +U+21fc:<-||-> +U+21fd:<- +U+21fe:-> +U+21ff:<-> +U+2200:FA +U+2201:C +U+2202:\partial +U+2203:TE +U+2204:TDNE +U+2205:{} +U+2206:Delta +U+2207:Nabla +U+2208:(- +U+2209:!(- +U+220a:(- +U+220b:-) +U+220c:!-) +U+220d:-) +U+220e " qed" +U+220f:\prod +U+2211:\sum +U+2212:- +U+2213:-/+ +U+2214:.+ +0x2f U+2215 +U+2216 " - " +U+2217:* +U+2218:Ob +U+2219:sb +U+221a " SQRT " +U+221b " ROOT3 " +U+221c " ROOT4 " +U+221d:0( +U+221e:infty +U+221f:-L +U+2220:-V +U+2225:PP +U+2226 " !PP " +U+2227:AND +U+2228:OR +U+2229:(U +U+222a:)U +U+222b "\134int " +U+222c "\134int\134int " +U+222d "\134int\134int\134int " +U+222e:Io +U+2234:.: +U+2235::. +U+2236::R +U+2237::: +U+2238:.- +U+2239:-: +U+223c "?1" # Why not use "~" in this and following tilde-like characters? I'll assume someone more math-literate than I did this, and leave them alone. +U+223e:CG +U+2241:!~ +U+2242:-~ +U+2243:?- +U+2244:!~- +U+2245:?= +U+2246:~!= +U+2247:!~= +# ALMOST EQUAL TO: +U+2248:~= +U+2249 " !~= " +U+224b:~3 +U+224c:=? +U+2250:=... +U+2253:HI +U+2254::= +U+2255:=: +U+2260:!= +U+2261:=3 +U+2262 " !=3 " +U+2263:=4 +U+2264:<= +U+2265:>= +U+2266:.LE. +U+2267:.GE. +U+2268:.LT.NOT.EQ. +U+2269:.GT.NOT.EQ. +U+226a:<< +U+226b:>> +U+226e:!< +U+226f:!> +U+2270:!<= +U+2271:!>= +U+2272:<~ +U+2273:>~ +U+2274:!<~ +U+2275:!>~ +U+2276 " <> " +U+2277 " >< " +U+2278 " !<> " +U+2279 " !>< " +U+2282:(C +U+2283:)C +U+2284 " !(C " +U+2285 " !)C " +U+2286:(_ +U+2287:)_ +U+2288:!(_ +U+2289:!)_ +U+228a:(!_ +U+228b:)!_ +U+228f:[ +U+2290:] +U+2291:[_ +U+2292:]_ +U+2295 "(+)" # CIRCLED PLUS +U+2296 "(-)" # CIRCLED MINUS +U+2297 "(x)" # CIRCLED TIMES +U+2298 "(/)" # CIRCLED DIVISION SLASH +U+2299 "(.)" # CIRCLED DOT OPERATOR +U+229A "(o)" # CIRCLED RING OPERATOR +U+229B "(*)" # CIRCLED ASTERISK OPERATOR +U+229C "(=)" # CIRCLED EQUALS +U+229D "(-)" # CIRCLED DASH +U+229E "[+]" # SQUARED PLUS +U+229F "[-]" # SQUARED MINUS +U+22A0 "[x]" # SQUARED TIMES +U+22A1 "[.]" # SQUARED DOT OPERATOR +U+22a5:-T +U+22A7 " MODELS " # MODELS +U+22A8 " TRUE " # TRUE +U+22A9 " FORCES " # FORCES +U+22AC " !PROVES " # DOES NOT PROVE +U+22AD " NOT TRUE " # NOT TRUE +U+22AE " !FORCES " # DOES NOT FORCE +U+22B2 " NORMAL SUBGROUP OF " +U+22B3 " CONTAINS AS NORMAL SUBGROUP " +U+22B4 " NORMAL SUBGROUP OF OR EQUAL TO " +U+22B5 " CONTAINS AS NORMAL SUBGROUP OR EQUAL TO " +U+22B8 " MULTIMAP " # MULTIMAP +U+22BA " INTERCALATE " # INTERCALATE +U+22BB " XOR " # XOR +U+22BC " NAND " # NAND +U+22C5 " DOT " # DOT OPERATOR +U+22c6 " STAR " +U+22d6:<. +U+22d7:>. +U+22d8:<<< +U+22d9:>>> +U+22da:<=|> +U+22db:>=|< +U+22dc:=< +U+22dd:=> +U+22e2:![_ +U+22e3:!]_ +U+22e4:[!_ +U+22e5:]!_ +U+22e6:<!~ +U+22e7:>!~ +U+22ee::3 +U+22ef:.3 +U+2302:Eh +U+2303:^ +U+2304:v +U+2307:~~ +U+2308:<7 +U+2309:>7 +U+230a:7< +U+230b:7> +U+2310:NI +U+2312:(A +U+2315:TR +U+2318:88 +U+231a:(-/) +U+231b " >i< " +U+2320:Iu +U+2321:Il +U+2322::( +U+2323::) +U+2324:|^| +U+2325 " OPT " +U+2326:[X> +U+2327:[X] +U+2328:[kbd] +U+2329:</ +U+232a:/> +U+232b:<X] +U+2387 " ALT " +U+2397:<-pp +U+2398:pp-> +U+2399:[PrSc] +U+239a:[ClSc] +U+23ce " CR " +U+23cf:_^_ +U+23da:GROUND +U+2423:Vs +U+2440:1h +U+2441:3h +U+2442:2h +U+2443:4h +U+2446:1j +U+2447:2j +U+2448:3j +U+2449:4j +U+2460:1-o +U+2461:2-o +U+2462:3-o +U+2463:4-o +U+2464:5-o +U+2465:6-o +U+2466:7-o +U+2467:8-o +U+2468:9-o +U+2469:10-o +U+246a:11-o +U+246b:12-o +U+246c:13-o +U+246d:14-o +U+246e:15-o +U+246f:16-o +U+2470:17-o +U+2471:18-o +U+2472:19-o +U+2473:20-o +U+2474:(1) +U+2475:(2) +U+2476:(3) +U+2477:(4) +U+2478:(5) +U+2479:(6) +U+247a:(7) +U+247b:(8) +U+247c:(9) +U+247d:(10) +U+247e:(11) +U+247f:(12) +U+2480:(13) +U+2481:(14) +U+2482:(15) +U+2483:(16) +U+2484:(17) +U+2485:(18) +U+2486:(19) +U+2487:(20) +U+2488:1. +U+2489:2. +U+248a:3. +U+248b:4. +U+248c:5. +U+248d:6. +U+248e:7. +U+248f:8. +U+2490:9. +U+2491:10. +U+2492:11. +U+2493:12. +U+2494:13. +U+2495:14. +U+2496:15. +U+2497:16. +U+2498:17. +U+2499:18. +U+249a:19. +U+249b:20. +U+249c:(a) +U+249d:(b) +U+249e:(c) +U+249f:(d) +U+24a0:(e) +U+24a1:(f) +U+24a2:(g) +U+24a3:(h) +U+24a4:(i) +U+24a5:(j) +U+24a6:(k) +U+24a7:(l) +U+24a8:(m) +U+24a9:(n) +U+24aa:(o) +U+24ab:(p) +U+24ac:(q) +U+24ad:(r) +U+24ae:(s) +U+24af:(t) +U+24b0:(u) +U+24b1:(v) +U+24b2:(w) +U+24b3:(x) +U+24b4:(y) +U+24b5:(z) +U+24b6:A-o +U+24b7:B-o +U+24b8:C-o +U+24b9:D-o +U+24ba:E-o +U+24bb:F-o +U+24bc:G-o +U+24bd:H-o +U+24be:I-o +U+24bf:J-o +U+24c0:K-o +U+24c1:L-o +U+24c2:M-o +U+24c3:N-o +U+24c4:O-o +U+24c5:P-o +U+24c6:Q-o +U+24c7:R-o +U+24c8:S-o +U+24c9:T-o +U+24ca:U-o +U+24cb:V-o +U+24cc:W-o +U+24cd:X-o +U+24ce:Y-o +U+24cf:Z-o +U+24d0:a-o +U+24d1:b-o +U+24d2:c-o +U+24d3:d-o +U+24d4:e-o +U+24d5:f-o +U+24d6:g-o +U+24d7:h-o +U+24d8:i-o +U+24d9:j-o +U+24da:k-o +U+24db:l-o +U+24dc:m-o +U+24dd:n-o +U+24de:o-o +U+24df:p-o +U+24e0:q-o +U+24e1:r-o +U+24e2:s-o +U+24e3:t-o +U+24e4:u-o +U+24e5:v-o +U+24e6:w-o +U+24e7:x-o +U+24e8:y-o +U+24e9:z-o +U+24ea:0-o +U+2500:- +U+2501:= +U+2502:| +U+2503:| +U+2504:- +U+2505:= +U+2506:| +U+2507:| +U+2508:- +U+2509:= +U+250a:| +U+250b:| +0x2b U+250c-U+256c # box drawings, use + +U+2571:/ +U+2572:\ +U+2580:TB +U+2584:LB +U+2588:FB +U+258c:lB +U+2590:RB +U+2591:.S +U+2592::S +U+2593:?S +U+25a0:fS +U+25a1:OS +U+25a2:RO +U+25a3:Rr +U+25a4:RF +U+25a5:RY +U+25a6:RH +U+25a7:RZ +U+25a8:RK +U+25a9:RX +U+25aa:sB +U+25ac:SR +U+25ad:Or +U+25b2:^ +U+25b3:uT +U+25b6:|> +U+25b7:Tr +U+25ba:|> +U+25bc:v +U+25bd:dT +U+25c0:<| +U+25c1:Tl +U+25c4:<| +U+25c6:Db +U+25c7:Dw +U+25ca:LZ +U+25cb:0m +U+25ce:0o +U+25cf:0M +U+25d0:0L +U+25d1:0R +U+25d8:Sn +U+25d9:Ic +U+25e2:Fd +U+25e3:Bd +U+25ef:Ci +# Miscellaneous Symbols +U+2600 "SU" # cf. U+263c +U+2601:cOo +U+2602:J +U+2603:"8" +U+2605:* +U+2606:* +U+2607:<v +U+2608:Rv +U+2609 "Sol" # cf. astrological symbols U+263c - U+2647, star-like U+2600 +U+260a:Asc. +U+260b:Desc. +U+260c:Conj. +U+260d:Opp. +U+260e:TEL +U+260f:tel +0x58 U+2611 U+2612 # checked ballot boxes -> x +U+2613:X +U+2614 "\"J\"" # umbrella with rain drops, quote marks are part of the symbol +U+2615:Joe +U+261a:<== +U+261b:==> +U+261c:<-- +U+261d:||^ +U+261e:--> +U+261f:||v +U+2620 "!X!" # or "POISON ", cf. U+2621 +U+2621 "!Z!" # previously "CAUTION " - or is it better to leave it spelled out in English? +U+2622 "!R!" # or "RADIOACTIVE ", cf. U+2621 +U+2623 "!B!" # or "BIOHAZARD ", cf. U+2621 +U+2624 "2TS" # cf. U+2695 +U+2626:t +U+2627:XP +U+2628:t +U+2629:+ +U+262a:(* +U+262d:'\,) +U+262e:(PEACE) +U+262f:Pd +U+2630:-HVN- +U+2631:-LAK- +U+2632:-FIR- +U+2633:-THR- +U+2634:-WND- +U+2635:-WTR- +U+2636:-MTN- +U+2637:-RTH- +U+2638:* +U+2639::-( +U+263a::-) +U+263b:(-: +U+263c "su" # previously "SU" - cf. U+2600, typical to have "dark" character in uppercase, eg. U+260e, U+260f +U+263d "Lun1" # Luna, 1st quarter +U+263e "Lun3" # Luna, 3rd quarter +U+263f:Mer +U+2640 "Ven" # previously "f." - this section labeled as astrological symbols +U+2641 "Ter" # Terra, to go with other Latin names +U+2642 "Mar" # previously "m." - this section labeled as astrological symbols +U+2643:Jup +U+2644:Sat +U+2645:Ura +U+2646:Nep +U+2647:Plu +U+2648 "Ari" # Standard astronomical abbreviation +U+2649 "Tau" +U+264a "Gem" +U+264b "Cnc" +U+264c "Leo" +U+264d "Vir" +U+264e "Lib" +U+264f "Sco" +U+2650 "Sgr" +U+2651 "Cap" +U+2652 "Aqr" +U+2653 "Psc" +U+2654 "k" # white chess king +U+2655:q +U+2656:r +U+2657:b +U+2658:n +U+2659:p +U+265a "K" # black chess king +U+265b:Q +U+265c:R +U+265d:B +U+265e:N +U+265f:P +U+2660 "cS" # black spade suit +U+2661:ch +U+2662:cd +U+2663:cC +U+2664:cs +U+2665:cH +U+2666:cD +U+2667:cc +#Musical symbols +U+2669:d +U+266a:d` +U+266b:d-d +U+266c:d=d +U+266d:b +U+266e:N +U+266f:# +U+2670:t +U+2671:t +#Recycling symbols +U+2672:/_\ +U+2673:/1\ +U+2674:/2\ +U+2675:/3\ +U+2676:/4\ +U+2677:/5\ +U+2678:/6\ +U+2679:/7\ +U+267a:/_\ +U+267b:/_\ +U+267c:/P\ +U+267d:/p\ +U+267e:(oo) +U+267f "oL" # wheelchair symbol +U+2680:d1 +U+2681:d2 +U+2682:d3 +U+2683:d4 +U+2684:d5 +U+2685:d6 +U+2686:(.) +U+2687:(:) +U+2688:((.)) +U+2689:((:)) +U+2690 " f " +U+2691 " F " +U+2692:'X` +U+2693 "+-)" # anchor (sideways) +U+2694:,X, +U+2695 "$" # cf. U+2624 +U+2696 "^T^" # scales of justice +U+269a "}T{" # staff of Hermes +U+269b ":*:" # atom symbol +U+269c "}|{" # fleur-de-lis +U+26a0 "!!!" # or "WARNING ", cf. U+2621 +U+26a1 "!V!" # or "VOLTAGE ", cf. U+2621 +U+26a2 "f.f." # two females, lesbian +U+26a3 "m.m." # two males, homosexual +U+26a4 "m.f." # male and female, bisexual +U+26a5 "mf." # cf. U+26A4 +U+26a6 "xm." # transgendered male +U+26a7 "xmf." # transgendered male/female +U+26aa:o +U+26ab:O +U+26ac:o +U+26ad:oo +U+26ae:o|o +U+26af:o-o +U+26b0 "/b/" # buried/coffin +U+26b1 "/c/" # cremated/urn +U+26b2 "n." # cf. U+26a2 - U+26a7 +# Dingbats +U+2702:8< +U+2704:>8 +U+2706:(TEL) +U+2708:+->- +U+2709 "[v]" # envelope +U+270c:mV, +0x58 U+2713 U+2714 U+2717 U+2718 # check marks -> x +U+2715: x +U+2716: X +U+2719:+ +U+271a:+ +U+271b:+ +U+271c:+ +U+271d:t +U+271e:t +U+271f:t +U+2720:-X +0x2a U+2721 U+272a U+272b U+272c U+272d U+272e U+272f U+2730 U+2731 U+2732 U+2733 U+2734 U+2735 U+2736 U+2737 U+2738 U+2739 U+273a U+273b U+273c U+273d +0x2a U+2742 U+2743 U+2744 U+2745 U+2746 U+2747 U+2748 U+2749 U+274a U+274b +U+2756:<x> +U+2758:| +U+2759:| +U+275a:| +U+275b:' +U+275c:' +U+275d:" +U+275e:" +U+2762:! +U+2763:! +U+2765:<3 +U+2768:( +U+2769:) +U+276a:( +U+276b:) +U+276c:< +U+276d:> +U+276e:< +U+276f:> +U+2770:< +U+2771:> +U+2772:[ +U+2773:] +U+2774:{ +U+2775:} +U+2776:((1)) +U+2777:((2)) +U+2778:((3)) +U+2779:((4)) +U+277a:((5)) +U+277b:((6)) +U+277c:((7)) +U+277d:((8)) +U+277e:((9)) +U+277f:((10)) +U+2780:(1) +U+2781:(2) +U+2782:(3) +U+2783:(4) +U+2784:(5) +U+2785:(6) +U+2786:(7) +U+2787:(8) +U+2788:(9) +U+2789:(10) +U+278a:((1)) +U+278b:((2)) +U+278c:((3)) +U+278d:((4)) +U+278e:((5)) +U+278f:((6)) +U+2790:((7)) +U+2791:((8)) +U+2792:((9)) +U+2793:((10)) +U+2794:-> +U+2798:\v +U+2799:-> +U+279a:/^ +U+279b:-> +U+279c:-> +U+279d:-> +U+279e:-> +U+279f:-> +U+27a0:-> +U+27a1:-> +U+27a2:> +U+27a3:> +U+27a4:> +U+27a5:-> +U+27a6:-> +U+27a7:-> +U+27a8:-> +U+27a9:-> +U+27aa:-> +U+27ab:-> +U+27ac:-> +U+27ad:-> +U+27ae:-> +U+27af:-> +U+27b0:-> +U+27b2:(->) +U+27b3:>>-> +U+27b4:vv\v +U+27b5:>>-> +U+27b6:^^/^ +U+27b7:vv\v +U+27b8:>>-> +U+27b9:^^/^ +U+27ba:-> +U+27bb:-> +U+27bc:-> +U+27bd:-> +U+27be:-> +# Supplemental Arrows A +U+27f0:^|||| +U+27f1:||||v +U+27f2:vO +U+27f3:Ov +U+27f4:(+)> +U+27f5:<--- +U+27f6:---> +U+27f7:<---> +U+27f8:<=== +U+27f9:===> +U+27fa:<===> +U+27fb:<---| +U+27fc:|---> +U+27fd:<===| +U+27fe:|===> +U+27ff:~~~> +# Supplemental Arrows B +U+2900:-|->> +U+2901:-||->> +U+2902:<=|= +U+2903:=|=> +U+2904:<=|=> +U+2905:|->> +U+2906:<=| +U+2907:|=> +U+2908:|-|v +U+2909:^|-| +U+290a:^||| +U+290b:|||v +U+290c:<- - +U+290d:- -> +U+290e:<- - - +U+290f:- - -> +U+2910:>- - -> +U+2911:-> +U+2913:|v_ +U+2914:>-|-> +U+2915:>-||-> +U+2916:>->> +U+2917:>-|->> +U+2918:>-||->> +U+2919:-< +U+291a:>- +U+291b:-<< +U+291c:>>- +U+291d:<><- +U+291e:-><> +U+291f:<><-| +U+2920:|-><> +U+2921:^\v +U+2922:v/^ +U+2923:^\, +U+2924:,/^ +U+2927:^X^ +U+292b:X +U+292c:X +U+292d:Xv^ +U+292e:X^v +U+292f:X ^ +U+2930:X v +U+2931:^X^ +U+2932:^X^ +U+2933:~> +U+2934:-^ +U+2935:-v +U+2938:)v +U+2939:(v +U+2945:->+ +U+2946:<-+ +U+2947:-x-> +U+2948:<-o-> +U+2949:^^|o +U+294a:<-> +U+294b:<-> +U+294c:^|v +U+294d:^|v +U+294e:<-> +U+294f:^|v +U+2950:<-> +U+2951:^|v +U+2952:|<- +U+2953:->| +U+2955:|v_ +U+2956:|<- +U+2957:->| +U+2959:|v_ +U+295a:<-| +U+295b:|-> +U+295c:^|_ +U+295e:<-| +U+295f:|-> +U+2960:^|_ +U+2962:<= +U+2963:^|| +U+2964:=> +U+2965:||v +U+2970:=) +U+2a00 "(.)" +U+2a01 "(+)" +U+2a02 "(x)" +U+2a09: * +U+2a0c "\134int\134int\134int\134int " +U+2a2f:x +U+2a30:.x +U+2a31:x_ +U+2a33:xx +U+2a34:(x +U+2a35:x) +U+2a37:((x)) +U+2a38:(/) +U+2a39:/+\ +U+2a3a:/-\ +U+2a3b:/x\ +U+2a74:::= +U+2a75:== +U+2a76:=== +U+2a77::=: +#Miscellaneous Symbols and Arrows +U+2b00:/^ +U+2b01:^\ +U+2b02:\v +U+2b03:v/ +U+2b04:<-> +U+2b05:<- +U+2b06:^| +U+2b07:|v +U+2b08:/^ +U+2b09:^\ +U+2b0a:\v +U+2b0b:v/ +U+2b0c:<-> +U+2b0d:^|v +U+2b0e:-v +U+2b0f:-^ +U+2b10:v- +U+2b11:^- +# Supplemental punctuation +U+2e0f:__ +U+2e1e:.~ +U+2e1f:~. +U+2e28:(( +U+2e29:)) +U+2e2a ":." +U+2e2b ".:" +U+2e2c "::" +U+2e2f:~ +# CJK area: +0x20 U+3000 # ideographic space +U+3001:,_ +U+3002:._ +U+3003:+" +U+3004:JIS +U+3005:*_ +U+3006:;_ +U+3007:0_ +U+300a:<+ +U+300b:>+ +U+300c:<' +U+300d:>' +U+300e:<" +U+300f:>" +U+3010:(" +U+3011:)" +U+3012:=T +U+3013:=_ +U+3014:(' +U+3015:)' +U+3016:(I +U+3017:)I +U+301a:[[ +U+301b:]] +U+301c:-? +U+3020:=T:) +0x20 U+303f +U+3041:A5 +U+3042:a5 +U+3043:I5 +U+3044:i5 +U+3045:U5 +U+3046:u5 +U+3047:E5 +U+3048:e5 +U+3049:O5 +U+304a:o5 +U+304b:ka +U+304c:ga +U+304d:ki +U+304e:gi +U+304f:ku +U+3050:gu +U+3051:ke +U+3052:ge +U+3053:ko +U+3054:go +U+3055:sa +U+3056:za +U+3057:si +U+3058:zi +U+3059:su +U+305a:zu +U+305b:se +U+305c:ze +U+305d:so +U+305e:zo +U+305f:ta +U+3060:da +U+3061:ti +U+3062:di +U+3063:tU +U+3064:tu +U+3065:du +U+3066:te +U+3067:de +U+3068:to +U+3069:do +U+306a:na +U+306b:ni +U+306c:nu +U+306d:ne +U+306e:no +U+306f:ha +U+3070:ba +U+3071:pa +U+3072:hi +U+3073:bi +U+3074:pi +U+3075:hu +U+3076:bu +U+3077:pu +U+3078:he +U+3079:be +U+307a:pe +U+307b:ho +U+307c:bo +U+307d:po +U+307e:ma +U+307f:mi +U+3080:mu +U+3081:me +U+3082:mo +U+3083:yA +U+3084:ya +U+3085:yU +U+3086:yu +U+3087:yO +U+3088:yo +U+3089:ra +U+308a:ri +U+308b:ru +U+308c:re +U+308d:ro +U+308e:wA +U+308f:wa +U+3090:wi +U+3091:we +U+3092:wo +U+3093:n5 +U+3094:vu +U+309b:"5 +U+309c:05 +U+309d:*5 +U+309e:+5 +U+30a1:a6 +U+30a2:A6 +U+30a3:i6 +U+30a4:I6 +U+30a5:u6 +U+30a6:U6 +U+30a7:e6 +U+30a8:E6 +U+30a9:o6 +U+30aa:O6 +U+30ab:Ka +U+30ac:Ga +U+30ad:Ki +U+30ae:Gi +U+30af:Ku +U+30b0:Gu +U+30b1:Ke +U+30b2:Ge +U+30b3:Ko +U+30b4:Go +U+30b5:Sa +U+30b6:Za +U+30b7:Si +U+30b8:Zi +U+30b9:Su +U+30ba:Zu +U+30bb:Se +U+30bc:Ze +U+30bd:So +U+30be:Zo +U+30bf:Ta +U+30c0:Da +U+30c1:Ti +U+30c2:Di +U+30c3:TU +U+30c4:Tu +U+30c5:Du +U+30c6:Te +U+30c7:De +U+30c8:To +U+30c9:Do +U+30ca:Na +U+30cb:Ni +U+30cc:Nu +U+30cd:Ne +U+30ce:No +U+30cf:Ha +U+30d0:Ba +U+30d1:Pa +U+30d2:Hi +U+30d3:Bi +U+30d4:Pi +U+30d5:Hu +U+30d6:Bu +U+30d7:Pu +U+30d8:He +U+30d9:Be +U+30da:Pe +U+30db:Ho +U+30dc:Bo +U+30dd:Po +U+30de:Ma +U+30df:Mi +U+30e0:Mu +U+30e1:Me +U+30e2:Mo +U+30e3:YA +U+30e4:Ya +U+30e5:YU +U+30e6:Yu +U+30e7:YO +U+30e8:Yo +U+30e9:Ra +U+30ea:Ri +U+30eb:Ru +U+30ec:Re +U+30ed:Ro +U+30ee:WA +U+30ef:Wa +U+30f0:Wi +U+30f1:We +U+30f2:Wo +U+30f3:N6 +U+30f4:Vu +U+30f5:KA +U+30f6:KE +U+30f7:Va +U+30f8:Vi +U+30f9:Ve +U+30fa:Vo +U+30fb:.6 +U+30fc:-6 +U+30fd:*6 +U+30fe:+6 +U+3105:b4 +U+3106:p4 +U+3107:m4 +U+3108:f4 +U+3109:d4 +U+310a:t4 +U+310b:n4 +U+310c:l4 +U+310d:g4 +U+310e:k4 +U+310f:h4 +U+3110:j4 +U+3111:q4 +U+3112:x4 +U+3113:zh +U+3114:ch +U+3115:sh +U+3116:r4 +U+3117:z4 +U+3118:c4 +U+3119:s4 +U+311a:a4 +U+311b:o4 +U+311c:e4 +U+311d:eh4 +U+311e:ai +U+311f:ei +U+3120:au +U+3121:ou +U+3122:an +U+3123:en +U+3124:aN +U+3125:eN +U+3126:er +U+3127:i4 +U+3128:u4 +U+3129:iu +U+312a:v4 +U+312b:nG +U+312c:gn +U+321c:(JU) +U+3220:1c +U+3221:2c +U+3222:3c +U+3223:4c +U+3224:5c +U+3225:6c +U+3226:7c +U+3227:8c +U+3228:9c +U+3229:10c +U+327f:KSC +U+33c2:am +U+33d8:pm +# +# +#There are four special ranges of characters that are represented only by +#their start and end characters <...> +# +# The CJK Ideographs Area (U+4E00 - U+9FFF) +# The Hangul Syllables Area (U+AC00 - U+D7A3) +# The Surrogates Area (U+D800 - U+DFFF) +# The Private Use Area (U+E000 - U+F8FF) +# +# +U+fb00:ff +U+fb01:fi +U+fb02:fl +U+fb03:ffi +U+fb04:ffl +U+fb05:St +U+fb06:st +U+fe7d:3+; +U+fe82:aM. +U+fe84:aH. +U+fe88:ah. +U+fe8d:a+- +U+fe8e:a+. +U+fe8f:b+- +U+fe90:b+. +U+fe91:b+, +U+fe92:b+; +U+fe93:tm- +U+fe94:tm. +U+fe95:t+- +U+fe96:t+. +U+fe97:t+, +U+fe98:t+; +U+fe99:tk- +U+fe9a:tk. +U+fe9b:tk, +U+fe9c:tk; +U+fe9d:g+- +U+fe9e:g+. +U+fe9f:g+, +U+fea0:g+; +U+fea1:hk- +U+fea2:hk. +U+fea3:hk, +U+fea4:hk; +U+fea5:x+- +U+fea6:x+. +U+fea7:x+, +U+fea8:x+; +U+fea9:d+- +U+feaa:d+. +U+feab:dk- +U+feac:dk. +U+fead:r+- +U+feae:r+. +U+feaf:z+- +U+feb0:z+. +U+feb1:s+- +U+feb2:s+. +U+feb3:s+, +U+feb4:s+; +U+feb5:sn- +U+feb6:sn. +U+feb7:sn, +U+feb8:sn; +U+feb9:c+- +U+feba:c+. +U+febb:c+, +U+febc:c+; +U+febd:dd- +U+febe:dd. +U+febf:dd, +U+fec0:dd; +U+fec1:tj- +U+fec2:tj. +U+fec3:tj, +U+fec4:tj; +U+fec5:zH- +U+fec6:zH. +U+fec7:zH, +U+fec8:zH; +U+fec9:e+- +U+feca:e+. +U+fecb:e+, +U+fecc:e+; +U+fecd:i+- +U+fece:i+. +U+fecf:i+, +U+fed0:i+; +U+fed1:f+- +U+fed2:f+. +U+fed3:f+, +U+fed4:f+; +U+fed5:q+- +U+fed6:q+. +U+fed7:q+, +U+fed8:q+; +U+fed9:k+- +U+feda:k+. +U+fedb:k+, +U+fedc:k+; +U+fedd:l+- +U+fede:l+. +U+fedf:l+, +U+fee0:l+; +U+fee1:m+- +U+fee2:m+. +U+fee3:m+, +U+fee4:m+; +U+fee5:n+- +U+fee6:n+. +U+fee7:n+, +U+fee8:n+; +U+fee9:h+- +U+feea:h+. +U+feeb:h+, +U+feec:h+; +U+feed:w+- +U+feee:w+. +U+feef:j+- +U+fef0:j+. +U+fef1:y+- +U+fef2:y+. +U+fef3:y+, +U+fef4:y+; +U+fef5:lM- +U+fef6:lM. +U+fef7:lH- +U+fef8:lH. +U+fef9:lh- +U+fefa:lh. +U+fefb:la- +U+fefc:la. + +# the reverse byte-order-mark: zero-width non-break space +U+feff "" + +0x21-0x7e U+ff01-U+ff5e +0x2e U+ff61 +0x22 U+ff62 U+ff63 +0x2c U+ff64 + +# Symbols for C0 and C1 control characters, in case they get through... +U+0000:NUL +U+0001:SH +U+0002:SX +U+0003:EX +U+0004:ET +U+0005:ENQ +U+0006:AK +U+0007:BL +U+0008:BS +U+0009:HT +U+000a:LF +U+000b:VT +U+000c:FF +U+000d:CR +U+000e:SO +U+000f:SI +U+0010:DL +U+0011:DC1 +U+0012:DC2 +U+0013:DC3 +U+0014:DC4 +U+0015:NAK +U+0016:SYN +U+0017:EB +U+0018:CN +U+0019:EM +U+001a:SB +U+001b:ESC +U+001c:FS +U+001d:GS +U+001e:RS +U+001f:US +U+007f:DT +# Most of these characters (80-9F) may be inflicted on us +# by MS FrontPages which uses Unicode notation such as ™ +# but there are no assigned letters in Unicode 128-159 range. +# It is assumed in the code that those codepoints are from windows-1252. +#U+0080:PA +#U+0081:HO +#U+0082:BH +#U+0083:NH +#U+0084:IN +#U+0085:NL +U+0085 "\012" +#U+0086:SA +#U+0087:ES +#U+0088:HS +#U+0089:HJ +#U+008a:VS +#U+008b:PD +#U+008c:PU +#U+008d:RI +#U+008e:SS2 +#U+008f:SS3 +#U+0090:DCS +#U+0091:P1 +#U+0092:P2 +#U+0093:TS +#U+0094:CC +#U+0095:MW +#U+0096:SG +#U+0097:EG +#U+0098:SS +#U+0099:GC +#U+009a:SC +#U+009b:CSI +#U+009c:ST +#U+009d:OC +#U+009e:PM +#U+009f:AC + +# Let's try to show a question mark for character that cannot +# be shown. U+fffd is used for invalid characters. +# It works, but let's stick with UHHH representation. - FM +#U+fffd "?" diff --git a/src/chrtrans/dmcs_uni.tbl b/src/chrtrans/dmcs_uni.tbl new file mode 100644 index 0000000..ab9edb2 --- /dev/null +++ b/src/chrtrans/dmcs_uni.tbl @@ -0,0 +1,233 @@ +#The MIME name of this charset. +Mdec-mcs + +#Name as a Display Charset (used on Options screen) +ODEC Multinational + +# +# Name: DEC Multinational (dec-mcs) [to unicode] +# Date: 29 October 1997 +# Author: Fote +# +# 1999-01-01 various corrections, verified against actual DEC VT220 +# Christian "naddy" Weisgerber <naddy@mips.rhein-neckar.de> +# +################## + +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +# +0x20-0x7E idem # ASCII + +0xA1 U+00A1 # inverted exclamation mark (¡) - iexcl +0xA2 U+00A2 # cent sign (¢) - cent +0xA3 U+00A3 # pound sign (£) - pound +0xA5 U+00A5 # yen sign (¥) - yen +# broken vertical bar (¦) - brvbar, brkbar +U+00A6:| +0xA7 U+00A7 # section sign (§) - sect +0xA8 U+00A4 # currency sign (¤) - curren +# spacing diaeresis (¨) - uml, die +U+00A8:" +0xA9 U+00A9 # copyright sign (©) - copy +0xAA U+00AA # feminine ordinal indicator (ª) - ordf +0xAB U+00AB # angle quotation mark, left («) - laquo +# negation sign (¬); - not +U+00AC:NOT +# soft hyphen (­) - shy +#U+00AD +# circled R registered sign (®) - reg +U+00AE:(R) +# spacing macron (¯) - hibar, macr +U+00AF:- +0xB0 U+00B0 # degree sign (°) - deg +0xB1 U+00B1 # plus-or-minus sign (±) - plusmn +0xB2 U+00B2 # superscript 2 (²) - sup2 +0xB3 U+00B3 # superscript 3 (³) - sup3 +#spacing acute (´) - acute +U+00B4:' +0xB5 U+00B5 # micro sign (µ) - micro +0xB6 U+00B6 # paragraph sign (¶) - para +0xB7 U+00B7 # middle dot (·) - middot +# spacing cedilla (¸) - cedil +U+00B8:, +0xB9 U+00B9 # superscript 1 (¹) - sup1 +0xBA U+00BA # masculine ordinal indicator (º) - ordm +0xBB U+00BB # angle quotation mark, right (») - raquo +0xBC U+00BC # fraction 1/4 (¼) - frac14 +0xBD U+00BD # fraction 1/2 (½) - frac12 +# fraction 3/4 (¾) - frac34 +U+00BE: 3/4 +0xBF U+00BF # inverted question mark (¿) - iquest +0xC0 U+00C0 # capital A, grave accent (À) - Agrave +0xC1 U+00C1 # capital A, acute accent (Á) - Aacute +0xC2 U+00C2 # capital A, circumflex accent (Â) - Acirc +0xC3 U+00C3 # capital A, tilde (Ã) - Atilde +0xC4 U+00C4 # capital A, dieresis or umlaut mark (Ä) - Auml +0xC5 U+00C5 # capital A, ring (Å) - Aring +0xC6 U+00C6 # capital AE diphthong (ligature) (Æ) - AElig +0xC7 U+00C7 # capital C, cedilla (Ç) - Ccedil +0xC8 U+00C8 # capital E, grave accent (È) - Egrave +0xC9 U+00C9 # capital E, acute accent (É) - Eacute +0xCA U+00CA # capital E, circumflex accent (Ê) - Ecirc +0xCB U+00CB # capital E, dieresis or umlaut mark (Ë) - Euml +0xCC U+00CC # capital I, grave accent (Ì) - Igrave +0xCD U+00CD # capital I, acute accent (Í) - Iacute +0xCE U+00CE # capital I, circumflex accent (Î) - Icirc +0xCF U+00CF # capital I, dieresis or umlaut mark (Ï) - Iuml +# capital Eth, Icelandic (Ð) - ETH */ +U+00D0:DH +# Dj # capital D with stroke - Dstrok +0xD1 U+00D1 # capital N, tilde (Ñ) - Ntilde +0xD2 U+00D2 # capital O, grave accent (Ò) - Ograve +0xD3 U+00D3 # capital O, acute accent (Ó) - Oacute +0xD4 U+00D4 # capital O, circumflex accent (Ô) - Ocirc +0xD5 U+00D5 # capital O, tilde (Õ) - Otilde +0xD6 U+00D6 # capital O, dieresis or umlaut mark (Ö) - Ouml +0xD7 U+0152 # capital OE ligature (Œ) - OElig +# multiplication sign (×) - times +U+00D7:* +0xD8 U+00D8 # capital O, slash (Ø) - Oslash +0xD9 U+00D9 # capital U, grave accent (Ù) - Ugrave +0xDA U+00DA # capital U, acute accent (Ú) - Uacute +0xDB U+00DB # capital U, circumflex accent (Û) - Ucirc +0xDC U+00DC # capital U, dieresis or umlaut mark (Ü) - Uuml +0xDD U+0178 # capital Y, dieresis or umlaut mark (Ÿ) - Yuml +# capital Y, acute accent (Ý) - Yacute +U+00DD:Y' +# capital THORN, Icelandic (Þ) - THORN */ +U+00DE:TH +0xDF U+00DF # small sharp s, German (sz ligature) (ß) - szlig +0xE0 U+00E0 # small a, grave accent (à) - agrave +0xE1 U+00E1 # small a, acute accent (á) - aacute +0xE2 U+00E2 # small a, circumflex accent (â) - acirc +0xE3 U+00E3 # small a, tilde (ã) - atilde +0xE4 U+00E4 # small a, dieresis or umlaut mark (ä) - auml +0xE5 U+00E5 # small a, ring (å) - aring +0xE6 U+00E6 # small ae diphthong (ligature) (æ) - aelig +0xE7 U+00E7 # small c, cedilla (ç) - ccedil +0xE8 U+00E8 # small e, grave accent (è) - egrave +0xE9 U+00E9 # small e, acute accent (é) - eacute +0xEA U+00EA # small e, circumflex accent (ê) - ecirc +0xEB U+00EB # small e, dieresis or umlaut mark (ë) - euml +0xEC U+00EC # small i, grave accent (ì) - igrave +0xED U+00ED # small i, acute accent (í) - iacute +0xEE U+00EE # small i, circumflex accent (î) - icirc +0xEF U+00EF # small i, dieresis or umlaut mark (ï) - iuml +# small eth, Icelandic (ð) - eth +U+00F0:dh +0xF1 U+00F1 # small n, tilde (ñ) - ntilde +0xF2 U+00F2 # small o, grave accent (ò) - ograve +0xF3 U+00F3 # small o, acute accent (ó) - oacute +0xF4 U+00F4 # small o, circumflex accent (ô) - ocirc +0xF5 U+00F5 # small o, tilde (õ) - otilde +0xF6 U+00F6 # small o, dieresis or umlaut mark (ö) - ouml +0xF7 U+0153 # small oe ligature (œ) - oelig +# division sign (÷) - divide +U+00F7:/ +0xF8 U+00F8 # small o, slash (ø) - oslash +0xF9 U+00F9 # small u, grave accent (ù) - ugrave +0xFA U+00FA # small u, acute accent (ú) - uacute +0xFB U+00FB # small u, circumflex accent (û) - ucirc +0xFC U+00FC # small u, dieresis or umlaut mark (ü) - uuml +0xFD U+00FF # small y, dieresis or umlaut mark (ÿ) - yuml +# small y, acute accent (ý) - yacute +U+00FD:y' +# small thorn, Icelandic (þ) - thorn +U+00FE:th +# +# TRADE MARK SIGN +U+2122:(TM) diff --git a/src/chrtrans/entities.h b/src/chrtrans/entities.h new file mode 100644 index 0000000..66f66ab --- /dev/null +++ b/src/chrtrans/entities.h @@ -0,0 +1,1414 @@ +/* + * $LynxId: entities.h,v 1.6 2020/01/21 21:31:36 tom Exp $ + * + * Entity Names to Unicode table + * ----------------------------- + * + * This is a one-way mapping to Unicode so chartrans implementation + * now process character entities like   the similar way it handles + * the numeric entities like {. + * The only call to this structure is via HTMLGetEntityUCValue(). + * + +Unlike the numeric entities ê which may be for any Unicode character, the +character references should be defined within HTML standards to get a +compatibility between browsers. + +Now we have a choice: use clean HTML4.0 entities list (and reject everything +others), or use a relaxed list with lots of synonyms and new symbols found at + +ftp://ftp.unicode.org/Public/MAPPINGS/VENDORS/MISC/SGML.TXT + +We hold both: #define ENTITIES_HTML40_ONLY for strict version, +otherwise relaxed. + + */ + +#include <UCkd.h> /* typedef u16 */ +typedef struct { + const char *name; /* sorted alphabetically (case-sensitive) */ + u16 code; +} UC_entity_info; + +static const UC_entity_info unicode_entities[] = +/* *INDENT-OFF* */ +#ifdef ENTITIES_HTML40_ONLY +/********************************************************************* + + The full list of character references defined as part of HTML 4.0. + http://www.w3.org/TR/PR-html40/sgml/entities.html + + Informal history: + * ISO Latin 1 entities for 160-255 range were introduced in HTML 2.0 + * few important entities were added, including <, >, &. + * Greek letters and some math symbols were finally added in HTML 4.0 + + Totally 252 entries (Nov 1997 HTML 4.0 draft), it is 1:1 mapping. + Please do not add more unless a new HTML version will be released, + try the #else table for experiments and fun... + +****/ +{ + {"AElig", 198}, /* latin capital letter AE = latin capital ligature AE, U+00C6 ISOlat1 */ + {"Aacute", 193}, /* latin capital letter A with acute, U+00C1 ISOlat1 */ + {"Acirc", 194}, /* latin capital letter A with circumflex, U+00C2 ISOlat1 */ + {"Agrave", 192}, /* latin capital letter A with grave = latin capital letter A grave, U+00C0 ISOlat1 */ + {"Alpha", 913}, /* greek capital letter alpha, U+0391 */ + {"Aring", 197}, /* latin capital letter A with ring above = latin capital letter A ring, U+00C5 ISOlat1 */ + {"Atilde", 195}, /* latin capital letter A with tilde, U+00C3 ISOlat1 */ + {"Auml", 196}, /* latin capital letter A with diaeresis, U+00C4 ISOlat1 */ + {"Beta", 914}, /* greek capital letter beta, U+0392 */ + {"Ccedil", 199}, /* latin capital letter C with cedilla, U+00C7 ISOlat1 */ + {"Chi", 935}, /* greek capital letter chi, U+03A7 */ + {"Dagger", 8225}, /* double dagger, U+2021 ISOpub */ + {"Delta", 916}, /* greek capital letter delta, U+0394 ISOgrk3 */ + {"ETH", 208}, /* latin capital letter ETH, U+00D0 ISOlat1 */ + {"Eacute", 201}, /* latin capital letter E with acute, U+00C9 ISOlat1 */ + {"Ecirc", 202}, /* latin capital letter E with circumflex, U+00CA ISOlat1 */ + {"Egrave", 200}, /* latin capital letter E with grave, U+00C8 ISOlat1 */ + {"Epsilon", 917}, /* greek capital letter epsilon, U+0395 */ + {"Eta", 919}, /* greek capital letter eta, U+0397 */ + {"Euml", 203}, /* latin capital letter E with diaeresis, U+00CB ISOlat1 */ + {"Gamma", 915}, /* greek capital letter gamma, U+0393 ISOgrk3 */ + {"Iacute", 205}, /* latin capital letter I with acute, U+00CD ISOlat1 */ + {"Icirc", 206}, /* latin capital letter I with circumflex, U+00CE ISOlat1 */ + {"Igrave", 204}, /* latin capital letter I with grave, U+00CC ISOlat1 */ + {"Iota", 921}, /* greek capital letter iota, U+0399 */ + {"Iuml", 207}, /* latin capital letter I with diaeresis, U+00CF ISOlat1 */ + {"Kappa", 922}, /* greek capital letter kappa, U+039A */ + {"Lambda", 923}, /* greek capital letter lambda, U+039B ISOgrk3 */ + {"Mu", 924}, /* greek capital letter mu, U+039C */ + {"Ntilde", 209}, /* latin capital letter N with tilde, U+00D1 ISOlat1 */ + {"Nu", 925}, /* greek capital letter nu, U+039D */ + {"OElig", 338}, /* latin capital ligature OE, U+0152 ISOlat2 */ + {"Oacute", 211}, /* latin capital letter O with acute, U+00D3 ISOlat1 */ + {"Ocirc", 212}, /* latin capital letter O with circumflex, U+00D4 ISOlat1 */ + {"Ograve", 210}, /* latin capital letter O with grave, U+00D2 ISOlat1 */ + {"Omega", 937}, /* greek capital letter omega, U+03A9 ISOgrk3 */ + {"Omicron", 927}, /* greek capital letter omicron, U+039F */ + {"Oslash", 216}, /* latin capital letter O with stroke = latin capital letter O slash, U+00D8 ISOlat1 */ + {"Otilde", 213}, /* latin capital letter O with tilde, U+00D5 ISOlat1 */ + {"Ouml", 214}, /* latin capital letter O with diaeresis, U+00D6 ISOlat1 */ + {"Phi", 934}, /* greek capital letter phi, U+03A6 ISOgrk3 */ + {"Pi", 928}, /* greek capital letter pi, U+03A0 ISOgrk3 */ + {"Prime", 8243}, /* double prime = seconds = inches, U+2033 ISOtech */ + {"Psi", 936}, /* greek capital letter psi, U+03A8 ISOgrk3 */ + {"Rho", 929}, /* greek capital letter rho, U+03A1 */ + {"Scaron", 352}, /* latin capital letter S with caron, U+0160 ISOlat2 */ +/* there is no Sigmaf, and no U+03A2 character either */ + {"Sigma", 931}, /* greek capital letter sigma, U+03A3 ISOgrk3 */ + {"THORN", 222}, /* latin capital letter THORN, U+00DE ISOlat1 */ + {"Tau", 932}, /* greek capital letter tau, U+03A4 */ + {"Theta", 920}, /* greek capital letter theta, U+0398 ISOgrk3 */ + {"Uacute", 218}, /* latin capital letter U with acute, U+00DA ISOlat1 */ + {"Ucirc", 219}, /* latin capital letter U with circumflex, U+00DB ISOlat1 */ + {"Ugrave", 217}, /* latin capital letter U with grave, U+00D9 ISOlat1 */ + {"Upsilon", 933}, /* greek capital letter upsilon, U+03A5 ISOgrk3 */ + {"Uuml", 220}, /* latin capital letter U with diaeresis, U+00DC ISOlat1 */ + {"Xi", 926}, /* greek capital letter xi, U+039E ISOgrk3 */ + {"Yacute", 221}, /* latin capital letter Y with acute, U+00DD ISOlat1 */ + {"Yuml", 376}, /* latin capital letter Y with diaeresis, U+0178 ISOlat2 */ + {"Zeta", 918}, /* greek capital letter zeta, U+0396 */ + {"aacute", 225}, /* latin small letter a with acute, U+00E1 ISOlat1 */ + {"acirc", 226}, /* latin small letter a with circumflex, U+00E2 ISOlat1 */ + {"acute", 180}, /* acute accent = spacing acute, U+00B4 ISOdia */ + {"aelig", 230}, /* latin small letter ae = latin small ligature ae, U+00E6 ISOlat1 */ + {"agrave", 224}, /* latin small letter a with grave = latin small letter a grave, U+00E0 ISOlat1 */ + {"alefsym", 8501}, /* alef symbol = first transfinite cardinal, U+2135 NEW */ +/* alef symbol is NOT the same as hebrew letter alef, U+05D0 although the same glyph could be used to depict both characters */ + {"alpha", 945}, /* greek small letter alpha, U+03B1 ISOgrk3 */ + {"amp", 38}, /* ampersand, U+0026 ISOnum */ + {"and", 8743}, /* logical and = wedge, U+2227 ISOtech */ + {"ang", 8736}, /* angle, U+2220 ISOamso */ + {"aring", 229}, /* latin small letter a with ring above = latin small letter a ring, U+00E5 ISOlat1 */ + {"asymp", 8776}, /* almost equal to = asymptotic to, U+2248 ISOamsr */ + {"atilde", 227}, /* latin small letter a with tilde, U+00E3 ISOlat1 */ + {"auml", 228}, /* latin small letter a with diaeresis, U+00E4 ISOlat1 */ + {"bdquo", 8222}, /* double low-9 quotation mark, U+201E NEW */ + {"beta", 946}, /* greek small letter beta, U+03B2 ISOgrk3 */ + {"brvbar", 166}, /* broken bar = broken vertical bar, U+00A6 ISOnum */ + {"bull", 8226}, /* bullet = black small circle, U+2022 ISOpub */ +/* bullet is NOT the same as bullet operator, U+2219 */ + {"cap", 8745}, /* intersection = cap, U+2229 ISOtech */ + {"ccedil", 231}, /* latin small letter c with cedilla, U+00E7 ISOlat1 */ + {"cedil", 184}, /* cedilla = spacing cedilla, U+00B8 ISOdia */ + {"cent", 162}, /* cent sign, U+00A2 ISOnum */ + {"chi", 967}, /* greek small letter chi, U+03C7 ISOgrk3 */ + {"circ", 710}, /* modifier letter circumflex accent, U+02C6 ISOpub */ + {"clubs", 9827}, /* black club suit = shamrock, U+2663 ISOpub */ + {"cong", 8773}, /* approximately equal to, U+2245 ISOtech */ + {"copy", 169}, /* copyright sign, U+00A9 ISOnum */ + {"crarr", 8629}, /* downwards arrow with corner leftwards = carriage return, U+21B5 NEW */ + {"cup", 8746}, /* union = cup, U+222A ISOtech */ + {"curren", 164}, /* currency sign, U+00A4 ISOnum */ + {"dArr", 8659}, /* downwards double arrow, U+21D3 ISOamsa */ + {"dagger", 8224}, /* dagger, U+2020 ISOpub */ + {"darr", 8595}, /* downwards arrow, U+2193 ISOnum */ + {"deg", 176}, /* degree sign, U+00B0 ISOnum */ + {"delta", 948}, /* greek small letter delta, U+03B4 ISOgrk3 */ + {"diams", 9830}, /* black diamond suit, U+2666 ISOpub */ + {"divide", 247}, /* division sign, U+00F7 ISOnum */ + {"eacute", 233}, /* latin small letter e with acute, U+00E9 ISOlat1 */ + {"ecirc", 234}, /* latin small letter e with circumflex, U+00EA ISOlat1 */ + {"egrave", 232}, /* latin small letter e with grave, U+00E8 ISOlat1 */ + {"empty", 8709}, /* empty set = null set = diameter, U+2205 ISOamso */ + {"emsp", 8195}, /* em space, U+2003 ISOpub */ + {"ensp", 8194}, /* en space, U+2002 ISOpub */ + {"epsilon", 949}, /* greek small letter epsilon, U+03B5 ISOgrk3 */ + {"equiv", 8801}, /* identical to, U+2261 ISOtech */ + {"eta", 951}, /* greek small letter eta, U+03B7 ISOgrk3 */ + {"eth", 240}, /* latin small letter eth, U+00F0 ISOlat1 */ + {"euml", 235}, /* latin small letter e with diaeresis, U+00EB ISOlat1 */ + {"euro", 8364}, /* euro sign, U+20AC NEW */ + {"exist", 8707}, /* there exists, U+2203 ISOtech */ + {"fnof", 402}, /* latin small f with hook = function = florin, U+0192 ISOtech */ + {"forall", 8704}, /* for all, U+2200 ISOtech */ + {"frac12", 189}, /* vulgar fraction one half = fraction one half, U+00BD ISOnum */ + {"frac14", 188}, /* vulgar fraction one quarter = fraction one quarter, U+00BC ISOnum */ + {"frac34", 190}, /* vulgar fraction three quarters = fraction three quarters, U+00BE ISOnum */ + {"frasl", 8260}, /* fraction slash, U+2044 NEW */ + {"gamma", 947}, /* greek small letter gamma, U+03B3 ISOgrk3 */ + {"ge", 8805}, /* greater-than or equal to, U+2265 ISOtech */ + {"gt", 62}, /* greater-than sign, U+003E ISOnum */ + {"hArr", 8660}, /* left right double arrow, U+21D4 ISOamsa */ + {"harr", 8596}, /* left right arrow, U+2194 ISOamsa */ + {"hearts", 9829}, /* black heart suit = valentine, U+2665 ISOpub */ + {"hellip", 8230}, /* horizontal ellipsis = three dot leader, U+2026 ISOpub */ + {"iacute", 237}, /* latin small letter i with acute, U+00ED ISOlat1 */ + {"icirc", 238}, /* latin small letter i with circumflex, U+00EE ISOlat1 */ + {"iexcl", 161}, /* inverted exclamation mark, U+00A1 ISOnum */ + {"igrave", 236}, /* latin small letter i with grave, U+00EC ISOlat1 */ + {"image", 8465}, /* blackletter capital I = imaginary part, U+2111 ISOamso */ + {"infin", 8734}, /* infinity, U+221E ISOtech */ + {"int", 8747}, /* integral, U+222B ISOtech */ + {"iota", 953}, /* greek small letter iota, U+03B9 ISOgrk3 */ + {"iquest", 191}, /* inverted question mark = turned question mark, U+00BF ISOnum */ + {"isin", 8712}, /* element of, U+2208 ISOtech */ + {"iuml", 239}, /* latin small letter i with diaeresis, U+00EF ISOlat1 */ + {"kappa", 954}, /* greek small letter kappa, U+03BA ISOgrk3 */ + {"lArr", 8656}, /* leftwards double arrow, U+21D0 ISOtech */ +/* Unicode does not say that lArr is the same as the 'is implied by' arrow + but also does not have any other character for that function. So ? lArr can + be used for 'is implied by' as ISOtech suggests */ + {"lambda", 955}, /* greek small letter lambda, U+03BB ISOgrk3 */ + {"lang", 9001}, /* left-pointing angle bracket = bra, U+2329 ISOtech */ +/* lang is NOT the same character as U+003C 'less than' or U+2039 'single left-pointing angle quotation mark' */ + {"laquo", 171}, /* left-pointing double angle quotation mark = left pointing guillemet, U+00AB ISOnum */ + {"larr", 8592}, /* leftwards arrow, U+2190 ISOnum */ + {"lceil", 8968}, /* left ceiling = apl upstile, U+2308 ISOamsc */ + {"ldquo", 8220}, /* left double quotation mark, U+201C ISOnum */ + {"le", 8804}, /* less-than or equal to, U+2264 ISOtech */ + {"lfloor", 8970}, /* left floor = apl downstile, U+230A ISOamsc */ + {"lowast", 8727}, /* asterisk operator, U+2217 ISOtech */ + {"loz", 9674}, /* lozenge, U+25CA ISOpub */ + {"lrm", 8206}, /* left-to-right mark, U+200E NEW RFC 2070 */ + {"lsaquo", 8249}, /* single left-pointing angle quotation mark, U+2039 ISO proposed */ +/* lsaquo is proposed but not yet ISO standardised */ + {"lsquo", 8216}, /* left single quotation mark, U+2018 ISOnum */ + {"lt", 60}, /* less-than sign, U+003C ISOnum */ + {"macr", 175}, /* macron = spacing macron = overline = APL overbar, U+00AF ISOdia */ + {"mdash", 8212}, /* em dash, U+2014 ISOpub */ + {"micro", 181}, /* micro sign, U+00B5 ISOnum */ + {"middot", 183}, /* middle dot = Georgian comma = Greek middle dot, U+00B7 ISOnum */ + {"minus", 8722}, /* minus sign, U+2212 ISOtech */ + {"mu", 956}, /* greek small letter mu, U+03BC ISOgrk3 */ + {"nabla", 8711}, /* nabla = backward difference, U+2207 ISOtech */ + {"nbsp", 160}, /* no-break space = non-breaking space, U+00A0 ISOnum */ + {"ndash", 8211}, /* en dash, U+2013 ISOpub */ + {"ne", 8800}, /* not equal to, U+2260 ISOtech */ + {"ni", 8715}, /* contains as member, U+220B ISOtech */ +/* should there be a more memorable name than 'ni'? */ + {"not", 172}, /* not sign = discretionary hyphen, U+00AC ISOnum */ + {"notin", 8713}, /* not an element of, U+2209 ISOtech */ + {"nsub", 8836}, /* not a subset of, U+2284 ISOamsn */ + {"ntilde", 241}, /* latin small letter n with tilde, U+00F1 ISOlat1 */ + {"nu", 957}, /* greek small letter nu, U+03BD ISOgrk3 */ + {"oacute", 243}, /* latin small letter o with acute, U+00F3 ISOlat1 */ + {"ocirc", 244}, /* latin small letter o with circumflex, U+00F4 ISOlat1 */ + {"oelig", 339}, /* latin small ligature oe, U+0153 ISOlat2 */ + {"ograve", 242}, /* latin small letter o with grave, U+00F2 ISOlat1 */ + {"oline", 8254}, /* overline = spacing overscore, U+203E NEW */ + {"omega", 969}, /* greek small letter omega, U+03C9 ISOgrk3 */ + {"omicron", 959}, /* greek small letter omicron, U+03BF NEW */ + {"oplus", 8853}, /* circled plus = direct sum, U+2295 ISOamsb */ + {"or", 8744}, /* logical or = vee, U+2228 ISOtech */ + {"ordf", 170}, /* feminine ordinal indicator, U+00AA ISOnum */ + {"ordm", 186}, /* masculine ordinal indicator, U+00BA ISOnum */ + {"oslash", 248}, /* latin small letter o with stroke, = latin small letter o slash, U+00F8 ISOlat1 */ + {"otilde", 245}, /* latin small letter o with tilde, U+00F5 ISOlat1 */ + {"otimes", 8855}, /* circled times = vector product, U+2297 ISOamsb */ + {"ouml", 246}, /* latin small letter o with diaeresis, U+00F6 ISOlat1 */ + {"para", 182}, /* pilcrow sign = paragraph sign, U+00B6 ISOnum */ + {"part", 8706}, /* partial differential, U+2202 ISOtech */ + {"permil", 8240}, /* per mille sign, U+2030 ISOtech */ + {"perp", 8869}, /* up tack = orthogonal to = perpendicular, U+22A5 ISOtech */ + {"phi", 966}, /* greek small letter phi, U+03C6 ISOgrk3 */ + {"pi", 960}, /* greek small letter pi, U+03C0 ISOgrk3 */ + {"piv", 982}, /* greek pi symbol, U+03D6 ISOgrk3 */ + {"plusmn", 177}, /* plus-minus sign = plus-or-minus sign, U+00B1 ISOnum */ + {"pound", 163}, /* pound sign, U+00A3 ISOnum */ + {"prime", 8242}, /* prime = minutes = feet, U+2032 ISOtech */ + {"prod", 8719}, /* n-ary product = product sign, U+220F ISOamsb */ +/* prod is NOT the same character as U+03A0 'greek capital letter pi' though the same glyph might be used for both */ + {"prop", 8733}, /* proportional to, U+221D ISOtech */ + {"psi", 968}, /* greek small letter psi, U+03C8 ISOgrk3 */ + {"quot", 34}, /* quotation mark = APL quote, U+0022 ISOnum */ + {"rArr", 8658}, /* rightwards double arrow, U+21D2 ISOtech */ +/* Unicode does not say this is the 'implies' character but does not have + another character with this function so ? + rArr can be used for 'implies' as ISOtech suggests */ + {"radic", 8730}, /* square root = radical sign, U+221A ISOtech */ + {"rang", 9002}, /* right-pointing angle bracket = ket, U+232A ISOtech */ +/* rang is NOT the same character as U+003E 'greater than' or U+203A 'single right-pointing angle quotation mark' */ + {"raquo", 187}, /* right-pointing double angle quotation mark = right pointing guillemet, U+00BB ISOnum */ + {"rarr", 8594}, /* rightwards arrow, U+2192 ISOnum */ + {"rceil", 8969}, /* right ceiling, U+2309 ISOamsc */ + {"rdquo", 8221}, /* right double quotation mark, U+201D ISOnum */ + {"real", 8476}, /* blackletter capital R = real part symbol, U+211C ISOamso */ + {"reg", 174}, /* registered sign = registered trade mark sign, U+00AE ISOnum */ + {"rfloor", 8971}, /* right floor, U+230B ISOamsc */ + {"rho", 961}, /* greek small letter rho, U+03C1 ISOgrk3 */ + {"rlm", 8207}, /* right-to-left mark, U+200F NEW RFC 2070 */ + {"rsaquo", 8250}, /* single right-pointing angle quotation mark, U+203A ISO proposed */ +/* rsaquo is proposed but not yet ISO standardised */ + {"rsquo", 8217}, /* right single quotation mark, U+2019 ISOnum */ + {"sbquo", 8218}, /* single low-9 quotation mark, U+201A NEW */ + {"scaron", 353}, /* latin small letter s with caron, U+0161 ISOlat2 */ + {"sdot", 8901}, /* dot operator, U+22C5 ISOamsb */ +/* dot operator is NOT the same character as U+00B7 middle dot */ + {"sect", 167}, /* section sign, U+00A7 ISOnum */ + {"shy", 173}, /* soft hyphen = discretionary hyphen, U+00AD ISOnum */ + {"sigma", 963}, /* greek small letter sigma, U+03C3 ISOgrk3 */ + {"sigmaf", 962}, /* greek small letter final sigma, U+03C2 ISOgrk3 */ + {"sim", 8764}, /* tilde operator = varies with = similar to, U+223C ISOtech */ +/* tilde operator is NOT the same character as the tilde, U+007E, although the same glyph might be used to represent both */ + {"spades", 9824}, /* black spade suit, U+2660 ISOpub */ +/* black here seems to mean filled as opposed to hollow */ + {"sub", 8834}, /* subset of, U+2282 ISOtech */ + {"sube", 8838}, /* subset of or equal to, U+2286 ISOtech */ + {"sum", 8721}, /* n-ary sumation, U+2211 ISOamsb */ +/* sum is NOT the same character as U+03A3 'greek capital letter sigma' though the same glyph might be used for both */ + {"sup", 8835}, /* superset of, U+2283 ISOtech */ +/* note that nsup, 'not a superset of, U+2283' is not covered by the Symbol + font encoding and is not included. Should it be, for symmetry? + It is in ISOamsn */ + {"sup1", 185}, /* superscript one = superscript digit one, U+00B9 ISOnum */ + {"sup2", 178}, /* superscript two = superscript digit two = squared, U+00B2 ISOnum */ + {"sup3", 179}, /* superscript three = superscript digit three = cubed, U+00B3 ISOnum */ + {"supe", 8839}, /* superset of or equal to, U+2287 ISOtech */ + {"szlig", 223}, /* latin small letter sharp s = ess-zed, U+00DF ISOlat1 */ + {"tau", 964}, /* greek small letter tau, U+03C4 ISOgrk3 */ + {"there4", 8756}, /* therefore, U+2234 ISOtech */ + {"theta", 952}, /* greek small letter theta, U+03B8 ISOgrk3 */ + {"thetasym", 977}, /* greek small letter theta symbol, U+03D1 NEW */ + {"thinsp", 8201}, /* thin space, U+2009 ISOpub */ + {"thorn", 254}, /* latin small letter thorn with, U+00FE ISOlat1 */ + {"tilde", 732}, /* small tilde, U+02DC ISOdia */ + {"times", 215}, /* multiplication sign, U+00D7 ISOnum */ + {"trade", 8482}, /* trade mark sign, U+2122 ISOnum */ + {"uArr", 8657}, /* upwards double arrow, U+21D1 ISOamsa */ + {"uacute", 250}, /* latin small letter u with acute, U+00FA ISOlat1 */ + {"uarr", 8593}, /* upwards arrow, U+2191 ISOnum */ + {"ucirc", 251}, /* latin small letter u with circumflex, U+00FB ISOlat1 */ + {"ugrave", 249}, /* latin small letter u with grave, U+00F9 ISOlat1 */ + {"uml", 168}, /* diaeresis = spacing diaeresis, U+00A8 ISOdia */ + {"upsih", 978}, /* greek upsilon with hook symbol, U+03D2 NEW */ + {"upsilon", 965}, /* greek small letter upsilon, U+03C5 ISOgrk3 */ + {"uuml", 252}, /* latin small letter u with diaeresis, U+00FC ISOlat1 */ + {"weierp", 8472}, /* script capital P = power set = Weierstrass p, U+2118 ISOamso */ + {"xi", 958}, /* greek small letter xi, U+03BE ISOgrk3 */ + {"yacute", 253}, /* latin small letter y with acute, U+00FD ISOlat1 */ + {"yen", 165}, /* yen sign = yuan sign, U+00A5 ISOnum */ + {"yuml", 255}, /* latin small letter y with diaeresis, U+00FF ISOlat1 */ + {"zeta", 950}, /* greek small letter zeta, U+03B6 ISOgrk3 */ + {"zwj", 8205}, /* zero width joiner, U+200D NEW RFC 2070 */ + {"zwnj", 8204}, /* zero width non-joiner, U+200C NEW RFC 2070 */ +}; + +#else /* not ENTITIES_HTML40_ONLY: */ +/*************************************************************************** + +This table prepared from ftp://ftp.unicode.org/MAPPINGS/VENDORS/MISC/SGML.TXT +original comment follows: + + +# Author: John Cowan <cowan@ccil.org> +# Date: 25 July 1997 +# +# The following table maps SGML character entities from various +# public sets (namely, ISOamsa, ISOamsb, ISOamsc, ISOamsn, ISOamso, +# ISOamsr, ISObox, ISOcyr1, ISOcyr2, ISOdia, ISOgrk1, ISOgrk2, +# ISOgrk3, ISOgrk4, ISOlat1, ISOlat2, ISOnum, ISOpub, ISOtech, +# HTMLspecial, HTMLsymbol) to corresponding Unicode characters. +# +# The table has four tab-separated columns: +# Column 1: SGML character entity name +# Column 2: SGML public entity set +# Column 3: Unicode 2.0 character code +# Column 4: Unicode 2.0 character name (UPPER CASE) +# Entries which don't have Unicode equivalents have "0x????" +# in Column 3 and a lower case description (from the public entity +# set DTD) in Column 4. The mapping is not reversible, because many +# distinctions are unified away in Unicode, particularly between +# mathematical symbols. + + + We just sort it and move column 2 away (line too long, sorry; + look at sgml.html in test/ directory for details). + +Changes: + * Add few (obsolete) synonyms for compatibility with Lynx/2.5 and up: + "brkbar" for "brvbar" 0x00A6 + "emdash" for "mdash" 0x2014 + "endash" for "ndash" 0x2013 + "hibar" for "macr" 0x00AF + BTW, lots of synonyms found in this table, we shouldn't worry about... + Totally around 1000 entries. + + +Modified by Jacob Poon <jacob.poon@utoronto.ca> + +This table is modified improve support of HTML 4.0 character entity references, +including Euro symbol support ("euro" 0x20AC). + +Known issues: + +The original table includes two different definitions of ◊ reference. +Since HTML 4.0 only uses U+25CA, the U+2727 definition is commented out, +until there is a good reason to put it back in. + +"b.delta" mapping fixed (was 0x03B3 = small gamma). + +At the end of the table, there are several unnumbered, commented references. +These are not defined in HTML 4.0, and will remain so until they are defined +in future SGML/HTML standards. + +The support for obsolete references are for backwards compatibility only. New +SGML/HTML documents should not depend on these references just because Lynx can +display them. + +****/ +{ + {"AElig", 0x00C6}, /* LATIN CAPITAL LETTER AE */ + {"Aacgr", 0x0386}, /* GREEK CAPITAL LETTER ALPHA WITH TONOS */ + {"Aacute", 0x00C1}, /* LATIN CAPITAL LETTER A WITH ACUTE */ + {"Abreve", 0x0102}, /* LATIN CAPITAL LETTER A WITH BREVE */ + {"Acirc", 0x00C2}, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX */ + {"Acy", 0x0410}, /* CYRILLIC CAPITAL LETTER A */ + {"Agr", 0x0391}, /* GREEK CAPITAL LETTER ALPHA */ + {"Agrave", 0x00C0}, /* LATIN CAPITAL LETTER A WITH GRAVE */ + {"Alpha", 0x0391}, /* GREEK CAPITAL LETTER ALPHA */ + {"Amacr", 0x0100}, /* LATIN CAPITAL LETTER A WITH MACRON */ + {"Aogon", 0x0104}, /* LATIN CAPITAL LETTER A WITH OGONEK */ + {"Aring", 0x00C5}, /* LATIN CAPITAL LETTER A WITH RING ABOVE */ + {"Atilde", 0x00C3}, /* LATIN CAPITAL LETTER A WITH TILDE */ + {"Auml", 0x00C4}, /* LATIN CAPITAL LETTER A WITH DIAERESIS */ + {"Barwed", 0x2306}, /* PERSPECTIVE */ + {"Bcy", 0x0411}, /* CYRILLIC CAPITAL LETTER BE */ + {"Beta", 0x0392}, /* GREEK CAPITAL LETTER BETA */ + {"Bgr", 0x0392}, /* GREEK CAPITAL LETTER BETA */ + {"CHcy", 0x0427}, /* CYRILLIC CAPITAL LETTER CHE */ + {"Cacute", 0x0106}, /* LATIN CAPITAL LETTER C WITH ACUTE */ + {"Cap", 0x22D2}, /* DOUBLE INTERSECTION */ + {"Ccaron", 0x010C}, /* LATIN CAPITAL LETTER C WITH CARON */ + {"Ccedil", 0x00C7}, /* LATIN CAPITAL LETTER C WITH CEDILLA */ + {"Ccirc", 0x0108}, /* LATIN CAPITAL LETTER C WITH CIRCUMFLEX */ + {"Cdot", 0x010A}, /* LATIN CAPITAL LETTER C WITH DOT ABOVE */ + {"Chi", 0x03A7}, /* GREEK CAPITAL LETTER CHI */ + {"Cup", 0x22D3}, /* DOUBLE UNION */ + {"DJcy", 0x0402}, /* CYRILLIC CAPITAL LETTER DJE */ + {"DScy", 0x0405}, /* CYRILLIC CAPITAL LETTER DZE */ + {"DZcy", 0x040F}, /* CYRILLIC CAPITAL LETTER DZHE */ + {"Dagger", 0x2021}, /* DOUBLE DAGGER */ + {"Dcaron", 0x010E}, /* LATIN CAPITAL LETTER D WITH CARON */ + {"Dcy", 0x0414}, /* CYRILLIC CAPITAL LETTER DE */ + {"Delta", 0x0394}, /* GREEK CAPITAL LETTER DELTA */ + {"Dgr", 0x0394}, /* GREEK CAPITAL LETTER DELTA */ + {"Dot", 0x00A8}, /* DIAERESIS */ + {"DotDot", 0x20DC}, /* COMBINING FOUR DOTS ABOVE */ + {"Dstrok", 0x0110}, /* LATIN CAPITAL LETTER D WITH STROKE */ + {"EEacgr", 0x0389}, /* GREEK CAPITAL LETTER ETA WITH TONOS */ + {"EEgr", 0x0397}, /* GREEK CAPITAL LETTER ETA */ + {"ENG", 0x014A}, /* LATIN CAPITAL LETTER ENG */ + {"ETH", 0x00D0}, /* LATIN CAPITAL LETTER ETH */ + {"Eacgr", 0x0388}, /* GREEK CAPITAL LETTER EPSILON WITH TONOS */ + {"Eacute", 0x00C9}, /* LATIN CAPITAL LETTER E WITH ACUTE */ + {"Ecaron", 0x011A}, /* LATIN CAPITAL LETTER E WITH CARON */ + {"Ecirc", 0x00CA}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX */ + {"Ecy", 0x042D}, /* CYRILLIC CAPITAL LETTER E */ + {"Edot", 0x0116}, /* LATIN CAPITAL LETTER E WITH DOT ABOVE */ + {"Egr", 0x0395}, /* GREEK CAPITAL LETTER EPSILON */ + {"Egrave", 0x00C8}, /* LATIN CAPITAL LETTER E WITH GRAVE */ + {"Emacr", 0x0112}, /* LATIN CAPITAL LETTER E WITH MACRON */ + {"Eogon", 0x0118}, /* LATIN CAPITAL LETTER E WITH OGONEK */ + {"Epsilon", 0x0395}, /* GREEK CAPITAL LETTER EPSILON */ + {"Eta", 0x0397}, /* GREEK CAPITAL LETTER ETA */ + {"Euml", 0x00CB}, /* LATIN CAPITAL LETTER E WITH DIAERESIS */ + {"Fcy", 0x0424}, /* CYRILLIC CAPITAL LETTER EF */ + {"GJcy", 0x0403}, /* CYRILLIC CAPITAL LETTER GJE */ + {"Gamma", 0x0393}, /* GREEK CAPITAL LETTER GAMMA */ + {"Gbreve", 0x011E}, /* LATIN CAPITAL LETTER G WITH BREVE */ + {"Gcedil", 0x0122}, /* LATIN CAPITAL LETTER G WITH CEDILLA */ + {"Gcirc", 0x011C}, /* LATIN CAPITAL LETTER G WITH CIRCUMFLEX */ + {"Gcy", 0x0413}, /* CYRILLIC CAPITAL LETTER GHE */ + {"Gdot", 0x0120}, /* LATIN CAPITAL LETTER G WITH DOT ABOVE */ + {"Gg", 0x22D9}, /* VERY MUCH GREATER-THAN */ + {"Ggr", 0x0393}, /* GREEK CAPITAL LETTER GAMMA */ + {"Gt", 0x226B}, /* MUCH GREATER-THAN */ + {"HARDcy", 0x042A}, /* CYRILLIC CAPITAL LETTER HARD SIGN */ + {"Hcirc", 0x0124}, /* LATIN CAPITAL LETTER H WITH CIRCUMFLEX */ + {"Hstrok", 0x0126}, /* LATIN CAPITAL LETTER H WITH STROKE */ + {"IEcy", 0x0415}, /* CYRILLIC CAPITAL LETTER IE */ + {"IJlig", 0x0132}, /* LATIN CAPITAL LIGATURE IJ */ + {"IOcy", 0x0401}, /* CYRILLIC CAPITAL LETTER IO */ + {"Iacgr", 0x038A}, /* GREEK CAPITAL LETTER IOTA WITH TONOS */ + {"Iacute", 0x00CD}, /* LATIN CAPITAL LETTER I WITH ACUTE */ + {"Icirc", 0x00CE}, /* LATIN CAPITAL LETTER I WITH CIRCUMFLEX */ + {"Icy", 0x0418}, /* CYRILLIC CAPITAL LETTER I */ + {"Idigr", 0x03AA}, /* GREEK CAPITAL LETTER IOTA WITH DIALYTIKA */ + {"Idot", 0x0130}, /* LATIN CAPITAL LETTER I WITH DOT ABOVE */ + {"Igr", 0x0399}, /* GREEK CAPITAL LETTER IOTA */ + {"Igrave", 0x00CC}, /* LATIN CAPITAL LETTER I WITH GRAVE */ + {"Imacr", 0x012A}, /* LATIN CAPITAL LETTER I WITH MACRON */ + {"Iogon", 0x012E}, /* LATIN CAPITAL LETTER I WITH OGONEK */ + {"Iota", 0x0399}, /* GREEK CAPITAL LETTER IOTA */ + {"Itilde", 0x0128}, /* LATIN CAPITAL LETTER I WITH TILDE */ + {"Iukcy", 0x0406}, /* CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN*/ + {"Iuml", 0x00CF}, /* LATIN CAPITAL LETTER I WITH DIAERESIS */ + {"Jcirc", 0x0134}, /* LATIN CAPITAL LETTER J WITH CIRCUMFLEX */ + {"Jcy", 0x0419}, /* CYRILLIC CAPITAL LETTER SHORT I */ + {"Jsercy", 0x0408}, /* CYRILLIC CAPITAL LETTER JE */ + {"Jukcy", 0x0404}, /* CYRILLIC CAPITAL LETTER UKRAINIAN IE */ + {"KHcy", 0x0425}, /* CYRILLIC CAPITAL LETTER HA */ + {"KHgr", 0x03A7}, /* GREEK CAPITAL LETTER CHI */ + {"KJcy", 0x040C}, /* CYRILLIC CAPITAL LETTER KJE */ + {"Kappa", 0x039A}, /* GREEK CAPITAL LETTER KAPPA */ + {"Kcedil", 0x0136}, /* LATIN CAPITAL LETTER K WITH CEDILLA */ + {"Kcy", 0x041A}, /* CYRILLIC CAPITAL LETTER KA */ + {"Kgr", 0x039A}, /* GREEK CAPITAL LETTER KAPPA */ + {"LJcy", 0x0409}, /* CYRILLIC CAPITAL LETTER LJE */ + {"Lacute", 0x0139}, /* LATIN CAPITAL LETTER L WITH ACUTE */ + {"Lambda", 0x039B}, /* GREEK CAPITAL LETTER LAMDA */ + {"Larr", 0x219E}, /* LEFTWARDS TWO HEADED ARROW */ + {"Lcaron", 0x013D}, /* LATIN CAPITAL LETTER L WITH CARON */ + {"Lcedil", 0x013B}, /* LATIN CAPITAL LETTER L WITH CEDILLA */ + {"Lcy", 0x041B}, /* CYRILLIC CAPITAL LETTER EL */ + {"Lgr", 0x039B}, /* GREEK CAPITAL LETTER LAMDA */ + {"Ll", 0x22D8}, /* VERY MUCH LESS-THAN */ + {"Lmidot", 0x013F}, /* LATIN CAPITAL LETTER L WITH MIDDLE DOT */ + {"Lstrok", 0x0141}, /* LATIN CAPITAL LETTER L WITH STROKE */ + {"Lt", 0x226A}, /* MUCH LESS-THAN */ + {"Mcy", 0x041C}, /* CYRILLIC CAPITAL LETTER EM */ + {"Mgr", 0x039C}, /* GREEK CAPITAL LETTER MU */ + {"Mu", 0x039C}, /* GREEK CAPITAL LETTER MU */ + {"NJcy", 0x040A}, /* CYRILLIC CAPITAL LETTER NJE */ + {"Nacute", 0x0143}, /* LATIN CAPITAL LETTER N WITH ACUTE */ + {"Ncaron", 0x0147}, /* LATIN CAPITAL LETTER N WITH CARON */ + {"Ncedil", 0x0145}, /* LATIN CAPITAL LETTER N WITH CEDILLA */ + {"Ncy", 0x041D}, /* CYRILLIC CAPITAL LETTER EN */ + {"Ngr", 0x039D}, /* GREEK CAPITAL LETTER NU */ + {"Ntilde", 0x00D1}, /* LATIN CAPITAL LETTER N WITH TILDE */ + {"Nu", 0x039D}, /* GREEK CAPITAL LETTER NU */ + {"OElig", 0x0152}, /* LATIN CAPITAL LIGATURE OE */ + {"OHacgr", 0x038F}, /* GREEK CAPITAL LETTER OMEGA WITH TONOS */ + {"OHgr", 0x03A9}, /* GREEK CAPITAL LETTER OMEGA */ + {"Oacgr", 0x038C}, /* GREEK CAPITAL LETTER OMICRON WITH TONOS */ + {"Oacute", 0x00D3}, /* LATIN CAPITAL LETTER O WITH ACUTE */ + {"Ocirc", 0x00D4}, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX */ + {"Ocy", 0x041E}, /* CYRILLIC CAPITAL LETTER O */ + {"Odblac", 0x0150}, /* LATIN CAPITAL LETTER O WITH DOUBLE ACUTE */ + {"Ogr", 0x039F}, /* GREEK CAPITAL LETTER OMICRON */ + {"Ograve", 0x00D2}, /* LATIN CAPITAL LETTER O WITH GRAVE */ + {"Omacr", 0x014C}, /* LATIN CAPITAL LETTER O WITH MACRON */ + {"Omega", 0x03A9}, /* GREEK CAPITAL LETTER OMEGA */ + {"Omicron", 0x039F}, /* GREEK CAPITAL LETTER OMICRON */ + {"Oslash", 0x00D8}, /* LATIN CAPITAL LETTER O WITH STROKE */ + {"Otilde", 0x00D5}, /* LATIN CAPITAL LETTER O WITH TILDE */ + {"Ouml", 0x00D6}, /* LATIN CAPITAL LETTER O WITH DIAERESIS */ + {"PHgr", 0x03A6}, /* GREEK CAPITAL LETTER PHI */ + {"PSgr", 0x03A8}, /* GREEK CAPITAL LETTER PSI */ + {"Pcy", 0x041F}, /* CYRILLIC CAPITAL LETTER PE */ + {"Pgr", 0x03A0}, /* GREEK CAPITAL LETTER PI */ + {"Phi", 0x03A6}, /* GREEK CAPITAL LETTER PHI */ + {"Pi", 0x03A0}, /* GREEK CAPITAL LETTER PI */ + {"Prime", 0x2033}, /* DOUBLE PRIME */ + {"Psi", 0x03A8}, /* GREEK CAPITAL LETTER PSI */ + {"Racute", 0x0154}, /* LATIN CAPITAL LETTER R WITH ACUTE */ + {"Rarr", 0x21A0}, /* RIGHTWARDS TWO HEADED ARROW */ + {"Rcaron", 0x0158}, /* LATIN CAPITAL LETTER R WITH CARON */ + {"Rcedil", 0x0156}, /* LATIN CAPITAL LETTER R WITH CEDILLA */ + {"Rcy", 0x0420}, /* CYRILLIC CAPITAL LETTER ER */ + {"Rgr", 0x03A1}, /* GREEK CAPITAL LETTER RHO */ + {"Rho", 0x03A1}, /* GREEK CAPITAL LETTER RHO */ + {"SHCHcy", 0x0429}, /* CYRILLIC CAPITAL LETTER SHCHA */ + {"SHcy", 0x0428}, /* CYRILLIC CAPITAL LETTER SHA */ + {"SOFTcy", 0x042C}, /* CYRILLIC CAPITAL LETTER SOFT SIGN */ + {"Sacute", 0x015A}, /* LATIN CAPITAL LETTER S WITH ACUTE */ + {"Scaron", 0x0160}, /* LATIN CAPITAL LETTER S WITH CARON */ + {"Scedil", 0x015E}, /* LATIN CAPITAL LETTER S WITH CEDILLA */ + {"Scirc", 0x015C}, /* LATIN CAPITAL LETTER S WITH CIRCUMFLEX */ + {"Scy", 0x0421}, /* CYRILLIC CAPITAL LETTER ES */ + {"Sgr", 0x03A3}, /* GREEK CAPITAL LETTER SIGMA */ + {"Sigma", 0x03A3}, /* GREEK CAPITAL LETTER SIGMA */ + {"Sub", 0x22D0}, /* DOUBLE SUBSET */ + {"Sup", 0x22D1}, /* DOUBLE SUPERSET */ + {"THORN", 0x00DE}, /* LATIN CAPITAL LETTER THORN */ + {"THgr", 0x0398}, /* GREEK CAPITAL LETTER THETA */ + {"TSHcy", 0x040B}, /* CYRILLIC CAPITAL LETTER TSHE */ + {"TScy", 0x0426}, /* CYRILLIC CAPITAL LETTER TSE */ + {"Tau", 0x03A4}, /* GREEK CAPITAL LETTER TAU */ + {"Tcaron", 0x0164}, /* LATIN CAPITAL LETTER T WITH CARON */ + {"Tcedil", 0x0162}, /* LATIN CAPITAL LETTER T WITH CEDILLA */ + {"Tcy", 0x0422}, /* CYRILLIC CAPITAL LETTER TE */ + {"Tgr", 0x03A4}, /* GREEK CAPITAL LETTER TAU */ + {"Theta", 0x0398}, /* GREEK CAPITAL LETTER THETA */ + {"Tstrok", 0x0166}, /* LATIN CAPITAL LETTER T WITH STROKE */ + {"Uacgr", 0x038E}, /* GREEK CAPITAL LETTER UPSILON WITH TONOS */ + {"Uacute", 0x00DA}, /* LATIN CAPITAL LETTER U WITH ACUTE */ + {"Ubrcy", 0x040E}, /* CYRILLIC CAPITAL LETTER SHORT U */ + {"Ubreve", 0x016C}, /* LATIN CAPITAL LETTER U WITH BREVE */ + {"Ucirc", 0x00DB}, /* LATIN CAPITAL LETTER U WITH CIRCUMFLEX */ + {"Ucy", 0x0423}, /* CYRILLIC CAPITAL LETTER U */ + {"Udblac", 0x0170}, /* LATIN CAPITAL LETTER U WITH DOUBLE ACUTE */ + {"Udigr", 0x03AB}, /* GREEK CAPITAL LETTER UPSILON WITH DIALYTIKA */ + {"Ugr", 0x03A5}, /* GREEK CAPITAL LETTER UPSILON */ + {"Ugrave", 0x00D9}, /* LATIN CAPITAL LETTER U WITH GRAVE */ + {"Umacr", 0x016A}, /* LATIN CAPITAL LETTER U WITH MACRON */ + {"Uogon", 0x0172}, /* LATIN CAPITAL LETTER U WITH OGONEK */ + {"Upsi", 0x03A5}, /* GREEK CAPITAL LETTER UPSILON */ + {"Upsilon", 0x03A5}, /* GREEK CAPITAL LETTER UPSILON */ + {"Uring", 0x016E}, /* LATIN CAPITAL LETTER U WITH RING ABOVE */ + {"Utilde", 0x0168}, /* LATIN CAPITAL LETTER U WITH TILDE */ + {"Uuml", 0x00DC}, /* LATIN CAPITAL LETTER U WITH DIAERESIS */ + {"Vcy", 0x0412}, /* CYRILLIC CAPITAL LETTER VE */ + {"Vdash", 0x22A9}, /* FORCES */ + {"Verbar", 0x2016}, /* DOUBLE VERTICAL LINE */ + {"Vvdash", 0x22AA}, /* TRIPLE VERTICAL BAR RIGHT TURNSTILE */ + {"Wcirc", 0x0174}, /* LATIN CAPITAL LETTER W WITH CIRCUMFLEX */ + {"Xgr", 0x039E}, /* GREEK CAPITAL LETTER XI */ + {"Xi", 0x039E}, /* GREEK CAPITAL LETTER XI */ + {"YAcy", 0x042F}, /* CYRILLIC CAPITAL LETTER YA */ + {"YIcy", 0x0407}, /* CYRILLIC CAPITAL LETTER YI */ + {"YUcy", 0x042E}, /* CYRILLIC CAPITAL LETTER YU */ + {"Yacute", 0x00DD}, /* LATIN CAPITAL LETTER Y WITH ACUTE */ + {"Ycirc", 0x0176}, /* LATIN CAPITAL LETTER Y WITH CIRCUMFLEX */ + {"Ycy", 0x042B}, /* CYRILLIC CAPITAL LETTER YERU */ + {"Yuml", 0x0178}, /* LATIN CAPITAL LETTER Y WITH DIAERESIS */ + {"ZHcy", 0x0416}, /* CYRILLIC CAPITAL LETTER ZHE */ + {"Zacute", 0x0179}, /* LATIN CAPITAL LETTER Z WITH ACUTE */ + {"Zcaron", 0x017D}, /* LATIN CAPITAL LETTER Z WITH CARON */ + {"Zcy", 0x0417}, /* CYRILLIC CAPITAL LETTER ZE */ + {"Zdot", 0x017B}, /* LATIN CAPITAL LETTER Z WITH DOT ABOVE */ + {"Zeta", 0x0396}, /* GREEK CAPITAL LETTER ZETA */ + {"Zgr", 0x0396}, /* GREEK CAPITAL LETTER ZETA */ + {"aacgr", 0x03AC}, /* GREEK SMALL LETTER ALPHA WITH TONOS */ + {"aacute", 0x00E1}, /* LATIN SMALL LETTER A WITH ACUTE */ + {"abreve", 0x0103}, /* LATIN SMALL LETTER A WITH BREVE */ + {"acirc", 0x00E2}, /* LATIN SMALL LETTER A WITH CIRCUMFLEX */ + {"acute", 0x00B4}, /* ACUTE ACCENT */ + {"acy", 0x0430}, /* CYRILLIC SMALL LETTER A */ + {"aelig", 0x00E6}, /* LATIN SMALL LETTER AE */ + {"agr", 0x03B1}, /* GREEK SMALL LETTER ALPHA */ + {"agrave", 0x00E0}, /* LATIN SMALL LETTER A WITH GRAVE */ + {"alefsym", 0x2135}, /* ALEF SYMBOL */ + {"aleph", 0x2135}, /* ALEF SYMBOL */ + {"alpha", 0x03B1}, /* GREEK SMALL LETTER ALPHA */ + {"amacr", 0x0101}, /* LATIN SMALL LETTER A WITH MACRON */ + {"amalg", 0x2210}, /* N-ARY COPRODUCT */ + {"amp", 0x0026}, /* AMPERSAND */ + {"and", 0x2227}, /* LOGICAL AND */ + {"ang", 0x2220}, /* ANGLE */ + {"ang90", 0x221F}, /* RIGHT ANGLE */ + {"angmsd", 0x2221}, /* MEASURED ANGLE */ + {"angsph", 0x2222}, /* SPHERICAL ANGLE */ + {"angst", 0x212B}, /* ANGSTROM SIGN */ + {"aogon", 0x0105}, /* LATIN SMALL LETTER A WITH OGONEK */ + {"ap", 0x2248}, /* ALMOST EQUAL TO */ + {"ape", 0x224A}, /* ALMOST EQUAL OR EQUAL TO */ + {"apos", 0x02BC}, /* MODIFIER LETTER APOSTROPHE */ + {"aring", 0x00E5}, /* LATIN SMALL LETTER A WITH RING ABOVE */ + {"ast", 0x002A}, /* ASTERISK */ + {"asymp", 0x2248}, /* ALMOST EQUAL TO */ + {"atilde", 0x00E3}, /* LATIN SMALL LETTER A WITH TILDE */ + {"auml", 0x00E4}, /* LATIN SMALL LETTER A WITH DIAERESIS */ + {"b.Delta", 0x0394}, /* GREEK CAPITAL LETTER DELTA */ + {"b.Gamma", 0x0393}, /* GREEK CAPITAL LETTER GAMMA */ + {"b.Lambda", 0x039B}, /* GREEK CAPITAL LETTER LAMDA */ + {"b.Omega", 0x03A9}, /* GREEK CAPITAL LETTER OMEGA */ + {"b.Phi", 0x03A6}, /* GREEK CAPITAL LETTER PHI */ + {"b.Pi", 0x03A0}, /* GREEK CAPITAL LETTER PI */ + {"b.Psi", 0x03A8}, /* GREEK CAPITAL LETTER PSI */ + {"b.Sigma", 0x03A3}, /* GREEK CAPITAL LETTER SIGMA */ + {"b.Theta", 0x0398}, /* GREEK CAPITAL LETTER THETA */ + {"b.Upsi", 0x03A5}, /* GREEK CAPITAL LETTER UPSILON */ + {"b.Xi", 0x039E}, /* GREEK CAPITAL LETTER XI */ + {"b.alpha", 0x03B1}, /* GREEK SMALL LETTER ALPHA */ + {"b.beta", 0x03B2}, /* GREEK SMALL LETTER BETA */ + {"b.chi", 0x03C7}, /* GREEK SMALL LETTER CHI */ + {"b.delta", 0x03B4}, /* GREEK SMALL LETTER DELTA */ + {"b.epsi", 0x03B5}, /* GREEK SMALL LETTER EPSILON */ + {"b.epsis", 0x03B5}, /* GREEK SMALL LETTER EPSILON */ + {"b.epsiv", 0x03B5}, /* GREEK SMALL LETTER EPSILON */ + {"b.eta", 0x03B7}, /* GREEK SMALL LETTER ETA */ + {"b.gamma", 0x03B3}, /* GREEK SMALL LETTER GAMMA */ + {"b.gammad", 0x03DC}, /* GREEK LETTER DIGAMMA */ + {"b.iota", 0x03B9}, /* GREEK SMALL LETTER IOTA */ + {"b.kappa", 0x03BA}, /* GREEK SMALL LETTER KAPPA */ + {"b.kappav", 0x03F0}, /* GREEK KAPPA SYMBOL */ + {"b.lambda", 0x03BB}, /* GREEK SMALL LETTER LAMDA */ + {"b.mu", 0x03BC}, /* GREEK SMALL LETTER MU */ + {"b.nu", 0x03BD}, /* GREEK SMALL LETTER NU */ + {"b.omega", 0x03CE}, /* GREEK SMALL LETTER OMEGA WITH TONOS */ + {"b.phis", 0x03C6}, /* GREEK SMALL LETTER PHI */ + {"b.phiv", 0x03D5}, /* GREEK PHI SYMBOL */ + {"b.pi", 0x03C0}, /* GREEK SMALL LETTER PI */ + {"b.piv", 0x03D6}, /* GREEK PI SYMBOL */ + {"b.psi", 0x03C8}, /* GREEK SMALL LETTER PSI */ + {"b.rho", 0x03C1}, /* GREEK SMALL LETTER RHO */ + {"b.rhov", 0x03F1}, /* GREEK RHO SYMBOL */ + {"b.sigma", 0x03C3}, /* GREEK SMALL LETTER SIGMA */ + {"b.sigmav", 0x03C2}, /* GREEK SMALL LETTER FINAL SIGMA */ + {"b.tau", 0x03C4}, /* GREEK SMALL LETTER TAU */ + {"b.thetas", 0x03B8}, /* GREEK SMALL LETTER THETA */ + {"b.thetav", 0x03D1}, /* GREEK THETA SYMBOL */ + {"b.upsi", 0x03C5}, /* GREEK SMALL LETTER UPSILON */ + {"b.xi", 0x03BE}, /* GREEK SMALL LETTER XI */ + {"b.zeta", 0x03B6}, /* GREEK SMALL LETTER ZETA */ + {"barwed", 0x22BC}, /* NAND */ + {"bcong", 0x224C}, /* ALL EQUAL TO */ + {"bcy", 0x0431}, /* CYRILLIC SMALL LETTER BE */ + {"bdquo", 0x201E}, /* DOUBLE LOW-9 QUOTATION MARK */ + {"becaus", 0x2235}, /* BECAUSE */ + {"bepsi", 0x220D}, /* SMALL CONTAINS AS MEMBER */ + {"bernou", 0x212C}, /* SCRIPT CAPITAL B */ + {"beta", 0x03B2}, /* GREEK SMALL LETTER BETA */ + {"beth", 0x2136}, /* BET SYMBOL */ + {"bgr", 0x03B2}, /* GREEK SMALL LETTER BETA */ + {"blank", 0x2423}, /* OPEN BOX */ + {"blk12", 0x2592}, /* MEDIUM SHADE */ + {"blk14", 0x2591}, /* LIGHT SHADE */ + {"blk34", 0x2593}, /* DARK SHADE */ + {"block", 0x2588}, /* FULL BLOCK */ + {"bottom", 0x22A5}, /* UP TACK */ + {"bowtie", 0x22C8}, /* BOWTIE */ + {"boxDL", 0x2557}, /* BOX DRAWINGS DOUBLE DOWN AND LEFT */ + {"boxDR", 0x2554}, /* BOX DRAWINGS DOUBLE DOWN AND RIGHT */ + {"boxDl", 0x2556}, /* BOX DRAWINGS DOWN DOUBLE AND LEFT SINGLE */ + {"boxDr", 0x2553}, /* BOX DRAWINGS DOWN DOUBLE AND RIGHT SINGLE */ + {"boxH", 0x2550}, /* BOX DRAWINGS DOUBLE HORIZONTAL */ + {"boxHD", 0x2566}, /* BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL */ + {"boxHU", 0x2569}, /* BOX DRAWINGS DOUBLE UP AND HORIZONTAL */ + {"boxHd", 0x2564}, /* BOX DRAWINGS DOWN SINGLE AND HORIZONTAL DOUBLE*/ + {"boxHu", 0x2567}, /* BOX DRAWINGS UP SINGLE AND HORIZONTAL DOUBLE */ + {"boxUL", 0x255D}, /* BOX DRAWINGS DOUBLE UP AND LEFT */ + {"boxUR", 0x255A}, /* BOX DRAWINGS DOUBLE UP AND RIGHT */ + {"boxUl", 0x255C}, /* BOX DRAWINGS UP DOUBLE AND LEFT SINGLE */ + {"boxUr", 0x2559}, /* BOX DRAWINGS UP DOUBLE AND RIGHT SINGLE */ + {"boxV", 0x2551}, /* BOX DRAWINGS DOUBLE VERTICAL */ + {"boxVH", 0x256C}, /* BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL */ + {"boxVL", 0x2563}, /* BOX DRAWINGS DOUBLE VERTICAL AND LEFT */ + {"boxVR", 0x2560}, /* BOX DRAWINGS DOUBLE VERTICAL AND RIGHT */ + {"boxVh", 0x256B}, /* BOX DRAWINGS VERTICAL DOUBLE AND HORIZONTAL SI*/ + {"boxVl", 0x2562}, /* BOX DRAWINGS VERTICAL DOUBLE AND LEFT SINGLE */ + {"boxVr", 0x255F}, /* BOX DRAWINGS VERTICAL DOUBLE AND RIGHT SINGLE */ + {"boxdL", 0x2555}, /* BOX DRAWINGS DOWN SINGLE AND LEFT DOUBLE */ + {"boxdR", 0x2552}, /* BOX DRAWINGS DOWN SINGLE AND RIGHT DOUBLE */ + {"boxdl", 0x2510}, /* BOX DRAWINGS LIGHT DOWN AND LEFT */ + {"boxdr", 0x250C}, /* BOX DRAWINGS LIGHT DOWN AND RIGHT */ + {"boxh", 0x2500}, /* BOX DRAWINGS LIGHT HORIZONTAL */ + {"boxhD", 0x2565}, /* BOX DRAWINGS DOWN DOUBLE AND HORIZONTAL SINGLE*/ + {"boxhU", 0x2568}, /* BOX DRAWINGS UP DOUBLE AND HORIZONTAL SINGLE */ + {"boxhd", 0x252C}, /* BOX DRAWINGS LIGHT DOWN AND HORIZONTAL */ + {"boxhu", 0x2534}, /* BOX DRAWINGS LIGHT UP AND HORIZONTAL */ + {"boxuL", 0x255B}, /* BOX DRAWINGS UP SINGLE AND LEFT DOUBLE */ + {"boxuR", 0x2558}, /* BOX DRAWINGS UP SINGLE AND RIGHT DOUBLE */ + {"boxul", 0x2518}, /* BOX DRAWINGS LIGHT UP AND LEFT */ + {"boxur", 0x2514}, /* BOX DRAWINGS LIGHT UP AND RIGHT */ + {"boxv", 0x2502}, /* BOX DRAWINGS LIGHT VERTICAL */ + {"boxvH", 0x256A}, /* BOX DRAWINGS VERTICAL SINGLE AND HORIZONTAL DO*/ + {"boxvL", 0x2561}, /* BOX DRAWINGS VERTICAL SINGLE AND LEFT DOUBLE */ + {"boxvR", 0x255E}, /* BOX DRAWINGS VERTICAL SINGLE AND RIGHT DOUBLE */ + {"boxvh", 0x253C}, /* BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL */ + {"boxvl", 0x2524}, /* BOX DRAWINGS LIGHT VERTICAL AND LEFT */ + {"boxvr", 0x251C}, /* BOX DRAWINGS LIGHT VERTICAL AND RIGHT */ + {"bprime", 0x2035}, /* REVERSED PRIME */ + {"breve", 0x02D8}, /* BREVE */ + {"brkbar", 0x00A6}, /* obsolete synonym for "brvbar" 0x00A6 */ + {"brvbar", 0x00A6}, /* BROKEN BAR */ + {"bsim", 0x223D}, /* REVERSED TILDE */ + {"bsime", 0x22CD}, /* REVERSED TILDE EQUALS */ + {"bsol", 0x005C}, /* REVERSE SOLIDUS */ + {"bull", 0x2022}, /* BULLET */ + {"bump", 0x224E}, /* GEOMETRICALLY EQUIVALENT TO */ + {"bumpe", 0x224F}, /* DIFFERENCE BETWEEN */ + {"cacute", 0x0107}, /* LATIN SMALL LETTER C WITH ACUTE */ + {"cap", 0x2229}, /* INTERSECTION */ + {"caret", 0x2041}, /* CARET INSERTION POINT */ + {"caron", 0x02C7}, /* CARON */ + {"ccaron", 0x010D}, /* LATIN SMALL LETTER C WITH CARON */ + {"ccedil", 0x00E7}, /* LATIN SMALL LETTER C WITH CEDILLA */ + {"ccirc", 0x0109}, /* LATIN SMALL LETTER C WITH CIRCUMFLEX */ + {"cdot", 0x010B}, /* LATIN SMALL LETTER C WITH DOT ABOVE */ + {"cedil", 0x00B8}, /* CEDILLA */ + {"cent", 0x00A2}, /* CENT SIGN */ + {"chcy", 0x0447}, /* CYRILLIC SMALL LETTER CHE */ + {"check", 0x2713}, /* CHECK MARK */ + {"chi", 0x03C7}, /* GREEK SMALL LETTER CHI */ + {"cir", 0x25CB}, /* WHITE CIRCLE */ + {"circ", 0x02C6}, /* MODIFIER LETTER CIRCUMFLEX ACCENT */ + {"cire", 0x2257}, /* RING EQUAL TO */ + {"clubs", 0x2663}, /* BLACK CLUB SUIT */ + {"colon", 0x003A}, /* COLON */ + {"colone", 0x2254}, /* COLON EQUALS */ + {"comma", 0x002C}, /* COMMA */ + {"commat", 0x0040}, /* COMMERCIAL AT */ + {"comp", 0x2201}, /* COMPLEMENT */ + {"compfn", 0x2218}, /* RING OPERATOR */ + {"cong", 0x2245}, /* APPROXIMATELY EQUAL TO */ + {"conint", 0x222E}, /* CONTOUR INTEGRAL */ + {"coprod", 0x2210}, /* N-ARY COPRODUCT */ + {"copy", 0x00A9}, /* COPYRIGHT SIGN */ + {"copysr", 0x2117}, /* SOUND RECORDING COPYRIGHT */ + {"crarr", 0x21B5}, /* DOWNWARDS ARROW WITH CORNER LEFTWARDS */ + {"cross", 0x2717}, /* BALLOT X */ + {"cuepr", 0x22DE}, /* EQUAL TO OR PRECEDES */ + {"cuesc", 0x22DF}, /* EQUAL TO OR SUCCEEDS */ + {"cularr", 0x21B6}, /* ANTICLOCKWISE TOP SEMICIRCLE ARROW */ + {"cup", 0x222A}, /* UNION */ + {"cupre", 0x227C}, /* PRECEDES OR EQUAL TO */ + {"curarr", 0x21B7}, /* CLOCKWISE TOP SEMICIRCLE ARROW */ + {"curren", 0x00A4}, /* CURRENCY SIGN */ + {"cuvee", 0x22CE}, /* CURLY LOGICAL OR */ + {"cuwed", 0x22CF}, /* CURLY LOGICAL AND */ + {"dArr", 0x21D3}, /* DOWNWARDS DOUBLE ARROW */ + {"dagger", 0x2020}, /* DAGGER */ + {"daleth", 0x2138}, /* DALET SYMBOL */ + {"darr", 0x2193}, /* DOWNWARDS ARROW */ + {"darr2", 0x21CA}, /* DOWNWARDS PAIRED ARROWS */ + {"dash", 0x2010}, /* HYPHEN */ + {"dashv", 0x22A3}, /* LEFT TACK */ + {"dblac", 0x02DD}, /* DOUBLE ACUTE ACCENT */ + {"dcaron", 0x010F}, /* LATIN SMALL LETTER D WITH CARON */ + {"dcy", 0x0434}, /* CYRILLIC SMALL LETTER DE */ + {"deg", 0x00B0}, /* DEGREE SIGN */ + {"delta", 0x03B4}, /* GREEK SMALL LETTER DELTA */ + {"dgr", 0x03B4}, /* GREEK SMALL LETTER DELTA */ + {"dharl", 0x21C3}, /* DOWNWARDS HARPOON WITH BARB LEFTWARDS */ + {"dharr", 0x21C2}, /* DOWNWARDS HARPOON WITH BARB RIGHTWARDS */ + {"diam", 0x22C4}, /* DIAMOND OPERATOR */ + {"diams", 0x2666}, /* BLACK DIAMOND SUIT */ + {"die", 0x00A8}, /* DIAERESIS */ + {"divide", 0x00F7}, /* DIVISION SIGN */ + {"divonx", 0x22C7}, /* DIVISION TIMES */ + {"djcy", 0x0452}, /* CYRILLIC SMALL LETTER DJE */ + {"dlarr", 0x2199}, /* SOUTH WEST ARROW */ + {"dlcorn", 0x231E}, /* BOTTOM LEFT CORNER */ + {"dlcrop", 0x230D}, /* BOTTOM LEFT CROP */ + {"dollar", 0x0024}, /* DOLLAR SIGN */ + {"dot", 0x02D9}, /* DOT ABOVE */ + {"drarr", 0x2198}, /* SOUTH EAST ARROW */ + {"drcorn", 0x231F}, /* BOTTOM RIGHT CORNER */ + {"drcrop", 0x230C}, /* BOTTOM RIGHT CROP */ + {"dscy", 0x0455}, /* CYRILLIC SMALL LETTER DZE */ + {"dstrok", 0x0111}, /* LATIN SMALL LETTER D WITH STROKE */ + {"dtri", 0x25BF}, /* WHITE DOWN-POINTING SMALL TRIANGLE */ + {"dtrif", 0x25BE}, /* BLACK DOWN-POINTING SMALL TRIANGLE */ + {"dzcy", 0x045F}, /* CYRILLIC SMALL LETTER DZHE */ + {"eDot", 0x2251}, /* GEOMETRICALLY EQUAL TO */ + {"eacgr", 0x03AD}, /* GREEK SMALL LETTER EPSILON WITH TONOS */ + {"eacute", 0x00E9}, /* LATIN SMALL LETTER E WITH ACUTE */ + {"ecaron", 0x011B}, /* LATIN SMALL LETTER E WITH CARON */ + {"ecir", 0x2256}, /* RING IN EQUAL TO */ + {"ecirc", 0x00EA}, /* LATIN SMALL LETTER E WITH CIRCUMFLEX */ + {"ecolon", 0x2255}, /* EQUALS COLON */ + {"ecy", 0x044D}, /* CYRILLIC SMALL LETTER E */ + {"edot", 0x0117}, /* LATIN SMALL LETTER E WITH DOT ABOVE */ + {"eeacgr", 0x03AE}, /* GREEK SMALL LETTER ETA WITH TONOS */ + {"eegr", 0x03B7}, /* GREEK SMALL LETTER ETA */ + {"efDot", 0x2252}, /* APPROXIMATELY EQUAL TO OR THE IMAGE OF */ + {"egr", 0x03B5}, /* GREEK SMALL LETTER EPSILON */ + {"egrave", 0x00E8}, /* LATIN SMALL LETTER E WITH GRAVE */ + {"egs", 0x22DD}, /* EQUAL TO OR GREATER-THAN */ + {"ell", 0x2113}, /* SCRIPT SMALL L */ + {"els", 0x22DC}, /* EQUAL TO OR LESS-THAN */ + {"emacr", 0x0113}, /* LATIN SMALL LETTER E WITH MACRON */ + {"emdash", 0x2014}, /* obsolete synonym for "mdash" 0x2014 */ + {"empty", 0x2205}, /* EMPTY SET */ + {"emsp", 0x2003}, /* EM SPACE */ + {"emsp13", 0x2004}, /* THREE-PER-EM SPACE */ + {"emsp14", 0x2005}, /* FOUR-PER-EM SPACE */ + {"endash", 0x2013}, /* obsolete synonym for "ndash" 0x2013 */ + {"eng", 0x014B}, /* LATIN SMALL LETTER ENG */ + {"ensp", 0x2002}, /* EN SPACE */ + {"eogon", 0x0119}, /* LATIN SMALL LETTER E WITH OGONEK */ + {"epsi", 0x03B5}, /* GREEK SMALL LETTER EPSILON */ + {"epsilon", 0x03B5}, /* GREEK SMALL LETTER EPSILON */ + {"epsis", 0x220A}, /* SMALL ELEMENT OF */ + {"equals", 0x003D}, /* EQUALS SIGN */ + {"equiv", 0x2261}, /* IDENTICAL TO */ + {"erDot", 0x2253}, /* IMAGE OF OR APPROXIMATELY EQUAL TO */ + {"esdot", 0x2250}, /* APPROACHES THE LIMIT */ + {"eta", 0x03B7}, /* GREEK SMALL LETTER ETA */ + {"eth", 0x00F0}, /* LATIN SMALL LETTER ETH */ + {"euml", 0x00EB}, /* LATIN SMALL LETTER E WITH DIAERESIS */ + {"euro", 0x20AC}, /* EURO SIGN */ + {"excl", 0x0021}, /* EXCLAMATION MARK */ + {"exist", 0x2203}, /* THERE EXISTS */ + {"fcy", 0x0444}, /* CYRILLIC SMALL LETTER EF */ + {"female", 0x2640}, /* FEMALE SIGN */ + {"ffilig", 0xFB03}, /* LATIN SMALL LIGATURE FFI */ + {"fflig", 0xFB00}, /* LATIN SMALL LIGATURE FF */ + {"ffllig", 0xFB04}, /* LATIN SMALL LIGATURE FFL */ + {"filig", 0xFB01}, /* LATIN SMALL LIGATURE FI */ + {"flat", 0x266D}, /* MUSIC FLAT SIGN */ + {"fllig", 0xFB02}, /* LATIN SMALL LIGATURE FL */ + {"fnof", 0x0192}, /* LATIN SMALL LETTER F WITH HOOK */ + {"forall", 0x2200}, /* FOR ALL */ + {"fork", 0x22D4}, /* PITCHFORK */ + {"frac12", 0x00BD}, /* VULGAR FRACTION ONE HALF */ + {"frac13", 0x2153}, /* VULGAR FRACTION ONE THIRD */ + {"frac14", 0x00BC}, /* VULGAR FRACTION ONE QUARTER */ + {"frac15", 0x2155}, /* VULGAR FRACTION ONE FIFTH */ + {"frac16", 0x2159}, /* VULGAR FRACTION ONE SIXTH */ + {"frac18", 0x215B}, /* VULGAR FRACTION ONE EIGHTH */ + {"frac23", 0x2154}, /* VULGAR FRACTION TWO THIRDS */ + {"frac25", 0x2156}, /* VULGAR FRACTION TWO FIFTHS */ + {"frac34", 0x00BE}, /* VULGAR FRACTION THREE QUARTERS */ + {"frac35", 0x2157}, /* VULGAR FRACTION THREE FIFTHS */ + {"frac38", 0x215C}, /* VULGAR FRACTION THREE EIGHTHS */ + {"frac45", 0x2158}, /* VULGAR FRACTION FOUR FIFTHS */ + {"frac56", 0x215A}, /* VULGAR FRACTION FIVE SIXTHS */ + {"frac58", 0x215D}, /* VULGAR FRACTION FIVE EIGHTHS */ + {"frac78", 0x215E}, /* VULGAR FRACTION SEVEN EIGHTHS */ + {"frasl", 0x2044}, /* FRACTION SLASH */ + {"frown", 0x2322}, /* FROWN */ + {"gE", 0x2267}, /* GREATER-THAN OVER EQUAL TO */ + {"gacute", 0x01F5}, /* LATIN SMALL LETTER G WITH ACUTE */ + {"gamma", 0x03B3}, /* GREEK SMALL LETTER GAMMA */ + {"gammad", 0x03DC}, /* GREEK LETTER DIGAMMA */ + {"gbreve", 0x011F}, /* LATIN SMALL LETTER G WITH BREVE */ + {"gcedil", 0x0123}, /* LATIN SMALL LETTER G WITH CEDILLA */ + {"gcirc", 0x011D}, /* LATIN SMALL LETTER G WITH CIRCUMFLEX */ + {"gcy", 0x0433}, /* CYRILLIC SMALL LETTER GHE */ + {"gdot", 0x0121}, /* LATIN SMALL LETTER G WITH DOT ABOVE */ + {"ge", 0x2265}, /* GREATER-THAN OR EQUAL TO */ + {"gel", 0x22DB}, /* GREATER-THAN EQUAL TO OR LESS-THAN */ + {"ges", 0x2265}, /* GREATER-THAN OR EQUAL TO */ + {"ggr", 0x03B3}, /* GREEK SMALL LETTER GAMMA */ + {"gimel", 0x2137}, /* GIMEL SYMBOL */ + {"gjcy", 0x0453}, /* CYRILLIC SMALL LETTER GJE */ + {"gl", 0x2277}, /* GREATER-THAN OR LESS-THAN */ + {"gnE", 0x2269}, /* GREATER-THAN BUT NOT EQUAL TO */ + {"gne", 0x2269}, /* GREATER-THAN BUT NOT EQUAL TO */ + {"gnsim", 0x22E7}, /* GREATER-THAN BUT NOT EQUIVALENT TO */ + {"grave", 0x0060}, /* GRAVE ACCENT */ + {"gsdot", 0x22D7}, /* GREATER-THAN WITH DOT */ + {"gsim", 0x2273}, /* GREATER-THAN OR EQUIVALENT TO */ + {"gt", 0x003E}, /* GREATER-THAN SIGN */ + {"gvnE", 0x2269}, /* GREATER-THAN BUT NOT EQUAL TO */ + {"hArr", 0x21D4}, /* LEFT RIGHT DOUBLE ARROW */ + {"hairsp", 0x200A}, /* HAIR SPACE */ + {"half", 0x00BD}, /* VULGAR FRACTION ONE HALF */ + {"hamilt", 0x210B}, /* SCRIPT CAPITAL H */ + {"hardcy", 0x044A}, /* CYRILLIC SMALL LETTER HARD SIGN */ + {"harr", 0x2194}, /* LEFT RIGHT ARROW */ + {"harrw", 0x21AD}, /* LEFT RIGHT WAVE ARROW */ + {"hcirc", 0x0125}, /* LATIN SMALL LETTER H WITH CIRCUMFLEX */ + {"hearts", 0x2665}, /* BLACK HEART SUIT */ + {"hellip", 0x2026}, /* HORIZONTAL ELLIPSIS */ + {"hibar", 0x00AF}, /* obsolete synonym for "macr" 0x00AF */ + {"horbar", 0x2015}, /* HORIZONTAL BAR */ + {"hstrok", 0x0127}, /* LATIN SMALL LETTER H WITH STROKE */ + {"hybull", 0x2043}, /* HYPHEN BULLET */ + {"hyphen", 0x002D}, /* HYPHEN-MINUS */ + {"iacgr", 0x03AF}, /* GREEK SMALL LETTER IOTA WITH TONOS */ + {"iacute", 0x00ED}, /* LATIN SMALL LETTER I WITH ACUTE */ + {"icirc", 0x00EE}, /* LATIN SMALL LETTER I WITH CIRCUMFLEX */ + {"icy", 0x0438}, /* CYRILLIC SMALL LETTER I */ + {"idiagr", 0x0390}, /* GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TON*/ + {"idigr", 0x03CA}, /* GREEK SMALL LETTER IOTA WITH DIALYTIKA */ + {"iecy", 0x0435}, /* CYRILLIC SMALL LETTER IE */ + {"iexcl", 0x00A1}, /* INVERTED EXCLAMATION MARK */ + {"iff", 0x21D4}, /* LEFT RIGHT DOUBLE ARROW */ + {"igr", 0x03B9}, /* GREEK SMALL LETTER IOTA */ + {"igrave", 0x00EC}, /* LATIN SMALL LETTER I WITH GRAVE */ + {"ijlig", 0x0133}, /* LATIN SMALL LIGATURE IJ */ + {"imacr", 0x012B}, /* LATIN SMALL LETTER I WITH MACRON */ + {"image", 0x2111}, /* BLACK-LETTER CAPITAL I */ + {"incare", 0x2105}, /* CARE OF */ + {"infin", 0x221E}, /* INFINITY */ + {"inodot", 0x0131}, /* LATIN SMALL LETTER DOTLESS I */ + {"int", 0x222B}, /* INTEGRAL */ + {"intcal", 0x22BA}, /* INTERCALATE */ + {"iocy", 0x0451}, /* CYRILLIC SMALL LETTER IO */ + {"iogon", 0x012F}, /* LATIN SMALL LETTER I WITH OGONEK */ + {"iota", 0x03B9}, /* GREEK SMALL LETTER IOTA */ + {"iquest", 0x00BF}, /* INVERTED QUESTION MARK */ + {"isin", 0x2208}, /* ELEMENT OF */ + {"itilde", 0x0129}, /* LATIN SMALL LETTER I WITH TILDE */ + {"iukcy", 0x0456}, /* CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I*/ + {"iuml", 0x00EF}, /* LATIN SMALL LETTER I WITH DIAERESIS */ + {"jcirc", 0x0135}, /* LATIN SMALL LETTER J WITH CIRCUMFLEX */ + {"jcy", 0x0439}, /* CYRILLIC SMALL LETTER SHORT I */ + {"jsercy", 0x0458}, /* CYRILLIC SMALL LETTER JE */ + {"jukcy", 0x0454}, /* CYRILLIC SMALL LETTER UKRAINIAN IE */ + {"kappa", 0x03BA}, /* GREEK SMALL LETTER KAPPA */ + {"kappav", 0x03F0}, /* GREEK KAPPA SYMBOL */ + {"kcedil", 0x0137}, /* LATIN SMALL LETTER K WITH CEDILLA */ + {"kcy", 0x043A}, /* CYRILLIC SMALL LETTER KA */ + {"kgr", 0x03BA}, /* GREEK SMALL LETTER KAPPA */ + {"kgreen", 0x0138}, /* LATIN SMALL LETTER KRA */ + {"khcy", 0x0445}, /* CYRILLIC SMALL LETTER HA */ + {"khgr", 0x03C7}, /* GREEK SMALL LETTER CHI */ + {"kjcy", 0x045C}, /* CYRILLIC SMALL LETTER KJE */ + {"lAarr", 0x21DA}, /* LEFTWARDS TRIPLE ARROW */ + {"lArr", 0x21D0}, /* LEFTWARDS DOUBLE ARROW */ + {"lE", 0x2266}, /* LESS-THAN OVER EQUAL TO */ + {"lacute", 0x013A}, /* LATIN SMALL LETTER L WITH ACUTE */ + {"lagran", 0x2112}, /* SCRIPT CAPITAL L */ + {"lambda", 0x03BB}, /* GREEK SMALL LETTER LAMDA */ + {"lang", 0x2329}, /* LEFT-POINTING ANGLE BRACKET */ + {"laquo", 0x00AB}, /* LEFT-POINTING DOUBLE ANGLE QUOTATION MARK */ + {"larr", 0x2190}, /* LEFTWARDS ARROW */ + {"larr2", 0x21C7}, /* LEFTWARDS PAIRED ARROWS */ + {"larrhk", 0x21A9}, /* LEFTWARDS ARROW WITH HOOK */ + {"larrlp", 0x21AB}, /* LEFTWARDS ARROW WITH LOOP */ + {"larrtl", 0x21A2}, /* LEFTWARDS ARROW WITH TAIL */ + {"lcaron", 0x013E}, /* LATIN SMALL LETTER L WITH CARON */ + {"lcedil", 0x013C}, /* LATIN SMALL LETTER L WITH CEDILLA */ + {"lceil", 0x2308}, /* LEFT CEILING */ + {"lcub", 0x007B}, /* LEFT CURLY BRACKET */ + {"lcy", 0x043B}, /* CYRILLIC SMALL LETTER EL */ + {"ldot", 0x22D6}, /* LESS-THAN WITH DOT */ + {"ldquo", 0x201C}, /* LEFT DOUBLE QUOTATION MARK */ + {"ldquor", 0x201E}, /* DOUBLE LOW-9 QUOTATION MARK */ + {"le", 0x2264}, /* LESS-THAN OR EQUAL TO */ + {"leg", 0x22DA}, /* LESS-THAN EQUAL TO OR GREATER-THAN */ + {"les", 0x2264}, /* LESS-THAN OR EQUAL TO */ + {"lfloor", 0x230A}, /* LEFT FLOOR */ + {"lg", 0x2276}, /* LESS-THAN OR GREATER-THAN */ + {"lgr", 0x03BB}, /* GREEK SMALL LETTER LAMDA */ + {"lhard", 0x21BD}, /* LEFTWARDS HARPOON WITH BARB DOWNWARDS */ + {"lharu", 0x21BC}, /* LEFTWARDS HARPOON WITH BARB UPWARDS */ + {"lhblk", 0x2584}, /* LOWER HALF BLOCK */ + {"ljcy", 0x0459}, /* CYRILLIC SMALL LETTER LJE */ + {"lmidot", 0x0140}, /* LATIN SMALL LETTER L WITH MIDDLE DOT */ + {"lnE", 0x2268}, /* LESS-THAN BUT NOT EQUAL TO */ + {"lne", 0x2268}, /* LESS-THAN BUT NOT EQUAL TO */ + {"lnsim", 0x22E6}, /* LESS-THAN BUT NOT EQUIVALENT TO */ + {"lowast", 0x2217}, /* ASTERISK OPERATOR */ + {"lowbar", 0x005F}, /* LOW LINE */ + {"loz", 0x25CA}, /* LOZENGE */ +/*{"loz", 0x2727}, WHITE FOUR POINTED STAR */ + /* Warning: Duplicated ◊ entry. HTML 4,0 defines it as U+25CA. */ + {"lozf", 0x2726}, /* BLACK FOUR POINTED STAR */ + {"lpar", 0x0028}, /* LEFT PARENTHESIS */ + {"lrarr2", 0x21C6}, /* LEFTWARDS ARROW OVER RIGHTWARDS ARROW */ + {"lrhar2", 0x21CB}, /* LEFTWARDS HARPOON OVER RIGHTWARDS HARPOON */ + {"lrm", 0x200E}, /* LEFT-TO-RIGHT MARK */ + {"lsaquo", 0x2039}, /* SINGLE LEFT-POINTING ANGLE QUOTATION MARK */ + {"lsh", 0x21B0}, /* UPWARDS ARROW WITH TIP LEFTWARDS */ + {"lsim", 0x2272}, /* LESS-THAN OR EQUIVALENT TO */ + {"lsqb", 0x005B}, /* LEFT SQUARE BRACKET */ + {"lsquo", 0x2018}, /* LEFT SINGLE QUOTATION MARK */ + {"lsquor", 0x201A}, /* SINGLE LOW-9 QUOTATION MARK */ + {"lstrok", 0x0142}, /* LATIN SMALL LETTER L WITH STROKE */ + {"lt", 0x003C}, /* LESS-THAN SIGN */ + {"lthree", 0x22CB}, /* LEFT SEMIDIRECT PRODUCT */ + {"ltimes", 0x22C9}, /* LEFT NORMAL FACTOR SEMIDIRECT PRODUCT */ + {"ltri", 0x25C3}, /* WHITE LEFT-POINTING SMALL TRIANGLE */ + {"ltrie", 0x22B4}, /* NORMAL SUBGROUP OF OR EQUAL TO */ + {"ltrif", 0x25C2}, /* BLACK LEFT-POINTING SMALL TRIANGLE */ + {"lvnE", 0x2268}, /* LESS-THAN BUT NOT EQUAL TO */ + {"macr", 0x00AF}, /* MACRON */ + {"male", 0x2642}, /* MALE SIGN */ + {"malt", 0x2720}, /* MALTESE CROSS */ + {"map", 0x21A6}, /* RIGHTWARDS ARROW FROM BAR */ + {"marker", 0x25AE}, /* BLACK VERTICAL RECTANGLE */ + {"mcy", 0x043C}, /* CYRILLIC SMALL LETTER EM */ + {"mdash", 0x2014}, /* EM DASH */ + {"mgr", 0x03BC}, /* GREEK SMALL LETTER MU */ + {"micro", 0x00B5}, /* MICRO SIGN */ + {"mid", 0x2223}, /* DIVIDES */ + {"middot", 0x00B7}, /* MIDDLE DOT */ + {"minus", 0x2212}, /* MINUS SIGN */ + {"minusb", 0x229F}, /* SQUARED MINUS */ + {"mldr", 0x2026}, /* HORIZONTAL ELLIPSIS */ + {"mnplus", 0x2213}, /* MINUS-OR-PLUS SIGN */ + {"models", 0x22A7}, /* MODELS */ + {"mu", 0x03BC}, /* GREEK SMALL LETTER MU */ + {"mumap", 0x22B8}, /* MULTIMAP */ + {"nVDash", 0x22AF}, /* NEGATED DOUBLE VERTICAL BAR DOUBLE RIGHT TURNS*/ + {"nVdash", 0x22AE}, /* DOES NOT FORCE */ + {"nabla", 0x2207}, /* NABLA */ + {"nacute", 0x0144}, /* LATIN SMALL LETTER N WITH ACUTE */ + {"nap", 0x2249}, /* NOT ALMOST EQUAL TO */ + {"napos", 0x0149}, /* LATIN SMALL LETTER N PRECEDED BY APOSTROPHE */ + {"natur", 0x266E}, /* MUSIC NATURAL SIGN */ + {"nbsp", 0x00A0}, /* NO-BREAK SPACE */ + {"ncaron", 0x0148}, /* LATIN SMALL LETTER N WITH CARON */ + {"ncedil", 0x0146}, /* LATIN SMALL LETTER N WITH CEDILLA */ + {"ncong", 0x2247}, /* NEITHER APPROXIMATELY NOR ACTUALLY EQUAL TO */ + {"ncy", 0x043D}, /* CYRILLIC SMALL LETTER EN */ + {"ndash", 0x2013}, /* EN DASH */ + {"ne", 0x2260}, /* NOT EQUAL TO */ + {"nearr", 0x2197}, /* NORTH EAST ARROW */ + {"nequiv", 0x2262}, /* NOT IDENTICAL TO */ + {"nexist", 0x2204}, /* THERE DOES NOT EXIST */ + {"nge", 0x2271}, /* NEITHER GREATER-THAN NOR EQUAL TO */ + {"nges", 0x2271}, /* NEITHER GREATER-THAN NOR EQUAL TO */ + {"ngr", 0x03BD}, /* GREEK SMALL LETTER NU */ + {"ngt", 0x226F}, /* NOT GREATER-THAN */ + {"nhArr", 0x21CE}, /* LEFT RIGHT DOUBLE ARROW WITH STROKE */ + {"nharr", 0x21AE}, /* LEFT RIGHT ARROW WITH STROKE */ + {"ni", 0x220B}, /* CONTAINS AS MEMBER */ + {"njcy", 0x045A}, /* CYRILLIC SMALL LETTER NJE */ + {"nlArr", 0x21CD}, /* LEFTWARDS DOUBLE ARROW WITH STROKE */ + {"nlarr", 0x219A}, /* LEFTWARDS ARROW WITH STROKE */ + {"nldr", 0x2025}, /* TWO DOT LEADER */ + {"nle", 0x2270}, /* NEITHER LESS-THAN NOR EQUAL TO */ + {"nles", 0x2270}, /* NEITHER LESS-THAN NOR EQUAL TO */ + {"nlt", 0x226E}, /* NOT LESS-THAN */ + {"nltri", 0x22EA}, /* NOT NORMAL SUBGROUP OF */ + {"nltrie", 0x22EC}, /* NOT NORMAL SUBGROUP OF OR EQUAL TO */ + {"nmid", 0x2224}, /* DOES NOT DIVIDE */ + {"not", 0x00AC}, /* NOT SIGN */ + {"notin", 0x2209}, /* NOT AN ELEMENT OF */ + {"npar", 0x2226}, /* NOT PARALLEL TO */ + {"npr", 0x2280}, /* DOES NOT PRECEDE */ + {"npre", 0x22E0}, /* DOES NOT PRECEDE OR EQUAL */ + {"nrArr", 0x21CF}, /* RIGHTWARDS DOUBLE ARROW WITH STROKE */ + {"nrarr", 0x219B}, /* RIGHTWARDS ARROW WITH STROKE */ + {"nrtri", 0x22EB}, /* DOES NOT CONTAIN AS NORMAL SUBGROUP */ + {"nrtrie", 0x22ED}, /* DOES NOT CONTAIN AS NORMAL SUBGROUP OR EQUAL */ + {"nsc", 0x2281}, /* DOES NOT SUCCEED */ + {"nsce", 0x22E1}, /* DOES NOT SUCCEED OR EQUAL */ + {"nsim", 0x2241}, /* NOT TILDE */ + {"nsime", 0x2244}, /* NOT ASYMPTOTICALLY EQUAL TO */ + {"nspar", 0x2226}, /* NOT PARALLEL TO */ + {"nsub", 0x2284}, /* NOT A SUBSET OF */ + {"nsubE", 0x2288}, /* NEITHER A SUBSET OF NOR EQUAL TO */ + {"nsube", 0x2288}, /* NEITHER A SUBSET OF NOR EQUAL TO */ + {"nsup", 0x2285}, /* NOT A SUPERSET OF */ + {"nsupE", 0x2289}, /* NEITHER A SUPERSET OF NOR EQUAL TO */ + {"nsupe", 0x2289}, /* NEITHER A SUPERSET OF NOR EQUAL TO */ + {"ntilde", 0x00F1}, /* LATIN SMALL LETTER N WITH TILDE */ + {"nu", 0x03BD}, /* GREEK SMALL LETTER NU */ + {"num", 0x0023}, /* NUMBER SIGN */ + {"numero", 0x2116}, /* NUMERO SIGN */ + {"numsp", 0x2007}, /* FIGURE SPACE */ + {"nvDash", 0x22AD}, /* NOT TRUE */ + {"nvdash", 0x22AC}, /* DOES NOT PROVE */ + {"nwarr", 0x2196}, /* NORTH WEST ARROW */ + {"oS", 0x24C8}, /* CIRCLED LATIN CAPITAL LETTER S */ + {"oacgr", 0x03CC}, /* GREEK SMALL LETTER OMICRON WITH TONOS */ + {"oacute", 0x00F3}, /* LATIN SMALL LETTER O WITH ACUTE */ + {"oast", 0x229B}, /* CIRCLED ASTERISK OPERATOR */ + {"ocir", 0x229A}, /* CIRCLED RING OPERATOR */ + {"ocirc", 0x00F4}, /* LATIN SMALL LETTER O WITH CIRCUMFLEX */ + {"ocy", 0x043E}, /* CYRILLIC SMALL LETTER O */ + {"odash", 0x229D}, /* CIRCLED DASH */ + {"odblac", 0x0151}, /* LATIN SMALL LETTER O WITH DOUBLE ACUTE */ + {"odot", 0x2299}, /* CIRCLED DOT OPERATOR */ + {"oelig", 0x0153}, /* LATIN SMALL LIGATURE OE */ + {"ogon", 0x02DB}, /* OGONEK */ + {"ogr", 0x03BF}, /* GREEK SMALL LETTER OMICRON */ + {"ograve", 0x00F2}, /* LATIN SMALL LETTER O WITH GRAVE */ + {"ohacgr", 0x03CE}, /* GREEK SMALL LETTER OMEGA WITH TONOS */ + {"ohgr", 0x03C9}, /* GREEK SMALL LETTER OMEGA */ + {"ohm", 0x2126}, /* OHM SIGN */ + {"olarr", 0x21BA}, /* ANTICLOCKWISE OPEN CIRCLE ARROW */ + {"oline", 0x203E}, /* OVERLINE */ + {"omacr", 0x014D}, /* LATIN SMALL LETTER O WITH MACRON */ + {"omega", 0x03C9}, /* GREEK SMALL LETTER OMEGA */ + {"omicron", 0x03BF}, /* GREEK SMALL LETTER OMICRON */ + {"ominus", 0x2296}, /* CIRCLED MINUS */ + {"oplus", 0x2295}, /* CIRCLED PLUS */ + {"or", 0x2228}, /* LOGICAL OR */ + {"orarr", 0x21BB}, /* CLOCKWISE OPEN CIRCLE ARROW */ + {"order", 0x2134}, /* SCRIPT SMALL O */ + {"ordf", 0x00AA}, /* FEMININE ORDINAL INDICATOR */ + {"ordm", 0x00BA}, /* MASCULINE ORDINAL INDICATOR */ + {"oslash", 0x00F8}, /* LATIN SMALL LETTER O WITH STROKE */ + {"osol", 0x2298}, /* CIRCLED DIVISION SLASH */ + {"otilde", 0x00F5}, /* LATIN SMALL LETTER O WITH TILDE */ + {"otimes", 0x2297}, /* CIRCLED TIMES */ + {"ouml", 0x00F6}, /* LATIN SMALL LETTER O WITH DIAERESIS */ + {"par", 0x2225}, /* PARALLEL TO */ + {"para", 0x00B6}, /* PILCROW SIGN */ + {"part", 0x2202}, /* PARTIAL DIFFERENTIAL */ + {"pcy", 0x043F}, /* CYRILLIC SMALL LETTER PE */ + {"percnt", 0x0025}, /* PERCENT SIGN */ + {"period", 0x002E}, /* FULL STOP */ + {"permil", 0x2030}, /* PER MILLE SIGN */ + {"perp", 0x22A5}, /* UP TACK */ + {"pgr", 0x03C0}, /* GREEK SMALL LETTER PI */ + {"phgr", 0x03C6}, /* GREEK SMALL LETTER PHI */ + {"phi", 0x03C6}, /* GREEK SMALL LETTER PHI */ + {"phis", 0x03C6}, /* GREEK SMALL LETTER PHI */ + {"phiv", 0x03D5}, /* GREEK PHI SYMBOL */ + {"phmmat", 0x2133}, /* SCRIPT CAPITAL M */ + {"phone", 0x260E}, /* BLACK TELEPHONE */ + {"pi", 0x03C0}, /* GREEK SMALL LETTER PI */ + {"piv", 0x03D6}, /* GREEK PI SYMBOL */ + {"planck", 0x210F}, /* PLANCK CONSTANT OVER TWO PI */ + {"plus", 0x002B}, /* PLUS SIGN */ + {"plusb", 0x229E}, /* SQUARED PLUS */ + {"plusdo", 0x2214}, /* DOT PLUS */ + {"plusmn", 0x00B1}, /* PLUS-MINUS SIGN */ + {"pound", 0x00A3}, /* POUND SIGN */ + {"pr", 0x227A}, /* PRECEDES */ + {"pre", 0x227C}, /* PRECEDES OR EQUAL TO */ + {"prime", 0x2032}, /* PRIME */ + {"prnsim", 0x22E8}, /* PRECEDES BUT NOT EQUIVALENT TO */ + {"prod", 0x220F}, /* N-ARY PRODUCT */ + {"prop", 0x221D}, /* PROPORTIONAL TO */ + {"prsim", 0x227E}, /* PRECEDES OR EQUIVALENT TO */ + {"psgr", 0x03C8}, /* GREEK SMALL LETTER PSI */ + {"psi", 0x03C8}, /* GREEK SMALL LETTER PSI */ + {"puncsp", 0x2008}, /* PUNCTUATION SPACE */ + {"quest", 0x003F}, /* QUESTION MARK */ + {"quot", 0x0022}, /* QUOTATION MARK */ + {"rAarr", 0x21DB}, /* RIGHTWARDS TRIPLE ARROW */ + {"rArr", 0x21D2}, /* RIGHTWARDS DOUBLE ARROW */ + {"racute", 0x0155}, /* LATIN SMALL LETTER R WITH ACUTE */ + {"radic", 0x221A}, /* SQUARE ROOT */ + {"rang", 0x232A}, /* RIGHT-POINTING ANGLE BRACKET */ + {"raquo", 0x00BB}, /* RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK */ + {"rarr", 0x2192}, /* RIGHTWARDS ARROW */ + {"rarr2", 0x21C9}, /* RIGHTWARDS PAIRED ARROWS */ + {"rarrhk", 0x21AA}, /* RIGHTWARDS ARROW WITH HOOK */ + {"rarrlp", 0x21AC}, /* RIGHTWARDS ARROW WITH LOOP */ + {"rarrtl", 0x21A3}, /* RIGHTWARDS ARROW WITH TAIL */ + {"rarrw", 0x219D}, /* RIGHTWARDS WAVE ARROW */ + {"rcaron", 0x0159}, /* LATIN SMALL LETTER R WITH CARON */ + {"rcedil", 0x0157}, /* LATIN SMALL LETTER R WITH CEDILLA */ + {"rceil", 0x2309}, /* RIGHT CEILING */ + {"rcub", 0x007D}, /* RIGHT CURLY BRACKET */ + {"rcy", 0x0440}, /* CYRILLIC SMALL LETTER ER */ + {"rdquo", 0x201D}, /* RIGHT DOUBLE QUOTATION MARK */ + {"rdquor", 0x201C}, /* LEFT DOUBLE QUOTATION MARK */ + {"real", 0x211C}, /* BLACK-LETTER CAPITAL R */ + {"rect", 0x25AD}, /* WHITE RECTANGLE */ + {"reg", 0x00AE}, /* REGISTERED SIGN */ + {"rfloor", 0x230B}, /* RIGHT FLOOR */ + {"rgr", 0x03C1}, /* GREEK SMALL LETTER RHO */ + {"rhard", 0x21C1}, /* RIGHTWARDS HARPOON WITH BARB DOWNWARDS */ + {"rharu", 0x21C0}, /* RIGHTWARDS HARPOON WITH BARB UPWARDS */ + {"rho", 0x03C1}, /* GREEK SMALL LETTER RHO */ + {"rhov", 0x03F1}, /* GREEK RHO SYMBOL */ + {"ring", 0x02DA}, /* RING ABOVE */ + {"rlarr2", 0x21C4}, /* RIGHTWARDS ARROW OVER LEFTWARDS ARROW */ + {"rlhar2", 0x21CC}, /* RIGHTWARDS HARPOON OVER LEFTWARDS HARPOON */ + {"rlm", 0x200F}, /* RIGHT-TO-LEFT MARK */ + {"rpar", 0x0029}, /* RIGHT PARENTHESIS */ + {"rsaquo", 0x203A}, /* SINGLE RIGHT-POINTING ANGLE QUOTATION MARK */ + {"rsh", 0x21B1}, /* UPWARDS ARROW WITH TIP RIGHTWARDS */ + {"rsqb", 0x005D}, /* RIGHT SQUARE BRACKET */ + {"rsquo", 0x2019}, /* RIGHT SINGLE QUOTATION MARK */ + {"rsquor", 0x2018}, /* LEFT SINGLE QUOTATION MARK */ + {"rthree", 0x22CC}, /* RIGHT SEMIDIRECT PRODUCT */ + {"rtimes", 0x22CA}, /* RIGHT NORMAL FACTOR SEMIDIRECT PRODUCT */ + {"rtri", 0x25B9}, /* WHITE RIGHT-POINTING SMALL TRIANGLE */ + {"rtrie", 0x22B5}, /* CONTAINS AS NORMAL SUBGROUP OR EQUAL TO */ + {"rtrif", 0x25B8}, /* BLACK RIGHT-POINTING SMALL TRIANGLE */ + {"rx", 0x211E}, /* PRESCRIPTION TAKE */ + {"sacute", 0x015B}, /* LATIN SMALL LETTER S WITH ACUTE */ + {"samalg", 0x2210}, /* N-ARY COPRODUCT */ + {"sbquo", 0x201A}, /* SINGLE LOW-9 QUOTATION MARK */ + {"sbsol", 0x005C}, /* REVERSE SOLIDUS */ + {"sc", 0x227B}, /* SUCCEEDS */ + {"scaron", 0x0161}, /* LATIN SMALL LETTER S WITH CARON */ + {"sccue", 0x227D}, /* SUCCEEDS OR EQUAL TO */ + {"sce", 0x227D}, /* SUCCEEDS OR EQUAL TO */ + {"scedil", 0x015F}, /* LATIN SMALL LETTER S WITH CEDILLA */ + {"scirc", 0x015D}, /* LATIN SMALL LETTER S WITH CIRCUMFLEX */ + {"scnsim", 0x22E9}, /* SUCCEEDS BUT NOT EQUIVALENT TO */ + {"scsim", 0x227F}, /* SUCCEEDS OR EQUIVALENT TO */ + {"scy", 0x0441}, /* CYRILLIC SMALL LETTER ES */ + {"sdot", 0x22C5}, /* DOT OPERATOR */ + {"sdotb", 0x22A1}, /* SQUARED DOT OPERATOR */ + {"sect", 0x00A7}, /* SECTION SIGN */ + {"semi", 0x003B}, /* SEMICOLON */ + {"setmn", 0x2216}, /* SET MINUS */ + {"sext", 0x2736}, /* SIX POINTED BLACK STAR */ + {"sfgr", 0x03C2}, /* GREEK SMALL LETTER FINAL SIGMA */ + {"sfrown", 0x2322}, /* FROWN */ + {"sgr", 0x03C3}, /* GREEK SMALL LETTER SIGMA */ + {"sharp", 0x266F}, /* MUSIC SHARP SIGN */ + {"shchcy", 0x0449}, /* CYRILLIC SMALL LETTER SHCHA */ + {"shcy", 0x0448}, /* CYRILLIC SMALL LETTER SHA */ + {"shy", 0x00AD}, /* SOFT HYPHEN */ + {"sigma", 0x03C3}, /* GREEK SMALL LETTER SIGMA */ + {"sigmaf", 0x03C2}, /* GREEK SMALL LETTER FINAL SIGMA */ + {"sigmav", 0x03C2}, /* GREEK SMALL LETTER FINAL SIGMA */ + {"sim", 0x223C}, /* TILDE OPERATOR */ + {"sime", 0x2243}, /* ASYMPTOTICALLY EQUAL TO */ + {"smile", 0x2323}, /* SMILE */ + {"softcy", 0x044C}, /* CYRILLIC SMALL LETTER SOFT SIGN */ + {"sol", 0x002F}, /* SOLIDUS */ + {"spades", 0x2660}, /* BLACK SPADE SUIT */ + {"spar", 0x2225}, /* PARALLEL TO */ + {"sqcap", 0x2293}, /* SQUARE CAP */ + {"sqcup", 0x2294}, /* SQUARE CUP */ + {"sqsub", 0x228F}, /* SQUARE IMAGE OF */ + {"sqsube", 0x2291}, /* SQUARE IMAGE OF OR EQUAL TO */ + {"sqsup", 0x2290}, /* SQUARE ORIGINAL OF */ + {"sqsupe", 0x2292}, /* SQUARE ORIGINAL OF OR EQUAL TO */ + {"squ", 0x25A1}, /* WHITE SQUARE */ + {"square", 0x25A1}, /* WHITE SQUARE */ + {"squf", 0x25AA}, /* BLACK SMALL SQUARE */ + {"ssetmn", 0x2216}, /* SET MINUS */ + {"ssmile", 0x2323}, /* SMILE */ + {"sstarf", 0x22C6}, /* STAR OPERATOR */ + {"star", 0x2606}, /* WHITE STAR */ + {"starf", 0x2605}, /* BLACK STAR */ + {"sub", 0x2282}, /* SUBSET OF */ + {"subE", 0x2286}, /* SUBSET OF OR EQUAL TO */ + {"sube", 0x2286}, /* SUBSET OF OR EQUAL TO */ + {"subnE", 0x228A}, /* SUBSET OF WITH NOT EQUAL TO */ + {"subne", 0x228A}, /* SUBSET OF WITH NOT EQUAL TO */ + {"sum", 0x2211}, /* N-ARY SUMMATION */ + {"sung", 0x266A}, /* EIGHTH NOTE */ + {"sup", 0x2283}, /* SUPERSET OF */ + {"sup1", 0x00B9}, /* SUPERSCRIPT ONE */ + {"sup2", 0x00B2}, /* SUPERSCRIPT TWO */ + {"sup3", 0x00B3}, /* SUPERSCRIPT THREE */ + {"supE", 0x2287}, /* SUPERSET OF OR EQUAL TO */ + {"supe", 0x2287}, /* SUPERSET OF OR EQUAL TO */ + {"supnE", 0x228B}, /* SUPERSET OF WITH NOT EQUAL TO */ + {"supne", 0x228B}, /* SUPERSET OF WITH NOT EQUAL TO */ + {"szlig", 0x00DF}, /* LATIN SMALL LETTER SHARP S */ + {"target", 0x2316}, /* POSITION INDICATOR */ + {"tau", 0x03C4}, /* GREEK SMALL LETTER TAU */ + {"tcaron", 0x0165}, /* LATIN SMALL LETTER T WITH CARON */ + {"tcedil", 0x0163}, /* LATIN SMALL LETTER T WITH CEDILLA */ + {"tcy", 0x0442}, /* CYRILLIC SMALL LETTER TE */ + {"tdot", 0x20DB}, /* COMBINING THREE DOTS ABOVE */ + {"telrec", 0x2315}, /* TELEPHONE RECORDER */ + {"tgr", 0x03C4}, /* GREEK SMALL LETTER TAU */ + {"there4", 0x2234}, /* THEREFORE */ + {"theta", 0x03B8}, /* GREEK SMALL LETTER THETA */ + {"thetas", 0x03B8}, /* GREEK SMALL LETTER THETA */ + {"thetasym", 0x03D1}, /* GREEK THETA SYMBOL */ + {"thetav", 0x03D1}, /* GREEK THETA SYMBOL */ + {"thgr", 0x03B8}, /* GREEK SMALL LETTER THETA */ + {"thinsp", 0x2009}, /* THIN SPACE */ + {"thkap", 0x2248}, /* ALMOST EQUAL TO */ + {"thksim", 0x223C}, /* TILDE OPERATOR */ + {"thorn", 0x00FE}, /* LATIN SMALL LETTER THORN */ + {"tilde", 0x02DC}, /* SMALL TILDE */ + {"times", 0x00D7}, /* MULTIPLICATION SIGN */ + {"timesb", 0x22A0}, /* SQUARED TIMES */ + {"top", 0x22A4}, /* DOWN TACK */ + {"tprime", 0x2034}, /* TRIPLE PRIME */ + {"trade", 0x2122}, /* TRADE MARK SIGN */ + {"trie", 0x225C}, /* DELTA EQUAL TO */ + {"tscy", 0x0446}, /* CYRILLIC SMALL LETTER TSE */ + {"tshcy", 0x045B}, /* CYRILLIC SMALL LETTER TSHE */ + {"tstrok", 0x0167}, /* LATIN SMALL LETTER T WITH STROKE */ + {"twixt", 0x226C}, /* BETWEEN */ + {"uArr", 0x21D1}, /* UPWARDS DOUBLE ARROW */ + {"uacgr", 0x03CD}, /* GREEK SMALL LETTER UPSILON WITH TONOS */ + {"uacute", 0x00FA}, /* LATIN SMALL LETTER U WITH ACUTE */ + {"uarr", 0x2191}, /* UPWARDS ARROW */ + {"uarr2", 0x21C8}, /* UPWARDS PAIRED ARROWS */ + {"ubrcy", 0x045E}, /* CYRILLIC SMALL LETTER SHORT U */ + {"ubreve", 0x016D}, /* LATIN SMALL LETTER U WITH BREVE */ + {"ucirc", 0x00FB}, /* LATIN SMALL LETTER U WITH CIRCUMFLEX */ + {"ucy", 0x0443}, /* CYRILLIC SMALL LETTER U */ + {"udblac", 0x0171}, /* LATIN SMALL LETTER U WITH DOUBLE ACUTE */ + {"udiagr", 0x03B0}, /* GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND */ + {"udigr", 0x03CB}, /* GREEK SMALL LETTER UPSILON WITH DIALYTIKA */ + {"ugr", 0x03C5}, /* GREEK SMALL LETTER UPSILON */ + {"ugrave", 0x00F9}, /* LATIN SMALL LETTER U WITH GRAVE */ + {"uharl", 0x21BF}, /* UPWARDS HARPOON WITH BARB LEFTWARDS */ + {"uharr", 0x21BE}, /* UPWARDS HARPOON WITH BARB RIGHTWARDS */ + {"uhblk", 0x2580}, /* UPPER HALF BLOCK */ + {"ulcorn", 0x231C}, /* TOP LEFT CORNER */ + {"ulcrop", 0x230F}, /* TOP LEFT CROP */ + {"umacr", 0x016B}, /* LATIN SMALL LETTER U WITH MACRON */ + {"uml", 0x00A8}, /* DIAERESIS */ + {"uogon", 0x0173}, /* LATIN SMALL LETTER U WITH OGONEK */ + {"uplus", 0x228E}, /* MULTISET UNION */ + {"upsi", 0x03C5}, /* GREEK SMALL LETTER UPSILON */ + {"upsih", 0x03D2}, /* GREEK UPSILON WITH HOOK SYMBOL */ + {"upsilon", 0x03C5}, /* GREEK SMALL LETTER UPSILON */ + {"urcorn", 0x231D}, /* TOP RIGHT CORNER */ + {"urcrop", 0x230E}, /* TOP RIGHT CROP */ + {"uring", 0x016F}, /* LATIN SMALL LETTER U WITH RING ABOVE */ + {"utilde", 0x0169}, /* LATIN SMALL LETTER U WITH TILDE */ + {"utri", 0x25B5}, /* WHITE UP-POINTING SMALL TRIANGLE */ + {"utrif", 0x25B4}, /* BLACK UP-POINTING SMALL TRIANGLE */ + {"uuml", 0x00FC}, /* LATIN SMALL LETTER U WITH DIAERESIS */ + {"vArr", 0x21D5}, /* UP DOWN DOUBLE ARROW */ + {"vDash", 0x22A8}, /* TRUE */ + {"varr", 0x2195}, /* UP DOWN ARROW */ + {"vcy", 0x0432}, /* CYRILLIC SMALL LETTER VE */ + {"vdash", 0x22A2}, /* RIGHT TACK */ + {"veebar", 0x22BB}, /* XOR */ + {"vellip", 0x22EE}, /* VERTICAL ELLIPSIS */ + {"verbar", 0x007C}, /* VERTICAL LINE */ + {"vltri", 0x22B2}, /* NORMAL SUBGROUP OF */ + {"vprime", 0x2032}, /* PRIME */ + {"vprop", 0x221D}, /* PROPORTIONAL TO */ + {"vrtri", 0x22B3}, /* CONTAINS AS NORMAL SUBGROUP */ + {"vsubnE", 0x228A}, /* SUBSET OF WITH NOT EQUAL TO */ + {"vsubne", 0x228A}, /* SUBSET OF WITH NOT EQUAL TO */ + {"vsupnE", 0x228B}, /* SUPERSET OF WITH NOT EQUAL TO */ + {"vsupne", 0x228B}, /* SUPERSET OF WITH NOT EQUAL TO */ + {"wcirc", 0x0175}, /* LATIN SMALL LETTER W WITH CIRCUMFLEX */ + {"wedgeq", 0x2259}, /* ESTIMATES */ + {"weierp", 0x2118}, /* SCRIPT CAPITAL P */ + {"wreath", 0x2240}, /* WREATH PRODUCT */ + {"xcirc", 0x25CB}, /* WHITE CIRCLE */ + {"xdtri", 0x25BD}, /* WHITE DOWN-POINTING TRIANGLE */ + {"xgr", 0x03BE}, /* GREEK SMALL LETTER XI */ + {"xhArr", 0x2194}, /* LEFT RIGHT ARROW */ + {"xharr", 0x2194}, /* LEFT RIGHT ARROW */ + {"xi", 0x03BE}, /* GREEK SMALL LETTER XI */ + {"xlArr", 0x21D0}, /* LEFTWARDS DOUBLE ARROW */ + {"xrArr", 0x21D2}, /* RIGHTWARDS DOUBLE ARROW */ + {"xutri", 0x25B3}, /* WHITE UP-POINTING TRIANGLE */ + {"yacute", 0x00FD}, /* LATIN SMALL LETTER Y WITH ACUTE */ + {"yacy", 0x044F}, /* CYRILLIC SMALL LETTER YA */ + {"ycirc", 0x0177}, /* LATIN SMALL LETTER Y WITH CIRCUMFLEX */ + {"ycy", 0x044B}, /* CYRILLIC SMALL LETTER YERU */ + {"yen", 0x00A5}, /* YEN SIGN */ + {"yicy", 0x0457}, /* CYRILLIC SMALL LETTER YI */ + {"yucy", 0x044E}, /* CYRILLIC SMALL LETTER YU */ + {"yuml", 0x00FF}, /* LATIN SMALL LETTER Y WITH DIAERESIS */ + {"zacute", 0x017A}, /* LATIN SMALL LETTER Z WITH ACUTE */ + {"zcaron", 0x017E}, /* LATIN SMALL LETTER Z WITH CARON */ + {"zcy", 0x0437}, /* CYRILLIC SMALL LETTER ZE */ + {"zdot", 0x017C}, /* LATIN SMALL LETTER Z WITH DOT ABOVE */ + {"zeta", 0x03B6}, /* GREEK SMALL LETTER ZETA */ + {"zgr", 0x03B6}, /* GREEK SMALL LETTER ZETA */ + {"zhcy", 0x0436}, /* CYRILLIC SMALL LETTER ZHE */ + {"zwj", 0x200D}, /* ZERO WIDTH JOINER */ + {"zwnj", 0x200C}, /* ZERO WIDTH NON-JOINER */ +/* {"epsiv", 0x????}, variant epsilon # ISOgrk3 */ +/* {"fjlig", 0x????}, fj ligature # ISOpub */ +/* {"gEl", 0x????}, greater-than, double equals, less-than # ISOamsr */ +/* {"gap", 0x????}, greater-than, approximately equal to # ISOamsr */ +/* {"gnap", 0x????}, greater-than, not approximately equal t# ISOamsn */ +/* {"jnodot", 0x????}, latin small letter dotless j # ISOamso */ +/* {"lEg", 0x????}, less-than, double equals, greater-than # ISOamsr */ +/* {"lap", 0x????}, less-than, approximately equal to # ISOamsr */ +/* {"lnap", 0x????}, less-than, not approximately equal to # ISOamsn */ +/* {"lpargt", 0x????}, left parenthesis, greater-than # ISOamsc */ +/* {"ngE", 0x????}, not greater-than, double equals # ISOamsn */ +/* {"nlE", 0x????}, not less-than, double equals # ISOamsn */ +/* {"nsmid", 0x????}, nshortmid # ISOamsn */ +/* {"prap", 0x????}, precedes, approximately equal to # ISOamsr */ +/* {"prnE", 0x????}, precedes, not double equal # ISOamsn */ +/* {"prnap", 0x????}, precedes, not approximately equal to # ISOamsn */ +/* {"rpargt", 0x????}, right parenthesis, greater-than # ISOamsc */ +/* {"scap", 0x????}, succeeds, approximately equal to # ISOamsr */ +/* {"scnE", 0x????}, succeeds, not double equals # ISOamsn */ +/* {"scnap", 0x????}, succeeds, not approximately equal to # ISOamsn */ +/* {"smid", 0x????}, shortmid # ISOamsr */ +}; + +#endif /* not ENTITIES_HTML40_ONLY */ +/* *INDENT-ON* */ diff --git a/src/chrtrans/hp_uni.tbl b/src/chrtrans/hp_uni.tbl new file mode 100644 index 0000000..5802551 --- /dev/null +++ b/src/chrtrans/hp_uni.tbl @@ -0,0 +1,212 @@ +# The MIME name of this charset. +Mhp-roman8 + +# Name as a Display Charset (used on Options screen) +OHP Roman8 + +# This is not the default font! +D0 + +# +# Name: HP Roman8 to Unicode +# Date: 1999-01-09 +# Authors: Christian "naddy" Weisgerber <naddy@mips.rhein-neckar.de> +# Mapping by Roman Czyborra, +# <URL:http://czyborra.com/charsets/codepages.html#HP-Roman8> +# + +0x20-0x7E idem # ASCII + +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE + +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xA2 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xA3 U+00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xA4 U+00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xA5 U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xA6 U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xA7 U+00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xA8 U+00B4 # ACUTE ACCENT +0xA9 U+02CB # MODIFIER LETTER GRAVE ACCENT +0xAA U+02C6 # MODIFIER LETTER CIRCUMFLEX ACCENT +0xAB U+00A8 # DIAERESIS +0xAC U+02DC # SMALL TILDE +0xAD U+00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xAE U+00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xAF U+20A4 # LIRA SIGN +0xB0 U+00AF # MACRON +0xB1 U+00DD # LATIN CAPITAL LETTER Y WITH ACUTE +0xB2 U+00FD # LATIN SMALL LETTER Y WITH ACUTE +0xB3 U+00B0 # DEGREE SIGN +0xB4 U+00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xB5 U+00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xB6 U+00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xB7 U+00F1 # LATIN SMALL LETTER N WITH TILDE +0xB8 U+00A1 # INVERTED EXCLAMATION MARK +0xB9 U+00BF # INVERTED QUESTION MARK +0xBA U+00A4 # CURRENCY SIGN +0xBB U+00A3 # POUND SIGN +0xBC U+00A5 # YEN SIGN +0xBD U+00A7 # SECTION SIGN +0xBE U+0192 # LATIN SMALL LETTER F WITH HOOK +0xBF U+00A2 # CENT SIGN +0xC0 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xC1 U+00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xC2 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xC3 U+00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xC4 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +0xC5 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0xC6 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +0xC7 U+00FA # LATIN SMALL LETTER U WITH ACUTE +0xC8 U+00E0 # LATIN SMALL LETTER A WITH GRAVE +0xC9 U+00E8 # LATIN SMALL LETTER E WITH GRAVE +0xCA U+00F2 # LATIN SMALL LETTER O WITH GRAVE +0xCB U+00F9 # LATIN SMALL LETTER U WITH GRAVE +0xCC U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xCD U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xCE U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xCF U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xD0 U+00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xD1 U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xD2 U+00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD3 U+00C6 # LATIN CAPITAL LETTER AE +0xD4 U+00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xD5 U+00ED # LATIN SMALL LETTER I WITH ACUTE +0xD6 U+00F8 # LATIN SMALL LETTER O WITH STROKE +0xD7 U+00E6 # LATIN SMALL LETTER AE +0xD8 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xD9 U+00EC # LATIN SMALL LETTER I WITH GRAVE +0xDA U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xDB U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDC U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xDD U+00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xDE U+00DF # LATIN SMALL LETTER SHARP S +0xDF U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xE0 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xE1 U+00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xE2 U+00E3 # LATIN SMALL LETTER A WITH TILDE +0xE3 U+00D0 # LATIN CAPITAL LETTER ETH +0xE4 U+00F0 # LATIN SMALL LETTER ETH +0xE5 U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xE6 U+00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xE7 U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xE8 U+00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xE9 U+00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xEA U+00F5 # LATIN SMALL LETTER O WITH TILDE +0xEB U+0160 # LATIN CAPITAL LETTER S WITH CARON +0xEC U+0161 # LATIN SMALL LETTER S WITH CARON +0xED U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xEE U+0178 # LATIN CAPITAL LETTER Y WITH DIAERESIS +0xEF U+00FF # LATIN SMALL LETTER Y WITH DIAERESIS +0xF0 U+00DE # LATIN CAPITAL LETTER THORN +0xF1 U+00FE # LATIN SMALL LETTER THORN +0xF2 U+00B7 # MIDDLE DOT +0xF3 U+00B5 # MICRO SIGN +0xF4 U+00B6 # PILCROW SIGN +0xF5 U+00BE # VULGAR FRACTION THREE QUARTERS +0xF6 U+2014 # EM DASH +0xF7 U+00BC # VULGAR FRACTION ONE QUARTER +0xF8 U+00BD # VULGAR FRACTION ONE HALF +0xF9 U+00AA # FEMININE ORDINAL INDICATOR +0xFA U+00BA # MASCULINE ORDINAL INDICATOR +0xFB U+00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xFC U+25A0 # BLACK SQUARE +0xFD U+00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xFE U+00B1 # PLUS-MINUS SIGN + +## EOF ## diff --git a/src/chrtrans/iso01_uni.tbl b/src/chrtrans/iso01_uni.tbl new file mode 100644 index 0000000..95e705e --- /dev/null +++ b/src/chrtrans/iso01_uni.tbl @@ -0,0 +1,334 @@ +# $LynxId: iso01_uni.tbl,v 1.11 2007/07/31 20:35:04 Tim.Larson Exp $ +# vile:tblmode: +# This file has been modified for lynx (see README.tables) + +#Shall this become the "default" translation? +#Meaning of that is currently not well defined. It is different +#from the default input or default output charset... +#but there has to be exactly one table marked as "default". +D0 +# +#The MIME name of this charset. +Miso-8859-1 + +#Name as a Display Charset (used on Options screen) +OWestern (ISO-8859-1) + +#Codepage number +C819 + +# +# Name: ISO/IEC 8859-1:1998 to Unicode +# Unicode version: 3.0 +# Table version: 1.0 +# Table format: Format A +# Date: 1999 July 27 +# Authors: Ken Whistler <kenw@sybase.com> +# +# Copyright (c) 1991-1999 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on optical media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-1:1998 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-1 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-1 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# +# Updated versions of this file may be found in: +# <ftp://ftp.unicode.org/Public/MAPPINGS/> +# +# Any comments or problems, contact <errata@unicode.org> +# Please note that <errata@unicode.org> is an archival address; +# notices will be checked, but do not expect an immediate response. +# +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw +# +0x20-0x7e idem +0xa0-0xff idem # iso 8859-1 special: trivial mapping to Unicode +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0xA0 U+00A0 U+2007 # NO-BREAK SPACE +#0xA1 U+00A1 # INVERTED EXCLAMATION MARK +#0xA2 U+00A2 # CENT SIGN +#0xA3 U+00A3 # POUND SIGN +#0xA4 U+00A4 # CURRENCY SIGN +#0xA5 U+00A5 # YEN SIGN +#0xA6 U+00A6 # BROKEN BAR +#0xA7 U+00A7 # SECTION SIGN +0xA8 U+00A8 U+0308 # DIAERESIS +#0xA9 U+00A9 # COPYRIGHT SIGN +#0xAA U+00AA # FEMININE ORDINAL INDICATOR +#0xAB U+00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +#0xAC U+00AC # NOT SIGN +#0xAD U+00AD # SOFT HYPHEN +#0xAE U+00AE # REGISTERED SIGN +0xAF U+00AF U+0304 # MACRON +0xB0 U+00B0 U+030a # DEGREE SIGN +#0xB1 U+00B1 # PLUS-MINUS SIGN +#0xB2 U+00B2 # SUPERSCRIPT TWO +#0xB3 U+00B3 # SUPERSCRIPT THREE +#0xB4 U+00B4 # ACUTE ACCENT +0xB5 U+00B5 U+03bc # MICRO SIGN +#0xB6 U+00B6 # PILCROW SIGN +0xB7 U+00B7 U+0307 U+0387 U+2027 # MIDDLE DOT +0xB8 U+00B8 U+0327 # CEDILLA +#0xB9 U+00B9 # SUPERSCRIPT ONE +#0xBA U+00BA # MASCULINE ORDINAL INDICATOR +#0xBB U+00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +#0xBC U+00BC # VULGAR FRACTION ONE QUARTER +#0xBD U+00BD # VULGAR FRACTION ONE HALF +#0xBE U+00BE # VULGAR FRACTION THREE QUARTERS +#0xBF U+00BF # INVERTED QUESTION MARK +#0xC0 U+00C0 # LATIN CAPITAL LETTER A WITH GRAVE +#0xC1 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +#0xC2 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +#0xC3 U+00C3 # LATIN CAPITAL LETTER A WITH TILDE +#0xC4 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +#0xC5 U+00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +#0xC6 U+00C6 # LATIN CAPITAL LETTER AE +#0xC7 U+00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +#0xC8 U+00C8 # LATIN CAPITAL LETTER E WITH GRAVE +#0xC9 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +#0xCA U+00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +#0xCB U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +#0xCC U+00CC # LATIN CAPITAL LETTER I WITH GRAVE +#0xCD U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +#0xCE U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +#0xCF U+00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +#0xD0 U+00D0 # LATIN CAPITAL LETTER ETH (Icelandic) +#0xD1 U+00D1 # LATIN CAPITAL LETTER N WITH TILDE +#0xD2 U+00D2 # LATIN CAPITAL LETTER O WITH GRAVE +#0xD3 U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +#0xD4 U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +#0xD5 U+00D5 # LATIN CAPITAL LETTER O WITH TILDE +#0xD6 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +#0xD7 U+00D7 # MULTIPLICATION SIGN +#0xD8 U+00D8 # LATIN CAPITAL LETTER O WITH STROKE +#0xD9 U+00D9 # LATIN CAPITAL LETTER U WITH GRAVE +#0xDA U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +#0xDB U+00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +#0xDC U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +#0xDD U+00DD # LATIN CAPITAL LETTER Y WITH ACUTE +#0xDE U+00DE # LATIN CAPITAL LETTER THORN (Icelandic) +#0xDF U+00DF # LATIN SMALL LETTER SHARP S (German) +#0xE0 U+00E0 # LATIN SMALL LETTER A WITH GRAVE +#0xE1 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +#0xE2 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +#0xE3 U+00E3 # LATIN SMALL LETTER A WITH TILDE +#0xE4 U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +#0xE5 U+00E5 # LATIN SMALL LETTER A WITH RING ABOVE +#0xE6 U+00E6 # LATIN SMALL LETTER AE +#0xE7 U+00E7 # LATIN SMALL LETTER C WITH CEDILLA +#0xE8 U+00E8 # LATIN SMALL LETTER E WITH GRAVE +#0xE9 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +#0xEA U+00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +#0xEB U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +#0xEC U+00EC # LATIN SMALL LETTER I WITH GRAVE +#0xED U+00ED # LATIN SMALL LETTER I WITH ACUTE +#0xEE U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +#0xEF U+00EF # LATIN SMALL LETTER I WITH DIAERESIS +#0xF0 U+00F0 # LATIN SMALL LETTER ETH (Icelandic) +#0xF1 U+00F1 # LATIN SMALL LETTER N WITH TILDE +#0xF2 U+00F2 # LATIN SMALL LETTER O WITH GRAVE +#0xF3 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +#0xF4 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +#0xF5 U+00F5 # LATIN SMALL LETTER O WITH TILDE +#0xF6 U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +#0xF7 U+00F7 # DIVISION SIGN +#0xF8 U+00F8 # LATIN SMALL LETTER O WITH STROKE +#0xF9 U+00F9 # LATIN SMALL LETTER U WITH GRAVE +#0xFA U+00FA # LATIN SMALL LETTER U WITH ACUTE +#0xFB U+00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +#0xFC U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +#0xFD U+00FD # LATIN SMALL LETTER Y WITH ACUTE +#0xFE U+00FE # LATIN SMALL LETTER THORN (Icelandic) +#0xFF U+00FF # LATIN SMALL LETTER Y WITH DIAERESIS + + +0xd0 U+0110 # Dstrok and ETH are nearly the same... + +U+219e "\253-" +U+21a0 "-\273" +U+21ab "<-\260" +U+21ac "\260->" +U+21b4 "\254v" +U+21b8 "\257^\134" +U+21c7 "\253=" +U+21c9 "=\273" +U+21e0 "<\267\267" +U+21e2 "\267\267>" +U+21f1 "|\257^\134" +U+21f6 "=-\273>" +U+2218 " \260 " # RING OPERATOR +U+221b " ROOT\263 " +U+2297 "(\327)" # CIRCLED TIMES +U+2299 "(\267)" # CIRCLED DOT OPERATOR +U+229A "(\260)" # CIRCLED RING OPERATOR +U+22A0 "[\327]" # SQUARED TIMES +U+22A1 "[\267]" # SQUARED DOT OPERATOR +U+22C5 " \267 " # DOT OPERATOR +U+2603 "\2508\250" +U+2609 "(\267)" +U+2614 "\250J\250" +U+262d "\264\134,)" +U+266b "d\257d" +U+2686 "(\267)" +U+2688 "((\267))" +U+2692 "\264X`" +U+2694 ",\327," +U+2697 "\360" +U+2698 "\316" +U+2699 "\244" +U+269b ":\244:" +U+2701 "8\264" +U+270c "mV\270" +U+2721 "\244" +U+273f "\244" +U+2740 "\244" +U+2741 "\244" +U+2761 "\266" +U+279f "\267->" +U+27a0 "\267->" +U+27a7 "\267>" +U+2900 "-|-\273" +U+2901 "-||-\273" +U+2905 "|-\273" +U+290c "<-\267" +U+290d "\267->" +U+290e "<-\267\267" +U+290f "\267\267->" +U+2910 ">\267\267-\273" +U+2911 "\267\267>" +U+2912 "\257^|" +U+2916 ">-\273" +U+2917 ">-|-\273" +U+2918 ">-||-\273" +U+291b "-\253" +U+291c "\273-" +U+2923 "^\134\270" +U+2924 "\270/^" +U+2942 "-><\267" +U+2943 "<-\267>" +U+2944 "\267><-" +U+2954 "\257^|" +U+2958 "\257^|" +U+295d "\257|v" +U+2961 "\257|v" diff --git a/src/chrtrans/iso02_uni.tbl b/src/chrtrans/iso02_uni.tbl new file mode 100644 index 0000000..fe3fc71 --- /dev/null +++ b/src/chrtrans/iso02_uni.tbl @@ -0,0 +1,265 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Miso-8859-2 + +#Name as a Display Charset (used on Options screen) +OEastern European (ISO-8859-2) + +#Codepage number +C912 + +# +# Name: ISO 8859-2:1999 to Unicode +# Unicode version: 3.0 +# Table version: 1.0 +# Table format: Format A +# Date: 1999 July 27 +# Authors: Ken Whistler <kenw@sybase.com> +# +# Copyright (c) 1991-1999 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on optical media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-2:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-2 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-2 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# +# Updated versions of this file may be found in: +# <ftp://ftp.unicode.org/Public/MAPPINGS/> +# +# Any comments or problems, contact <errata@unicode.org> +# Please note that <errata@unicode.org> is an archival address; +# notices will be checked, but do not expect an immediate response. +# +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw +# +0x20-0x7e idem +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+0104 # LATIN CAPITAL LETTER A WITH OGONEK +0xA2 U+02D8 U+0306 # BREVE +0xA3 U+0141 # LATIN CAPITAL LETTER L WITH STROKE +0xA4 U+00A4 # CURRENCY SIGN +0xA5 U+013D # LATIN CAPITAL LETTER L WITH CARON +0xA6 U+015A # LATIN CAPITAL LETTER S WITH ACUTE +0xA7 U+00A7 # SECTION SIGN +0xA8 U+00A8 U+0308 # DIAERESIS +0xA9 U+0160 U+0428 # LATIN CAPITAL LETTER S WITH CARON +0xAA U+015E # LATIN CAPITAL LETTER S WITH CEDILLA +0xAB U+0164 # LATIN CAPITAL LETTER T WITH CARON +0xAC U+0179 # LATIN CAPITAL LETTER Z WITH ACUTE +0xAD U+00AD # SOFT HYPHEN +0xAE U+017D U+0416 # LATIN CAPITAL LETTER Z WITH CARON +0xAF U+017B # LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xB0 U+00B0 U+030a # DEGREE SIGN +0xB1 U+0105 # LATIN SMALL LETTER A WITH OGONEK +0xB2 U+02DB U+0328 # OGONEK +0xB3 U+0142 # LATIN SMALL LETTER L WITH STROKE +0xB4 U+00B4 # ACUTE ACCENT +0xB5 U+013E # LATIN SMALL LETTER L WITH CARON +0xB6 U+015B # LATIN SMALL LETTER S WITH ACUTE +0xB7 U+02C7 U+030c # CARON +0xB8 U+00B8 U+0327 # CEDILLA +0xB9 U+0161 U+0448 # LATIN SMALL LETTER S WITH CARON +0xBA U+015F # LATIN SMALL LETTER S WITH CEDILLA +0xBB U+0165 # LATIN SMALL LETTER T WITH CARON +0xBC U+017A # LATIN SMALL LETTER Z WITH ACUTE +0xBD U+02DD U+030b # DOUBLE ACUTE ACCENT +0xBE U+017E U+0436 # LATIN SMALL LETTER Z WITH CARON +0xBF U+017C # LATIN SMALL LETTER Z WITH DOT ABOVE +0xC0 U+0154 # LATIN CAPITAL LETTER R WITH ACUTE +0xC1 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 U+0102 # LATIN CAPITAL LETTER A WITH BREVE +0xC4 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+0139 # LATIN CAPITAL LETTER L WITH ACUTE +0xC6 U+0106 # LATIN CAPITAL LETTER C WITH ACUTE +0xC7 U+00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 U+010C U+0427 # LATIN CAPITAL LETTER C WITH CARON +0xC9 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+0118 # LATIN CAPITAL LETTER E WITH OGONEK +0xCB U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC U+011A # LATIN CAPITAL LETTER E WITH CARON +0xCD U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF U+010E # LATIN CAPITAL LETTER D WITH CARON +0xD0 U+0110 # LATIN CAPITAL LETTER D WITH STROKE +0xD1 U+0143 # LATIN CAPITAL LETTER N WITH ACUTE +0xD2 U+0147 # LATIN CAPITAL LETTER N WITH CARON +0xD3 U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 U+0150 # LATIN CAPITAL LETTER O WITH DOUBLE ACUTE +0xD6 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+00D7 U+00b7 # MULTIPLICATION SIGN +0xD8 U+0158 # LATIN CAPITAL LETTER R WITH CARON +0xD9 U+016E # LATIN CAPITAL LETTER U WITH RING ABOVE +0xDA U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB U+0170 # LATIN CAPITAL LETTER U WITH DOUBLE ACUTE +0xDC U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+00DD # LATIN CAPITAL LETTER Y WITH ACUTE +0xDE U+0162 # LATIN CAPITAL LETTER T WITH CEDILLA +0xDF U+00DF # LATIN SMALL LETTER SHARP S +0xE0 U+0155 # LATIN SMALL LETTER R WITH ACUTE +0xE1 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 U+0103 # LATIN SMALL LETTER A WITH BREVE +0xE4 U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+013A # LATIN SMALL LETTER L WITH ACUTE +0xE6 U+0107 # LATIN SMALL LETTER C WITH ACUTE +0xE7 U+00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 U+010D U+02a7 U+0447 # LATIN SMALL LETTER C WITH CARON +0xE9 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA U+0119 # LATIN SMALL LETTER E WITH OGONEK +0xEB U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+011B # LATIN SMALL LETTER E WITH CARON +0xED U+00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF U+010F # LATIN SMALL LETTER D WITH CARON +0xF0 U+0111 # LATIN SMALL LETTER D WITH STROKE +0xF1 U+0144 # LATIN SMALL LETTER N WITH ACUTE +0xF2 U+0148 # LATIN SMALL LETTER N WITH CARON +0xF3 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 U+0151 # LATIN SMALL LETTER O WITH DOUBLE ACUTE +0xF6 U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+00F7 # DIVISION SIGN +0xF8 U+0159 # LATIN SMALL LETTER R WITH CARON +0xF9 U+016F # LATIN SMALL LETTER U WITH RING ABOVE +0xFA U+00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB U+0171 # LATIN SMALL LETTER U WITH DOUBLE ACUTE +0xFC U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+00FD # LATIN SMALL LETTER Y WITH ACUTE +0xFE U+0163 # LATIN SMALL LETTER T WITH CEDILLA +0xFF U+02D9 U+0307 U+0387 # DOT ABOVE + + +0xd0 U+00d0 # Dstrok and ETH are nearly the same... + +U+2218 " \260 " # RING OPERATOR +U+2297 "(\327)" # CIRCLED TIMES +U+2299 "(\377)" # CIRCLED DOT OPERATOR +U+229A "(\260)" # CIRCLED RING OPERATOR +U+22A0 "[\327]" # SQUARED TIMES +U+22A1 "[\377]" # SQUARED DOT OPERATOR +U+22C5 " \377 " # DOT OPERATOR diff --git a/src/chrtrans/iso03_uni.tbl b/src/chrtrans/iso03_uni.tbl new file mode 100644 index 0000000..aafce8d --- /dev/null +++ b/src/chrtrans/iso03_uni.tbl @@ -0,0 +1,255 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Miso-8859-3 + +#Name as a Display Charset (used on Options screen) +OLatin 3 (ISO-8859-3) + +#Codepage number +C913 + +# +# Name: ISO/IEC 8859-3:1999 to Unicode +# Unicode version: 3.0 +# Table version: 1.0 +# Table format: Format A +# Date: 1999 July 27 +# Authors: Ken Whistler <kenw@sybase.com> +# +# Copyright (c) 1991-1999 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on optical media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-3:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-3 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-3 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# +# Updated versions of this file may be found in: +# <ftp://ftp.unicode.org/Public/MAPPINGS/> +# +# Any comments or problems, contact <errata@unicode.org> +# Please note that <errata@unicode.org> is an archival address; +# notices will be checked, but do not expect an immediate response. +# +0x20-0x7e idem +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+0126 # LATIN CAPITAL LETTER H WITH STROKE +0xA2 U+02D8 # BREVE +0xA3 U+00A3 # POUND SIGN +0xA4 U+00A4 # CURRENCY SIGN +0xA6 U+0124 # LATIN CAPITAL LETTER H WITH CIRCUMFLEX +0xA7 U+00A7 # SECTION SIGN +0xA8 U+00A8 # DIAERESIS +0xA9 U+0130 # LATIN CAPITAL LETTER I WITH DOT ABOVE +0xAA U+015E # LATIN CAPITAL LETTER S WITH CEDILLA +0xAB U+011E # LATIN CAPITAL LETTER G WITH BREVE +0xAC U+0134 # LATIN CAPITAL LETTER J WITH CIRCUMFLEX +0xAD U+00AD # SOFT HYPHEN +0xAF U+017B # LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xB0 U+00B0 # DEGREE SIGN +0xB1 U+0127 # LATIN SMALL LETTER H WITH STROKE +0xB2 U+00B2 # SUPERSCRIPT TWO +0xB3 U+00B3 # SUPERSCRIPT THREE +0xB4 U+00B4 # ACUTE ACCENT +0xB5 U+00B5 # MICRO SIGN +0xB6 U+0125 # LATIN SMALL LETTER H WITH CIRCUMFLEX +0xB7 U+00B7 # MIDDLE DOT +0xB8 U+00B8 # CEDILLA +0xB9 U+0131 # LATIN SMALL LETTER DOTLESS I +0xBA U+015F # LATIN SMALL LETTER S WITH CEDILLA +0xBB U+011F # LATIN SMALL LETTER G WITH BREVE +0xBC U+0135 # LATIN SMALL LETTER J WITH CIRCUMFLEX +0xBD U+00BD # VULGAR FRACTION ONE HALF +0xBF U+017C # LATIN SMALL LETTER Z WITH DOT ABOVE +0xC0 U+00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC4 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+010A # LATIN CAPITAL LETTER C WITH DOT ABOVE +0xC6 U+0108 # LATIN CAPITAL LETTER C WITH CIRCUMFLEX +0xC7 U+00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 U+00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC U+00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF U+00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD1 U+00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xD2 U+00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 U+0120 # LATIN CAPITAL LETTER G WITH DOT ABOVE +0xD6 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+00D7 # MULTIPLICATION SIGN +0xD8 U+011C # LATIN CAPITAL LETTER G WITH CIRCUMFLEX +0xD9 U+00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB U+00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+016C # LATIN CAPITAL LETTER U WITH BREVE +0xDE U+015C # LATIN CAPITAL LETTER S WITH CIRCUMFLEX +0xDF U+00DF # LATIN SMALL LETTER SHARP S +0xE0 U+00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE4 U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+010B # LATIN SMALL LETTER C WITH DOT ABOVE +0xE6 U+0109 # LATIN SMALL LETTER C WITH CIRCUMFLEX +0xE7 U+00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 U+00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA U+00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+00EC # LATIN SMALL LETTER I WITH GRAVE +0xED U+00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF U+00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF1 U+00F1 # LATIN SMALL LETTER N WITH TILDE +0xF2 U+00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 U+0121 # LATIN SMALL LETTER G WITH DOT ABOVE +0xF6 U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+00F7 # DIVISION SIGN +0xF8 U+011D # LATIN SMALL LETTER G WITH CIRCUMFLEX +0xF9 U+00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA U+00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB U+00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+016D # LATIN SMALL LETTER U WITH BREVE +0xFE U+015D # LATIN SMALL LETTER S WITH CIRCUMFLEX +0xFF U+02D9 # DOT ABOVE + + +# unassigned 8859-3 codepoints: +# 0xa5 unused +# 0xae unused +# 0xbe unused +# 0xc3 unused +# 0xd0 unused +# 0xe3 unused +# 0xf0 unused + diff --git a/src/chrtrans/iso04_uni.tbl b/src/chrtrans/iso04_uni.tbl new file mode 100644 index 0000000..a1c3ed0 --- /dev/null +++ b/src/chrtrans/iso04_uni.tbl @@ -0,0 +1,252 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Miso-8859-4 + +#Name as a Display Charset (used on Options screen) +OLatin 4 (ISO-8859-4) + +#Codepage number +C914 + +# +# Name: ISO/IEC 8859-4:1998 to Unicode +# Unicode version: 3.0 +# Table version: 1.0 +# Table format: Format A +# Date: 1999 July 27 +# Authors: Ken Whistler <kenw@sybase.com> +# +# Copyright (c) 1991-1999 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on optical media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-4:1998 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-4 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-4 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# +# Updated versions of this file may be found in: +# <ftp://ftp.unicode.org/Public/MAPPINGS/> +# +# Any comments or problems, contact <errata@unicode.org> +# Please note that <errata@unicode.org> is an archival address; +# notices will be checked, but do not expect an immediate response. +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+0104 # LATIN CAPITAL LETTER A WITH OGONEK +0xA2 U+0138 # LATIN SMALL LETTER KRA +0xA3 U+0156 # LATIN CAPITAL LETTER R WITH CEDILLA +0xA4 U+00A4 # CURRENCY SIGN +0xA5 U+0128 # LATIN CAPITAL LETTER I WITH TILDE +0xA6 U+013B # LATIN CAPITAL LETTER L WITH CEDILLA +0xA7 U+00A7 # SECTION SIGN +0xA8 U+00A8 # DIAERESIS +0xA9 U+0160 # LATIN CAPITAL LETTER S WITH CARON +0xAA U+0112 # LATIN CAPITAL LETTER E WITH MACRON +0xAB U+0122 # LATIN CAPITAL LETTER G WITH CEDILLA +0xAC U+0166 # LATIN CAPITAL LETTER T WITH STROKE +0xAD U+00AD # SOFT HYPHEN +0xAE U+017D # LATIN CAPITAL LETTER Z WITH CARON +0xAF U+00AF # MACRON +0xB0 U+00B0 # DEGREE SIGN +0xB1 U+0105 # LATIN SMALL LETTER A WITH OGONEK +0xB2 U+02DB # OGONEK +0xB3 U+0157 # LATIN SMALL LETTER R WITH CEDILLA +0xB4 U+00B4 # ACUTE ACCENT +0xB5 U+0129 # LATIN SMALL LETTER I WITH TILDE +0xB6 U+013C # LATIN SMALL LETTER L WITH CEDILLA +0xB7 U+02C7 # CARON +0xB8 U+00B8 # CEDILLA +0xB9 U+0161 # LATIN SMALL LETTER S WITH CARON +0xBA U+0113 # LATIN SMALL LETTER E WITH MACRON +0xBB U+0123 # LATIN SMALL LETTER G WITH CEDILLA +0xBC U+0167 # LATIN SMALL LETTER T WITH STROKE +0xBD U+014A # LATIN CAPITAL LETTER ENG +0xBE U+017E # LATIN SMALL LETTER Z WITH CARON +0xBF U+014B # LATIN SMALL LETTER ENG +0xC0 U+0100 # LATIN CAPITAL LETTER A WITH MACRON +0xC1 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 U+00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 U+00C6 # LATIN CAPITAL LETTER AE +0xC7 U+012E # LATIN CAPITAL LETTER I WITH OGONEK +0xC8 U+010C # LATIN CAPITAL LETTER C WITH CARON +0xC9 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+0118 # LATIN CAPITAL LETTER E WITH OGONEK +0xCB U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC U+0116 # LATIN CAPITAL LETTER E WITH DOT ABOVE +0xCD U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF U+012A # LATIN CAPITAL LETTER I WITH MACRON +0xD0 U+0110 # LATIN CAPITAL LETTER D WITH STROKE +0xD1 U+0145 # LATIN CAPITAL LETTER N WITH CEDILLA +0xD2 U+014C # LATIN CAPITAL LETTER O WITH MACRON +0xD3 U+0136 # LATIN CAPITAL LETTER K WITH CEDILLA +0xD4 U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 U+00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+00D7 # MULTIPLICATION SIGN +0xD8 U+00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 U+0172 # LATIN CAPITAL LETTER U WITH OGONEK +0xDA U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB U+00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+0168 # LATIN CAPITAL LETTER U WITH TILDE +0xDE U+016A # LATIN CAPITAL LETTER U WITH MACRON +0xDF U+00DF # LATIN SMALL LETTER SHARP S +0xE0 U+0101 # LATIN SMALL LETTER A WITH MACRON +0xE1 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 U+00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 U+00E6 # LATIN SMALL LETTER AE +0xE7 U+012F # LATIN SMALL LETTER I WITH OGONEK +0xE8 U+010D # LATIN SMALL LETTER C WITH CARON +0xE9 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA U+0119 # LATIN SMALL LETTER E WITH OGONEK +0xEB U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+0117 # LATIN SMALL LETTER E WITH DOT ABOVE +0xED U+00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF U+012B # LATIN SMALL LETTER I WITH MACRON +0xF0 U+0111 # LATIN SMALL LETTER D WITH STROKE +0xF1 U+0146 # LATIN SMALL LETTER N WITH CEDILLA +0xF2 U+014D # LATIN SMALL LETTER O WITH MACRON +0xF3 U+0137 # LATIN SMALL LETTER K WITH CEDILLA +0xF4 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 U+00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+00F7 # DIVISION SIGN +0xF8 U+00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 U+0173 # LATIN SMALL LETTER U WITH OGONEK +0xFA U+00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB U+00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+0169 # LATIN SMALL LETTER U WITH TILDE +0xFE U+016B # LATIN SMALL LETTER U WITH MACRON +0xFF U+02D9 # DOT ABOVE + + +0xd0 U+00d0 # Dstrok and ETH are nearly the same... diff --git a/src/chrtrans/iso05_uni.tbl b/src/chrtrans/iso05_uni.tbl new file mode 100644 index 0000000..afbb213 --- /dev/null +++ b/src/chrtrans/iso05_uni.tbl @@ -0,0 +1,259 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Miso-8859-5 + +#Name as a Display Charset (used on Options screen) +OCyrillic (ISO-8859-5) + +#Codepage number +C915 + +# +# Name: ISO 8859-5:1999 to Unicode +# Unicode version: 3.0 +# Table version: 1.0 +# Table format: Format A +# Date: 1999 July 27 +# Authors: Ken Whistler <kenw@sybase.com> +# +# Copyright (c) 1991-1999 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on optical media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-5:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-5 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-5 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# +# Updated versions of this file may be found in: +# <ftp://ftp.unicode.org/Public/MAPPINGS/> +# +# Any comments or problems, contact <errata@unicode.org> +# Please note that <errata@unicode.org> is an archival address; +# notices will be checked, but do not expect an immediate response. +# +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw +# +0x20-0x7e idem +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+0401 # CYRILLIC CAPITAL LETTER IO +0xA2 U+0402 # CYRILLIC CAPITAL LETTER DJE +0xA3 U+0403 # CYRILLIC CAPITAL LETTER GJE +0xA4 U+0404 # CYRILLIC CAPITAL LETTER UKRAINIAN IE +0xA5 U+0405 # CYRILLIC CAPITAL LETTER DZE +0xA6 U+0406 U+0130 # CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I +0xA7 U+0407 U+03AA # CYRILLIC CAPITAL LETTER YI +0xA8 U+0408 # CYRILLIC CAPITAL LETTER JE +0xA9 U+0409 # CYRILLIC CAPITAL LETTER LJE +0xAA U+040A # CYRILLIC CAPITAL LETTER NJE +0xAB U+040B # CYRILLIC CAPITAL LETTER TSHE +0xAC U+040C # CYRILLIC CAPITAL LETTER KJE +0xAD U+00AD # SOFT HYPHEN +0xAE U+040E # CYRILLIC CAPITAL LETTER SHORT U +0xAF U+040F # CYRILLIC CAPITAL LETTER DZHE +0xB0 U+0410 # CYRILLIC CAPITAL LETTER A +0xB1 U+0411 # CYRILLIC CAPITAL LETTER BE +0xB2 U+0412 # CYRILLIC CAPITAL LETTER VE +0xB3 U+0413 U+0393 # CYRILLIC CAPITAL LETTER GHE +0xB4 U+0414 # CYRILLIC CAPITAL LETTER DE +0xB5 U+0415 # CYRILLIC CAPITAL LETTER IE +0xB6 U+0416 U+017d # CYRILLIC CAPITAL LETTER ZHE +0xB7 U+0417 # CYRILLIC CAPITAL LETTER ZE +0xB8 U+0418 # CYRILLIC CAPITAL LETTER I +0xB9 U+0419 # CYRILLIC CAPITAL LETTER SHORT I +0xBA U+041A # CYRILLIC CAPITAL LETTER KA +0xBB U+041B U+039b # CYRILLIC CAPITAL LETTER EL +0xBC U+041C # CYRILLIC CAPITAL LETTER EM +0xBD U+041D # CYRILLIC CAPITAL LETTER EN +0xBE U+041E # CYRILLIC CAPITAL LETTER O +0xBF U+041F U+03a0 # CYRILLIC CAPITAL LETTER PE +0xC0 U+0420 # CYRILLIC CAPITAL LETTER ER +0xC1 U+0421 # CYRILLIC CAPITAL LETTER ES +0xC2 U+0422 # CYRILLIC CAPITAL LETTER TE +0xC3 U+0423 # CYRILLIC CAPITAL LETTER U +0xC4 U+0424 U+03a6 # CYRILLIC CAPITAL LETTER EF +0xC5 U+0425 # CYRILLIC CAPITAL LETTER HA +0xC6 U+0426 # CYRILLIC CAPITAL LETTER TSE +0xC7 U+0427 U+010c # CYRILLIC CAPITAL LETTER CHE +0xC8 U+0428 U+0160 # CYRILLIC CAPITAL LETTER SHA +0xC9 U+0429 # CYRILLIC CAPITAL LETTER SHCHA +0xCA U+042A # CYRILLIC CAPITAL LETTER HARD SIGN +0xCB U+042B # CYRILLIC CAPITAL LETTER YERU +0xCC U+042C # CYRILLIC CAPITAL LETTER SOFT SIGN +0xCD U+042D # CYRILLIC CAPITAL LETTER E +0xCE U+042E # CYRILLIC CAPITAL LETTER YU +0xCF U+042F # CYRILLIC CAPITAL LETTER YA +0xD0 U+0430 # CYRILLIC SMALL LETTER A +0xD1 U+0431 # CYRILLIC SMALL LETTER BE +0xD2 U+0432 # CYRILLIC SMALL LETTER VE +0xD3 U+0433 # CYRILLIC SMALL LETTER GHE +0xD4 U+0434 # CYRILLIC SMALL LETTER DE +0xD5 U+0435 # CYRILLIC SMALL LETTER IE +0xD6 U+0436 U+017e # CYRILLIC SMALL LETTER ZHE +0xD7 U+0437 # CYRILLIC SMALL LETTER ZE +0xD8 U+0438 # CYRILLIC SMALL LETTER I +0xD9 U+0439 # CYRILLIC SMALL LETTER SHORT I +0xDA U+043A # CYRILLIC SMALL LETTER KA +0xDB U+043B U+03bb # CYRILLIC SMALL LETTER EL +0xDC U+043C # CYRILLIC SMALL LETTER EM +0xDD U+043D # CYRILLIC SMALL LETTER EN +0xDE U+043E # CYRILLIC SMALL LETTER O +0xDF U+043F U+03c0 # CYRILLIC SMALL LETTER PE +0xE0 U+0440 # CYRILLIC SMALL LETTER ER +0xE1 U+0441 # CYRILLIC SMALL LETTER ES +0xE2 U+0442 # CYRILLIC SMALL LETTER TE +0xE3 U+0443 # CYRILLIC SMALL LETTER U +0xE4 U+0444 U+03c6 # CYRILLIC SMALL LETTER EF +0xE5 U+0445 # CYRILLIC SMALL LETTER HA +0xE6 U+0446 # CYRILLIC SMALL LETTER TSE +0xE7 U+0447 U+010d # CYRILLIC SMALL LETTER CHE +0xE8 U+0448 U+0161 # CYRILLIC SMALL LETTER SHA +0xE9 U+0449 # CYRILLIC SMALL LETTER SHCHA +0xEA U+044A # CYRILLIC SMALL LETTER HARD SIGN +0xEB U+044B U+0131 # CYRILLIC SMALL LETTER YERU +0xEC U+044C # CYRILLIC SMALL LETTER SOFT SIGN +0xED U+044D # CYRILLIC SMALL LETTER E +0xEE U+044E # CYRILLIC SMALL LETTER YU +0xEF U+044F # CYRILLIC SMALL LETTER YA +0xF0 U+2116 # NUMERO SIGN +0xF1 U+0451 # CYRILLIC SMALL LETTER IO +0xF2 U+0452 # CYRILLIC SMALL LETTER DJE +0xF3 U+0453 # CYRILLIC SMALL LETTER GJE +0xF4 U+0454 # CYRILLIC SMALL LETTER UKRAINIAN IE +0xF5 U+0455 # CYRILLIC SMALL LETTER DZE +0xF6 U+0456 # CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I +0xF7 U+0457 U+03CA # CYRILLIC SMALL LETTER YI +0xF8 U+0458 # CYRILLIC SMALL LETTER JE +0xF9 U+0459 # CYRILLIC SMALL LETTER LJE +0xFA U+045A # CYRILLIC SMALL LETTER NJE +0xFB U+045B # CYRILLIC SMALL LETTER TSHE +0xFC U+045C # CYRILLIC SMALL LETTER KJE +0xFD U+00A7 # SECTION SIGN +0xFE U+045E # CYRILLIC SMALL LETTER SHORT U +0xFF U+045F # CYRILLIC SMALL LETTER DZHE + +U+0400 "`\265" +U+040d "`\270" +U+0450 "`\325" +U+045d "`\330" diff --git a/src/chrtrans/iso06_uni.tbl b/src/chrtrans/iso06_uni.tbl new file mode 100644 index 0000000..e4ef995 --- /dev/null +++ b/src/chrtrans/iso06_uni.tbl @@ -0,0 +1,208 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Miso-8859-6 + +#Name as a Display Charset (used on Options screen). +OArabic (ISO-8859-6) + +#Codepage number +C1089 + +# +# Name: ISO 8859-6:1999 to Unicode +# Unicode version: 3.0 +# Table version: 1.0 +# Table format: Format A +# Date: 1999 July 27 +# Authors: Ken Whistler <kenw@sybase.com> +# +# Copyright (c) 1991-1999 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on optical media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-6:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-6 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-6 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# 0x30..0x39 remapped to the ASCII digits (U+0030..U+0039) instead +# of the Arabic digits (U+0660..U+0669). +# +# Updated versions of this file may be found in: +# <ftp://ftp.unicode.org/Public/MAPPINGS/> +# +# Any comments or problems, contact <errata@unicode.org> +# Please note that <errata@unicode.org> is an archival address; +# notices will be checked, but do not expect an immediate response. +# +0x20-0x7e idem +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0xA0 U+00A0 # NO-BREAK SPACE +0xA4 U+00A4 # CURRENCY SIGN +0xAC U+060C # ARABIC COMMA +0xAD U+00AD # SOFT HYPHEN +0xBB U+061B # ARABIC SEMICOLON +0xBF U+061F # ARABIC QUESTION MARK +0xC1 U+0621 # ARABIC LETTER HAMZA +0xC2 U+0622 # ARABIC LETTER ALEF WITH MADDA ABOVE +0xC3 U+0623 # ARABIC LETTER ALEF WITH HAMZA ABOVE +0xC4 U+0624 # ARABIC LETTER WAW WITH HAMZA ABOVE +0xC5 U+0625 # ARABIC LETTER ALEF WITH HAMZA BELOW +0xC6 U+0626 # ARABIC LETTER YEH WITH HAMZA ABOVE +0xC7 U+0627 # ARABIC LETTER ALEF +0xC8 U+0628 # ARABIC LETTER BEH +0xC9 U+0629 # ARABIC LETTER TEH MARBUTA +0xCA U+062A # ARABIC LETTER TEH +0xCB U+062B # ARABIC LETTER THEH +0xCC U+062C # ARABIC LETTER JEEM +0xCD U+062D # ARABIC LETTER HAH +0xCE U+062E # ARABIC LETTER KHAH +0xCF U+062F # ARABIC LETTER DAL +0xD0 U+0630 # ARABIC LETTER THAL +0xD1 U+0631 # ARABIC LETTER REH +0xD2 U+0632 # ARABIC LETTER ZAIN +0xD3 U+0633 # ARABIC LETTER SEEN +0xD4 U+0634 # ARABIC LETTER SHEEN +0xD5 U+0635 # ARABIC LETTER SAD +0xD6 U+0636 # ARABIC LETTER DAD +0xD7 U+0637 # ARABIC LETTER TAH +0xD8 U+0638 # ARABIC LETTER ZAH +0xD9 U+0639 # ARABIC LETTER AIN +0xDA U+063A # ARABIC LETTER GHAIN +0xE0 U+0640 # ARABIC TATWEEL +0xE1 U+0641 # ARABIC LETTER FEH +0xE2 U+0642 # ARABIC LETTER QAF +0xE3 U+0643 # ARABIC LETTER KAF +0xE4 U+0644 # ARABIC LETTER LAM +0xE5 U+0645 # ARABIC LETTER MEEM +0xE6 U+0646 # ARABIC LETTER NOON +0xE7 U+0647 # ARABIC LETTER HEH +0xE8 U+0648 # ARABIC LETTER WAW +0xE9 U+0649 # ARABIC LETTER ALEF MAKSURA +0xEA U+064A # ARABIC LETTER YEH +0xEB U+064B # ARABIC FATHATAN +0xEC U+064C # ARABIC DAMMATAN +0xED U+064D # ARABIC KASRATAN +0xEE U+064E # ARABIC FATHA +0xEF U+064F # ARABIC DAMMA +0xF0 U+0650 # ARABIC KASRA +0xF1 U+0651 # ARABIC SHADDA +0xF2 U+0652 # ARABIC SUKUN diff --git a/src/chrtrans/iso07_uni.tbl b/src/chrtrans/iso07_uni.tbl new file mode 100644 index 0000000..87372a9 --- /dev/null +++ b/src/chrtrans/iso07_uni.tbl @@ -0,0 +1,275 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Miso-8859-7 + +#Name as a Display Charset (used on Options screen) +OGreek (ISO-8859-7) + +#Codepage number +C813 + +# +# Name: ISO 8859-7:2003 to Unicode +# Unicode version: 4.0 +# Table version: 2.0 +# Table format: Format A +# Date: 2003-Nov-12 +# Authors: Ken Whistler <kenw@sybase.com> +# +# Copyright (c) 1991-2003 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on optical media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO 8859-7:2003 characters map into Unicode. +# +# ISO 8859-7:1987 is equivalent to ISO-IR-126, ELOT 928, +# and ECMA 118. ISO 8859-7:2003 adds two currency signs +# and one other character not in the earlier standard. +# +# Format: Three tab-separated columns +# Column #1 is the ISO 8859-7 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO 8859-7 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# Remap 0xA1 to U+2018 (instead of 0x02BD) to match text of 8859-7 +# Remap 0xA2 to U+2019 (instead of 0x02BC) to match text of 8859-7 +# +# 2.0 version updates 1.0 version by adding mappings for the +# three newly added characters 0xA4, 0xA5, 0xAA. +# +# Updated versions of this file may be found in: +# <http://www.unicode.org/Public/MAPPINGS/> +# +# Any comments or problems, contact the Unicode Consortium at: +# <http://www.unicode.org/reporting.html> +# +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw +# +0x20-0x7e idem +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0xA0 U+00A0 # NO-BREAK SPACE +# +# The following two changed in ISO 8859:1999 +# +# Remap 0xA1 to U+2018 (instead of U+02BD) +# Remap 0xA2 to U+2019 (instead of U+02BC) +# +# Keep the old ones as primary for now. Also added old U+037[12] +# found in existing linux kbd files and in RFC 1345 for compatibility. +# - kw 1999-10-29 +0xA1 U+02BD U+2018 U+0371 # MODIFIER LETTER REVERSED COMMA +0xA2 U+02BC U+2019 U+0372 # MODIFIER LETTER APOSTROPHE +0xA3 U+00A3 # POUND SIGN +0xA4 U+20AC # EURO SIGN +0xA5 U+20AF # DRACHMA SIGN +0xA6 U+00A6 # BROKEN BAR +0xA7 U+00A7 # SECTION SIGN +0xA8 U+00A8 # DIAERESIS +0xA9 U+00A9 # COPYRIGHT SIGN +0xAA U+037A # GREEK YPOGEGRAMMENI +0xAB U+00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC # NOT SIGN +0xAD U+00AD # SOFT HYPHEN +0xAF U+2015 # HORIZONTAL BAR +0xB0 U+00B0 # DEGREE SIGN +0xB1 U+00B1 # PLUS-MINUS SIGN +0xB2 U+00B2 # SUPERSCRIPT TWO +0xB3 U+00B3 # SUPERSCRIPT THREE +0xB4 U+0384 # GREEK TONOS +0xB5 U+0385 # GREEK DIALYTIKA TONOS +0xB6 U+0386 U+1fbb # GREEK CAPITAL LETTER ALPHA WITH TONOS +0xB7 U+00B7 U+0307 U+0387 U+2027 # MIDDLE DOT +0xB8 U+0388 U+1fc9 # GREEK CAPITAL LETTER EPSILON WITH TONOS +0xB9 U+0389 U+1fcb # GREEK CAPITAL LETTER ETA WITH TONOS +0xBA U+038A U+1fdb # GREEK CAPITAL LETTER IOTA WITH TONOS +0xBB U+00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+038C U+1ff9 # GREEK CAPITAL LETTER OMICRON WITH TONOS +0xBD U+00BD # VULGAR FRACTION ONE HALF +0xBE U+038E U+1feb # GREEK CAPITAL LETTER UPSILON WITH TONOS +0xBF U+038F U+1ffb # GREEK CAPITAL LETTER OMEGA WITH TONOS +0xC0 U+0390 # GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS +0xC1 U+0391 # GREEK CAPITAL LETTER ALPHA +0xC2 U+0392 # GREEK CAPITAL LETTER BETA +0xC3 U+0393 U+0413 # GREEK CAPITAL LETTER GAMMA +0xC4 U+0394 # GREEK CAPITAL LETTER DELTA +0xC5 U+0395 # GREEK CAPITAL LETTER EPSILON +0xC6 U+0396 # GREEK CAPITAL LETTER ZETA +0xC7 U+0397 # GREEK CAPITAL LETTER ETA +0xC8 U+0398 # GREEK CAPITAL LETTER THETA +0xC9 U+0399 # GREEK CAPITAL LETTER IOTA +0xCA U+039A # GREEK CAPITAL LETTER KAPPA +0xCB U+039B U+041b # GREEK CAPITAL LETTER LAMDA +0xCC U+039C # GREEK CAPITAL LETTER MU +0xCD U+039D # GREEK CAPITAL LETTER NU +0xCE U+039E # GREEK CAPITAL LETTER XI +0xCF U+039F # GREEK CAPITAL LETTER OMICRON +0xD0 U+03A0 U+041f # GREEK CAPITAL LETTER PI +0xD1 U+03A1 # GREEK CAPITAL LETTER RHO +0xD3 U+03A3 # GREEK CAPITAL LETTER SIGMA +0xD4 U+03A4 # GREEK CAPITAL LETTER TAU +0xD5 U+03A5 # GREEK CAPITAL LETTER UPSILON +0xD6 U+03A6 U+0424 # GREEK CAPITAL LETTER PHI +0xD7 U+03A7 U+0425 # GREEK CAPITAL LETTER CHI +0xD8 U+03A8 # GREEK CAPITAL LETTER PSI +0xD9 U+03A9 # GREEK CAPITAL LETTER OMEGA +0xDA U+03AA # GREEK CAPITAL LETTER IOTA WITH DIALYTIKA +0xDB U+03AB # GREEK CAPITAL LETTER UPSILON WITH DIALYTIKA +0xDC U+03AC U+1f71 # GREEK SMALL LETTER ALPHA WITH TONOS +0xDD U+03AD U+1f73 # GREEK SMALL LETTER EPSILON WITH TONOS +0xDE U+03AE U+1f75 # GREEK SMALL LETTER ETA WITH TONOS +0xDF U+03AF U+1f77 # GREEK SMALL LETTER IOTA WITH TONOS +0xE0 U+03B0 # GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS +0xE1 U+03B1 # GREEK SMALL LETTER ALPHA +0xE2 U+03B2 # GREEK SMALL LETTER BETA +0xE3 U+03B3 U+0263 # GREEK SMALL LETTER GAMMA +0xE4 U+03B4 U+00f0 # GREEK SMALL LETTER DELTA +0xE5 U+03B5 # GREEK SMALL LETTER EPSILON +0xE6 U+03B6 # GREEK SMALL LETTER ZETA +0xE7 U+03B7 # GREEK SMALL LETTER ETA +0xE8 U+03B8 # GREEK SMALL LETTER THETA +0xE9 U+03B9 U+0131 # GREEK SMALL LETTER IOTA +0xEA U+03BA # GREEK SMALL LETTER KAPPA +0xEB U+03BB # GREEK SMALL LETTER LAMDA +0xEC U+03BC U+00b5 # GREEK SMALL LETTER MU +0xED U+03BD # GREEK SMALL LETTER NU +0xEE U+03BE # GREEK SMALL LETTER XI +0xEF U+03BF # GREEK SMALL LETTER OMICRON +0xF0 U+03C0 # GREEK SMALL LETTER PI +0xF1 U+03C1 # GREEK SMALL LETTER RHO +0xF2 U+03C2 # GREEK SMALL LETTER FINAL SIGMA +0xF3 U+03C3 # GREEK SMALL LETTER SIGMA +0xF4 U+03C4 # GREEK SMALL LETTER TAU +0xF5 U+03C5 U+028a # GREEK SMALL LETTER UPSILON +0xF6 U+03C6 # GREEK SMALL LETTER PHI +0xF7 U+03C7 # GREEK SMALL LETTER CHI +0xF8 U+03C8 # GREEK SMALL LETTER PSI +0xF9 U+03C9 # GREEK SMALL LETTER OMEGA +0xFA U+03CA # GREEK SMALL LETTER IOTA WITH DIALYTIKA +0xFB U+03CB U+00fc # GREEK SMALL LETTER UPSILON WITH DIALYTIKA +0xFC U+03CC U+1f79 # GREEK SMALL LETTER OMICRON WITH TONOS +0xFD U+03CD U+1f7b # GREEK SMALL LETTER UPSILON WITH TONOS +0xFE U+03CE U+1f7d # GREEK SMALL LETTER OMEGA WITH TONOS + +U+2218 " \260 " # RING OPERATOR +U+2209 " !\345 " +U+221b " ROOT\263 " +U+229A "(\260)" # CIRCLED RING OPERATOR +U+02a4 "d\346" +U+20af "\304\361\367" diff --git a/src/chrtrans/iso08_uni.tbl b/src/chrtrans/iso08_uni.tbl new file mode 100644 index 0000000..d162396 --- /dev/null +++ b/src/chrtrans/iso08_uni.tbl @@ -0,0 +1,229 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Miso-8859-8 + +#Name as a Display Charset (used on Options screen). +OHebrew (ISO-8859-8) + +#Codepage number +C916 + +# +# Name: ISO/IEC 8859-8:1999 to Unicode +# Unicode version: 3.0 +# Table version: 1.1 +# Table format: Format A +# Date: 2000-Jan-03 +# Authors: Ken Whistler <kenw@sybase.com> +# +# Copyright (c) 1991-1999 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on optical media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-8:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-8 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-8 order. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# 1.1 version updates to the published 8859-8:1999, correcting +# the mapping of 0xAF and adding mappings for LRM and RLM. +# +# Updated versions of this file may be found in: +# <ftp://ftp.unicode.org/Public/MAPPINGS/> +# +# Any comments or problems, contact <errata@unicode.org> +# Please note that <errata@unicode.org> is an archival address; +# notices will be checked, but do not expect an immediate response. +# +0x20-0x7e idem +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0xA0 U+00A0 # NO-BREAK SPACE +0xA2 U+00A2 # CENT SIGN +0xA3 U+00A3 # POUND SIGN +0xA4 U+00A4 # CURRENCY SIGN +0xA5 U+00A5 # YEN SIGN +0xA6 U+00A6 # BROKEN BAR +0xA7 U+00A7 # SECTION SIGN +0xA8 U+00A8 # DIAERESIS +0xA9 U+00A9 # COPYRIGHT SIGN +0xAA U+00D7 # MULTIPLICATION SIGN +0xAB U+00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC # NOT SIGN +0xAD U+00AD # SOFT HYPHEN +0xAE U+00AE # REGISTERED SIGN +0xAF U+00AF # MACRON +0xB0 U+00B0 # DEGREE SIGN +0xB1 U+00B1 # PLUS-MINUS SIGN +0xB2 U+00B2 # SUPERSCRIPT TWO +0xB3 U+00B3 # SUPERSCRIPT THREE +0xB4 U+00B4 # ACUTE ACCENT +0xB5 U+00B5 # MICRO SIGN +0xB6 U+00B6 # PILCROW SIGN +0xB7 U+00B7 # MIDDLE DOT +0xB8 U+00B8 # CEDILLA +0xB9 U+00B9 # SUPERSCRIPT ONE +0xBA U+00F7 # DIVISION SIGN +0xBB U+00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+00BC # VULGAR FRACTION ONE QUARTER +0xBD U+00BD # VULGAR FRACTION ONE HALF +0xBE U+00BE # VULGAR FRACTION THREE QUARTERS +0xDF U+2017 # DOUBLE LOW LINE +0xE0 U+05D0 # HEBREW LETTER ALEF +0xE1 U+05D1 # HEBREW LETTER BET +0xE2 U+05D2 # HEBREW LETTER GIMEL +0xE3 U+05D3 # HEBREW LETTER DALET +0xE4 U+05D4 # HEBREW LETTER HE +0xE5 U+05D5 # HEBREW LETTER VAV +0xE6 U+05D6 # HEBREW LETTER ZAYIN +0xE7 U+05D7 # HEBREW LETTER HET +0xE8 U+05D8 # HEBREW LETTER TET +0xE9 U+05D9 # HEBREW LETTER YOD +0xEA U+05DA # HEBREW LETTER FINAL KAF +0xEB U+05DB # HEBREW LETTER KAF +0xEC U+05DC # HEBREW LETTER LAMED +0xED U+05DD # HEBREW LETTER FINAL MEM +0xEE U+05DE # HEBREW LETTER MEM +0xEF U+05DF # HEBREW LETTER FINAL NUN +0xF0 U+05E0 # HEBREW LETTER NUN +0xF1 U+05E1 # HEBREW LETTER SAMEKH +0xF2 U+05E2 # HEBREW LETTER AYIN +0xF3 U+05E3 # HEBREW LETTER FINAL PE +0xF4 U+05E4 # HEBREW LETTER PE +0xF5 U+05E5 # HEBREW LETTER FINAL TSADI +0xF6 U+05E6 # HEBREW LETTER TSADI +0xF7 U+05E7 # HEBREW LETTER QOF +0xF8 U+05E8 # HEBREW LETTER RESH +0xF9 U+05E9 # HEBREW LETTER SHIN +0xFA U+05EA # HEBREW LETTER TAV +0xFD U+200E # LEFT-TO-RIGHT MARK +0xFE U+200F # RIGHT-TO-LEFT MARK + + +#Hebrew points - map to empty string +U+05B0-U+05C2: + +#HEBREW LETTER DOUBLE VAV #U+05F0:åå +U+05F0 "\345\345" +#HEBREW LETTER VAV YOD #U+05F1:éå +U+05F1 "\351\345" +#HEBREW LETTER DOUBLE YOD #U+05F2:éé +U+05F2 "\351\351" + diff --git a/src/chrtrans/iso09_uni.tbl b/src/chrtrans/iso09_uni.tbl new file mode 100644 index 0000000..87afe48 --- /dev/null +++ b/src/chrtrans/iso09_uni.tbl @@ -0,0 +1,266 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Miso-8859-9 + +#Name as a Display Charset (used on Options screen) +OTurkish (ISO-8859-9) + +#Codepage number +C920 + +# +# Name: ISO/IEC 8859-9:1999 to Unicode +# Unicode version: 3.0 +# Table version: 1.0 +# Table format: Format A +# Date: 1999 July 27 +# Authors: Ken Whistler <kenw@sybase.com> +# +# Copyright (c) 1991-1999 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on magnetic media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# ISO/IEC 8859-9:1999 characters map into Unicode. +# +# Format: Three tab-separated columns +# Column #1 is the ISO/IEC 8859-9 code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 the Unicode name (follows a comment sign, '#') +# +# The entries are in ISO/IEC 8859-9 order. +# +# ISO/IEC 8859-9 is also equivalent to ISO-IR-148. +# +# Version history +# 1.0 version updates 0.1 version by adding mappings for all +# control characters. +# +# Updated versions of this file may be found in: +# <ftp://ftp.unicode.org/Public/MAPPINGS/> +# +# Any comments or problems, contact <errata@unicode.org> +# Please note that <errata@unicode.org> is an archival address; +# notices will be checked, but do not expect an immediate response. +# +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw +# +0x20-0x7e idem +0x49 U+042b +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+00A1 # INVERTED EXCLAMATION MARK +0xA2 U+00A2 # CENT SIGN +0xA3 U+00A3 # POUND SIGN +0xA4 U+00A4 # CURRENCY SIGN +0xA5 U+00A5 # YEN SIGN +0xA6 U+00A6 # BROKEN BAR +0xA7 U+00A7 # SECTION SIGN +0xA8 U+00A8 U+0308 # DIAERESIS +0xA9 U+00A9 # COPYRIGHT SIGN +0xAA U+00AA # FEMININE ORDINAL INDICATOR +0xAB U+00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC # NOT SIGN +0xAD U+00AD # SOFT HYPHEN +0xAE U+00AE # REGISTERED SIGN +0xAF U+00AF U+0304 # MACRON +0xB0 U+00B0 U+030a # DEGREE SIGN +0xB1 U+00B1 # PLUS-MINUS SIGN +0xB2 U+00B2 # SUPERSCRIPT TWO +0xB3 U+00B3 # SUPERSCRIPT THREE +0xB4 U+00B4 # ACUTE ACCENT +0xB5 U+00B5 U+03bc # MICRO SIGN +0xB6 U+00B6 # PILCROW SIGN +0xB7 U+00B7 U+0307 U+0387 # MIDDLE DOT +0xB8 U+00B8 U+0327 # CEDILLA +0xB9 U+00B9 # SUPERSCRIPT ONE +0xBA U+00BA # MASCULINE ORDINAL INDICATOR +0xBB U+00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+00BC # VULGAR FRACTION ONE QUARTER +0xBD U+00BD # VULGAR FRACTION ONE HALF +0xBE U+00BE # VULGAR FRACTION THREE QUARTERS +0xBF U+00BF # INVERTED QUESTION MARK +0xC0 U+00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 U+00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 U+00C6 # LATIN CAPITAL LETTER AE +0xC7 U+00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 U+00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC U+00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF U+00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 U+011E # LATIN CAPITAL LETTER G WITH BREVE +0xD1 U+00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xD2 U+00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 U+00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+00D7 # MULTIPLICATION SIGN +0xD8 U+00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 U+00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB U+00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+0130 U+0418 U+0406 # LATIN CAPITAL LETTER I WITH DOT ABOVE +0xDE U+015E U+0428 # LATIN CAPITAL LETTER S WITH CEDILLA +0xDF U+00DF # LATIN SMALL LETTER SHARP S +0xE0 U+00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 U+00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 U+00E6 # LATIN SMALL LETTER AE +0xE7 U+00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 U+00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA U+00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+00EC # LATIN SMALL LETTER I WITH GRAVE +0xED U+00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF U+00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF0 U+011F # LATIN SMALL LETTER G WITH BREVE +0xF1 U+00F1 # LATIN SMALL LETTER N WITH TILDE +0xF2 U+00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 U+00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+00F7 # DIVISION SIGN +0xF8 U+00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 U+00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA U+00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB U+00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+0131 U+03b9 U+044b # LATIN SMALL LETTER DOTLESS I +0xFE U+015F U+0448 # LATIN SMALL LETTER S WITH CEDILLA +0xFF U+00FF # LATIN SMALL LETTER Y WITH DIAERESIS + +U+2218 " \260 " # RING OPERATOR +U+221b " ROOT\263 " +U+2297 "(\327)" # CIRCLED TIMES +U+2299 "(\267)" # CIRCLED DOT OPERATOR +U+229A "(\260)" # CIRCLED RING OPERATOR +U+22A0 "[\327]" # SQUARED TIMES +U+22A1 "[\267]" # SQUARED DOT OPERATOR +U+22C5 " \267 " # DOT OPERATOR diff --git a/src/chrtrans/iso10_uni.tbl b/src/chrtrans/iso10_uni.tbl new file mode 100644 index 0000000..edd59e2 --- /dev/null +++ b/src/chrtrans/iso10_uni.tbl @@ -0,0 +1,153 @@ +# +# Unicode mapping table for the fonts iso10.* +# [use: unicode_start iso10.f16 iso10] +# +#This is not default font! +D0 + +#The MIME name of this charset. + +Miso-8859-10 + +#Name as a Display Charset (used on Options screen) +ONorth European (ISO-8859-10) + +# Name: ISO 8859-10 Latin 6 (1998) to Unicode +# Date: 2005-12-15 +# Authors: Thomas E Dickey from +# http://czyborra.com/charsets/iso8859.html +# (ISO 8859 Alphabet Soup) + +0x20-0x7e idem + +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+0104 # LATIN CAPITAL LETTER A WITH OGONEK +0xA2 U+0112 # LATIN CAPITAL LETTER E WITH MACRON +0xA3 U+0122 # LATIN CAPITAL LETTER G WITH CEDILLA +0xA4 U+012A # LATIN CAPITAL LETTER I WITH MACRON +0xA5 U+0128 # LATIN CAPITAL LETTER I WITH TILDE +0xA6 U+0136 # LATIN CAPITAL LETTER K WITH CEDILLA +0xA7 U+00A7 # SECTION SIGN +0xA8 U+013B # LATIN CAPITAL LETTER L WITH CEDILLA +0xA9 U+0110 # LATIN CAPITAL LETTER D WITH STROKE +0xAA U+0160 # LATIN CAPITAL LETTER S WITH CARON +0xAB U+0166 # LATIN CAPITAL LETTER T WITH STROKE +0xAC U+017D # LATIN CAPITAL LETTER Z WITH CARON +0xAD U+00AD # SOFT HYPHEN +0xAE U+016A # LATIN CAPITAL LETTER U WITH MACRON +0xAF U+014A # LATIN CAPITAL LETTER ENG +0xB0 U+00B0 # DEGREE SIGN +0xB1 U+0105 # LATIN SMALL LETTER A WITH OGONEK +0xB2 U+0113 # LATIN SMALL LETTER E WITH MACRON +0xB3 U+0123 # LATIN SMALL LETTER G WITH CEDILLA +0xB4 U+012B # LATIN SMALL LETTER I WITH MACRON +0xB5 U+0129 # LATIN SMALL LETTER I WITH TILDE +0xB6 U+0137 # LATIN SMALL LETTER K WITH CEDILLA +0xB7 U+00B7 # MIDDLE DOT +0xB8 U+013C # LATIN SMALL LETTER L WITH CEDILLA +0xB9 U+0111 # LATIN SMALL LETTER D WITH STROKE +0xBA U+0161 # LATIN SMALL LETTER S WITH CARON +0xBB U+0167 # LATIN SMALL LETTER T WITH STROKE +0xBC U+017E # LATIN SMALL LETTER Z WITH CARON +0xBD U+2015 # HORIZONTAL BAR +0xBE U+016B # LATIN SMALL LETTER U WITH MACRON +0xBF U+014B # LATIN SMALL LETTER ENG +0xC0 U+0100 # LATIN CAPITAL LETTER A WITH MACRON + +0xC1-0xC6 idem + +#0xC1 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +#0xC2 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +#0xC3 U+00C3 # LATIN CAPITAL LETTER A WITH TILDE +#0xC4 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +#0xC5 U+00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +#0xC6 U+00C6 # LATIN CAPITAL LETTER AE + +0xC7 U+012E # LATIN CAPITAL LETTER I WITH OGONEK +0xC8 U+010C # LATIN CAPITAL LETTER C WITH CARON +0xC9 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+0118 # LATIN CAPITAL LETTER E WITH OGONEK +0xCB U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC U+0116 # LATIN CAPITAL LETTER E WITH DOT ABOVE + +0xCD-0xD0 idem + +#0xCD U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +#0xCE U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +#0xCF U+00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +#0xD0 U+00D0 # LATIN CAPITAL LETTER ETH + +0xD1 U+0145 # LATIN CAPITAL LETTER N WITH CEDILLA +0xD2 U+014C # LATIN CAPITAL LETTER O WITH MACRON + +0xD3-0xD6 idem + +#0xD3 U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +#0xD4 U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +#0xD5 U+00D5 # LATIN CAPITAL LETTER O WITH TILDE +#0xD6 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS + +0xD7 U+0168 # LATIN CAPITAL LETTER U WITH TILDE +0xD8 U+00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 U+0172 # LATIN CAPITAL LETTER U WITH OGONEK + +0xDA-0xDE idem + +#0xDA U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +#0xDB U+00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +#0xDC U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +#0xDD U+00DD # LATIN CAPITAL LETTER Y WITH ACUTE +#0xDE U+00DE # LATIN CAPITAL LETTER THORN + +0xDf U+00DF # LATIN SMALL LETTER SHARP S +0xE0 U+0101 # LATIN SMALL LETTER A WITH MACRON + +0xE1-0xE6 idem + +#0xE1 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +#0xE2 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +#0xE3 U+00E3 # LATIN SMALL LETTER A WITH TILDE +#0xE4 U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +#0xE5 U+00E5 # LATIN SMALL LETTER A WITH RING ABOVE +#0xE6 U+00E6 # LATIN SMALL LETTER AE + +0xE7 U+012F # LATIN SMALL LETTER I WITH OGONEK +0xE8 U+010D # LATIN SMALL LETTER C WITH CARON +0xE9 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA U+0119 # LATIN SMALL LETTER E WITH OGONEK +0xEB U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+0117 # LATIN SMALL LETTER E WITH DOT ABOVE + +0xED-0xF0 idem + +#0xED U+00ED # LATIN SMALL LETTER I WITH ACUTE +#0xEE U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +#0xEF U+00EF # LATIN SMALL LETTER I WITH DIAERESIS +#0xF0 U+00f0 # LATIN SMALL LETTER ETH + +0xF1 U+0146 # LATIN SMALL LETTER N WITH CEDILLA +0xF2 U+014D # LATIN SMALL LETTER O WITH MACRON + +0xF3-0xF6 idem + +#0xF3 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +#0xF4 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +#0xF5 U+00F5 # LATIN SMALL LETTER O WITH TILDE +#0xF6 U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS + +0xF7 U+0169 # LATIN SMALL LETTER U WITH TILDE +0xF8 U+00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 U+0173 # LATIN SMALL LETTER U WITH OGONEK + +0xFA-0xFE idem + +#0xFA U+00FA # LATIN SMALL LETTER U WITH ACUTE +#0xFB U+00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +#0xFC U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +#0xFD U+00FD # LATIN SMALL LETTER Y WITH ACUTE +#0xFE U+00FE # LATIN SMALL LETTER THORN + +0xFF U+0138 # LATIN SMALL LETTER KRA + +# TRADE MARK SIGN: +U+2122:(TM) diff --git a/src/chrtrans/iso13_uni.tbl b/src/chrtrans/iso13_uni.tbl new file mode 100644 index 0000000..33e3b8f --- /dev/null +++ b/src/chrtrans/iso13_uni.tbl @@ -0,0 +1,114 @@ +# The MIME name of this charset. +Miso-8859-13 + +# Name as a Display Charset (used on Options screen) +OBaltic Rim (ISO-8859-13) + +# This is not the default font! +D0 + +# +# Name: ISO 8859-13 Latin 7 (1998) to Unicode +# Date: 2005-12-15 +# Authors: Thomas E Dickey from +# http://czyborra.com/charsets/iso8859.html +# (ISO 8859 Alphabet Soup) + +0x20-0x7E idem # ASCII + +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+201D # RIGHT DOUBLE QUOTATION MARK +0xA2 U+00A2 # CENT SIGN +0xA3 U+00A3 # POUND SIGN +0xA4 U+00A4 # CURRENCY SIGN +0xA5 U+201E # DOUBLE LOW-9 QUOTATION MARK +0xA6 U+00A6 # BROKEN BAR +0xA7 U+00A7 # SECTION SIGN +0xA8 U+00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xA9 U+00A9 # COPYRIGHT SIGN +0xAA U+0156 # LATIN CAPITAL LETTER R WITH CEDILLA +0xAB U+00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC # NOT SIGN +0xAD U+00AD # SOFT HYPHEN +0xAE U+00AE # REGISTERED SIGN +0xAF U+00C6 # LATIN CAPITAL LETTER AE +0xB0 U+00B0 # DEGREE SIGN +0xB1 U+00B1 # PLUS-MINUS SIGN +0xB2 U+00B2 # SUPERSCRIPT TWO +0xB3 U+00B3 # SUPERSCRIPT THREE +0xB4 U+201C # LEFT DOUBLE QUOTATION MARK +0xB5 U+00B5 # MICRO SIGN +0xB6 U+00B6 # PILCROW SIGN +0xB7 U+00B7 # MIDDLE DOT +0xB8 U+00F8 # LATIN SMALL LETTER O WITH STROKE +0xB9 U+00B9 # SUPERSCRIPT ONE +0xBA U+0157 # LATIN SMALL LETTER R WITH CEDILLA +0xBB U+00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+00BC # VULGAR FRACTION ONE QUARTER +0xBD U+00BD # VULGAR FRACTION ONE HALF +0xBE U+00BE # VULGAR FRACTION THREE QUARTERS +0xBF U+00E6 # LATIN SMALL LETTER AE +0xC0 U+0104 # LATIN CAPITAL LETTER A WITH OGONEK +0xC1 U+012E # LATIN CAPITAL LETTER I WITH OGONEK +0xC2 U+0100 # LATIN CAPITAL LETTER A WITH MACRON +0xC3 U+0106 # LATIN CAPITAL LETTER C WITH ACUTE +0xC4 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 U+0118 # LATIN CAPITAL LETTER E WITH OGONEK +0xC7 U+0112 # LATIN CAPITAL LETTER E WITH MACRON +0xC8 U+010C # LATIN CAPITAL LETTER C WITH CARON +0xC9 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+0179 # LATIN CAPITAL LETTER Z WITH ACUTE +0xCB U+0116 # LATIN CAPITAL LETTER E WITH DOT ABOVE +0xCC U+0122 # LATIN CAPITAL LETTER G WITH CEDILLA +0xCD U+0136 # LATIN CAPITAL LETTER K WITH CEDILLA +0xCE U+012A # LATIN CAPITAL LETTER I WITH MACRON +0xCF U+013B # LATIN CAPITAL LETTER L WITH CEDILLA +0xD0 U+0160 # LATIN CAPITAL LETTER S WITH CARON +0xD1 U+0143 # LATIN CAPITAL LETTER N WITH ACUTE +0xD2 U+0145 # LATIN CAPITAL LETTER N WITH CEDILLA +0xD3 U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 U+014C # LATIN CAPITAL LETTER O WITH MACRON +0xD5 U+00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+00D7 # MULTIPLICATION SIGN +0xD8 U+0172 # LATIN CAPITAL LETTER U WITH OGONEK +0xD9 U+0141 # LATIN CAPITAL LETTER L WITH STROKE +0xDA U+015A # LATIN CAPITAL LETTER S WITH ACUTE +0xDB U+016A # LATIN CAPITAL LETTER U WITH MACRON +0xDC U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+017B # LATIN CAPITAL LETTER Z WITH DOT ABOVE +0xDE U+017D # LATIN CAPITAL LETTER Z WITH CARON +0xDF U+00DF # LATIN SMALL LETTER SHARP S (German) +0xE0 U+0105 # LATIN SMALL LETTER A WITH OGONEK +0xE1 U+012F # LATIN SMALL LETTER I WITH OGONEK +0xE2 U+0101 # LATIN SMALL LETTER A WITH MACRON +0xE3 U+0107 # LATIN SMALL LETTER C WITH ACUTE +0xE4 U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 U+0119 # LATIN SMALL LETTER E WITH OGONEK +0xE7 U+0113 # LATIN SMALL LETTER E WITH MACRON +0xE8 U+010D # LATIN SMALL LETTER C WITH CARON +0xE9 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA U+017A # LATIN SMALL LETTER Z WITH ACUTE +0xEB U+0117 # LATIN SMALL LETTER E WITH DOT ABOVE +0xEC U+0123 # LATIN SMALL LETTER G WITH CEDILLA +0xED U+0137 # LATIN SMALL LETTER K WITH CEDILLA +0xEE U+012B # LATIN SMALL LETTER I WITH MACRON +0xEF U+013C # LATIN SMALL LETTER L WITH CEDILLA +0xF0 U+0161 # LATIN SMALL LETTER S WITH CARON +0xF1 U+0144 # LATIN SMALL LETTER N WITH ACUTE +0xF2 U+0146 # LATIN SMALL LETTER N WITH CEDILLA +0xF3 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 U+014D # LATIN SMALL LETTER O WITH MACRON +0xF5 U+00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+00F7 # DIVISION SIGN +0xF8 U+0173 # LATIN SMALL LETTER U WITH OGONEK +0xF9 U+0142 # LATIN SMALL LETTER L WITH STROKE +0xFA U+015B # LATIN SMALL LETTER S WITH ACUTE +0xFB U+016B # LATIN SMALL LETTER U WITH MACRON +0xFC U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+017C # LATIN SMALL LETTER Z WITH DOT ABOVE +0xFE U+017E # LATIN SMALL LETTER Z WITH CARON +0xFF U+2019 # RIGHT SINGLE QUOTATION MARK diff --git a/src/chrtrans/iso14_uni.tbl b/src/chrtrans/iso14_uni.tbl new file mode 100644 index 0000000..630a946 --- /dev/null +++ b/src/chrtrans/iso14_uni.tbl @@ -0,0 +1,114 @@ +# The MIME name of this charset. +Miso-8859-14 + +# Name as a Display Charset (used on Options screen) +OCeltic (ISO-8859-14) + +# This is not the default font! +D0 + +# +# Name: ISO 8859-13 Latin 8 (1998) to Unicode +# Date: 2005-12-15 +# Authors: Thomas E Dickey from +# http://czyborra.com/charsets/iso8859.html +# (ISO 8859 Alphabet Soup) + +0x20-0x7E idem # ASCII + +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+1E02 # LATIN CAPITAL LETTER B WITH DOT ABOVE +0xA2 U+1E03 # LATIN SMALL LETTER B WITH DOT ABOVE +0xA3 U+00A3 # POUND SIGN +0xA4 U+010A # LATIN CAPITAL LETTER C WITH DOT ABOVE +0xA5 U+010B # LATIN SMALL LETTER C WITH DOT ABOVE +0xA6 U+1E0A # LATIN CAPITAL LETTER D WITH DOT ABOVE +0xA7 U+00A7 # SECTION SIGN +0xA8 U+1E80 # LATIN CAPITAL LETTER W WITH GRAVE +0xA9 U+00A9 # COPYRIGHT SIGN +0xAA U+1E82 # LATIN CAPITAL LETTER W WITH ACUTE +0xAB U+1E0B # LATIN SMALL LETTER D WITH DOT ABOVE +0xAC U+1EF2 # LATIN CAPITAL LETTER Y WITH GRAVE +0xAD U+00AD # SOFT HYPHEN +0xAE U+00AE # REGISTERED SIGN +0xAF U+0178 # LATIN CAPITAL LETTER Y WITH DIAERESIS +0xB0 U+1E1E # LATIN CAPITAL LETTER F WITH DOT ABOVE +0xB1 U+1E1F # LATIN SMALL LETTER F WITH DOT ABOVE +0xB2 U+0120 # LATIN CAPITAL LETTER G WITH DOT ABOVE +0xB3 U+0121 # LATIN SMALL LETTER G WITH DOT ABOVE +0xB4 U+1E40 # LATIN CAPITAL LETTER M WITH DOT ABOVE +0xB5 U+1E41 # LATIN SMALL LETTER M WITH DOT ABOVE +0xB6 U+00B6 # PILCROW SIGN +0xB7 U+1E56 # LATIN CAPITAL LETTER P WITH DOT ABOVE +0xB8 U+1E81 # LATIN SMALL LETTER W WITH GRAVE +0xB9 U+1E57 # LATIN SMALL LETTER P WITH DOT ABOVE +0xBA U+1E83 # LATIN SMALL LETTER W WITH ACUTE +0xBB U+1E60 # LATIN CAPITAL LETTER S WITH DOT ABOVE +0xBC U+1EF3 # LATIN SMALL LETTER Y WITH GRAVE +0xBD U+1E84 # LATIN CAPITAL LETTER W WITH DIAERESIS +0xBE U+1E85 # LATIN SMALL LETTER W WITH DIAERESIS +0xBF U+1E61 # LATIN SMALL LETTER S WITH DOT ABOVE +0xC0 U+00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 U+00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 U+00C6 # LATIN CAPITAL LETTER AE +0xC7 U+00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 U+00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC U+00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF U+00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 U+0174 # LATIN CAPITAL LETTER W WITH CIRCUMFLEX +0xD1 U+00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xD2 U+00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 U+00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+1E6A # LATIN CAPITAL LETTER T WITH DOT ABOVE +0xD8 U+00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 U+00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB U+00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+00DD # LATIN CAPITAL LETTER Y WITH ACUTE +0xDE U+0176 # LATIN CAPITAL LETTER Y WITH CIRCUMFLEX +0xDF U+00DF # LATIN SMALL LETTER SHARP S +0xE0 U+00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 U+00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 U+00E6 # LATIN SMALL LETTER AE +0xE7 U+00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 U+00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA U+00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+00EC # LATIN SMALL LETTER I WITH GRAVE +0xED U+00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF U+00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF0 U+0175 # LATIN SMALL LETTER W WITH CIRCUMFLEX +0xF1 U+00F1 # LATIN SMALL LETTER N WITH TILDE +0xF2 U+00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 U+00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+1E6B # LATIN SMALL LETTER T WITH DOT ABOVE +0xF8 U+00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 U+00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA U+00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB U+00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+00FD # LATIN SMALL LETTER Y WITH ACUTE +0xFE U+0177 # LATIN SMALL LETTER Y WITH CIRCUMFLEX +0xFF U+00FF # LATIN SMALL LETTER Y WITH DIAERESIS diff --git a/src/chrtrans/iso15_uni.tbl b/src/chrtrans/iso15_uni.tbl new file mode 100644 index 0000000..398affd --- /dev/null +++ b/src/chrtrans/iso15_uni.tbl @@ -0,0 +1,216 @@ +# The MIME name of this charset. +Miso-8859-15 + +# Name as a Display Charset (used on Options screen) +OWestern (ISO-8859-15) + +# This is not the default font! +D0 + +#Codepage number +#? + +# +# Name: ISO 8859-15 Latin 9 (1998) to Unicode +# Date: 1999-01-01 +# Authors: Christian "naddy" Weisgerber <naddy@mips.rhein-neckar.de> +# +# Remarks: Latin 9 is identical to Latin 1 except for +# code positions A4, A6, A8, B4, B8, BC, BD, BE + +0x20-0x7E idem # ASCII + +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE + +0xA0 U+00A0 # NO-BREAK SPACE +0xA1 U+00A1 # INVERTED EXCLAMATION MARK +0xA2 U+00A2 # CENT SIGN +0xA3 U+00A3 # POUND SIGN +0xA4 U+20AC # EURO SIGN +0xA5 U+00A5 # YEN SIGN +0xA6 U+0160 # LATIN CAPITAL LETTER S WITH CARON +0xA7 U+00A7 # SECTION SIGN +0xA8 U+0161 # LATIN SMALL LETTER S WITH CARON +0xA9 U+00A9 # COPYRIGHT SIGN +0xAA U+00AA # FEMININE ORDINAL INDICATOR +0xAB U+00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xAC U+00AC # NOT SIGN +0xAD U+00AD # SOFT HYPHEN +0xAE U+00AE # REGISTERED SIGN +0xAF U+00AF # MACRON +0xB0 U+00B0 # DEGREE SIGN +0xB1 U+00B1 # PLUS-MINUS SIGN +0xB2 U+00B2 # SUPERSCRIPT TWO +0xB3 U+00B3 # SUPERSCRIPT THREE +0xB4 U+017D # LATIN CAPITAL LETTER Z WITH CARON +0xB5 U+00B5 # MICRO SIGN +0xB6 U+00B6 # PILCROW SIGN +0xB7 U+00B7 # MIDDLE DOT +0xB8 U+017E # LATIN SMALL LETTER Z WITH CARON +0xB9 U+00B9 # SUPERSCRIPT ONE +0xBA U+00BA # MASCULINE ORDINAL INDICATOR +0xBB U+00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xBC U+0152 # LATIN CAPITAL LIGATURE OE +0xBD U+0153 # LATIN SMALL LIGATURE OE +0xBE U+0178 # LATIN CAPITAL LETTER Y WITH DIAERESIS +0xBF U+00BF # INVERTED QUESTION MARK +0xC0 U+00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xC1 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xC2 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xC3 U+00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xC4 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0xC5 U+00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0xC6 U+00C6 # LATIN CAPITAL LETTER AE +0xC7 U+00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0xC8 U+00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xC9 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0xCA U+00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xCB U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xCC U+00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xCD U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xCE U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xCF U+00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xD0 U+00D0 # LATIN CAPITAL LETTER ETH +0xD1 U+00D1 # LATIN CAPITAL LETTER N WITH TILDE +0xD2 U+00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xD3 U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xD4 U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xD5 U+00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xD6 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0xD7 U+00D7 # MULTIPLICATION SIGN +0xD8 U+00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xD9 U+00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xDA U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xDB U+00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xDC U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0xDD U+00DD # LATIN CAPITAL LETTER Y WITH ACUTE +0xDE U+00DE # LATIN CAPITAL LETTER THORN +0xDF U+00DF # LATIN SMALL LETTER SHARP S +0xE0 U+00E0 # LATIN SMALL LETTER A WITH GRAVE +0xE1 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +0xE2 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xE3 U+00E3 # LATIN SMALL LETTER A WITH TILDE +0xE4 U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0xE5 U+00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0xE6 U+00E6 # LATIN SMALL LETTER AE +0xE7 U+00E7 # LATIN SMALL LETTER C WITH CEDILLA +0xE8 U+00E8 # LATIN SMALL LETTER E WITH GRAVE +0xE9 U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0xEA U+00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xEB U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +0xEC U+00EC # LATIN SMALL LETTER I WITH GRAVE +0xED U+00ED # LATIN SMALL LETTER I WITH ACUTE +0xEE U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xEF U+00EF # LATIN SMALL LETTER I WITH DIAERESIS +0xF0 U+00F0 # LATIN SMALL LETTER ETH +0xF1 U+00F1 # LATIN SMALL LETTER N WITH TILDE +0xF2 U+00F2 # LATIN SMALL LETTER O WITH GRAVE +0xF3 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +0xF4 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xF5 U+00F5 # LATIN SMALL LETTER O WITH TILDE +0xF6 U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0xF7 U+00F7 # DIVISION SIGN +0xF8 U+00F8 # LATIN SMALL LETTER O WITH STROKE +0xF9 U+00F9 # LATIN SMALL LETTER U WITH GRAVE +0xFA U+00FA # LATIN SMALL LETTER U WITH ACUTE +0xFB U+00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xFC U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xFD U+00FD # LATIN SMALL LETTER Y WITH ACUTE +0xFE U+00FE # LATIN SMALL LETTER THORN +0xFF U+00FF # LATIN SMALL LETTER Y WITH DIAERESIS + +## EOF ## diff --git a/src/chrtrans/jcuken_kb.h b/src/chrtrans/jcuken_kb.h new file mode 100644 index 0000000..5f42d26 --- /dev/null +++ b/src/chrtrans/jcuken_kb.h @@ -0,0 +1,22 @@ +static LYKbLayout_t kb_layout_jcuken[128] = +{ + 0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007, /* 00..07 */ + 0x0008, 0x0009, 0x000A, 0x000B, 0x000C, 0x000D, 0x000E, 0x000F, /* 08..0F */ + 0x0010, 0x0011, 0x0012, 0x0013, 0x0014, 0x0015, 0x0016, 0x0017, /* 10..17 */ + 0x0018, 0x0019, 0x001A, 0x001B, 0x001C, 0x001D, 0x001E, 0x001F, /* 18..1F */ + + 0x0020, 0x0021, 0x042D, 0x002F, 0x0024, 0x003A, 0x002E, 0x044D, /* 20..27 */ + 0x003F, 0x0025, 0x003B, 0x002B, 0x0431, 0x002D, 0x044E, 0x0451, /* 28..2F */ + 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 0x0037, /* 30..37 */ + 0x0038, 0x0039, 0x0416, 0x0436, 0x0411, 0x003D, 0x042E, 0x0401, /* 38..3F */ + + 0x0022, 0x0424, 0x0418, 0x0421, 0x0412, 0x0423, 0x0410, 0x041F, /* 40..47 */ + 0x0420, 0x0428, 0x041E, 0x041B, 0x0414, 0x042C, 0x0422, 0x0429, /* 48..4F */ + 0x0417, 0x0419, 0x041A, 0x042B, 0x0415, 0x0413, 0x041C, 0x0426, /* 50..57 */ + 0x0427, 0x041D, 0x042F, 0x0445, 0x005C, 0x044A, 0x002C, 0x005F, /* 58..5F */ + + 0x0029, 0x0444, 0x0438, 0x0441, 0x0432, 0x0443, 0x0430, 0x043F, /* 60..67 */ + 0x0440, 0x0448, 0x043E, 0x043B, 0x0434, 0x044C, 0x0442, 0x0449, /* 68..6F */ + 0x0437, 0x0439, 0x043A, 0x044B, 0x0435, 0x0433, 0x043C, 0x0446, /* 70..77 */ + 0x0447, 0x043D, 0x044F, 0x0425, 0x007C, 0x042A, 0x0028, 0x007F /* 78..7F */ +}; diff --git a/src/chrtrans/koi8r_uni.tbl b/src/chrtrans/koi8r_uni.tbl new file mode 100644 index 0000000..8bf4001 --- /dev/null +++ b/src/chrtrans/koi8r_uni.tbl @@ -0,0 +1,147 @@ +# Options screen name for this character set +OCyrillic (KOI8-R) + +# MIME name for this charset +Mkoi8-r + +#Codepage number +C878 + +0x20-0x7f idem +# Based on a table received from "Glenn E. Thobe" <thobe@lafn.org> +# (verified against RFC1489). +# +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw +# +#hex unicode # description +#--- U+---- # --------------- +0x80 U+2500 # FORMS LIGHT HORIZONTAL +0x81 U+2502 # FORMS LIGHT VERTICAL +0x82 U+250C # FORMS LIGHT DOWN AND RIGHT +0x83 U+2510 # FORMS LIGHT DOWN AND LEFT +0x84 U+2514 # FORMS LIGHT UP AND RIGHT +0x85 U+2518 # FORMS LIGHT UP AND LEFT +0x86 U+251C # FORMS LIGHT VERTICAL AND RIGHT +0x87 U+2524 # FORMS LIGHT VERTICAL AND LEFT +0x88 U+252C # FORMS LIGHT DOWN AND HORIZONTAL +0x89 U+2534 # FORMS LIGHT UP AND HORIZONTAL +0x8A U+253C # FORMS LIGHT VERTICAL AND HORIZONTAL +0x8B U+2580 # UPPER HALF BLOCK +0x8C U+2584 # LOWER HALF BLOCK +0x8D U+2588 # FULL BLOCK +0x8E U+258C # LEFT HALF BLOCK +0x8F U+2590 # RIGHT HALF BLOCK +0x90 U+2591 # LIGHT SHADE +0x91 U+2592 # MEDIUM SHADE +0x92 U+2593 # DARK SHADE +0x93 U+2320 # TOP HALF INTEGRAL +0x94 U+25A0 # BLACK SMALL SQUARE +0x95 U+2219 # BULLET OPERATOR +0x96 U+221A # SQUARE ROOT +0x97 U+2248 # ALMOST EQUAL TO +0x98 U+2264 # LESS THAN OR EQUAL TO +0x99 U+2265 # GREATER THAN OR EQUAL TO +0x9A U+00A0 # NON-BREAKING SPACE +0x9B U+2321 # BOTTOM HALF INTEGRAL +0x9C U+00B0 # DEGREE SIGN +0x9D U+00B2 # SUPERSCRIPT DIGIT TWO +0x9E U+00B7 U+2027 # MIDDLE DOT +0x9F U+00F7 # DIVISION SIGN +0xA0 U+2550 # FORMS DOUBLE HORIZONTAL +0xA1 U+2551 # FORMS DOUBLE VERTICAL +0xA2 U+2552 # FORMS DOWN SINGLE AND RIGHT DOUBLE +0xA3 U+0451 # SMA IO +0xA4 U+2553 # FORMS DOWN DOUBLE AND RIGHT SINGLE +0xA5 U+2554 # FORMS DOUBLE DOWN AND RIGHT +0xA6 U+2555 # FORMS DOWN SINGLE AND LEFT DOUBLE +0xA7 U+2556 # FORMS DOWN DOUBLE AND LEFT SINGLE +0xA8 U+2557 # FORMS DOUBLE DOWN AND LEFT +0xA9 U+2558 # FORMS UP SINGLE AND RIGHT DOUBLE +0xAA U+2559 # FORMS UP DOUBLE AND RIGHT SINGLE +0xAB U+255A # FORMS DOUBLE UP AND RIGHT +0xAC U+255B # FORMS UP SINGLE AND LEFT DOUBLE +0xAD U+255C # FORMS UP DOUBLE AND LEFT SINGLE +0xAE U+255D # FORMS DOUBLE UP AND LEFT +0xAF U+255E # FORMS VERTICAL SINGLE AND RIGHT DOUBLE +0xB0 U+255F # FORMS VERTICAL DOUBLE AND RIGHT SINGLE +0xB1 U+2560 # FORMS DOUBLE VERTICAL AND RIGHT +0xB2 U+2561 # FORMS VERTICAL SINGLE AND LEFT DOUBLE +0xB3 U+0401 # CAP IO +0xB4 U+2562 # FORMS VERTICAL DOUBLE AND LEFT SINGLE +0xB5 U+2563 # FORMS DOUBLE VERTICAL AND LEFT +0xB6 U+2564 # FORMS DOWN SINGLE AND HORIZONTAL DOUBLE +0xB7 U+2565 # FORMS DOWN DOUBLE AND HORIZONTAL SINGLE +0xB8 U+2566 # FORMS DOUBLE DOWN AND HORIZONTAL +0xB9 U+2567 # FORMS UP SINGLE AND HORIZONTAL DOUBLE +0xBA U+2568 # FORMS UP DOUBLE AND HORIZONTAL SINGLE +0xBB U+2569 # FORMS DOUBLE UP AND HORIZONTAL +0xBC U+256A # FORMS VERTICAL SINGLE AND HORIZONTAL DOUBLE +0xBD U+256B # FORMS VERTICAL DOUBLE AND HORIZONTAL SINGLE +0xBE U+256C # FORMS DOUBLE VERTICAL AND HORIZONTAL +0xBF U+00A9 # COPYRIGHT SIGN +0xC0 U+044E # SMA IU +0xC1 U+0430 # SMA A +0xC2 U+0431 # SMA BE +0xC3 U+0446 # SMA TSE +0xC4 U+0434 # SMA DE +0xC5 U+0435 # SMA IE +0xC6 U+0444 U+03c6 # SMA EF +0xC7 U+0433 # SMA GE +0xC8 U+0445 # SMA KHA +0xC9 U+0438 # SMA II +0xCA U+0439 # SMA SHORT II +0xCB U+043A # SMA KA +0xCC U+043B U+03bb # SMA EL +0xCD U+043C # SMA EM +0xCE U+043D # SMA EN +0xCF U+043E # SMA O +0xD0 U+043F U+03c0 # SMA PE +0xD1 U+044F # SMA IA +0xD2 U+0440 # SMA ER +0xD3 U+0441 # SMA ES +0xD4 U+0442 # SMA TE +0xD5 U+0443 # SMA U +0xD6 U+0436 U+017e # SMA ZHE +0xD7 U+0432 # SMA VE +0xD8 U+044C # SMA SOFT SIGN +0xD9 U+044B U+0131 # SMA YERI +0xDA U+0437 # SMA ZE +0xDB U+0448 U+0161 # SMA SHA +0xDC U+044D # SMA REVERSED E +0xDD U+0449 # SMA SHCHA +0xDE U+0447 U+010d # SMA CHE +0xDF U+044A # SMA HARD SIGN +0xE0 U+042E # CAP IU +0xE1 U+0410 # CAP A +0xE2 U+0411 # CAP BE +0xE3 U+0426 # CAP TSE +0xE4 U+0414 # CAP DE +0xE5 U+0415 # CAP IE +0xE6 U+0424 U+03a6 # CAP EF +0xE7 U+0413 U+0393 # CAP GE +0xE8 U+0425 # CAP KHA +0xE9 U+0418 # CAP II +0xEA U+0419 # CAP SHORT II +0xEB U+041A # CAP KA +0xEC U+041B U+039b # CAP EL +0xED U+041C # CAP EM +0xEE U+041D # CAP EN +0xEF U+041E # CAP O +0xF0 U+041F U+03a0 # CAP PE +0xF1 U+042F # CAP IA +0xF2 U+0420 # CAP ER +0xF3 U+0421 # CAP ES +0xF4 U+0422 # CAP TE +0xF5 U+0423 # CAP U +0xF6 U+0416 U+017d # CAP ZHE +0xF7 U+0412 # CAP VE +0xF8 U+042C # CAP SOFT SIGN +0xF9 U+042B # CAP YERI +0xFA U+0417 # CAP ZE +0xFB U+0428 U+0160 # CAP SHA +0xFC U+042D # CAP REVERSED E +0xFD U+0429 # CAP SHCHA +0xFE U+0427 U+010c # CAP CHE +0xFF U+042A # CAP HARD SIGN + diff --git a/src/chrtrans/koi8u_uni.tbl b/src/chrtrans/koi8u_uni.tbl new file mode 100644 index 0000000..2c13845 --- /dev/null +++ b/src/chrtrans/koi8u_uni.tbl @@ -0,0 +1,154 @@ +# Options screen name for this character set +OUkrainian Cyrillic (KOI8-U) + +# MIME name for this charset +Mkoi8-u + +#Codepage number +#? + +0x20-0x7f idem +# Based on a table received from "Denis V. Dmitrienko" <denis@null.net> +# (verified against RFC2319). +# KOI8-U home page: <http://www.net.ua/KOI8-U> +# +# Quoted from RFC2319: +# The upper part of the KOI8-U Character Set contains all Russian +# letters defined in KOI8-R and four Ukrainian letters (#164, #180 - +# ukr. ie, #166, #182 - ukr. i, #167, #183 - ukr. yi, #173, #189 - ukr. +# ghe with upturn) which locations are compliant with ISO-IR-111. +# +# BOX DRAWINGS elements in the other positions (that are not used by +# Ukrainian letters) are the same as in KOI8-R character set. +# +# +#hex unicode # description +#--- U+---- # --------------- +0x80 U+2500 # BOX DRAWINGS LIGHT HORIZONTAL +0x81 U+2502 # BOX DRAWINGS LIGHT VERTICAL +0x82 U+250C # BOX DRAWINGS LIGHT DOWN AND RIGHT +0x83 U+2510 # BOX DRAWINGS LIGHT DOWN AND LEFT +0x84 U+2514 # BOX DRAWINGS LIGHT UP AND RIGHT +0x85 U+2518 # BOX DRAWINGS LIGHT UP AND LEFT +0x86 U+251C # BOX DRAWINGS LIGHT VERTICAL AND RIGHT +0x87 U+2524 # BOX DRAWINGS LIGHT VERTICAL AND LEFT +0x88 U+252C # BOX DRAWINGS LIGHT DOWN AND HORIZONTAL +0x89 U+2534 # BOX DRAWINGS LIGHT UP AND HORIZONTAL +0x8A U+253C # BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL +0x8B U+2580 # UPPER HALF BLOCK +0x8C U+2584 # LOWER HALF BLOCK +0x8D U+2588 # FULL BLOCK +0x8E U+258C # LEFT HALF BLOCK +0x8F U+2590 # RIGHT HALF BLOCK +0x90 U+2591 # LIGHT SHADE +0x91 U+2592 # MEDIUM SHADE +0x92 U+2593 # DARK SHADE +0x93 U+2320 # TOP HALF INTEGRAL +0x94 U+25A0 # BLACK SQUARE +0x95 U+2219 # BULLET OPERATOR +0x96 U+221A # SQUARE ROOT +0x97 U+2248 # ALMOST EQUAL TO +0x98 U+2264 # LESS THAN OR EQUAL TO +0x99 U+2265 # GREATER THAN OR EQUAL TO +0x9A U+00A0 # NO-BREAK SPACE +0x9B U+2321 # BOTTOM HALF INTEGRAL +0x9C U+00B0 # DEGREE SIGN +0x9D U+00B2 # SUPERSCRIPT TWO +0x9E U+00B7 # MIDDLE DOT +0x9F U+00F7 # DIVISION SIGN +0xA0 U+2550 # BOX DRAWINGS DOUBLE HORIZONTAL +0xA1 U+2551 # BOX DRAWINGS DOUBLE VERTICAL +0xA2 U+2552 # BOX DRAWINGS DOWN SINGLE AND RIGHT DOUBLE +0xA3 U+0451 # CYRILLIC SMALL LETTER IO +0xA4 U+0454 # CYRILLIC SMALL LETTER UKRAINIAN IE +0xA5 U+2554 # BOX DRAWINGS DOUBLE DOWN AND RIGHT +0xA6 U+0456 # CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I +0xA7 U+0457 # CYRILLIC SMALL LETTER YI (UKRAINIAN) +0xA8 U+2557 # BOX DRAWINGS DOUBLE DOWN AND LEFT +0xA9 U+2558 # BOX DRAWINGS UP SINGLE AND RIGHT DOUBLE +0xAA U+2559 # BOX DRAWINGS UP DOUBLE AND RIGHT SINGLE +0xAB U+255A # BOX DRAWINGS DOUBLE UP AND RIGHT +0xAC U+255B # BOX DRAWINGS UP SINGLE AND LEFT DOUBLE +0xAD U+0491 # CYRILLIC SMALL LETTER GHE WITH UPTURN +0xAE U+255D # BOX DRAWINGS DOUBLE UP AND LEFT +0xAF U+255E # BOX DRAWINGS VERTICAL SINGLE AND RIGHT DOUBLE +0xB0 U+255F # BOX DRAWINGS VERTICAL DOUBLE AND RIGHT SINGLE +0xB1 U+2560 # BOX DRAWINGS DOUBLE VERTICAL AND RIGHT +0xB2 U+2561 # BOX DRAWINGS VERTICAL SINGLE AND LEFT DOUBLE +0xB3 U+0401 # CYRILLIC CAPITAL LETTER IO +0xB4 U+0404 # CYRILLIC CAPITAL LETTER UKRAINIAN IE +0xB5 U+2563 # BOX DRAWINGS DOUBLE VERTICAL AND LEFT +0xB6 U+0406 # CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I +0xB7 U+0407 # CYRILLIC CAPITAL LETTER YI (UKRAINIAN) +0xB8 U+2566 # BOX DRAWINGS DOUBLE DOWN AND HORIZONTAL +0xB9 U+2567 # BOX DRAWINGS UP SINGLE AND HORIZONTAL DOUBLE +0xBA U+2568 # BOX DRAWINGS UP DOUBLE AND HORIZONTAL SINGLE +0xBB U+2569 # BOX DRAWINGS DOUBLE UP AND HORIZONTAL +0xBC U+256A # BOX DRAWINGS VERTICAL SINGLE AND HORIZONTAL DOUBLE +0xBD U+0490 # CYRILLIC CAPITAL LETTER GHE WITH UPTURN +0xBE U+256C # BOX DRAWINGS DOUBLE VERTICAL AND HORIZONTAL +0xBF U+00A9 # COPYRIGHT SIGN +0xC0 U+044E # CYRILLIC SMALL LETTER YU +0xC1 U+0430 # CYRILLIC SMALL LETTER A +0xC2 U+0431 # CYRILLIC SMALL LETTER BE +0xC3 U+0446 # CYRILLIC SMALL LETTER TSE +0xC4 U+0434 # CYRILLIC SMALL LETTER DE +0xC5 U+0435 # CYRILLIC SMALL LETTER IE +0xC6 U+0444 # CYRILLIC SMALL LETTER EF +0xC7 U+0433 # CYRILLIC SMALL LETTER GHE +0xC8 U+0445 # CYRILLIC SMALL LETTER KHA +0xC9 U+0438 # CYRILLIC SMALL LETTER I +0xCA U+0439 # CYRILLIC SMALL LETTER SHORT I +0xCB U+043A # CYRILLIC SMALL LETTER KA +0xCC U+043B # CYRILLIC SMALL LETTER EL +0xCD U+043C # CYRILLIC SMALL LETTER EM +0xCE U+043D # CYRILLIC SMALL LETTER EN +0xCF U+043E # CYRILLIC SMALL LETTER O +0xD0 U+043F # CYRILLIC SMALL LETTER PE +0xD1 U+044F # CYRILLIC SMALL LETTER YA +0xD2 U+0440 # CYRILLIC SMALL LETTER ER +0xD3 U+0441 # CYRILLIC SMALL LETTER ES +0xD4 U+0442 # CYRILLIC SMALL LETTER TE +0xD5 U+0443 # CYRILLIC SMALL LETTER U +0xD6 U+0436 # CYRILLIC SMALL LETTER ZHE +0xD7 U+0432 # CYRILLIC SMALL LETTER VE +0xD8 U+044C # CYRILLIC SMALL LETTER SOFT SIGN +0xD9 U+044B # CYRILLIC SMALL LETTER YERU +0xDA U+0437 # CYRILLIC SMALL LETTER ZE +0xDB U+0448 # CYRILLIC SMALL LETTER SHA +0xDC U+044D # CYRILLIC SMALL LETTER E +0xDD U+0449 # CYRILLIC SMALL LETTER SHCHA +0xDE U+0447 # CYRILLIC SMALL LETTER CHE +0xDF U+044A # CYRILLIC SMALL LETTER HARD SIGN +0xE0 U+042E # CYRILLIC CAPITAL LETTER YU +0xE1 U+0410 # CYRILLIC CAPITAL LETTER A +0xE2 U+0411 # CYRILLIC CAPITAL LETTER BE +0xE3 U+0426 # CYRILLIC CAPITAL LETTER TSE +0xE4 U+0414 # CYRILLIC CAPITAL LETTER DE +0xE5 U+0415 # CYRILLIC CAPITAL LETTER IE +0xE6 U+0424 # CYRILLIC CAPITAL LETTER EF +0xE7 U+0413 # CYRILLIC CAPITAL LETTER GHE +0xE8 U+0425 # CYRILLIC CAPITAL LETTER KHA +0xE9 U+0418 # CYRILLIC CAPITAL LETTER I +0xEA U+0419 # CYRILLIC CAPITAL LETTER SHORT I +0xEB U+041A # CYRILLIC CAPITAL LETTER KA +0xEC U+041B # CYRILLIC CAPITAL LETTER EL +0xED U+041C # CYRILLIC CAPITAL LETTER EM +0xEE U+041D # CYRILLIC CAPITAL LETTER EN +0xEF U+041E # CYRILLIC CAPITAL LETTER O +0xF0 U+041F # CYRILLIC CAPITAL LETTER PE +0xF1 U+042F # CYRILLIC CAPITAL LETTER YA +0xF2 U+0420 # CYRILLIC CAPITAL LETTER ER +0xF3 U+0421 # CYRILLIC CAPITAL LETTER ES +0xF4 U+0422 # CYRILLIC CAPITAL LETTER TE +0xF5 U+0423 # CYRILLIC CAPITAL LETTER U +0xF6 U+0416 # CYRILLIC CAPITAL LETTER ZHE +0xF7 U+0412 # CYRILLIC CAPITAL LETTER VE +0xF8 U+042C # CYRILLIC CAPITAL LETTER SOFT SIGN +0xF9 U+042B # CYRILLIC CAPITAL LETTER YERU +0xFA U+0417 # CYRILLIC CAPITAL LETTER ZE +0xFB U+0428 # CYRILLIC CAPITAL LETTER SHA +0xFC U+042D # CYRILLIC CAPITAL LETTER E +0xFD U+0429 # CYRILLIC CAPITAL LETTER SHCHA +0xFE U+0427 # CYRILLIC CAPITAL LETTER CHE +0xFF U+042A # CYRILLIC CAPITAL LETTER HARD SIGN diff --git a/src/chrtrans/mac_uni.tbl b/src/chrtrans/mac_uni.tbl new file mode 100644 index 0000000..2564701 --- /dev/null +++ b/src/chrtrans/mac_uni.tbl @@ -0,0 +1,284 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mmacintosh + +#Name as a Display Charset (used on Options screen) +OMacintosh (8 bit) + +# +# Name: cp10000_MacRoman to Unicode table +# Unicode version: 2.0 +# Table version: 2.00 +# Table format: Format A +# Date: 04/24/96 +# Authors: Lori Brownell <loribr@microsoft.com> +# K.D. Chang <a-kchang@microsoft.com> +# General notes: none +# +# Format: Three tab-separated columns +# Column #1 is the cp10000_MacRoman code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in cp10000_MacRoman order +# +# Lines with more than one Unicode (U+XXXX) value contain additional +# replacement mappings added for lynx. - kw +# +0x20-0x7f idem +# +#0x20 U+0020 # SPACE +#0x21 U+0021 # EXCLAMATION MARK +#0x22 U+0022 # QUOTATION MARK +#0x23 U+0023 # NUMBER SIGN +#0x24 U+0024 # DOLLAR SIGN +#0x25 U+0025 # PERCENT SIGN +#0x26 U+0026 # AMPERSAND +#0x27 U+0027 # APOSTROPHE +#0x28 U+0028 # LEFT PARENTHESIS +#0x29 U+0029 # RIGHT PARENTHESIS +#0x2A U+002A # ASTERISK +#0x2B U+002B # PLUS SIGN +#0x2C U+002C # COMMA +#0x2D U+002D # HYPHEN-MINUS +#0x2E U+002E # FULL STOP +#0x2F U+002F # SOLIDUS +#0x30 U+0030 # DIGIT ZERO +#0x31 U+0031 # DIGIT ONE +#0x32 U+0032 # DIGIT TWO +#0x33 U+0033 # DIGIT THREE +#0x34 U+0034 # DIGIT FOUR +#0x35 U+0035 # DIGIT FIVE +#0x36 U+0036 # DIGIT SIX +#0x37 U+0037 # DIGIT SEVEN +#0x38 U+0038 # DIGIT EIGHT +#0x39 U+0039 # DIGIT NINE +#0x3A U+003A # COLON +#0x3B U+003B # SEMICOLON +#0x3C U+003C # LESS-THAN SIGN +#0x3D U+003D # EQUALS SIGN +#0x3E U+003E # GREATER-THAN SIGN +#0x3F U+003F # QUESTION MARK +#0x40 U+0040 # COMMERCIAL AT +#0x41 U+0041 # LATIN CAPITAL LETTER A +#0x42 U+0042 # LATIN CAPITAL LETTER B +#0x43 U+0043 # LATIN CAPITAL LETTER C +#0x44 U+0044 # LATIN CAPITAL LETTER D +#0x45 U+0045 # LATIN CAPITAL LETTER E +#0x46 U+0046 # LATIN CAPITAL LETTER F +#0x47 U+0047 # LATIN CAPITAL LETTER G +#0x48 U+0048 # LATIN CAPITAL LETTER H +#0x49 U+0049 # LATIN CAPITAL LETTER I +#0x4A U+004A # LATIN CAPITAL LETTER J +#0x4B U+004B # LATIN CAPITAL LETTER K +#0x4C U+004C # LATIN CAPITAL LETTER L +#0x4D U+004D # LATIN CAPITAL LETTER M +#0x4E U+004E # LATIN CAPITAL LETTER N +#0x4F U+004F # LATIN CAPITAL LETTER O +#0x50 U+0050 # LATIN CAPITAL LETTER P +#0x51 U+0051 # LATIN CAPITAL LETTER Q +#0x52 U+0052 # LATIN CAPITAL LETTER R +#0x53 U+0053 # LATIN CAPITAL LETTER S +#0x54 U+0054 # LATIN CAPITAL LETTER T +#0x55 U+0055 # LATIN CAPITAL LETTER U +#0x56 U+0056 # LATIN CAPITAL LETTER V +#0x57 U+0057 # LATIN CAPITAL LETTER W +#0x58 U+0058 # LATIN CAPITAL LETTER X +#0x59 U+0059 # LATIN CAPITAL LETTER Y +#0x5A U+005A # LATIN CAPITAL LETTER Z +#0x5B U+005B # LEFT SQUARE BRACKET +#0x5C U+005C # REVERSE SOLIDUS +#0x5D U+005D # RIGHT SQUARE BRACKET +#0x5E U+005E # CIRCUMFLEX ACCENT +#0x5F U+005F # LOW LINE +#0x60 U+0060 # GRAVE ACCENT +#0x61 U+0061 # LATIN SMALL LETTER A +#0x62 U+0062 # LATIN SMALL LETTER B +#0x63 U+0063 # LATIN SMALL LETTER C +#0x64 U+0064 # LATIN SMALL LETTER D +#0x65 U+0065 # LATIN SMALL LETTER E +#0x66 U+0066 # LATIN SMALL LETTER F +#0x67 U+0067 # LATIN SMALL LETTER G +#0x68 U+0068 # LATIN SMALL LETTER H +#0x69 U+0069 # LATIN SMALL LETTER I +#0x6A U+006A # LATIN SMALL LETTER J +#0x6B U+006B # LATIN SMALL LETTER K +#0x6C U+006C # LATIN SMALL LETTER L +#0x6D U+006D # LATIN SMALL LETTER M +#0x6E U+006E # LATIN SMALL LETTER N +#0x6F U+006F # LATIN SMALL LETTER O +#0x70 U+0070 # LATIN SMALL LETTER P +#0x71 U+0071 # LATIN SMALL LETTER Q +#0x72 U+0072 # LATIN SMALL LETTER R +#0x73 U+0073 # LATIN SMALL LETTER S +#0x74 U+0074 # LATIN SMALL LETTER T +#0x75 U+0075 # LATIN SMALL LETTER U +#0x76 U+0076 # LATIN SMALL LETTER V +#0x77 U+0077 # LATIN SMALL LETTER W +#0x78 U+0078 # LATIN SMALL LETTER X +#0x79 U+0079 # LATIN SMALL LETTER Y +#0x7A U+007A # LATIN SMALL LETTER Z +#0x7B U+007B # LEFT CURLY BRACKET +#0x7C U+007C # VERTICAL LINE +#0x7D U+007D # RIGHT CURLY BRACKET +#0x7E U+007E # TILDE +0x80 U+00C4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0x81 U+00C5 # LATIN CAPITAL LETTER A WITH RING ABOVE +0x82 U+00C7 # LATIN CAPITAL LETTER C WITH CEDILLA +0x83 U+00C9 # LATIN CAPITAL LETTER E WITH ACUTE +0x84 U+00D1 # LATIN CAPITAL LETTER N WITH TILDE +0x85 U+00D6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0x86 U+00DC # LATIN CAPITAL LETTER U WITH DIAERESIS +0x87 U+00E1 # LATIN SMALL LETTER A WITH ACUTE +0x88 U+00E0 # LATIN SMALL LETTER A WITH GRAVE +0x89 U+00E2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0x8A U+00E4 # LATIN SMALL LETTER A WITH DIAERESIS +0x8B U+00E3 # LATIN SMALL LETTER A WITH TILDE +0x8C U+00E5 # LATIN SMALL LETTER A WITH RING ABOVE +0x8D U+00E7 # LATIN SMALL LETTER C WITH CEDILLA +0x8E U+00E9 # LATIN SMALL LETTER E WITH ACUTE +0x8F U+00E8 # LATIN SMALL LETTER E WITH GRAVE +0x90 U+00EA # LATIN SMALL LETTER E WITH CIRCUMFLEX +0x91 U+00EB # LATIN SMALL LETTER E WITH DIAERESIS +0x92 U+00ED # LATIN SMALL LETTER I WITH ACUTE +0x93 U+00EC # LATIN SMALL LETTER I WITH GRAVE +0x94 U+00EE # LATIN SMALL LETTER I WITH CIRCUMFLEX +0x95 U+00EF # LATIN SMALL LETTER I WITH DIAERESIS +0x96 U+00F1 # LATIN SMALL LETTER N WITH TILDE +0x97 U+00F3 # LATIN SMALL LETTER O WITH ACUTE +0x98 U+00F2 # LATIN SMALL LETTER O WITH GRAVE +0x99 U+00F4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0x9A U+00F6 # LATIN SMALL LETTER O WITH DIAERESIS +0x9B U+00F5 # LATIN SMALL LETTER O WITH TILDE +0x9C U+00FA # LATIN SMALL LETTER U WITH ACUTE +0x9D U+00F9 # LATIN SMALL LETTER U WITH GRAVE +0x9E U+00FB # LATIN SMALL LETTER U WITH CIRCUMFLEX +0x9F U+00FC # LATIN SMALL LETTER U WITH DIAERESIS +0xA0 U+2020 # DAGGER +0xA1 U+00B0 # DEGREE SIGN +0xA2 U+00A2 # CENT SIGN +0xA3 U+00A3 # POUND SIGN +0xA4 U+00A7 # SECTION SIGN +0xA5 U+2022 # BULLET +0xA6 U+00B6 # PILCROW SIGN +0xA7 U+00DF # LATIN SMALL LETTER SHARP S +0xA8 U+00AE # REGISTERED SIGN +0xA9 U+00A9 # COPYRIGHT SIGN +0xAA U+2122 # TRADE MARK SIGN +0xAB U+00B4 # ACUTE ACCENT +0xAC U+00A8 # DIAERESIS +0xAD U+2260 # NOT EQUAL TO +0xAE U+00C6 # LATIN CAPITAL LIGATURE AE +0xAF U+00D8 # LATIN CAPITAL LETTER O WITH STROKE +0xB0 U+221E # INFINITY +0xB1 U+00B1 # PLUS-MINUS SIGN +0xB2 U+2264 # LESS-THAN OR EQUAL TO +0xB3 U+2265 # GREATER-THAN OR EQUAL TO +0xB4 U+00A5 # YEN SIGN +0xB5 U+00B5 # MICRO SIGN +0xB6 U+2202 # PARTIAL DIFFERENTIAL +0xB7 U+2211 # N-ARY SUMMATION +0xB8 U+220F # N-ARY PRODUCT +0xB9 U+03C0 # GREEK SMALL LETTER PI +0xBA U+222B # INTEGRAL +0xBB U+00AA # FEMININE ORDINAL INDICATOR +0xBC U+00BA # MASCULINE ORDINAL INDICATOR +0xBD U+2126 # OHM SIGN +0xBE U+00E6 # LATIN SMALL LIGATURE AE +0xBF U+00F8 # LATIN SMALL LETTER O WITH STROKE +0xC0 U+00BF # INVERTED QUESTION MARK +0xC1 U+00A1 # INVERTED EXCLAMATION MARK +0xC2 U+00AC # NOT SIGN +0xC3 U+221A # SQUARE ROOT +0xC4 U+0192 # LATIN SMALL LETTER F WITH HOOK +0xC5 U+2248 # ALMOST EQUAL TO +0xC6 U+2206 # INCREMENT +0xC7 U+00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xC8 U+00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xC9 U+2026 # HORIZONTAL ELLIPSIS +0xCA U+00A0 # NO-BREAK SPACE +0xCB U+00C0 # LATIN CAPITAL LETTER A WITH GRAVE +0xCC U+00C3 # LATIN CAPITAL LETTER A WITH TILDE +0xCD U+00D5 # LATIN CAPITAL LETTER O WITH TILDE +0xCE U+0152 # LATIN CAPITAL LIGATURE OE +0xCF U+0153 # LATIN SMALL LIGATURE OE +0xD0 U+2013 # EN DASH +0xD1 U+2014 # EM DASH +0xD2 U+201C # LEFT DOUBLE QUOTATION MARK +0xD3 U+201D # RIGHT DOUBLE QUOTATION MARK +0xD4 U+2018 # LEFT SINGLE QUOTATION MARK +0xD5 U+2019 # RIGHT SINGLE QUOTATION MARK +0xD6 U+00F7 # DIVISION SIGN +0xD7 U+25CA # LOZENGE +0xD8 U+00FF # LATIN SMALL LETTER Y WITH DIAERESIS +0xD9 U+0178 # LATIN CAPITAL LETTER Y WITH DIAERESIS +0xDA U+2044 # FRACTION SLASH +0xDB U+00A4 # CURRENCY SIGN +0xDC U+2039 # SINGLE LEFT-POINTING ANGLE QUOTATION MARK +0xDD U+203A # SINGLE RIGHT-POINTING ANGLE QUOTATION MARK +0xDE U+FB01 # LATIN SMALL LIGATURE FI +0xDF U+FB02 # LATIN SMALL LIGATURE FL +0xE0 U+2021 # DOUBLE DAGGER +0xE1 U+00B7 U+0307 U+0387 U+2027 # MIDDLE DOT +0xE2 U+201A # SINGLE LOW-9 QUOTATION MARK +0xE3 U+201E # DOUBLE LOW-9 QUOTATION MARK +0xE4 U+2030 # PER MILLE SIGN +0xE5 U+00C2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0xE6 U+00CA # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0xE7 U+00C1 # LATIN CAPITAL LETTER A WITH ACUTE +0xE8 U+00CB # LATIN CAPITAL LETTER E WITH DIAERESIS +0xE9 U+00C8 # LATIN CAPITAL LETTER E WITH GRAVE +0xEA U+00CD # LATIN CAPITAL LETTER I WITH ACUTE +0xEB U+00CE # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0xEC U+00CF # LATIN CAPITAL LETTER I WITH DIAERESIS +0xED U+00CC # LATIN CAPITAL LETTER I WITH GRAVE +0xEE U+00D3 # LATIN CAPITAL LETTER O WITH ACUTE +0xEF U+00D4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0xF0 # UNDEFINED +0xF1 U+00D2 # LATIN CAPITAL LETTER O WITH GRAVE +0xF2 U+00DA # LATIN CAPITAL LETTER U WITH ACUTE +0xF3 U+00DB # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0xF4 U+00D9 # LATIN CAPITAL LETTER U WITH GRAVE +0xF5 U+0131 # LATIN SMALL LETTER DOTLESS I +0xF6 U+02C6 # MODIFIER LETTER CIRCUMFLEX ACCENT +0xF7 U+02DC # SMALL TILDE +0xF8 U+00AF # MACRON +0xF9 U+02D8 # BREVE +0xFA U+02D9 # DOT ABOVE +0xFB U+02DA # RING ABOVE +0xFC U+00B8 # CEDILLA +0xFD U+02DD # DOUBLE ACUTE ACCENT +0xFE U+02DB # OGONEK +0xFF U+02C7 # CARON +# +# broken vertical bar (¦) - brvbar, brkbar +U+00A6:| +# superscript 3 (³) - sup3 +U+00B3:^3 +# superscript 2 (²) - sup2 +U+00B2:^2 +# superscript 1 (¹) - sup1 +U+00B9:^1 +# fraction 1/4 (¼) - frac14 +U+00BC: 1/4 +# fraction 1/2 (½) - frac12 +U+00BD: 1/2 +# fraction 3/4 (¾) - frac34 +U+00BE: 3/4 +# capital Eth, Icelandic (Ð) - ETH +U+00D0:DH +# Dj # capital D with stroke - Dstrok +# capital Y, acute accent (Ý) - Yacute +U+00DD:Y' +# capital THORN, Icelandic (Þ) - THORN +U+00DE:P +# multiplication sign (×) - times +U+00D7:* +# small eth, Icelandic (ð) - eth +U+00F0:dh +# small y, acute accent (ý) - yacute +U+00FD:y' +# small thorn, Icelandic (þ) - thorn +U+00FE:p +# diff --git a/src/chrtrans/make-msc.bat b/src/chrtrans/make-msc.bat new file mode 100644 index 0000000..81d615e --- /dev/null +++ b/src/chrtrans/make-msc.bat @@ -0,0 +1,6 @@ +@rem $LynxId: make-msc.bat,v 1.6 2008/02/18 00:34:44 tom Exp $
+@echo off
+
+nmake -f makefile.msc %1 %2 %3 %4 %5 %6 %7 %8 %9
+
+if exist makeuctb.exe call makehdrs
diff --git a/src/chrtrans/makefile.bcb b/src/chrtrans/makefile.bcb new file mode 100644 index 0000000..134a395 --- /dev/null +++ b/src/chrtrans/makefile.bcb @@ -0,0 +1,123 @@ +# +# Borland C++ IDE generated makefile +# +# 1997/11/09 (Sun) 14:29:50 +# +.AUTODEPEND + + +# +# Borland C++ tools +# +IMPLIB = Implib +BCC32 = Bcc32 +BccW32.cfg +TLINK32 = TLink32 +TLIB = TLib +BRC32 = Brc32 +TASM32 = Tasm32 +# +# macros +# +BCB = $(MAKEDIR)/.. +BCC_INC = $(BCB)/INCLUDE + +# +# Options +# + +INCLUDES = -I.;../..;../../WWW/LIBRARY/IMPLEMENTATION;$(BCC_INC) +DEFS =-DNO_FILIO_H;NO_UNISTD_H;_WINDOWS;DOSPATH +LNIEAT_dbmakeuctbdexe = -x + +# +# Dependency List +# +Dep_char = .\makeuctb.exe + +char : BccW32.cfg $(Dep_char) + echo MakeNode + +Dep_dbmakeuctbdexe = .\makeuctb.obj + +.\makeuctb.exe : $(Dep_dbmakeuctbdexe) + $(BCC32) makeuctb.obj + +### +.\makeuctb.obj : makeuctb.c + $(BCC32) -P- -c $(DEFS) $(INCLUDES) -o$@ makeuctb.c + +# Compiler configuration file +BccW32.cfg : + Copy &&| +-R +-v +-vi +-H +-H=lynx.csm +-w- +-A- +-wcpt +-wrpt +-wrng +-w-voi +-w-ret +-w-sus +-w-dup +-w-big +-w-ext +-w-zdi +-w-bei +-w-obi +-w-ofp +-w-eas +-w-hid +-w-ncf +-w-ibc +-w-dsz +-w-nst +-w-mpc +-w-mpd +-w-ntd +-w-nvf +-w-hch +-w-inl +-w-lin +-w-lvc +-w-pia +-w-def +-w-nod +-w-pro +-w-rvl +-w-ccc +-w-aus +-w-par +-w-rch +-w-eff +-w-ill +-w-ias +-w-msg +-WC +-Ot +-d- +-K +-a- +-w-stu +-wbbf +-w-dpu +-wcln +-wsig +-wucp +-g200 +-H- +-v- +| $@ + +clean : + -del *_uni.h + -del *_suni.h + -del *.exe + -del *.map + -del *.obj + -del *.tds + -del BccW32.cfg + -del /f/s/q *.i diff --git a/src/chrtrans/makefile.dos b/src/chrtrans/makefile.dos new file mode 100644 index 0000000..07c141f --- /dev/null +++ b/src/chrtrans/makefile.dos @@ -0,0 +1,135 @@ +# +# Makefile for the makeuctb and unicode tables +# for use with DJGPP. +# +# Type make to build makeuctb and all character translation maps. +# Type make fontmap to build makeuctb and translation map iso8859-1. +# Type make makeuctb.exe to build makeuctb only. +# Type make clean to remove makeuctb and character translation maps. +# Type make distclean to remove makeuctb, character translation maps +# and .bak files. +# +CFLAGS = $(MCFLAGS) + +CC = gcc +MCFLAGS = -O2 -DDOSPATH \ +-I. \ +-I../../WWW/Library/Implementation \ +-I/djgpp/watt32/inc +-I../.. + +.SUFFIXES: .tbl +# +# This file contains the font map for the default (hardware) font +# + +FONTMAP_INC = iso01_un.h + +TABLES= \ + cp1250_uni.h \ + cp1251_uni.h \ + cp1252_uni.h \ + cp1253_uni.h \ + cp1255_uni.h \ + cp1256_uni.h \ + cp1257_uni.h \ + cp437_uni.h \ + cp737_uni.h \ + cp775_uni.h \ + cp850_uni.h \ + cp852_uni.h \ + cp857_uni.h \ + cp862_uni.h \ + cp864_uni.h \ + cp866_uni.h \ + cp866u_uni.h \ + cp869_uni.h \ + def7_uni.h \ + dmcs_uni.h \ + hp_uni.h \ + iso01_uni.h \ + iso02_uni.h \ + iso03_uni.h \ + iso04_uni.h \ + iso05_uni.h \ + iso06_uni.h \ + iso07_uni.h \ + iso08_uni.h \ + iso09_uni.h \ + iso10_uni.h \ + iso13_uni.h \ + iso14_uni.h \ + iso15_uni.h \ + koi8r_uni.h \ + koi8u_uni.h \ + mac_uni.h \ + mnem2_suni.h \ + mnem_suni.h \ + next_uni.h \ + pt154_uni.h \ + rfc_suni.h \ + utf8_uni.h \ + viscii_uni.h + +default: $(TABLES) + +fontmap: $(FONTMAP_INC) + +makeuctb.exe: makeuctb.c UCkd.h + $(CC) $(CFLAGS) -o makeuctb.exe makeuctb.c + strip makeuctb.exe + +.tbl.h: + ./makeuctb $*.tbl + +cp1250_uni.h: cp1250_uni.tbl makeuctb.exe +cp1251_uni.h: cp1251_uni.tbl makeuctb.exe +cp1252_uni.h: cp1252_uni.tbl makeuctb.exe +cp1253_uni.h: cp1253_uni.tbl makeuctb.exe +cp1255_uni.h: cp1255_uni.tbl makeuctb.exe +cp1256_uni.h: cp1256_uni.tbl makeuctb.exe +cp1257_uni.h: cp1257_uni.tbl makeuctb.exe +cp437_uni.h: cp437_uni.tbl makeuctb.exe +cp737_uni.h: cp737_uni.tbl makeuctb.exe +cp775_uni.h: cp775_uni.tbl makeuctb.exe +cp850_uni.h: cp850_uni.tbl makeuctb.exe +cp852_uni.h: cp852_uni.tbl makeuctb.exe +cp857_uni.h: cp857_uni.tbl makeuctb.exe +cp862_uni.h: cp862_uni.tbl makeuctb.exe +cp864_uni.h: cp864_uni.tbl makeuctb.exe +cp866_uni.h: cp866_uni.tbl makeuctb.exe +cp866u_uni.h: cp866u_uni.tbl makeuctb.exe +cp869_uni.h: cp869_uni.tbl makeuctb.exe +def7_uni.h: def7_uni.tbl makeuctb.exe +dmcs_uni.h: dmcs_uni.tbl makeuctb.exe +hp_uni.h: hp_uni.tbl makeuctb.exe +iso01_uni.h: iso01_uni.tbl makeuctb.exe +iso02_uni.h: iso02_uni.tbl makeuctb.exe +iso03_uni.h: iso03_uni.tbl makeuctb.exe +iso04_uni.h: iso04_uni.tbl makeuctb.exe +iso05_uni.h: iso05_uni.tbl makeuctb.exe +iso06_uni.h: iso06_uni.tbl makeuctb.exe +iso07_uni.h: iso07_uni.tbl makeuctb.exe +iso08_uni.h: iso08_uni.tbl makeuctb.exe +iso09_uni.h: iso09_uni.tbl makeuctb.exe +iso10_uni.h: iso10_uni.tbl makeuctb.exe +iso13_uni.h: iso13_uni.tbl makeuctb.exe +iso14_uni.h: iso14_uni.tbl makeuctb.exe +iso15_uni.h: iso15_uni.tbl makeuctb.exe +koi8r_uni.h: koi8r_uni.tbl makeuctb.exe +koi8u_uni.h: koi8u_uni.tbl makeuctb.exe +mac_uni.h: mac_uni.tbl makeuctb.exe +mnem2_suni.h: mnem2_suni.tbl makeuctb.exe +mnem_suni.h: mnem_suni.tbl makeuctb.exe +next_uni.h: next_uni.tbl makeuctb.exe +pt154_uni.h: pt154_uni.tbl makeuctb.exe +rfc_suni.h: rfc_suni.tbl makeuctb.exe +utf8_uni.h: utf8_uni.tbl makeuctb.exe +viscii_uni.h: viscii_uni.tbl makeuctb.exe + +clean: + rm -f makeuctb.exe makeuctb *.o *un.h *u.h *c.h *i.h + +distclean: clean + -rm -f *.bak + diff --git a/src/chrtrans/makefile.in b/src/chrtrans/makefile.in new file mode 100644 index 0000000..938d36d --- /dev/null +++ b/src/chrtrans/makefile.in @@ -0,0 +1,199 @@ +# $LynxId: makefile.in,v 1.44 2020/08/28 18:45:03 tom Exp $ +# +# Makefile for the makeuctb and unicode tables. +# +# This may not yet work for the general case. +# Only some dependencies included. +# +SHELL = @CONFIG_SHELL@ + +prefix = @prefix@ +exec_prefix = @exec_prefix@ +datarootdir = @datarootdir@ +top_srcdir = @top_srcdir@ +srcdir = @srcdir@ +VPATH = $(srcdir) + +top_builddir = ../.. + +CC = @CC@ +CPP = @CPP@ +CFLAGS = @CFLAGS@ @EXTRA_CPPFLAGS@ +CPPFLAGS = @CPPFLAGS@ + +WWWINC = WWW/Library/Implementation + +SITE_DEFS = # FIXME: set in parent makefile + +BUILD_CC = @BUILD_CC@ +BUILD_CPP = @BUILD_CPP@ +BUILD_CFLAGS = @BUILD_CFLAGS@ +BUILD_CPPFLAGS = @BUILD_CPPFLAGS@ @DEFS@ + +BUILD_LIBS = @BUILD_LIBS@ +BUILD_LDFLAGS = @BUILD_LDFLAGS@ + +x = @EXEEXT@ +o = .@BUILD_OBJEXT@ +BUILD_EXEEXT = @BUILD_EXEEXT@ + +CPP_OPTS = \ + -I$(top_builddir) \ + -I$(top_srcdir)/src \ + -I$(top_srcdir)/src/chrtrans \ + -I$(top_srcdir)/$(WWWINC) \ + -I$(top_srcdir)/ \ + $(SITE_DEFS) $(BUILD_CPPFLAGS) +CC_OPTS = $(CPP_OPTS) $(BUILD_CFLAGS) + +LINT = @LINT@ +LINTOPTS = @LINT_OPTS@ + +CTAGS = @CTAGS@ + +# +# This file contains the font map for the default (hardware) font +# + +FONTMAP_INC = iso01_uni.h# default, if not set by recursive call + +### fastdep: $(FONTMAP_INC) + +MAKEUCTB = makeuctb$(BUILD_EXEEXT) + +TABLES= \ + cp1250_uni.h \ + cp1251_uni.h \ + cp1252_uni.h \ + cp1253_uni.h \ + cp1255_uni.h \ + cp1256_uni.h \ + cp1257_uni.h \ + cp437_uni.h \ + cp737_uni.h \ + cp775_uni.h \ + cp850_uni.h \ + cp852_uni.h \ + cp857_uni.h \ + cp862_uni.h \ + cp864_uni.h \ + cp866_uni.h \ + cp866u_uni.h \ + cp869_uni.h \ + def7_uni.h \ + dmcs_uni.h \ + hp_uni.h \ + iso01_uni.h \ + iso02_uni.h \ + iso03_uni.h \ + iso04_uni.h \ + iso05_uni.h \ + iso06_uni.h \ + iso07_uni.h \ + iso08_uni.h \ + iso09_uni.h \ + iso10_uni.h \ + iso13_uni.h \ + iso14_uni.h \ + iso15_uni.h \ + koi8r_uni.h \ + koi8u_uni.h \ + mac_uni.h \ + mnem2_suni.h \ + mnem_suni.h \ + next_uni.h \ + next_uni.h \ + pt154_uni.h \ + rfc_suni.h \ + utf8_uni.h \ + viscii_uni.h + +default: $(FONTMAP_INC) + +all tables: $(TABLES) + +OBJS = makeuctb$o +C_SRC = $(OBJS:$o=.c) + +$(MAKEUCTB) : $(OBJS) + $(BUILD_CC) $(CC_OPTS) $(BUILD_LDFLAGS) -o $@ $(OBJS) $(BUILD_LIBS) + +makeuctb$o : $(srcdir)/UCkd.h $(srcdir)/makeuctb.c + +.SUFFIXES : $o .tbl .i .h + +.c$o: + @RULE_CC@ + @ECHO_CC@$(BUILD_CC) $(CC_OPTS) -c $(srcdir)/$*.c + +.c.i: + @RULE_CC@ + @ECHO_CC@$(BUILD_CPP) $(CPP_OPTS) $(srcdir)/$*.c >$@ + +.tbl.h: + ./$(MAKEUCTB) $(srcdir)/$*.tbl $*.h + +# table files listed here once again to get the make dependencies +# right, in case makeuctb was recompiled. +cp1250_uni.h: $(srcdir)/cp1250_uni.tbl $(MAKEUCTB) +cp1251_uni.h: $(srcdir)/cp1251_uni.tbl $(MAKEUCTB) +cp1252_uni.h: $(srcdir)/cp1252_uni.tbl $(MAKEUCTB) +cp1253_uni.h: $(srcdir)/cp1253_uni.tbl $(MAKEUCTB) +cp1255_uni.h: $(srcdir)/cp1255_uni.tbl $(MAKEUCTB) +cp1256_uni.h: $(srcdir)/cp1256_uni.tbl $(MAKEUCTB) +cp1257_uni.h: $(srcdir)/cp1257_uni.tbl $(MAKEUCTB) +cp437_uni.h: $(srcdir)/cp437_uni.tbl $(MAKEUCTB) +cp737_uni.h: $(srcdir)/cp737_uni.tbl $(MAKEUCTB) +cp775_uni.h: $(srcdir)/cp775_uni.tbl $(MAKEUCTB) +cp850_uni.h: $(srcdir)/cp850_uni.tbl $(MAKEUCTB) +cp852_uni.h: $(srcdir)/cp852_uni.tbl $(MAKEUCTB) +cp857_uni.h: $(srcdir)/cp857_uni.tbl $(MAKEUCTB) +cp862_uni.h: $(srcdir)/cp862_uni.tbl $(MAKEUCTB) +cp864_uni.h: $(srcdir)/cp864_uni.tbl $(MAKEUCTB) +cp866_uni.h: $(srcdir)/cp866_uni.tbl $(MAKEUCTB) +cp866u_uni.h: $(srcdir)/cp866u_uni.tbl $(MAKEUCTB) +cp869_uni.h: $(srcdir)/cp869_uni.tbl $(MAKEUCTB) +def7_uni.h: $(srcdir)/def7_uni.tbl $(MAKEUCTB) +dmcs_uni.h: $(srcdir)/dmcs_uni.tbl $(MAKEUCTB) +hp_uni.h: $(srcdir)/hp_uni.tbl $(MAKEUCTB) +iso01_uni.h: $(srcdir)/iso01_uni.tbl $(MAKEUCTB) +iso02_uni.h: $(srcdir)/iso02_uni.tbl $(MAKEUCTB) +iso03_uni.h: $(srcdir)/iso03_uni.tbl $(MAKEUCTB) +iso04_uni.h: $(srcdir)/iso04_uni.tbl $(MAKEUCTB) +iso05_uni.h: $(srcdir)/iso05_uni.tbl $(MAKEUCTB) +iso06_uni.h: $(srcdir)/iso06_uni.tbl $(MAKEUCTB) +iso07_uni.h: $(srcdir)/iso07_uni.tbl $(MAKEUCTB) +iso08_uni.h: $(srcdir)/iso08_uni.tbl $(MAKEUCTB) +iso09_uni.h: $(srcdir)/iso09_uni.tbl $(MAKEUCTB) +iso10_uni.h: $(srcdir)/iso10_uni.tbl $(MAKEUCTB) +iso13_uni.h: $(srcdir)/iso13_uni.tbl $(MAKEUCTB) +iso14_uni.h: $(srcdir)/iso14_uni.tbl $(MAKEUCTB) +iso15_uni.h: $(srcdir)/iso15_uni.tbl $(MAKEUCTB) +koi8r_uni.h: $(srcdir)/koi8r_uni.tbl $(MAKEUCTB) +koi8u_uni.h: $(srcdir)/koi8u_uni.tbl $(MAKEUCTB) +mac_uni.h: $(srcdir)/mac_uni.tbl $(MAKEUCTB) +mnem2_suni.h: $(srcdir)/mnem2_suni.tbl $(MAKEUCTB) +mnem_suni.h: $(srcdir)/mnem_suni.tbl $(MAKEUCTB) +next_uni.h: $(srcdir)/next_uni.tbl $(MAKEUCTB) +pt154_uni.h: $(srcdir)/pt154_uni.tbl $(MAKEUCTB) +rfc_suni.h: $(srcdir)/rfc_suni.tbl $(MAKEUCTB) +utf8_uni.h: $(srcdir)/utf8_uni.tbl $(MAKEUCTB) +viscii_uni.h: $(srcdir)/viscii_uni.tbl $(MAKEUCTB) + +lint: + $(LINT) $(LINTOPTS) $(CPP_OPTS) $(C_SRC) 2>&1 |tee $(top_builddir)/lint.chrtrans + +tags: + $(CTAGS) *.[ch] + +clean: + rm -f $(MAKEUCTB) *$o *uni.h *uni2.h *.i + +distclean: clean + -rm -rf obsolete + rm -f core *.bak *.sav *~ *.h_old + +depend : + makedepend -fmakefile -- $(CPP_OPTS) -- $(C_SRC) + +# DO NOT DELETE THIS LINE -- make depend depends on it. diff --git a/src/chrtrans/makefile.msc b/src/chrtrans/makefile.msc new file mode 100644 index 0000000..48d074a --- /dev/null +++ b/src/chrtrans/makefile.msc @@ -0,0 +1,137 @@ +# +# Makefile for Microsoft Visual C++ 4.2 or later +# + +CC = cl +LD = link + +INCLUDES = /I "." /I ".." /I "..\.." /I "..\..\WWW\Library\Implementation" /I "..\..\lib" +DEFS = /D "WIN32_LEAN_AND_MEAN" /D "NDEBUG" /D "__WIN32__" /D "WIN32" /D "_CONSOLE" /D "_MBCS" /D "_WIN32" /D "NO_FILEIO_H" /D "NO_UNISTD_H" /D "_WINDOWS" /D "DOSPATH" +CFLAGS = /nologo /MT /W3 /EHsc /O2 /c + +MACHINE = i386 +LDFLAGS = /nologo /subsystem:console /incremental:no /machine:$(MACHINE) +LIBS = user32.lib wsock32.lib + +COMPILE = $(CC) $(CFLAGS) $(INCLUDES) $(DEFS) +LINK = $(LD) $(LDFLAGS) /out:$@ + + +.SUFFIXES: .tbl +# +# This file contains the font map for the default (hardware) font +# + +FONTMAP_INC = iso01_un.h + +TABLES= \ + cp1250_uni.h \ + cp1251_uni.h \ + cp1252_uni.h \ + cp1253_uni.h \ + cp1255_uni.h \ + cp1256_uni.h \ + cp1257_uni.h \ + cp437_uni.h \ + cp737_uni.h \ + cp775_uni.h \ + cp850_uni.h \ + cp852_uni.h \ + cp857_uni.h \ + cp862_uni.h \ + cp864_uni.h \ + cp866_uni.h \ + cp866u_uni.h \ + cp869_uni.h \ + def7_uni.h \ + dmcs_uni.h \ + hp_uni.h \ + iso01_uni.h \ + iso02_uni.h \ + iso03_uni.h \ + iso04_uni.h \ + iso05_uni.h \ + iso06_uni.h \ + iso07_uni.h \ + iso08_uni.h \ + iso09_uni.h \ + iso10_uni.h \ + iso13_uni.h \ + iso14_uni.h \ + iso15_uni.h \ + koi8r_uni.h \ + koi8u_uni.h \ + mac_uni.h \ + mnem2_suni.h \ + mnem_suni.h \ + next_uni.h \ + pt154_uni.h \ + rfc_suni.h \ + utf8_uni.h \ + viscii_uni.h + +default: $(TABLES) + +fontmap: $(FONTMAP_INC) + +makeuctb.exe : makeuctb.obj + $(LINK) makeuctb.obj $(LIBS) + +makeuctb.obj : makeuctb.c + $(COMPILE) makeuctb.c + +.tbl.h: + makeuctb $*.tbl + +cp1250_uni.h: cp1250_uni.tbl makeuctb.exe +cp1251_uni.h: cp1251_uni.tbl makeuctb.exe +cp1252_uni.h: cp1252_uni.tbl makeuctb.exe +cp1253_uni.h: cp1253_uni.tbl makeuctb.exe +cp1255_uni.h: cp1255_uni.tbl makeuctb.exe +cp1256_uni.h: cp1256_uni.tbl makeuctb.exe +cp1257_uni.h: cp1257_uni.tbl makeuctb.exe +cp437_uni.h: cp437_uni.tbl makeuctb.exe +cp737_uni.h: cp737_uni.tbl makeuctb.exe +cp775_uni.h: cp775_uni.tbl makeuctb.exe +cp850_uni.h: cp850_uni.tbl makeuctb.exe +cp852_uni.h: cp852_uni.tbl makeuctb.exe +cp857_uni.h: cp857_uni.tbl makeuctb.exe +cp862_uni.h: cp862_uni.tbl makeuctb.exe +cp864_uni.h: cp864_uni.tbl makeuctb.exe +cp866_uni.h: cp866_uni.tbl makeuctb.exe +cp866u_uni.h: cp866u_uni.tbl makeuctb.exe +cp869_uni.h: cp869_uni.tbl makeuctb.exe +def7_uni.h: def7_uni.tbl makeuctb.exe +dmcs_uni.h: dmcs_uni.tbl makeuctb.exe +hp_uni.h: hp_uni.tbl makeuctb.exe +iso01_uni.h: iso01_uni.tbl makeuctb.exe +iso02_uni.h: iso02_uni.tbl makeuctb.exe +iso03_uni.h: iso03_uni.tbl makeuctb.exe +iso04_uni.h: iso04_uni.tbl makeuctb.exe +iso05_uni.h: iso05_uni.tbl makeuctb.exe +iso06_uni.h: iso06_uni.tbl makeuctb.exe +iso07_uni.h: iso07_uni.tbl makeuctb.exe +iso08_uni.h: iso08_uni.tbl makeuctb.exe +iso09_uni.h: iso09_uni.tbl makeuctb.exe +iso10_uni.h: iso10_uni.tbl makeuctb.exe +iso13_uni.h: iso13_uni.tbl makeuctb.exe +iso14_uni.h: iso14_uni.tbl makeuctb.exe +iso15_uni.h: iso15_uni.tbl makeuctb.exe +koi8r_uni.h: koi8r_uni.tbl makeuctb.exe +koi8u_uni.h: koi8u_uni.tbl makeuctb.exe +mac_uni.h: mac_uni.tbl makeuctb.exe +mnem2_suni.h: mnem2_suni.tbl makeuctb.exe +mnem_suni.h: mnem_suni.tbl makeuctb.exe +next_uni.h: next_uni.tbl makeuctb.exe +pt154_uni.h: pt154_uni.tbl makeuctb.exe +rfc_suni.h: rfc_suni.tbl makeuctb.exe +utf8_uni.h: utf8_uni.tbl makeuctb.exe +viscii_uni.h: viscii_uni.tbl makeuctb.exe + +clean : + - erase *.obj + - erase *.exe + - for %%i in ( $(TABLES) ) do erase %%i + +distclean : clean + - erase *.bak diff --git a/src/chrtrans/makehdrs.bat b/src/chrtrans/makehdrs.bat new file mode 100644 index 0000000..2edbef7 --- /dev/null +++ b/src/chrtrans/makehdrs.bat @@ -0,0 +1,50 @@ +@rem $LynxId: makehdrs.bat,v 1.1 2007/06/28 21:50:29 tom Exp $
+@echo If .tbl files are added or removed you will need to hand edit
+@echo this batch file.
+@echo .
+@echo off
+
+makeuctb cp1250_uni.tbl
+makeuctb cp1251_uni.tbl
+makeuctb cp1252_uni.tbl
+makeuctb cp1253_uni.tbl
+makeuctb cp1255_uni.tbl
+makeuctb cp1256_uni.tbl
+makeuctb cp1257_uni.tbl
+makeuctb cp437_uni.tbl
+makeuctb cp737_uni.tbl
+makeuctb cp775_uni.tbl
+makeuctb cp850_uni.tbl
+makeuctb cp852_uni.tbl
+makeuctb cp857_uni.tbl
+makeuctb cp862_uni.tbl
+makeuctb cp864_uni.tbl
+makeuctb cp866_uni.tbl
+makeuctb cp866u_uni.tbl
+makeuctb cp869_uni.tbl
+makeuctb def7_uni.tbl
+makeuctb dmcs_uni.tbl
+makeuctb hp_uni.tbl
+makeuctb iso01_uni.tbl
+makeuctb iso02_uni.tbl
+makeuctb iso03_uni.tbl
+makeuctb iso04_uni.tbl
+makeuctb iso05_uni.tbl
+makeuctb iso06_uni.tbl
+makeuctb iso07_uni.tbl
+makeuctb iso08_uni.tbl
+makeuctb iso09_uni.tbl
+makeuctb iso10_uni.tbl
+makeuctb iso13_uni.tbl
+makeuctb iso14_uni.tbl
+makeuctb iso15_uni.tbl
+makeuctb koi8r_uni.tbl
+makeuctb koi8u_uni.tbl
+makeuctb mac_uni.tbl
+makeuctb mnem2_suni.tbl
+makeuctb mnem_suni.tbl
+makeuctb next_uni.tbl
+makeuctb pt154_uni.tbl
+makeuctb rfc_suni.tbl
+makeuctb utf8_uni.tbl
+makeuctb viscii_uni.tbl
diff --git a/src/chrtrans/makeuctb.c b/src/chrtrans/makeuctb.c new file mode 100644 index 0000000..9035292 --- /dev/null +++ b/src/chrtrans/makeuctb.c @@ -0,0 +1,914 @@ +/* + * $LynxId: makeuctb.c,v 1.51 2020/09/03 22:09:36 tom Exp $ + * + * makeuctb.c, derived from conmakehash.c - kw + * + * Original comments from conmakehash.c: + * + * Create arrays for initializing the kernel folded tables (using a hash + * table turned out to be to limiting...) Unfortunately we can't simply + * preinitialize the tables at compile time since kfree() cannot accept + * memory not allocated by kmalloc(), and doing our own memory management + * just for this seems like massive overkill. + * + * Copyright (C) 1995 H. Peter Anvin + * + * This program is a part of the Linux kernel, and may be freely + * copied under the terms of the GNU General Public License (GPL), + * version 2, or at your option any later version. + */ + +#ifndef HAVE_CONFIG_H +/* override HTUtils.h fallbacks for cross-compiling */ +#undef HAVE_LSTAT +#undef NO_FILIO_H +#define HAVE_LSTAT 1 +#define NO_FILIO_H 1 +#endif + +#define DONT_USE_GETTEXT +#define DONT_USE_SOCKS5 +#include <UCDefs.h> +#include <UCkd.h> + +#ifdef LY_FIND_LEAKS /* CF_ARG_ENABLE(find-leaks) */ +#define FreeLeak(p) FREE(p) +#else +#define FreeLeak(p) /* nothing */ +#endif + +#define L_CURL '{' +#define R_CURL '}' + +/* + * Don't try to use LYexit() since this is a standalone file. + */ +#ifdef exit +#undef exit +#endif /* exit */ + +#include <stdlib.h> +#include <string.h> +#include <ctype.h> + +#define MAX_FONTLEN 256 + +/* + * We don't deal with UCS4 here. - KW + */ +typedef u16 unicode; + +static FILE *chdr = 0; + +/* + * Since we may be writing the formatted file to stdout, ensure that we flush + * everything before leaving, since some old (and a few not-so-old) platforms + * do not properly implement POSIX 'exit()'. + */ +static void done(int code) GCC_NORETURN; + +static void done(int code) +{ + if (chdr != 0) { + fflush(chdr); + fclose(chdr); + } + fflush(stderr); + exit(code); +} + +static void usage(void) +{ + static const char *tbl[] = + { + "Usage: makeuctb [parameters]", + "", + "Utility to convert .tbl into .h files for Lynx compilation.", + "", + "Parameters (all are optional):", + " 1: the input file (normally {filename}.tbl, but \"-\" for stdin", + " 2: the output file (normally {filename}.tbl but \"-\" for stdout", + " 3: charset mime name", + " 4: charset display name" + }; + unsigned n; + + for (n = 0; n < TABLESIZE(tbl); n++) { + fprintf(stderr, "%s\n", tbl[n]); + }; + done(EX_USAGE); +} + +#ifdef USE_ASCII_CTYPES +static int ascii_tolower(int i) +{ + if (91 > i && i > 64) + return (i + 32); + else + return i; +} +#endif + +/* copied from HTString.c, not everybody has strncasecmp */ +int strncasecomp(const char *a, const char *b, int n) +{ + const char *p; + const char *q; + + for (p = a, q = b;; p++, q++) { + int diff; + + if (p == (a + n)) + return 0; /* Match up to n characters */ + if (!(*p && *q)) + return (*p - *q); + diff = TOLOWER(*p) - TOLOWER(*q); + if (diff) + return diff; + } + /*NOTREACHED */ +} + +static int getunicode(char **p0) +{ + char *p = *p0; + + while (*p == ' ' || *p == '\t') + p++; + + if (*p == '-') { + return -2; + } else if (*p != 'U' || p[1] != '+' || + !isxdigit(UCH(p[2])) || + !isxdigit(UCH(p[3])) || + !isxdigit(UCH(p[4])) || + !isxdigit(UCH(p[5])) || + isxdigit(UCH(p[6]))) { + return -1; + } + *p0 = p + 6; + return (int) strtol((p + 2), 0, 16); +} + +/* + * Massive overkill, but who cares? + */ +static unicode unitable[MAX_FONTLEN][255]; +static int unicount[MAX_FONTLEN]; + +static struct unimapdesc_str themap_str = +{0, NULL, 0, 0}; + +static const char *tblname; +static const char *hdrname; + +static int RawOrEnc = 0; +static int Raw_found = 0; /* whether explicit R directive found */ +static int CodePage = 0; + +#define MAX_UNIPAIRS 4500 + +static void addpair_str(char *str, int un) +{ + int i = 0; + + if (un <= 0xfffe) { + if (!themap_str.entry_ct) { + /* + * Initialize the map for replacement strings. + */ + themap_str.entries = (struct unipair_str *) calloc(MAX_UNIPAIRS, + sizeof(struct unipair_str)); + + if (!themap_str.entries) { + fprintf(stderr, + "%s: Out of memory\n", tblname); + done(EX_DATAERR); + } + } else { + /* + * Check that it isn't a duplicate. + */ + for (i = 0; i < themap_str.entry_ct; i++) { + if (themap_str.entries[i].unicode == un) { + FreeLeak(themap_str.entries[i].replace_str); + themap_str.entries[i].replace_str = str; + return; + } + } + } + + /* + * Add to list. + */ + if (themap_str.entry_ct > MAX_UNIPAIRS - 1) { + fprintf(stderr, + "ERROR: Only %d unicode replacement strings permitted!\n", + MAX_UNIPAIRS); + done(EX_DATAERR); + } + FreeLeak(themap_str.entries[themap_str.entry_ct].replace_str); + themap_str.entries[themap_str.entry_ct].unicode = (u16) un; + themap_str.entries[themap_str.entry_ct].replace_str = str; + themap_str.entry_ct++; + } + /* otherwise: ignore */ +} + +static void addpair(int fp, int un) +{ + int i; + + if (!Raw_found) { /* enc not (yet) explicitly given with 'R' */ + if (fp >= 128) { + if (RawOrEnc != UCT_ENC_8BIT && RawOrEnc <= UCT_ENC_8859) { + if (fp < 160) { /* cannot be 8859 */ + RawOrEnc = UCT_ENC_8BIT; + } else if (fp != 160 && fp != 173) { + RawOrEnc = UCT_ENC_8859; /* hmmm.. more tests needed? */ + } else if (unicount[fp] == 0 && fp != un) { + /* first unicode for fp doesn't map to itself */ + RawOrEnc = UCT_ENC_8BIT; + } else { + RawOrEnc = UCT_ENC_8859; /* hmmm.. more tests needed? */ + } + } + } + } + if (un <= 0xfffe) { + /* + * Check that it isn't a duplicate. + */ + for (i = 0; i < unicount[fp]; i++) { + if (unitable[fp][i] == un) { + return; + } + } + + /* + * Add to list. + */ + if (unicount[fp] > 254) { + fprintf(stderr, "ERROR: Only 255 unicodes/glyph permitted!\n"); + done(EX_DATAERR); + } + unitable[fp][unicount[fp]] = (u16) un; + unicount[fp]++; + } + /* otherwise: ignore */ +} + +static char this_MIMEcharset[UC_MAXLEN_MIMECSNAME + 1]; +static char this_LYNXcharset[UC_MAXLEN_LYNXCSNAME + 1]; +static char id_append[UC_MAXLEN_ID_APPEND + 1] = "_"; +static int this_isDefaultMap = -1; +static int useDefaultMap = 1; +static int lowest_eight = 999; + +int main(int argc, char **argv) +{ + static const char *first_ifdefs[] = + { + "/*", + " * Compile-in this chunk of code unless we've turned it off specifically", + " * or in general (id=%s).", + " */", + "", + "#ifndef INCL_CHARSET%s", + "#define INCL_CHARSET%s 1", + "", + "/*ifdef NO_CHARSET*/", + "#ifdef NO_CHARSET", + "#undef NO_CHARSET", + "#endif", + "#define NO_CHARSET 0 /* force default to always be active */", + "", + "/*ifndef NO_CHARSET%s*/", + "#ifndef NO_CHARSET%s", + "", + "#if ALL_CHARSETS", + "#define NO_CHARSET%s 0", + "#else", + "#define NO_CHARSET%s 1", + "#endif", + "", + "#endif /* ndef(NO_CHARSET%s) */", + "", + "#if NO_CHARSET%s", + "#define UC_CHARSET_SETUP%s /*nothing*/", + "#else" + }; + static const char *last_ifdefs[] = + { + "", + "#endif /* NO_CHARSET%s */", + "", + "#endif /* INCL_CHARSET%s */" + }; + + FILE *ctbl; + char buffer[65536]; + char *outname = 0; + unsigned n; + int fontlen; + int i, nuni, nent; + int fp0 = 0, fp1 = 0, un0, un1; + char *p, *p1; + char *tbuf = NULL, ch; + + if (argc < 2 || argc > 5) { + usage(); + } + + if (!strcmp(argv[1], "-")) { + ctbl = stdin; + tblname = "stdin"; + } else { + ctbl = fopen(tblname = argv[1], "r"); + if (!ctbl) { + perror(tblname); + done(EX_NOINPUT); + } + } + + if (argc > 2) { + if (!strcmp(argv[2], "-")) { + chdr = stdout; + hdrname = "stdout"; + } else { + hdrname = argv[2]; + } + } else if (ctbl == stdin) { + chdr = stdout; + hdrname = "stdout"; + } else if ((outname = (char *) malloc(strlen(tblname) + 3)) != 0) { + strcpy(outname, tblname); + hdrname = outname; + if ((p = strrchr(outname, '.')) == 0) + p = outname + strlen(outname); + strcpy(p, ".h"); + } else { + perror("malloc"); + done(EX_NOINPUT); + } + + if (chdr == 0) { + chdr = fopen(hdrname, "w"); + if (!chdr) { + perror(hdrname); + done(EX_NOINPUT); + } + } + + /* + * For now we assume the default font is always 256 characters. + */ + fontlen = 256; + + /* + * Initialize table. + */ + for (i = 0; i < fontlen; i++) { + unicount[i] = 0; + } + + /* + * Now we comes to the tricky part. Parse the input table. + */ + while (fgets(buffer, (int) sizeof(buffer), ctbl) != NULL) { + if ((p = StrChr(buffer, '\n')) != NULL) { + *p = '\0'; + } else { + fprintf(stderr, + "%s: Warning: line too long or incomplete.\n", + tblname); + } + + /* + * Syntax accepted: + * <fontpos> <unicode> <unicode> ... + * <fontpos> <unicode range> <unicode range> ... + * <fontpos> idem + * <range> idem + * <range> <unicode range> + * <unicode> :<replace> + * <unicode range> :<replace> + * <unicode> "<C replace>" + * <unicode range> "<C replace>" + * + * where <range> ::= <fontpos>-<fontpos> + * and <unicode> ::= U+<h><h><h><h> + * and <h> ::= <hexadecimal digit> + * and <replace> any string not containing '\n' or '\0' + * and <C replace> any string with C backslash escapes. + */ + p = buffer; + while (*p == ' ' || *p == '\t') { + p++; + } + if (!(*p) || *p == '#') { + /* + * Skip comment or blank line. + */ + continue; + } + + switch (*p) { + /* + * Raw Unicode? I.e. needs some special + * processing. One digit code. + */ + case 'R': + if (p[1] == 'a' || p[1] == 'A') { + buffer[sizeof(buffer) - 1] = '\0'; + if (!strncasecomp(p, "RawOrEnc", 8)) { + p += 8; + } + } + p++; + while (*p == ' ' || *p == '\t') { + p++; + } + RawOrEnc = (int) strtol(p, 0, 10); + Raw_found = 1; + continue; + + /* + * Is this the default table? + */ + case 'D': + if (p[1] == 'e' || p[1] == 'E') { + buffer[sizeof(buffer) - 1] = '\0'; + if (!strncasecomp(p, "Default", 7)) { + p += 7; + } + } + p++; + while (*p == ' ' || *p == '\t') { + p++; + } + this_isDefaultMap = (*p == '1' || TOLOWER(*p) == 'y'); + continue; + + /* + * Is this the default table? + */ + case 'F': + if (p[1] == 'a' || p[1] == 'A') { + buffer[sizeof(buffer) - 1] = '\0'; + if (!strncasecomp(p, "FallBack", 8)) { + p += 8; + } + } + p++; + while (*p == ' ' || *p == '\t') { + p++; + } + useDefaultMap = (*p == '1' || TOLOWER(*p) == 'y'); + continue; + + case 'M': + if (p[1] == 'i' || p[1] == 'I') { + buffer[sizeof(buffer) - 1] = '\0'; + if (!strncasecomp(p, "MIMEName", 8)) { + p += 8; + } + } + p++; + while (*p == ' ' || *p == '\t') { + p++; + } + sscanf(p, "%40s", this_MIMEcharset); + continue; + + /* + * Display charset name for options screen. + */ + case 'O': + if (p[1] == 'p' || p[1] == 'P') { + buffer[sizeof(buffer) - 1] = '\0'; + if (!strncasecomp(p, "OptionName", 10)) { + p += 10; + } + } + p++; + while (*p == ' ' || *p == '\t') { + p++; + } + for (i = 0; *p && i < UC_MAXLEN_LYNXCSNAME; p++, i++) { + this_LYNXcharset[i] = *p; + } + this_LYNXcharset[i] = '\0'; + continue; + + /* + * Codepage number. Three or four digit code. + */ + case 'C': + if (p[1] == 'o' || p[1] == 'O') { + buffer[sizeof(buffer) - 1] = '\0'; + if (!strncasecomp(p, "CodePage", 8)) { + p += 8; + } + } + p++; + while (*p == ' ' || *p == '\t') { + p++; + } + CodePage = (int) strtol(p, 0, 10); + continue; + } + + if (*p == 'U') { + un0 = getunicode(&p); + if (un0 < 0) { + fprintf(stderr, "Bad input line: %s\n", buffer); + done(EX_DATAERR); + fprintf(stderr, + "%s: Bad Unicode range corresponding to font position range 0x%x-0x%x\n", + tblname, fp0, fp1); + done(EX_DATAERR); + } + un1 = un0; + while (*p == ' ' || *p == '\t') { + p++; + } + if (*p == '-') { + p++; + while (*p == ' ' || *p == '\t') { + p++; + } + un1 = getunicode(&p); + if (un1 < 0 || un1 < un0) { + fprintf(stderr, + "%s: Bad Unicode range U+%x-U+%x\n", + tblname, un0, un1); + fprintf(stderr, "Bad input line: %s\n", buffer); + done(EX_DATAERR); + } + while (*p == ' ' || *p == '\t') { + p++; + } + } + + if (*p != ':' && *p != '"') { + fprintf(stderr, "No ':' or '\"' where expected: %s\n", + buffer); + continue; + } + + /* + * Allocate a string large enough for the worst-case use in the + * loop using sprintf. + */ + tbuf = (char *) malloc(5 * strlen(p)); + + if (!(p1 = tbuf)) { + fprintf(stderr, "%s: Out of memory\n", tblname); + done(EX_DATAERR); + } + if (*p == '"') { + /* + * Handle "<C replace>". + * Copy chars verbatim until first '"' not \-escaped or + * end of buffer. + */ + int escaped = 0; + + ch = 0; + for (++p; *p != '\0'; p++) { + ch = *p; + if (escaped) { + escaped = 0; + } else if (ch == '"') { + break; + } else if (ch == '\\') { + escaped = 1; + } + *p1++ = ch; + } + if (escaped || ch != '"') { + fprintf(stderr, "Warning: String not terminated: %s\n", + buffer); + if (escaped) + *p1++ = '\n'; + } + } else { + /* + * We had ':'. + */ + for (++p; *p != '\0'; p++, p1++) { + ch = *p; + if (UCH(ch) < 32 || ch == '\\' || ch == '\"' || + UCH(ch) >= 127) { + sprintf(p1, "\\%.3o", UCH(ch)); + p1 += 3; + } else { + *p1 = ch; + } + } + } + *p1 = '\0'; + for (i = un0; i <= un1; i++) { + addpair_str(tbuf, i); + } + continue; + } + + /* + * Input line (after skipping spaces) doesn't start with one + * of the specially recognized characters, so try to interpret + * it as starting with a fontpos. + */ + fp0 = (int) strtol(p, &p1, 0); + if (p1 == p) { + fprintf(stderr, "Bad input line: %s\n", buffer); + done(EX_DATAERR); + } + p = p1; + + while (*p == ' ' || *p == '\t') { + p++; + } + if (*p == '-') { + p++; + fp1 = (int) strtol(p, &p1, 0); + if (p1 == p) { + fprintf(stderr, "Bad input line: %s\n", buffer); + done(EX_DATAERR); + } + p = p1; + } else { + fp1 = 0; + } + + if (fp0 < 0 || fp0 >= fontlen) { + fprintf(stderr, + "%s: Glyph number (0x%x) larger than font length\n", + tblname, fp0); + done(EX_DATAERR); + } + if (fp1 && (fp1 < fp0 || fp1 >= fontlen)) { + fprintf(stderr, + "%s: Bad end of range (0x%x)\n", + tblname, fp1); + done(EX_DATAERR); + } + + if (fp1) { + /* + * We have a range; expect the word "idem" + * or a Unicode range of the same length. + */ + while (*p == ' ' || *p == '\t') { + p++; + } + if (!StrNCmp(p, "idem", 4)) { + for (i = fp0; i <= fp1; i++) { + addpair(i, i); + } + p += 4; + } else { + un0 = getunicode(&p); + while (*p == ' ' || *p == '\t') { + p++; + } + if (*p != '-') { + fprintf(stderr, + "%s: Corresponding to a range of font positions,", + tblname); + fprintf(stderr, + " there should be a Unicode range.\n"); + done(EX_DATAERR); + } + p++; + un1 = getunicode(&p); + if (un0 < 0 || un1 < 0) { + fprintf(stderr, + "%s: Bad Unicode range corresponding to font position range 0x%x-0x%x\n", + tblname, fp0, fp1); + done(EX_DATAERR); + } + if (un1 - un0 != fp1 - fp0) { + fprintf(stderr, + "%s: Unicode range U+%x-U+%x not of the same length", + tblname, un0, un1); + fprintf(stderr, + " as font position range 0x%x-0x%x\n", + fp0, fp1); + done(EX_DATAERR); + } + for (i = fp0; i <= fp1; i++) { + addpair(i, un0 - fp0 + i); + } + } + } else { + /* + * No range; expect a list of unicode values + * or unicode ranges for a single font position, + * or the word "idem" + */ + while (*p == ' ' || *p == '\t') { + p++; + } + if (!StrNCmp(p, "idem", 4)) { + addpair(fp0, fp0); + p += 4; + } + while ((un0 = getunicode(&p)) >= 0) { + addpair(fp0, un0); + while (*p == ' ' || *p == '\t') { + p++; + } + if (*p == '-') { + p++; + un1 = getunicode(&p); + if (un1 < un0) { + fprintf(stderr, + "%s: Bad Unicode range 0x%x-0x%x\n", + tblname, un0, un1); + done(EX_DATAERR); + } + for (un0++; un0 <= un1; un0++) { + addpair(fp0, un0); + } + } + } + } + while (*p == ' ' || *p == '\t') { + p++; + } + if (*p && *p != '#') { + fprintf(stderr, "%s: trailing junk (%s) ignored\n", tblname, p); + } + } + + /* + * Okay, we hit EOF, now output tables. + */ + fclose(ctbl); + + /* + * Compute total size of Unicode list. + */ + nuni = 0; + for (i = 0; i < fontlen; i++) { + nuni += unicount[i]; + } + + if (argc > 3) { + StrNCpy(this_MIMEcharset, argv[3], UC_MAXLEN_MIMECSNAME); + } else if (this_MIMEcharset[0] == '\0') { + StrNCpy(this_MIMEcharset, tblname, UC_MAXLEN_MIMECSNAME); + if ((p = StrChr(this_MIMEcharset, '.')) != 0) { + *p = '\0'; + } + } + for (p = this_MIMEcharset; *p; p++) { + *p = (char) TOLOWER(*p); + } + if (argc > 4) { + StrNCpy(this_LYNXcharset, argv[4], UC_MAXLEN_LYNXCSNAME); + } else if (this_LYNXcharset[0] == '\0') { + memcpy(this_LYNXcharset, this_MIMEcharset, UC_MAXLEN_LYNXCSNAME); + } + + if (this_isDefaultMap == -1) { + this_isDefaultMap = !StrNCmp(this_MIMEcharset, "iso-8859-1", 10); + } + fprintf(stderr, + "makeuctb: %s: %stranslation map", + this_MIMEcharset, (this_isDefaultMap ? "default " : "")); + if (this_isDefaultMap == 1) { + *id_append = '\0'; + } else { + for (i = 0, p = this_MIMEcharset; + *p && (i < UC_MAXLEN_ID_APPEND - 1); + p++, i++) { + id_append[i + 1] = (char) (isalnum(UCH(*p)) ? *p : '_'); + } + id_append[i + 1] = '\0'; + } + fprintf(stderr, " (%s).\n", id_append); + + for (n = 0; n < TABLESIZE(first_ifdefs); n++) { + fprintf(chdr, first_ifdefs[n], id_append); + fprintf(chdr, "\n"); + } + + fprintf(chdr, "\n\ +/*\n\ + * uni_hash.tbl\n\ + *\n\ + * Do not edit this file; it was automatically generated by\n\ + *\n\ + * %s %s\n\ + *\n\ + */\n\ +\n\ +static const u8 dfont_unicount%s[%d] = \n\ +%c\n\t", argv[0], argv[1], id_append, fontlen, L_CURL); + + for (i = 0; i < fontlen; i++) { + if (i >= 128 && unicount[i] > 0 && i < lowest_eight) { + lowest_eight = i; + } + fprintf(chdr, "%3d", unicount[i]); + if (i == (fontlen - 1)) { + fprintf(chdr, "\n%c;\n", R_CURL); + } else if ((i % 8) == 7) { + fprintf(chdr, ",\n\t"); + } else { + fprintf(chdr, ", "); + } + } + + /* + * If lowest_eightbit is anything else but 999, + * this can't be 7-bit only. + */ + if (lowest_eight != 999 && !RawOrEnc) { + RawOrEnc = UCT_ENC_8BIT; + } + + if (nuni) { + fprintf(chdr, "\nstatic const u16 dfont_unitable%s[%d] = \n%c\n\t", + id_append, nuni, L_CURL); + } else { + fprintf(chdr, + "\nstatic const u16 dfont_unitable%s[1] = {0}; /* dummy */\n", id_append); + } + + fp0 = 0; + nent = 0; + for (i = 0; i < nuni; i++) { + while (nent >= unicount[fp0]) { + fp0++; + nent = 0; + } + fprintf(chdr, "0x%04x", unitable[fp0][nent++]); + if (i == (nuni - 1)) { + fprintf(chdr, "\n%c;\n", R_CURL); + } else if ((i % 8) == 7) { + fprintf(chdr, ",\n\t"); + } else { + fprintf(chdr, ", "); + } + } + + if (themap_str.entry_ct) { + fprintf(chdr, "\n\ +static struct unipair_str repl_map%s[%d] = \n\ +%c\n\t", id_append, themap_str.entry_ct, L_CURL); + } else { + fprintf(chdr, "\n\ +/* static struct unipair_str repl_map%s[]; */\n", id_append); + } + + for (i = 0; i < themap_str.entry_ct; i++) { + fprintf(chdr, "%c0x%x,\"%s\"%c", + L_CURL, + themap_str.entries[i].unicode, + themap_str.entries[i].replace_str, + R_CURL); + if (i == (themap_str.entry_ct - 1)) { + fprintf(chdr, "\n%c;\n", R_CURL); + } else if ((i % 4) == 3) { + fprintf(chdr, ",\n\t"); + } else { + fprintf(chdr, ", "); + } + } + if (themap_str.entry_ct) { + fprintf(chdr, "\n\ +static const struct unimapdesc_str dfont_replacedesc%s = %c%d,repl_map%s,", + id_append, L_CURL, themap_str.entry_ct, id_append); + } else { + fprintf(chdr, "\n\ +static const struct unimapdesc_str dfont_replacedesc%s = %c0,NULL,", id_append, L_CURL); + } + fprintf(chdr, "%d,%d%c;\n", + this_isDefaultMap ? 1 : 0, + (useDefaultMap && !this_isDefaultMap) ? 1 : 0, + R_CURL); + + fprintf(chdr, "#define UC_CHARSET_SETUP%s UC_Charset_Setup(\ +\"%s\",\\\n\"%s\",\\\n\ +dfont_unicount%s,dfont_unitable%s,%d,\\\n\ +dfont_replacedesc%s,%d,%d,%d)\n", + id_append, this_MIMEcharset, this_LYNXcharset, + id_append, id_append, nuni, id_append, lowest_eight, RawOrEnc, CodePage); + + for (n = 0; n < TABLESIZE(last_ifdefs); n++) { + fprintf(chdr, last_ifdefs[n], id_append); + fprintf(chdr, "\n"); + } + + done(EX_OK); +#ifdef LY_FIND_LEAKS /* CF_ARG_ENABLE(find-leaks) */ + for (i = 0; i < themap_str.entry_ct; i++) { + FreeLeak(themap_str.entries[i].replace_str); + } +#endif + return 0; +} diff --git a/src/chrtrans/makew32.bat b/src/chrtrans/makew32.bat new file mode 100644 index 0000000..da47601 --- /dev/null +++ b/src/chrtrans/makew32.bat @@ -0,0 +1,13 @@ +@rem $LynxId: makew32.bat,v 1.6 2007/06/28 21:07:24 tom Exp $
+@echo off
+
+if "%1"=="" goto normal
+make -l -f makefile.bcb %1
+goto done
+
+:normal
+make -l -f makefile.bcb
+
+call makehdrs
+
+:done
diff --git a/src/chrtrans/mnem2_suni.tbl b/src/chrtrans/mnem2_suni.tbl new file mode 100644 index 0000000..3fc122c --- /dev/null +++ b/src/chrtrans/mnem2_suni.tbl @@ -0,0 +1,1865 @@ +#The MIME name of this charset. +# (this file was renamed from mnemonic_suni.tbl) +Mmnemonic + +#Name as a Display Charset (used on Options screen) +O RFC 1345 Mnemonic + +# Don't fall back to default table for unicode -> 8bit +Fallback NO + +# U+0020:&SP +U+0021:! +U+0022:" +U+0023:&Nb +U+0024:&DO +U+0025:% +U+0026:&& +U+0027:' +U+0028:( +U+0029:) +U+002a:* +U+002b:+ +U+002c:, +U+002d:- +U+002e:. +U+002f:/ +U+0030:0 +U+0031:1 +U+0032:2 +U+0033:3 +U+0034:4 +U+0035:5 +U+0036:6 +U+0037:7 +U+0038:8 +U+0039:9 +U+003a:: +U+003b:; +U+003c:< +U+003d:= +U+003e:> +U+003f:? +U+0040:&At +U+0041:A +U+0042:B +U+0043:C +U+0044:D +U+0045:E +U+0046:F +U+0047:G +U+0048:H +U+0049:I +U+004a:J +U+004b:K +U+004c:L +U+004d:M +U+004e:N +U+004f:O +U+0050:P +U+0051:Q +U+0052:R +U+0053:S +U+0054:T +U+0055:U +U+0056:V +U+0057:W +U+0058:X +U+0059:Y +U+005a:Z +U+005b:&<( +U+005c:&// +U+005d:&)> +U+005e:&'> +U+005f:_ +U+0060:&'! +U+0061:a +U+0062:b +U+0063:c +U+0064:d +U+0065:e +U+0066:f +U+0067:g +U+0068:h +U+0069:i +U+006a:j +U+006b:k +U+006c:l +U+006d:m +U+006e:n +U+006f:o +U+0070:p +U+0071:q +U+0072:r +U+0073:s +U+0074:t +U+0075:u +U+0076:v +U+0077:w +U+0078:x +U+0079:y +U+007a:z +U+007b:&(! +U+007c:&!! +U+007d:&!) +U+007e:&'? +U+00a0:&NS +U+00a1:&!I +U+00a2:&Ct +U+00a3:&Pd +U+00a4:&Cu +U+00a5:&Ye +U+00a6:&BB +U+00a7:&SE +U+00a8:&': +U+00a9:&Co +U+00aa:&-a +U+00ab:&<< +U+00ac:&NO +U+00ad:&-- +U+00ae:&Rg +U+00af:&'m +U+00b0:&DG +U+00b1:&+- +U+00b2:&2S +U+00b3:&3S +U+00b4:&'' +U+00b5:&My +U+00b6:&PI +U+00b7:&.M +U+00b8:&', +U+00b9:&1S +U+00ba:&-o +U+00bb:&>> +U+00bc:&14 +U+00bd:&12 +U+00be:&34 +U+00bf:&?I +U+00c0:&A! +U+00c1:&A' +U+00c2:&A> +U+00c3:&A? +U+00c4:&A: +U+00c5:&AA +U+00c6:&AE +U+00c7:&C, +U+00c8:&E! +U+00c9:&E' +U+00ca:&E> +U+00cb:&E: +U+00cc:&I! +U+00cd:&I' +U+00ce:&I> +U+00cf:&I: +U+00d0:&D- +U+00d1:&N? +U+00d2:&O! +U+00d3:&O' +U+00d4:&O> +U+00d5:&O? +U+00d6:&O: +U+00d7:&*X +U+00d8:&O/ +U+00d9:&U! +U+00da:&U' +U+00db:&U> +U+00dc:&U: +U+00dd:&Y' +U+00de:&TH +U+00df:&ss +U+00e0:&a! +U+00e1:&a' +U+00e2:&a> +U+00e3:&a? +U+00e4:&a: +U+00e5:&aa +U+00e6:&ae +U+00e7:&c, +U+00e8:&e! +U+00e9:&e' +U+00ea:&e> +U+00eb:&e: +U+00ec:&i! +U+00ed:&i' +U+00ee:&i> +U+00ef:&i: +U+00f0:&d- +U+00f1:&n? +U+00f2:&o! +U+00f3:&o' +U+00f4:&o> +U+00f5:&o? +U+00f6:&o: +U+00f7:&-: +U+00f8:&o/ +U+00f9:&u! +U+00fa:&u' +U+00fb:&u> +U+00fc:&u: +U+00fd:&y' +U+00fe:&th +U+00ff:&y: +U+0100:&A- +U+0101:&a- +U+0102:&A( +U+0103:&a( +U+0104:&A; +U+0105:&a; +U+0106:&C' +U+0107:&c' +U+0108:&C> +U+0109:&c> +U+010a:&C. +U+010b:&c. +U+010c:&C< +U+010d:&c< +U+010e:&D< +U+010f:&d< +U+0110:&D/ +U+0111:&d/ +U+0112:&E- +U+0113:&e- +U+0114:&E( +U+0115:&e( +U+0116:&E. +U+0117:&e. +U+0118:&E; +U+0119:&e; +U+011a:&E< +U+011b:&e< +U+011c:&G> +U+011d:&g> +U+011e:&G( +U+011f:&g( +U+0120:&G. +U+0121:&g. +U+0122:&G, +U+0123:&g, +U+0124:&H> +U+0125:&h> +U+0126:&H/ +U+0127:&h/ +U+0128:&I? +U+0129:&i? +U+012a:&I- +U+012b:&i- +U+012c:&I( +U+012d:&i( +U+012e:&I; +U+012f:&i; +U+0130:&I. +U+0131:&i. +U+0132:&IJ +U+0133:&ij +U+0134:&J> +U+0135:&j> +U+0136:&K, +U+0137:&k, +U+0138:&kk +U+0139:&L' +U+013a:&l' +U+013b:&L, +U+013c:&l, +U+013d:&L< +U+013e:&l< +U+013f:&L. +U+0140:&l. +U+0141:&L/ +U+0142:&l/ +U+0143:&N' +U+0144:&n' +U+0145:&N, +U+0146:&n, +U+0147:&N< +U+0148:&n< +U+0149:&'n +U+014a:&NG +U+014b:&ng +U+014c:&O- +U+014d:&o- +U+014e:&O( +U+014f:&o( +U+0150:&O" +U+0151:&o" +U+0152:&OE +U+0153:&oe +U+0154:&R' +U+0155:&r' +U+0156:&R, +U+0157:&r, +U+0158:&R< +U+0159:&r< +U+015a:&S' +U+015b:&s' +U+015c:&S> +U+015d:&s> +U+015e:&S, +U+015f:&s, +U+0160:&S< +U+0161:&s< +U+0162:&T, +U+0163:&t, +U+0164:&T< +U+0165:&t< +U+0166:&T/ +U+0167:&t/ +U+0168:&U? +U+0169:&u? +U+016a:&U- +U+016b:&u- +U+016c:&U( +U+016d:&u( +U+016e:&U0 +U+016f:&u0 +U+0170:&U" +U+0171:&u" +U+0172:&U; +U+0173:&u; +U+0174:&W> +U+0175:&w> +U+0176:&Y> +U+0177:&y> +U+0178:&Y: +U+0179:&Z' +U+017a:&z' +U+017b:&Z. +U+017c:&z. +U+017d:&Z< +U+017e:&z< +U+01a0:&O9 +U+01a1:&o9 +U+01a2:&OI +U+01a3:&oi +U+01a6:&yr +U+01af:&U9 +U+01b0:&u9 +U+01b5:&Z/ +U+01b6:&z/ +U+01b7:&ED +U+01cd:&A< +U+01ce:&a< +U+01cf:&I< +U+01d0:&i< +U+01d1:&O< +U+01d2:&o< +U+01d3:&U< +U+01d4:&u< +U+01d5:&_U:-_ +U+01d6:&_u:-_ +U+01d7:&_U:'_ +U+01d8:&_u:'_ +U+01d9:&_U:<_ +U+01da:&_u:<_ +U+01db:&_U:!_ +U+01dc:&_u:!_ +U+01de:&A1 +U+01df:&a1 +U+01e0:&A7 +U+01e1:&a7 +U+01e2:&A3 +U+01e3:&a3 +U+01e4:&G/ +U+01e5:&g/ +U+01e6:&G< +U+01e7:&g< +U+01e8:&K< +U+01e9:&k< +U+01ea:&O; +U+01eb:&o; +U+01ec:&O1 +U+01ed:&o1 +U+01ee:&EZ +U+01ef:&ez +U+01f0:&j< +U+01f4:&G' +U+01f5:&g' +U+01fa:&_AA'_ +U+01fb:&_aa'_ +U+01fc:&_AE'_ +U+01fd:&_ae'_ +U+01fe:&_O/'_ +U+01ff:&_o/'_ +U+02bf:&;S +U+02c7:&'< +U+02d8:&'( +U+02d9:&'. +U+02da:&'0 +U+02db:&'; +U+02dd:&'" +U+0386:&A% +U+0388:&E% +U+0389:&Y% +U+038a:&I% +U+038c:&O% +U+038e:&U% +U+038f:&W% +U+0390:&i3 +U+0391:&A* +U+0392:&B* +U+0393:&G* +U+0394:&D* +U+0395:&E* +U+0396:&Z* +U+0397:&Y* +U+0398:&H* +U+0399:&I* +U+039a:&K* +U+039b:&L* +U+039c:&M* +U+039d:&N* +U+039e:&C* +U+039f:&O* +U+03a0:&P* +U+03a1:&R* +U+03a3:&S* +U+03a4:&T* +U+03a5:&U* +U+03a6:&F* +U+03a7:&X* +U+03a8:&Q* +U+03a9:&W* +U+03aa:&J* +U+03ab:&V* +U+03ac:&a% +U+03ad:&e% +U+03ae:&y% +U+03af:&i% +U+03b0:&u3 +U+03b1:&a* +U+03b2:&b* +U+03b3:&g* +U+03b4:&d* +U+03b5:&e* +U+03b6:&z* +U+03b7:&y* +U+03b8:&h* +U+03b9:&i* +U+03ba:&k* +U+03bb:&l* +U+03bc:&m* +U+03bd:&n* +U+03be:&c* +U+03bf:&o* +U+03c0:&p* +U+03c1:&r* +U+03c2:&*s +U+03c3:&s* +U+03c4:&t* +U+03c5:&u* +U+03c6:&f* +U+03c7:&x* +U+03c8:&q* +U+03c9:&w* +U+03ca:&j* +U+03cb:&v* +U+03cc:&o% +U+03cd:&u% +U+03ce:&w% +U+03d8:&'G +U+03d9:&,G +U+03da:&T3 +U+03db:&t3 +U+03dc:&M3 +U+03dd:&m3 +U+03de:&K3 +U+03df:&k3 +U+03e0:&P3 +U+03e1:&p3 +U+03f4:&'% +U+03f5:&j3 +U+0401:&IO +U+0402:&D% +U+0403:&G% +U+0404:&IE +U+0405:&DS +U+0406:&II +U+0407:&YI +U+0408:&J% +U+0409:&LJ +U+040a:&NJ +U+040b:&Ts +U+040c:&KJ +U+040e:&V% +U+040f:&DZ +U+0410:&A= +U+0411:&B= +U+0412:&V= +U+0413:&G= +U+0414:&D= +U+0415:&E= +U+0416:&Z% +U+0417:&Z= +U+0418:&I= +U+0419:&J= +U+041a:&K= +U+041b:&L= +U+041c:&M= +U+041d:&N= +U+041e:&O= +U+041f:&P= +U+0420:&R= +U+0421:&S= +U+0422:&T= +U+0423:&U= +U+0424:&F= +U+0425:&H= +U+0426:&C= +U+0427:&C% +U+0428:&S% +U+0429:&Sc +U+042a:&=" +U+042b:&Y= +U+042c:&%" +U+042d:&JE +U+042e:&JU +U+042f:&JA +U+0430:&a= +U+0431:&b= +U+0432:&v= +U+0433:&g= +U+0434:&d= +U+0435:&e= +U+0436:&z% +U+0437:&z= +U+0438:&i= +U+0439:&j= +U+043a:&k= +U+043b:&l= +U+043c:&m= +U+043d:&n= +U+043e:&o= +U+043f:&p= +U+0440:&r= +U+0441:&s= +U+0442:&t= +U+0443:&u= +U+0444:&f= +U+0445:&h= +U+0446:&c= +U+0447:&c% +U+0448:&s% +U+0449:&sc +U+044a:&=' +U+044b:&y= +U+044c:&%' +U+044d:&je +U+044e:&ju +U+044f:&ja +U+0451:&io +U+0452:&d% +U+0453:&g% +U+0454:&ie +U+0455:&ds +U+0456:&ii +U+0457:&yi +U+0458:&j% +U+0459:&lj +U+045a:&nj +U+045b:&ts +U+045c:&kj +U+045e:&v% +U+045f:&dz +U+0462:&Y3 +U+0463:&y3 +U+046a:&O3 +U+046b:&o3 +U+0472:&F3 +U+0473:&f3 +U+0474:&V3 +U+0475:&v3 +U+0480:&C3 +U+0481:&c3 +U+0490:&G3 +U+0491:&g3 +U+05d0:&A+ +U+05d1:&B+ +U+05d2:&G+ +U+05d3:&D+ +U+05d4:&H+ +U+05d5:&W+ +U+05d6:&Z+ +U+05d7:&X+ +U+05d8:&Tj +U+05d9:&J+ +U+05da:&K% +U+05db:&K+ +U+05dc:&L+ +U+05dd:&M% +U+05de:&M+ +U+05df:&N% +U+05e0:&N+ +U+05e1:&S+ +U+05e2:&E+ +U+05e3:&P% +U+05e4:&P+ +U+05e5:&Zj +U+05e6:&ZJ +U+05e7:&Q+ +U+05e8:&R+ +U+05e9:&Sh +U+05ea:&T+ +U+060c:&,+ +U+061b:&;+ +U+061f:&?+ +U+0621:&H' +U+0622:&aM +U+0623:&aH +U+0624:&wH +U+0625:&ah +U+0626:&yH +U+0627:&a+ +U+0628:&b+ +U+0629:&tm +U+062a:&t+ +U+062b:&tk +U+062c:&g+ +U+062d:&hk +U+062e:&x+ +U+062f:&d+ +U+0630:&dk +U+0631:&r+ +U+0632:&z+ +U+0633:&s+ +U+0634:&sn +U+0635:&c+ +U+0636:&dd +U+0637:&tj +U+0638:&zH +U+0639:&e+ +U+063a:&i+ +U+0640:&++ +U+0641:&f+ +U+0642:&q+ +U+0643:&k+ +U+0644:&l+ +U+0645:&m+ +U+0646:&n+ +U+0647:&h+ +U+0648:&w+ +U+0649:&j+ +U+064a:&y+ +U+064b:&:+ +U+064c:&"+ +U+064d:&=+ +U+064e:&/+ +U+064f:&'+ +U+0650:&1+ +U+0651:&3+ +U+0652:&0+ +U+0670:&aS +U+067e:&p+ +U+06a4:&v+ +U+06af:&gf +U+06f0:&0a +U+06f1:&1a +U+06f2:&2a +U+06f3:&3a +U+06f4:&4a +U+06f5:&5a +U+06f6:&6a +U+06f7:&7a +U+06f8:&8a +U+06f9:&9a +U+1e00:&_A-0_ +U+1e01:&_a-0_ +U+1e02:&B. +U+1e03:&b. +U+1e04:&_B-._ +U+1e05:&_b-._ +U+1e06:&B_ +U+1e07:&b_ +U+1e08:&_C,'_ +U+1e09:&_c,'_ +U+1e0a:&D. +U+1e0b:&d. +U+1e0c:&_D-._ +U+1e0d:&_d-._ +U+1e0e:&D_ +U+1e0f:&d_ +U+1e10:&D, +U+1e11:&d, +U+1e12:&_D->_ +U+1e13:&_d->_ +U+1e14:&_E-!_ +U+1e15:&_e-!_ +U+1e16:&_E-'_ +U+1e17:&_e-'_ +U+1e18:&_E->_ +U+1e19:&_e->_ +U+1e1a:&_E-?_ +U+1e1b:&_e-?_ +U+1e1c:&_E,(_ +U+1e1d:&_e,(_ +U+1e1e:&F. +U+1e1f:&f. +U+1e20:&G- +U+1e21:&g- +U+1e22:&H. +U+1e23:&h. +U+1e24:&_H-._ +U+1e25:&_h-._ +U+1e26:&H: +U+1e27:&h: +U+1e28:&H, +U+1e29:&h, +U+1e2a:&_H-(_ +U+1e2b:&_h-(_ +U+1e2c:&_I-?_ +U+1e2d:&_i-?_ +U+1e2e:&_I:'_ +U+1e2f:&_i:'_ +U+1e30:&K' +U+1e31:&k' +U+1e32:&_K-._ +U+1e33:&_k-._ +U+1e34:&K_ +U+1e35:&k_ +U+1e36:&_L-._ +U+1e37:&_l-._ +U+1e38:&_L--._ +U+1e39:&_l--._ +U+1e3a:&L_ +U+1e3b:&l_ +U+1e3c:&_L->_ +U+1e3d:&_l->_ +U+1e3e:&M' +U+1e3f:&m' +U+1e40:&M. +U+1e41:&m. +U+1e42:&_M-._ +U+1e43:&_m-._ +U+1e44:&N. +U+1e45:&n. +U+1e46:&_N-._ +U+1e47:&_n-._ +U+1e48:&N_ +U+1e49:&n_ +U+1e4a:&_N->_ +U+1e4b:&_N->_ +U+1e4c:&_O?'_ +U+1e4d:&_o?'_ +U+1e4e:&_O?:_ +U+1e4f:&_o?:_ +U+1e50:&_O-!_ +U+1e51:&_o-!_ +U+1e52:&_O-'_ +U+1e53:&_o-'_ +U+1e54:&P' +U+1e55:&p' +U+1e56:&P. +U+1e57:&p. +U+1e58:&R. +U+1e59:&r. +U+1e5a:&_R-._ +U+1e5b:&_r-._ +U+1e5c:&_R--._ +U+1e5d:&_r--._ +U+1e5e:&R_ +U+1e5f:&r_ +U+1e60:&S. +U+1e61:&s. +U+1e62:&_S-._ +U+1e63:&_s-._ +U+1e64:&_S'._ +U+1e65:&_s'._ +U+1e66:&_S<._ +U+1e67:&_s<._ +U+1e68:&_S.-._ +U+1e69:&_S.-._ +U+1e6a:&T. +U+1e6b:&t. +U+1e6c:&_T-._ +U+1e6d:&_t-._ +U+1e6e:&T_ +U+1e6f:&t_ +U+1e70:&_T->_ +U+1e71:&_t->_ +U+1e72:&_U--:_ +U+1e73:&_u--:_ +U+1e74:&_U-?_ +U+1e75:&_u-?_ +U+1e76:&_U->_ +U+1e77:&_u->_ +U+1e78:&_U?'_ +U+1e79:&_u?'_ +U+1e7a:&_U-:_ +U+1e7b:&_u-:_ +U+1e7c:&V? +U+1e7d:&v? +U+1e7e:&_V-._ +U+1e7f:&_v-._ +U+1e80:&W! +U+1e81:&w! +U+1e82:&W' +U+1e83:&w' +U+1e84:&W: +U+1e85:&w: +U+1e86:&W. +U+1e87:&w. +U+1e88:&_W-._ +U+1e89:&_w-._ +U+1e8a:&X. +U+1e8b:&x. +U+1e8c:&X: +U+1e8d:&x: +U+1e8e:&Y. +U+1e8f:&y. +U+1e90:&Z> +U+1e91:&z> +U+1e92:&_Z-._ +U+1e93:&_z-._ +U+1e94:&Z_ +U+1e95:&z_ +U+1e96:&h_ +U+1e97:&t: +U+1e98:&w0 +U+1e99:&y0 +U+1ea0:&_A-._ +U+1ea1:&_a-._ +U+1ea2:&A2 +U+1ea3:&a2 +U+1ea4:&_A>'_ +U+1ea5:&_a>'_ +U+1ea6:&_A>!_ +U+1ea7:&_a>!_ +U+1ea8:&_A>2_ +U+1ea9:&_a>2_ +U+1eaa:&_A>?_ +U+1eab:&_a>?_ +U+1eac:&_A>-._ +U+1ead:&_a>-._ +U+1eae:&_A('_ +U+1eaf:&_a('_ +U+1eb0:&_A(!_ +U+1eb1:&_a(!_ +U+1eb2:&_A(2_ +U+1eb3:&_a(2_ +U+1eb4:&_A(?_ +U+1eb5:&_a(?_ +U+1eb6:&_A(-._ +U+1eb7:&_a(-._ +U+1eb8:&_E-._ +U+1eb9:&_e-._ +U+1eba:&E2 +U+1ebb:&e2 +U+1ebc:&E? +U+1ebd:&e? +U+1ebe:&_E>'_ +U+1ebf:&_e>'_ +U+1ec0:&_E>!_ +U+1ec1:&_e>!_ +U+1ec2:&_E>2_ +U+1ec3:&_e>2_ +U+1ec4:&_E>?_ +U+1ec5:&_e>?_ +U+1ec6:&_E>-._ +U+1ec7:&_e>-._ +U+1ec8:&I2 +U+1ec9:&i2 +U+1eca:&_I-._ +U+1ecb:&_i-._ +U+1ecc:&_O-._ +U+1ecd:&_o-._ +U+1ece:&O2 +U+1ecf:&o2 +U+1ed0:&_O>'_ +U+1ed1:&_o>'_ +U+1ed2:&_O>!_ +U+1ed3:&_o>!_ +U+1ed4:&_O>2_ +U+1ed5:&_o>2_ +U+1ed6:&_O>?_ +U+1ed7:&_o>?_ +U+1ed8:&_O>-._ +U+1ed9:&_o>-._ +U+1eda:&_O9'_ +U+1edb:&_o9'_ +U+1edc:&_O9!_ +U+1edd:&_o9!_ +U+1ede:&_O92_ +U+1edf:&_o92_ +U+1ee0:&_O9?_ +U+1ee1:&_o9?_ +U+1ee2:&_O9-._ +U+1ee3:&_o9-._ +U+1ee4:&_U-._ +U+1ee5:&_u-._ +U+1ee6:&U2 +U+1ee7:&u2 +U+1ee8:&_U9'_ +U+1ee9:&_u9'_ +U+1eea:&_U9!_ +U+1eeb:&_u9!_ +U+1eec:&_U92_ +U+1eed:&_u92_ +U+1eee:&_U9?_ +U+1eef:&_u9?_ +U+1ef0:&_U9-._ +U+1ef1:&_u9-._ +U+1ef2:&Y! +U+1ef3:&y! +U+1ef4:&_Y-._ +U+1ef5:&_y-._ +U+1ef6:&Y2 +U+1ef7:&y2 +U+1ef8:&Y? +U+1ef9:&y? +U+1f00:&;' +U+1f01:&,' +U+1f02:&;! +U+1f03:&,! +U+1f04:&?; +U+1f05:&?, +U+1f06:&!: +U+1f07:&?: +U+2002:&1N +U+2003:&1M +U+2004:&3M +U+2005:&4M +U+2006:&6M +U+2009:&1T +U+200a:&1H +U+2010:&-1 +U+2013:&-N +U+2014:&-M +U+2015:&-3 +U+2016:&!2 +U+2017:&=2 +U+2018:&'6 +U+2019:&'9 +U+201a:&.9 +U+201b:&9' +U+201c:&"6 +U+201d:&"9 +U+201e:&:9 +U+201f:&9" +U+2020:&/- +U+2021:&/= +U+2025:&.. +U+2030:&%0 +U+2032:&1' +U+2033:&2' +U+2034:&3' +U+2035:&1" +U+2036:&2" +U+2037:&3" +U+2038:&Ca +U+2039:&<1 +U+203a:&>1 +U+203b:&:X +U+203c:&_!*2_ +U+203e:&'- +U+2044:&/f +U+2070:&0S +U+2074:&4S +U+2075:&5S +U+2076:&6S +U+2077:&7S +U+2078:&8S +U+2079:&9S +U+207a:&+S +U+207b:&-S +U+207c:&=S +U+207d:&(S +U+207e:&)S +U+207f:&nS +U+2080:&0s +U+2081:&1s +U+2082:&2s +U+2083:&3s +U+2084:&4s +U+2085:&5s +U+2086:&6s +U+2087:&7s +U+2088:&8s +U+2089:&9s +U+208a:&+s +U+208b:&-s +U+208c:&=s +U+208d:&(s +U+208e:&)s +U+20a4:&Li +U+20a7:&Pt +U+20a9:&W= +U+2103:&oC +U+2105:&co +U+2109:&oF +U+2116:&N0 +U+2117:&PO +U+211e:&Rx +U+2120:&SM +U+2122:&TM +U+2126:&Om +U+212b:&AO +U+2153:&13 +U+2154:&23 +U+2155:&15 +U+2156:&25 +U+2157:&35 +U+2158:&45 +U+2159:&16 +U+215a:&56 +U+215b:&18 +U+215c:&38 +U+215d:&58 +U+215e:&78 +U+2160:&1R +U+2161:&2R +U+2162:&3R +U+2163:&4R +U+2164:&5R +U+2165:&6R +U+2166:&7R +U+2167:&8R +U+2168:&9R +U+2169:&aR +U+216a:&bR +U+216b:&cR +U+216c:&_50R_ +U+216d:&_100R_ +U+216e:&_500R_ +U+216f:&_1000R_ +U+2170:&1r +U+2171:&2r +U+2172:&3r +U+2173:&4r +U+2174:&5r +U+2175:&6r +U+2176:&7r +U+2177:&8r +U+2178:&9r +U+2179:&ar +U+217a:&br +U+217b:&cr +U+217c:&_50r_ +U+217d:&_100r_ +U+217e:&_500r_ +U+217f:&_1000r_ +U+2180:&_1000RCD_ +U+2181:&_5000R_ +U+2182:&_10000R_ +U+2190:&<- +U+2191:&-! +U+2192:&-> +U+2193:&-v +U+2194:&<> +U+2195:&UD +U+2196:&_<!!_ +U+2197:&_//>_ +U+2198:&_!!>_ +U+2199:&_<//_ +U+21d0:&<= +U+21d2:&=> +U+21d4:&== +U+2200:&FA +U+2202:&dP +U+2203:&TE +U+2205:&/0 +U+2206:&DE +U+2207:&NB +U+2208:&(- +U+220b:&-) +U+220f:&*P +U+2211:&+Z +U+2212:&-2 +U+2213:&-+ +U+2217:&*- +U+2218:&Ob +U+2219:&Sb +U+221a:&RT +U+221d:&0( +U+221e:&00 +U+221f:&-L +U+2220:&-V +U+2225:&PP +U+2227:&AN +U+2228:&OR +U+2229:&(U +U+222a:&)U +U+222b:&In +U+222c:&DI +U+222e:&Io +U+2234:&.: +U+2235:&:. +U+2236:&:R +U+2237:&:: +U+223c:&?1 +U+223e:&CG +U+2243:&?- +U+2245:&?= +U+2248:&?2 +U+224c:&=? +U+2253:&HI +U+2260:&!= +U+2261:&=3 +U+2264:&=< +U+2265:&>= +U+226a:&<* +U+226b:&*> +U+226e:&!< +U+226f:&!> +U+2282:&(C +U+2283:&)C +U+2286:&(_ +U+2287:&)_ +U+2299:&0. +U+229a:&02 +U+22a5:&-T +U+22c5:&.P +U+22ee:&:3 +U+22ef:&.3 +U+2302:&Eh +U+2308:&<7 +U+2309:&>7 +U+230a:&7< +U+230b:&7> +U+2310:&NI +U+2312:&(A +U+2315:&TR +U+2320:&Iu +U+2321:&Il +U+2329:&</ +U+232a:&/> +U+2423:&Vs +U+2440:&1h +U+2441:&3h +U+2442:&2h +U+2443:&4h +U+2446:&1j +U+2447:&2j +U+2448:&3j +U+2449:&4j +U+2460:&_1-o_ +U+2461:&_2-o_ +U+2462:&_3-o_ +U+2463:&_4-o_ +U+2464:&_5-o_ +U+2465:&_6-o_ +U+2466:&_7-o_ +U+2467:&_8-o_ +U+2468:&_9-o_ +U+2469:&_10-o_ +U+246a:&_11-o_ +U+246b:&_12-o_ +U+246c:&_13-o_ +U+246d:&_14-o_ +U+246e:&_15-o_ +U+246f:&_16-o_ +U+2470:&_17-o_ +U+2471:&_18-o_ +U+2472:&_19-o_ +U+2473:&_20-o_ +U+2474:&_(1)_ +U+2475:&_(2)_ +U+2476:&_(3)_ +U+2477:&_(4)_ +U+2478:&_(5)_ +U+2479:&_(6)_ +U+247a:&_(7)_ +U+247b:&_(8)_ +U+247c:&_(9)_ +U+247d:&_(10)_ +U+247e:&_(11)_ +U+247f:&_(12)_ +U+2480:&_(13)_ +U+2481:&_(14)_ +U+2482:&_(15)_ +U+2483:&_(16)_ +U+2484:&_(17)_ +U+2485:&_(18)_ +U+2486:&_(19)_ +U+2487:&_(20)_ +U+2488:&1. +U+2489:&2. +U+248a:&3. +U+248b:&4. +U+248c:&5. +U+248d:&6. +U+248e:&7. +U+248f:&8. +U+2490:&9. +U+2491:&_10._ +U+2492:&_11._ +U+2493:&_12._ +U+2494:&_13._ +U+2495:&_14._ +U+2496:&_15._ +U+2497:&_16._ +U+2498:&_17._ +U+2499:&_18._ +U+249a:&_19._ +U+249b:&_20._ +U+249c:&_(a)_ +U+249d:&_(b)_ +U+249e:&_(c)_ +U+249f:&_(d)_ +U+24a0:&_(e)_ +U+24a1:&_(f)_ +U+24a2:&_(g)_ +U+24a3:&_(h)_ +U+24a4:&_(i)_ +U+24a5:&_(j)_ +U+24a6:&_(k)_ +U+24a7:&_(l)_ +U+24a8:&_(m)_ +U+24a9:&_(n)_ +U+24aa:&_(o)_ +U+24ab:&_(p)_ +U+24ac:&_(q)_ +U+24ad:&_(r)_ +U+24ae:&_(s)_ +U+24af:&_(t)_ +U+24b0:&_(u)_ +U+24b1:&_(v)_ +U+24b2:&_(w)_ +U+24b3:&_(x)_ +U+24b4:&_(y)_ +U+24b5:&_(z)_ +U+24b6:&_A-o_ +U+24b7:&_B-o_ +U+24b8:&_C-o_ +U+24b9:&_D-o_ +U+24ba:&_E-o_ +U+24bb:&_F-o_ +U+24bc:&_G-o_ +U+24bd:&_H-o_ +U+24be:&_I-o_ +U+24bf:&_J-o_ +U+24c0:&_K-o_ +U+24c1:&_L-o_ +U+24c2:&_M-o_ +U+24c3:&_N-o_ +U+24c4:&_O-o_ +U+24c5:&_P-o_ +U+24c6:&_Q-o_ +U+24c7:&_R-o_ +U+24c8:&_S-o_ +U+24c9:&_T-o_ +U+24ca:&_U-o_ +U+24cb:&_V-o_ +U+24cc:&_W-o_ +U+24cd:&_X-o_ +U+24ce:&_Y-o_ +U+24cf:&_Z-o_ +U+24d0:&_a-o_ +U+24d1:&_b-o_ +U+24d2:&_c-o_ +U+24d3:&_d-o_ +U+24d4:&_e-o_ +U+24d5:&_f-o_ +U+24d6:&_g-o_ +U+24d7:&_h-o_ +U+24d8:&_i-o_ +U+24d9:&_j-o_ +U+24da:&_k-o_ +U+24db:&_l-o_ +U+24dc:&_m-o_ +U+24dd:&_n-o_ +U+24de:&_o-o_ +U+24df:&_p-o_ +U+24e0:&_q-o_ +U+24e1:&_r-o_ +U+24e2:&_s-o_ +U+24e3:&_t-o_ +U+24e4:&_u-o_ +U+24e5:&_v-o_ +U+24e6:&_w-o_ +U+24e7:&_x-o_ +U+24e8:&_y-o_ +U+24e9:&_z-o_ +U+24ea:&_0-o_ +U+2500:&hh +U+2501:&HH +U+2502:&vv +U+2503:&VV +U+2504:&3- +U+2505:&3_ +U+2506:&3! +U+2507:&3/ +U+2508:&4- +U+2509:&4_ +U+250a:&4! +U+250b:&4/ +U+250c:&dr +U+250d:&dR +U+250e:&Dr +U+250f:&DR +U+2510:&dl +U+2511:&dL +U+2512:&Dl +U+2513:&LD +U+2514:&ur +U+2515:&uR +U+2516:&Ur +U+2517:&UR +U+2518:&ul +U+2519:&uL +U+251a:&Ul +U+251b:&UL +U+251c:&vr +U+251d:&vR +U+251e:&_Udr_ +U+251f:&_uDr_ +U+2520:&Vr +U+2521:&_UdR_ +U+2522:&_uDR_ +U+2523:&VR +U+2524:&vl +U+2525:&vL +U+2526:&_Udl_ +U+2527:&_uDl_ +U+2528:&Vl +U+2529:&_UdL_ +U+252a:&_uDL_ +U+252b:&VL +U+252c:&dh +U+252d:&_dLr_ +U+252e:&_dlR_ +U+252f:&dH +U+2530:&Dh +U+2531:&_DLr_ +U+2532:&_DlR_ +U+2533:&DH +U+2534:&uh +U+2535:&_uLr_ +U+2536:&_ulR_ +U+2537:&uH +U+2538:&Uh +U+2539:&_ULr_ +U+253a:&_UlR_ +U+253b:&UH +U+253c:&vh +U+253d:&_vLr_ +U+253e:&_vlR_ +U+253f:&vH +U+2540:&_Udh_ +U+2541:&_uDh_ +U+2542:&Vh +U+2543:&_UdLr_ +U+2544:&_UdlR_ +U+2545:&_uDLr_ +U+2546:&_uDlR_ +U+2547:&_UdH_ +U+2548:&_uDH_ +U+2549:&_VLr_ +U+254a:&_VlR_ +U+254b:&VH +U+2571:&FD +U+2572:&BD +U+2580:&TB +U+2584:&LB +U+2588:&FB +U+258c:&lB +U+2590:&RB +U+2591:&.S +U+2592:&:S +U+2593:&?S +U+25a0:&fS +U+25a1:&OS +U+25a2:&RO +U+25a3:&Rr +U+25a4:&RF +U+25a5:&RY +U+25a6:&RH +U+25a7:&RZ +U+25a8:&RK +U+25a9:&RX +U+25aa:&sB +U+25ac:&SR +U+25ad:&Or +U+25b2:&UT +U+25b3:&uT +U+25b6:&PR +U+25b7:&Tr +U+25bc:&Dt +U+25bd:&dT +U+25c0:&PL +U+25c1:&Tl +U+25c6:&Db +U+25c7:&Dw +U+25ca:&LZ +U+25cb:&0m +U+25ce:&0o +U+25cf:&0M +U+25d0:&0L +U+25d1:&0R +U+25d8:&Sn +U+25d9:&Ic +U+25e2:&Fd +U+25e3:&Bd +U+2605:&*2 +U+2606:&*1 +U+260e:&_TEL_ +U+260f:&_tel_ +U+261c:&<H +U+261e:&>H +U+263a:&0u +U+263b:&0U +U+263c:&SU +U+2640:&Fm +U+2642:&Ml +U+2660:&cS +U+2661:&cH +U+2662:&cD +U+2663:&cC +U+2664:&_cS-_ +U+2665:&_cH-_ +U+2666:&_cD-_ +U+2667:&_cC-_ +U+2669:&Md +U+266a:&M8 +U+266b:&M2 +U+266c:&_M16_ +U+266d:&Mb +U+266e:&Mx +U+266f:&MX +U+2713:&OK +U+2717:&XX +U+2720:&-X +U+3000:&IS +U+3001:&,_ +U+3002:&._ +U+3003:&+" +U+3004:&+_ +U+3005:&*_ +U+3006:&;_ +U+3007:&0_ +U+300a:&<+ +U+300b:&>+ +U+300c:&<' +U+300d:&>' +U+300e:&<" +U+300f:&>" +U+3010:&(" +U+3011:&)" +U+3012:&=T +U+3013:&=_ +U+3014:&(' +U+3015:&)' +U+3016:&(I +U+3017:&)I +U+301c:&-? +U+3020:&_=T:)_ +U+3041:&A5 +U+3042:&a5 +U+3043:&I5 +U+3044:&i5 +U+3045:&U5 +U+3046:&u5 +U+3047:&E5 +U+3048:&e5 +U+3049:&O5 +U+304a:&o5 +U+304b:&ka +U+304c:&ga +U+304d:&ki +U+304e:&gi +U+304f:&ku +U+3050:&gu +U+3051:&ke +U+3052:&ge +U+3053:&ko +U+3054:&go +U+3055:&sa +U+3056:&za +U+3057:&si +U+3058:&zi +U+3059:&su +U+305a:&zu +U+305b:&se +U+305c:&ze +U+305d:&so +U+305e:&zo +U+305f:&ta +U+3060:&da +U+3061:&ti +U+3062:&di +U+3063:&tU +U+3064:&tu +U+3065:&du +U+3066:&te +U+3067:&de +U+3068:&to +U+3069:&do +U+306a:&na +U+306b:&ni +U+306c:&nu +U+306d:&ne +U+306e:&no +U+306f:&ha +U+3070:&ba +U+3071:&pa +U+3072:&hi +U+3073:&bi +U+3074:&pi +U+3075:&hu +U+3076:&bu +U+3077:&pu +U+3078:&he +U+3079:&be +U+307a:&pe +U+307b:&ho +U+307c:&bo +U+307d:&po +U+307e:&ma +U+307f:&mi +U+3080:&mu +U+3081:&me +U+3082:&mo +U+3083:&yA +U+3084:&ya +U+3085:&yU +U+3086:&yu +U+3087:&yO +U+3088:&yo +U+3089:&ra +U+308a:&ri +U+308b:&ru +U+308c:&re +U+308d:&ro +U+308e:&wA +U+308f:&wa +U+3090:&wi +U+3091:&we +U+3092:&wo +U+3093:&n5 +U+3094:&vu +U+309b:&"5 +U+309c:&05 +U+309d:&*5 +U+309e:&+5 +U+30a1:&a6 +U+30a2:&A6 +U+30a3:&i6 +U+30a4:&I6 +U+30a5:&u6 +U+30a6:&U6 +U+30a7:&e6 +U+30a8:&E6 +U+30a9:&o6 +U+30aa:&O6 +U+30ab:&Ka +U+30ac:&Ga +U+30ad:&Ki +U+30ae:&Gi +U+30af:&Ku +U+30b0:&Gu +U+30b1:&Ke +U+30b2:&Ge +U+30b3:&Ko +U+30b4:&Go +U+30b5:&Sa +U+30b6:&Za +U+30b7:&Si +U+30b8:&Zi +U+30b9:&Su +U+30ba:&Zu +U+30bb:&Se +U+30bc:&Ze +U+30bd:&So +U+30be:&Zo +U+30bf:&Ta +U+30c0:&Da +U+30c1:&Ti +U+30c2:&Di +U+30c3:&TU +U+30c4:&Tu +U+30c5:&Du +U+30c6:&Te +U+30c7:&De +U+30c8:&To +U+30c9:&Do +U+30ca:&Na +U+30cb:&Ni +U+30cc:&Nu +U+30cd:&Ne +U+30ce:&No +U+30cf:&Ha +U+30d0:&Ba +U+30d1:&Pa +U+30d2:&Hi +U+30d3:&Bi +U+30d4:&Pi +U+30d5:&Hu +U+30d6:&Bu +U+30d7:&Pu +U+30d8:&He +U+30d9:&Be +U+30da:&Pe +U+30db:&Ho +U+30dc:&Bo +U+30dd:&Po +U+30de:&Ma +U+30df:&Mi +U+30e0:&Mu +U+30e1:&Me +U+30e2:&Mo +U+30e3:&YA +U+30e4:&Ya +U+30e5:&YU +U+30e6:&Yu +U+30e7:&YO +U+30e8:&Yo +U+30e9:&Ra +U+30ea:&Ri +U+30eb:&Ru +U+30ec:&Re +U+30ed:&Ro +U+30ee:&WA +U+30ef:&Wa +U+30f0:&Wi +U+30f1:&We +U+30f2:&Wo +U+30f3:&N6 +U+30f4:&Vu +U+30f5:&KA +U+30f6:&KE +U+30f7:&Va +U+30f8:&Vi +U+30f9:&Ve +U+30fa:&Vo +U+30fb:&.6 +U+30fc:&-6 +U+30fd:&*6 +U+30fe:&+6 +U+3105:&b4 +U+3106:&p4 +U+3107:&m4 +U+3108:&f4 +U+3109:&d4 +U+310a:&t4 +U+310b:&n4 +U+310c:&l4 +U+310d:&g4 +U+310e:&k4 +U+310f:&h4 +U+3110:&j4 +U+3111:&q4 +U+3112:&x4 +U+3113:&zh +U+3114:&ch +U+3115:&sh +U+3116:&r4 +U+3117:&z4 +U+3118:&c4 +U+3119:&s4 +U+311a:&a4 +U+311b:&o4 +U+311c:&e4 +U+311d:&_eh4_ +U+311e:&ai +U+311f:&ei +U+3120:&au +U+3121:&ou +U+3122:&an +U+3123:&en +U+3124:&aN +U+3125:&eN +U+3126:&er +U+3127:&i4 +U+3128:&u4 +U+3129:&iu +U+312a:&v4 +U+312b:&nG +U+312c:&gn +U+321c:&_(JU)_ +U+3220:&1c +U+3221:&2c +U+3222:&3c +U+3223:&4c +U+3224:&5c +U+3225:&6c +U+3226:&7c +U+3227:&8c +U+3228:&9c +U+3229:&_10c_ +U+327f:&_KSC_ +U+fb00:&ff +U+fb01:&fi +U+fb02:&fl +U+fb03:&_ffi_ +U+fb04:&_ffl_ +U+fb05:&ft +U+fb06:&st +U+fe7d:&_3+;_ +U+fe82:&_aM._ +U+fe84:&_aH._ +U+fe8d:&_a+-_ +U+fe8e:&_a+._ +U+fe8f:&_b+-_ +U+fe90:&_b+,_ +U+fe91:&_b+;_ +U+fe92:&_b+._ +U+fe93:&_tm-_ +U+fe94:&_tm._ +U+fe95:&_t+-_ +U+fe96:&_t+,_ +U+fe97:&_t+;_ +U+fe98:&_t+._ +U+fe99:&_tk-_ +U+fe9a:&_tk,_ +U+fe9b:&_tk;_ +U+fe9c:&_tk._ +U+fe9d:&_g+-_ +U+fe9e:&_g+,_ +U+fe9f:&_g+;_ +U+fea0:&_g+._ +U+fea1:&_hk-_ +U+fea2:&_hk,_ +U+fea3:&_hk;_ +U+fea4:&_hk._ +U+fea5:&_x+-_ +U+fea6:&_x+,_ +U+fea7:&_x+;_ +U+fea8:&_x+._ +U+fea9:&_d+-_ +U+feaa:&_d+._ +U+feab:&_dk-_ +U+feac:&_dk._ +U+fead:&_r+-_ +U+feae:&_r+._ +U+feaf:&_z+-_ +U+feb0:&_z+._ +U+feb1:&_s+-_ +U+feb2:&_s+,_ +U+feb3:&_s+;_ +U+feb4:&_s+._ +U+feb5:&_sn-_ +U+feb6:&_sn,_ +U+feb7:&_sn;_ +U+feb8:&_sn._ +U+feb9:&_c+-_ +U+feba:&_c+,_ +U+febb:&_c+;_ +U+febc:&_c+._ +U+febd:&_dd-_ +U+febe:&_dd,_ +U+febf:&_dd;_ +U+fec0:&_dd._ +U+fec1:&_tj-_ +U+fec2:&_tj,_ +U+fec3:&_tj;_ +U+fec4:&_tj._ +U+fec5:&_zH-_ +U+fec6:&_zH,_ +U+fec7:&_zH;_ +U+fec8:&_zH._ +U+fec9:&_e+-_ +U+feca:&_e+,_ +U+fecb:&_e+;_ +U+fecc:&_e+._ +U+fecd:&_i+-_ +U+fece:&_i+,_ +U+fecf:&_i+;_ +U+fed0:&_i+._ +U+fed1:&_f+-_ +U+fed2:&_f+,_ +U+fed3:&_f+;_ +U+fed4:&_f+._ +U+fed5:&_q+-_ +U+fed6:&_q+,_ +U+fed7:&_q+;_ +U+fed8:&_q+._ +U+fed9:&_k+-_ +U+feda:&_k+,_ +U+fedb:&_k+;_ +U+fedc:&_k+._ +U+fedd:&_l+-_ +U+fede:&_l+,_ +U+fedf:&_l+;_ +U+fee0:&_l+._ +U+fee1:&_m+-_ +U+fee2:&_m+,_ +U+fee3:&_m+;_ +U+fee4:&_m+._ +U+fee5:&_n+-_ +U+fee6:&_n+,_ +U+fee7:&_n+;_ +U+fee8:&_n+._ +U+fee9:&_h+-_ +U+feea:&_h+,_ +U+feeb:&_h+;_ +U+feec:&_h+._ +U+feed:&_w+-_ +U+feee:&_w+._ +U+feef:&_j+-_ +U+fef0:&_j+._ +U+fef1:&_y+-_ +U+fef2:&_y+,_ +U+fef3:&_y+;_ +U+fef4:&_y+._ +U+fef5:&_lM-_ +U+fef6:&_lM._ +U+fef7:&_lH-_ +U+fef8:&_lH._ +U+fef9:&_lh-_ +U+fefa:&_lh._ +U+fefb:&_la-_ +U+fefc:&_la._ +U+0000:&NU +U+0001:&SH +U+0002:&SX +U+0003:&EX +U+0004:&ET +U+0005:&EQ +U+0006:&AK +U+0007:&BL +U+0008:&BS +U+0009:&HT +# U+000a:&LF +U+000b:&VT +U+000c:&FF +U+000d:&CR +U+000e:&SO +U+000f:&SI +U+0010:&DL +U+0011:&D1 +U+0012:&D2 +U+0013:&D3 +U+0014:&D4 +U+0015:&NK +U+0016:&SY +U+0017:&EB +U+0018:&CN +U+0019:&EM +U+001a:&SB +U+001b:&EC +U+001c:&FS +U+001d:&GS +U+001e:&RS +U+001f:&US +U+007f:&DT +U+0080:&PA +U+0081:&HO +U+0082:&BH +U+0083:&NH +U+0084:&IN +U+0085:&NL +U+0086:&SA +U+0087:&ES +U+0088:&HS +U+0089:&HJ +U+008a:&VS +U+008b:&PD +U+008c:&PU +U+008d:&RI +U+008e:&S2 +U+008f:&S3 +U+0090:&DC +U+0091:&P1 +U+0092:&P2 +U+0093:&TS +U+0094:&CC +U+0095:&MW +U+0096:&SG +U+0097:&EG +U+0098:&SS +U+0099:&GC +U+009a:&SC +U+009b:&CI +U+009c:&ST +U+009d:&OC +U+009e:&PM +U+009f:&AC +# Characters in Private Use Area (e000-f8ff) do not have ussigned numbers +# according Unicode 2.0 diff --git a/src/chrtrans/mnem_suni.tbl b/src/chrtrans/mnem_suni.tbl new file mode 100644 index 0000000..02bd8ea --- /dev/null +++ b/src/chrtrans/mnem_suni.tbl @@ -0,0 +1,1861 @@ +#The MIME name of this charset. +Mmnem + +#Name as a Display Charset (used on Options screen) +ORFC1345 Mnem + +# U+0020: SP +U+0021:! +U+0022:" +U+0023: Nb +U+0024: DO +U+0025:% +U+0026:&& +U+0027:' +U+0028:( +U+0029:) +U+002a:* +U+002b:+ +U+002c:, +U+002d:- +U+002e:. +U+002f:/ +U+0030:0 +U+0031:1 +U+0032:2 +U+0033:3 +U+0034:4 +U+0035:5 +U+0036:6 +U+0037:7 +U+0038:8 +U+0039:9 +U+003a:: +U+003b:; +U+003c:< +U+003d:= +U+003e:> +U+003f:? +U+0040: At +U+0041:A +U+0042:B +U+0043:C +U+0044:D +U+0045:E +U+0046:F +U+0047:G +U+0048:H +U+0049:I +U+004a:J +U+004b:K +U+004c:L +U+004d:M +U+004e:N +U+004f:O +U+0050:P +U+0051:Q +U+0052:R +U+0053:S +U+0054:T +U+0055:U +U+0056:V +U+0057:W +U+0058:X +U+0059:Y +U+005a:Z +U+005b: <( +U+005c: // +U+005d: )> +U+005e: '> +U+005f:_ +U+0060: '! +U+0061:a +U+0062:b +U+0063:c +U+0064:d +U+0065:e +U+0066:f +U+0067:g +U+0068:h +U+0069:i +U+006a:j +U+006b:k +U+006c:l +U+006d:m +U+006e:n +U+006f:o +U+0070:p +U+0071:q +U+0072:r +U+0073:s +U+0074:t +U+0075:u +U+0076:v +U+0077:w +U+0078:x +U+0079:y +U+007a:z +U+007b: (! +U+007c: !! +U+007d: !) +U+007e: '? +U+00a0: NS +U+00a1: !I +U+00a2: Ct +U+00a3: Pd +U+00a4: Cu +U+00a5: Ye +U+00a6: BB +U+00a7: SE +U+00a8: ': +U+00a9: Co +U+00aa: -a +U+00ab: << +U+00ac: NO +U+00ad: -- +U+00ae: Rg +U+00af: 'm +U+00b0: DG +U+00b1: +- +U+00b2: 2S +U+00b3: 3S +U+00b4: '' +U+00b5: My +U+00b6: PI +U+00b7: .M +U+00b8: ', +U+00b9: 1S +U+00ba: -o +U+00bb: >> +U+00bc: 14 +U+00bd: 12 +U+00be: 34 +U+00bf: ?I +U+00c0: A! +U+00c1: A' +U+00c2: A> +U+00c3: A? +U+00c4: A: +U+00c5: AA +U+00c6: AE +U+00c7: C, +U+00c8: E! +U+00c9: E' +U+00ca: E> +U+00cb: E: +U+00cc: I! +U+00cd: I' +U+00ce: I> +U+00cf: I: +U+00d0: D- +U+00d1: N? +U+00d2: O! +U+00d3: O' +U+00d4: O> +U+00d5: O? +U+00d6: O: +U+00d7: *X +U+00d8: O/ +U+00d9: U! +U+00da: U' +U+00db: U> +U+00dc: U: +U+00dd: Y' +U+00de: TH +U+00df: ss +U+00e0: a! +U+00e1: a' +U+00e2: a> +U+00e3: a? +U+00e4: a: +U+00e5: aa +U+00e6: ae +U+00e7: c, +U+00e8: e! +U+00e9: e' +U+00ea: e> +U+00eb: e: +U+00ec: i! +U+00ed: i' +U+00ee: i> +U+00ef: i: +U+00f0: d- +U+00f1: n? +U+00f2: o! +U+00f3: o' +U+00f4: o> +U+00f5: o? +U+00f6: o: +U+00f7: -: +U+00f8: o/ +U+00f9: u! +U+00fa: u' +U+00fb: u> +U+00fc: u: +U+00fd: y' +U+00fe: th +U+00ff: y: +U+0100: A- +U+0101: a- +U+0102: A( +U+0103: a( +U+0104: A; +U+0105: a; +U+0106: C' +U+0107: c' +U+0108: C> +U+0109: c> +U+010a: C. +U+010b: c. +U+010c: C< +U+010d: c< +U+010e: D< +U+010f: d< +U+0110: D/ +U+0111: d/ +U+0112: E- +U+0113: e- +U+0114: E( +U+0115: e( +U+0116: E. +U+0117: e. +U+0118: E; +U+0119: e; +U+011a: E< +U+011b: e< +U+011c: G> +U+011d: g> +U+011e: G( +U+011f: g( +U+0120: G. +U+0121: g. +U+0122: G, +U+0123: g, +U+0124: H> +U+0125: h> +U+0126: H/ +U+0127: h/ +U+0128: I? +U+0129: i? +U+012a: I- +U+012b: i- +U+012c: I( +U+012d: i( +U+012e: I; +U+012f: i; +U+0130: I. +U+0131: i. +U+0132: IJ +U+0133: ij +U+0134: J> +U+0135: j> +U+0136: K, +U+0137: k, +U+0138: kk +U+0139: L' +U+013a: l' +U+013b: L, +U+013c: l, +U+013d: L< +U+013e: l< +U+013f: L. +U+0140: l. +U+0141: L/ +U+0142: l/ +U+0143: N' +U+0144: n' +U+0145: N, +U+0146: n, +U+0147: N< +U+0148: n< +U+0149: 'n +U+014a: NG +U+014b: ng +U+014c: O- +U+014d: o- +U+014e: O( +U+014f: o( +U+0150: O" +U+0151: o" +U+0152: OE +U+0153: oe +U+0154: R' +U+0155: r' +U+0156: R, +U+0157: r, +U+0158: R< +U+0159: r< +U+015a: S' +U+015b: s' +U+015c: S> +U+015d: s> +U+015e: S, +U+015f: s, +U+0160: S< +U+0161: s< +U+0162: T, +U+0163: t, +U+0164: T< +U+0165: t< +U+0166: T/ +U+0167: t/ +U+0168: U? +U+0169: u? +U+016a: U- +U+016b: u- +U+016c: U( +U+016d: u( +U+016e: U0 +U+016f: u0 +U+0170: U" +U+0171: u" +U+0172: U; +U+0173: u; +U+0174: W> +U+0175: w> +U+0176: Y> +U+0177: y> +U+0178: Y: +U+0179: Z' +U+017a: z' +U+017b: Z. +U+017c: z. +U+017d: Z< +U+017e: z< +U+01a0: O9 +U+01a1: o9 +U+01a2: OI +U+01a3: oi +U+01a6: yr +U+01af: U9 +U+01b0: u9 +U+01b5: Z/ +U+01b6: z/ +U+01b7: ED +U+01cd: A< +U+01ce: a< +U+01cf: I< +U+01d0: i< +U+01d1: O< +U+01d2: o< +U+01d3: U< +U+01d4: u< +U+01d5: _U:-_ +U+01d6: _u:-_ +U+01d7: _U:'_ +U+01d8: _u:'_ +U+01d9: _U:<_ +U+01da: _u:<_ +U+01db: _U:!_ +U+01dc: _u:!_ +U+01de: A1 +U+01df: a1 +U+01e0: A7 +U+01e1: a7 +U+01e2: A3 +U+01e3: a3 +U+01e4: G/ +U+01e5: g/ +U+01e6: G< +U+01e7: g< +U+01e8: K< +U+01e9: k< +U+01ea: O; +U+01eb: o; +U+01ec: O1 +U+01ed: o1 +U+01ee: EZ +U+01ef: ez +U+01f0: j< +U+01f4: G' +U+01f5: g' +U+01fa: _AA'_ +U+01fb: _aa'_ +U+01fc: _AE'_ +U+01fd: _ae'_ +U+01fe: _O/'_ +U+01ff: _o/'_ +U+02bf: ;S +U+02c7: '< +U+02d8: '( +U+02d9: '. +U+02da: '0 +U+02db: '; +U+02dd: '" +U+0386: A% +U+0388: E% +U+0389: Y% +U+038a: I% +U+038c: O% +U+038e: U% +U+038f: W% +U+0390: i3 +U+0391: A* +U+0392: B* +U+0393: G* +U+0394: D* +U+0395: E* +U+0396: Z* +U+0397: Y* +U+0398: H* +U+0399: I* +U+039a: K* +U+039b: L* +U+039c: M* +U+039d: N* +U+039e: C* +U+039f: O* +U+03a0: P* +U+03a1: R* +U+03a3: S* +U+03a4: T* +U+03a5: U* +U+03a6: F* +U+03a7: X* +U+03a8: Q* +U+03a9: W* +U+03aa: J* +U+03ab: V* +U+03ac: a% +U+03ad: e% +U+03ae: y% +U+03af: i% +U+03b0: u3 +U+03b1: a* +U+03b2: b* +U+03b3: g* +U+03b4: d* +U+03b5: e* +U+03b6: z* +U+03b7: y* +U+03b8: h* +U+03b9: i* +U+03ba: k* +U+03bb: l* +U+03bc: m* +U+03bd: n* +U+03be: c* +U+03bf: o* +U+03c0: p* +U+03c1: r* +U+03c2: *s +U+03c3: s* +U+03c4: t* +U+03c5: u* +U+03c6: f* +U+03c7: x* +U+03c8: q* +U+03c9: w* +U+03ca: j* +U+03cb: v* +U+03cc: o% +U+03cd: u% +U+03ce: w% +U+03d8: 'G +U+03d9: ,G +U+03da: T3 +U+03db: t3 +U+03dc: M3 +U+03dd: m3 +U+03de: K3 +U+03df: k3 +U+03e0: P3 +U+03e1: p3 +U+03f4: '% +U+03f5: j3 +U+0401: IO +U+0402: D% +U+0403: G% +U+0404: IE +U+0405: DS +U+0406: II +U+0407: YI +U+0408: J% +U+0409: LJ +U+040a: NJ +U+040b: Ts +U+040c: KJ +U+040e: V% +U+040f: DZ +U+0410: A= +U+0411: B= +U+0412: V= +U+0413: G= +U+0414: D= +U+0415: E= +U+0416: Z% +U+0417: Z= +U+0418: I= +U+0419: J= +U+041a: K= +U+041b: L= +U+041c: M= +U+041d: N= +U+041e: O= +U+041f: P= +U+0420: R= +U+0421: S= +U+0422: T= +U+0423: U= +U+0424: F= +U+0425: H= +U+0426: C= +U+0427: C% +U+0428: S% +U+0429: Sc +U+042a: =" +U+042b: Y= +U+042c: %" +U+042d: JE +U+042e: JU +U+042f: JA +U+0430: a= +U+0431: b= +U+0432: v= +U+0433: g= +U+0434: d= +U+0435: e= +U+0436: z% +U+0437: z= +U+0438: i= +U+0439: j= +U+043a: k= +U+043b: l= +U+043c: m= +U+043d: n= +U+043e: o= +U+043f: p= +U+0440: r= +U+0441: s= +U+0442: t= +U+0443: u= +U+0444: f= +U+0445: h= +U+0446: c= +U+0447: c% +U+0448: s% +U+0449: sc +U+044a: =' +U+044b: y= +U+044c: %' +U+044d: je +U+044e: ju +U+044f: ja +U+0451: io +U+0452: d% +U+0453: g% +U+0454: ie +U+0455: ds +U+0456: ii +U+0457: yi +U+0458: j% +U+0459: lj +U+045a: nj +U+045b: ts +U+045c: kj +U+045e: v% +U+045f: dz +U+0462: Y3 +U+0463: y3 +U+046a: O3 +U+046b: o3 +U+0472: F3 +U+0473: f3 +U+0474: V3 +U+0475: v3 +U+0480: C3 +U+0481: c3 +U+0490: G3 +U+0491: g3 +U+05d0: A+ +U+05d1: B+ +U+05d2: G+ +U+05d3: D+ +U+05d4: H+ +U+05d5: W+ +U+05d6: Z+ +U+05d7: X+ +U+05d8: Tj +U+05d9: J+ +U+05da: K% +U+05db: K+ +U+05dc: L+ +U+05dd: M% +U+05de: M+ +U+05df: N% +U+05e0: N+ +U+05e1: S+ +U+05e2: E+ +U+05e3: P% +U+05e4: P+ +U+05e5: Zj +U+05e6: ZJ +U+05e7: Q+ +U+05e8: R+ +U+05e9: Sh +U+05ea: T+ +U+060c: ,+ +U+061b: ;+ +U+061f: ?+ +U+0621: H' +U+0622: aM +U+0623: aH +U+0624: wH +U+0625: ah +U+0626: yH +U+0627: a+ +U+0628: b+ +U+0629: tm +U+062a: t+ +U+062b: tk +U+062c: g+ +U+062d: hk +U+062e: x+ +U+062f: d+ +U+0630: dk +U+0631: r+ +U+0632: z+ +U+0633: s+ +U+0634: sn +U+0635: c+ +U+0636: dd +U+0637: tj +U+0638: zH +U+0639: e+ +U+063a: i+ +U+0640: ++ +U+0641: f+ +U+0642: q+ +U+0643: k+ +U+0644: l+ +U+0645: m+ +U+0646: n+ +U+0647: h+ +U+0648: w+ +U+0649: j+ +U+064a: y+ +U+064b: :+ +U+064c: "+ +U+064d: =+ +U+064e: /+ +U+064f: '+ +U+0650: 1+ +U+0651: 3+ +U+0652: 0+ +U+0670: aS +U+067e: p+ +U+06a4: v+ +U+06af: gf +U+06f0: 0a +U+06f1: 1a +U+06f2: 2a +U+06f3: 3a +U+06f4: 4a +U+06f5: 5a +U+06f6: 6a +U+06f7: 7a +U+06f8: 8a +U+06f9: 9a +U+1e00: _A-0_ +U+1e01: _a-0_ +U+1e02: B. +U+1e03: b. +U+1e04: _B-._ +U+1e05: _b-._ +U+1e06: B_ +U+1e07: b_ +U+1e08: _C,'_ +U+1e09: _c,'_ +U+1e0a: D. +U+1e0b: d. +U+1e0c: _D-._ +U+1e0d: _d-._ +U+1e0e: D_ +U+1e0f: d_ +U+1e10: D, +U+1e11: d, +U+1e12: _D->_ +U+1e13: _d->_ +U+1e14: _E-!_ +U+1e15: _e-!_ +U+1e16: _E-'_ +U+1e17: _e-'_ +U+1e18: _E->_ +U+1e19: _e->_ +U+1e1a: _E-?_ +U+1e1b: _e-?_ +U+1e1c: _E,(_ +U+1e1d: _e,(_ +U+1e1e: F. +U+1e1f: f. +U+1e20: G- +U+1e21: g- +U+1e22: H. +U+1e23: h. +U+1e24: _H-._ +U+1e25: _h-._ +U+1e26: H: +U+1e27: h: +U+1e28: H, +U+1e29: h, +U+1e2a: _H-(_ +U+1e2b: _h-(_ +U+1e2c: _I-?_ +U+1e2d: _i-?_ +U+1e2e: _I:'_ +U+1e2f: _i:'_ +U+1e30: K' +U+1e31: k' +U+1e32: _K-._ +U+1e33: _k-._ +U+1e34: K_ +U+1e35: k_ +U+1e36: _L-._ +U+1e37: _l-._ +U+1e38: _L--._ +U+1e39: _l--._ +U+1e3a: L_ +U+1e3b: l_ +U+1e3c: _L->_ +U+1e3d: _l->_ +U+1e3e: M' +U+1e3f: m' +U+1e40: M. +U+1e41: m. +U+1e42: _M-._ +U+1e43: _m-._ +U+1e44: N. +U+1e45: n. +U+1e46: _N-._ +U+1e47: _n-._ +U+1e48: N_ +U+1e49: n_ +U+1e4a: _N->_ +U+1e4b: _N->_ +U+1e4c: _O?'_ +U+1e4d: _o?'_ +U+1e4e: _O?:_ +U+1e4f: _o?:_ +U+1e50: _O-!_ +U+1e51: _o-!_ +U+1e52: _O-'_ +U+1e53: _o-'_ +U+1e54: P' +U+1e55: p' +U+1e56: P. +U+1e57: p. +U+1e58: R. +U+1e59: r. +U+1e5a: _R-._ +U+1e5b: _r-._ +U+1e5c: _R--._ +U+1e5d: _r--._ +U+1e5e: R_ +U+1e5f: r_ +U+1e60: S. +U+1e61: s. +U+1e62: _S-._ +U+1e63: _s-._ +U+1e64: _S'._ +U+1e65: _s'._ +U+1e66: _S<._ +U+1e67: _s<._ +U+1e68: _S.-._ +U+1e69: _S.-._ +U+1e6a: T. +U+1e6b: t. +U+1e6c: _T-._ +U+1e6d: _t-._ +U+1e6e: T_ +U+1e6f: t_ +U+1e70: _T->_ +U+1e71: _t->_ +U+1e72: _U--:_ +U+1e73: _u--:_ +U+1e74: _U-?_ +U+1e75: _u-?_ +U+1e76: _U->_ +U+1e77: _u->_ +U+1e78: _U?'_ +U+1e79: _u?'_ +U+1e7a: _U-:_ +U+1e7b: _u-:_ +U+1e7c: V? +U+1e7d: v? +U+1e7e: _V-._ +U+1e7f: _v-._ +U+1e80: W! +U+1e81: w! +U+1e82: W' +U+1e83: w' +U+1e84: W: +U+1e85: w: +U+1e86: W. +U+1e87: w. +U+1e88: _W-._ +U+1e89: _w-._ +U+1e8a: X. +U+1e8b: x. +U+1e8c: X: +U+1e8d: x: +U+1e8e: Y. +U+1e8f: y. +U+1e90: Z> +U+1e91: z> +U+1e92: _Z-._ +U+1e93: _z-._ +U+1e94: Z_ +U+1e95: z_ +U+1e96: h_ +U+1e97: t: +U+1e98: w0 +U+1e99: y0 +U+1ea0: _A-._ +U+1ea1: _a-._ +U+1ea2: A2 +U+1ea3: a2 +U+1ea4: _A>'_ +U+1ea5: _a>'_ +U+1ea6: _A>!_ +U+1ea7: _a>!_ +U+1ea8: _A>2_ +U+1ea9: _a>2_ +U+1eaa: _A>?_ +U+1eab: _a>?_ +U+1eac: _A>-._ +U+1ead: _a>-._ +U+1eae: _A('_ +U+1eaf: _a('_ +U+1eb0: _A(!_ +U+1eb1: _a(!_ +U+1eb2: _A(2_ +U+1eb3: _a(2_ +U+1eb4: _A(?_ +U+1eb5: _a(?_ +U+1eb6: _A(-._ +U+1eb7: _a(-._ +U+1eb8: _E-._ +U+1eb9: _e-._ +U+1eba: E2 +U+1ebb: e2 +U+1ebc: E? +U+1ebd: e? +U+1ebe: _E>'_ +U+1ebf: _e>'_ +U+1ec0: _E>!_ +U+1ec1: _e>!_ +U+1ec2: _E>2_ +U+1ec3: _e>2_ +U+1ec4: _E>?_ +U+1ec5: _e>?_ +U+1ec6: _E>-._ +U+1ec7: _e>-._ +U+1ec8: I2 +U+1ec9: i2 +U+1eca: _I-._ +U+1ecb: _i-._ +U+1ecc: _O-._ +U+1ecd: _o-._ +U+1ece: O2 +U+1ecf: o2 +U+1ed0: _O>'_ +U+1ed1: _o>'_ +U+1ed2: _O>!_ +U+1ed3: _o>!_ +U+1ed4: _O>2_ +U+1ed5: _o>2_ +U+1ed6: _O>?_ +U+1ed7: _o>?_ +U+1ed8: _O>-._ +U+1ed9: _o>-._ +U+1eda: _O9'_ +U+1edb: _o9'_ +U+1edc: _O9!_ +U+1edd: _o9!_ +U+1ede: _O92_ +U+1edf: _o92_ +U+1ee0: _O9?_ +U+1ee1: _o9?_ +U+1ee2: _O9-._ +U+1ee3: _o9-._ +U+1ee4: _U-._ +U+1ee5: _u-._ +U+1ee6: U2 +U+1ee7: u2 +U+1ee8: _U9'_ +U+1ee9: _u9'_ +U+1eea: _U9!_ +U+1eeb: _u9!_ +U+1eec: _U92_ +U+1eed: _u92_ +U+1eee: _U9?_ +U+1eef: _u9?_ +U+1ef0: _U9-._ +U+1ef1: _u9-._ +U+1ef2: Y! +U+1ef3: y! +U+1ef4: _Y-._ +U+1ef5: _y-._ +U+1ef6: Y2 +U+1ef7: y2 +U+1ef8: Y? +U+1ef9: y? +U+1f00: ;' +U+1f01: ,' +U+1f02: ;! +U+1f03: ,! +U+1f04: ?; +U+1f05: ?, +U+1f06: !: +U+1f07: ?: +U+2002: 1N +U+2003: 1M +U+2004: 3M +U+2005: 4M +U+2006: 6M +U+2009: 1T +U+200a: 1H +U+2010: -1 +U+2013: -N +U+2014: -M +U+2015: -3 +U+2016: !2 +U+2017: =2 +U+2018: '6 +U+2019: '9 +U+201a: .9 +U+201b: 9' +U+201c: "6 +U+201d: "9 +U+201e: :9 +U+201f: 9" +U+2020: /- +U+2021: /= +U+2025: .. +U+2030: %0 +U+2032: 1' +U+2033: 2' +U+2034: 3' +U+2035: 1" +U+2036: 2" +U+2037: 3" +U+2038: Ca +U+2039: <1 +U+203a: >1 +U+203b: :X +U+203c: _!*2_ +U+203e: '- +U+2044: /f +U+2070: 0S +U+2074: 4S +U+2075: 5S +U+2076: 6S +U+2077: 7S +U+2078: 8S +U+2079: 9S +U+207a: +S +U+207b: -S +U+207c: =S +U+207d: (S +U+207e: )S +U+207f: nS +U+2080: 0s +U+2081: 1s +U+2082: 2s +U+2083: 3s +U+2084: 4s +U+2085: 5s +U+2086: 6s +U+2087: 7s +U+2088: 8s +U+2089: 9s +U+208a: +s +U+208b: -s +U+208c: =s +U+208d: (s +U+208e: )s +U+20a4: Li +U+20a7: Pt +U+20a9: W= +U+2103: oC +U+2105: co +U+2109: oF +U+2116: N0 +U+2117: PO +U+211e: Rx +U+2120: SM +U+2122: TM +U+2126: Om +U+212b: AO +U+2153: 13 +U+2154: 23 +U+2155: 15 +U+2156: 25 +U+2157: 35 +U+2158: 45 +U+2159: 16 +U+215a: 56 +U+215b: 18 +U+215c: 38 +U+215d: 58 +U+215e: 78 +U+2160: 1R +U+2161: 2R +U+2162: 3R +U+2163: 4R +U+2164: 5R +U+2165: 6R +U+2166: 7R +U+2167: 8R +U+2168: 9R +U+2169: aR +U+216a: bR +U+216b: cR +U+216c: _50R_ +U+216d: _100R_ +U+216e: _500R_ +U+216f: _1000R_ +U+2170: 1r +U+2171: 2r +U+2172: 3r +U+2173: 4r +U+2174: 5r +U+2175: 6r +U+2176: 7r +U+2177: 8r +U+2178: 9r +U+2179: ar +U+217a: br +U+217b: cr +U+217c: _50r_ +U+217d: _100r_ +U+217e: _500r_ +U+217f: _1000r_ +U+2180: _1000RCD_ +U+2181: _5000R_ +U+2182: _10000R_ +U+2190: <- +U+2191: -! +U+2192: -> +U+2193: -v +U+2194: <> +U+2195: UD +U+2196: _<!!_ +U+2197: _//>_ +U+2198: _!!>_ +U+2199: _<//_ +U+21d0: <= +U+21d2: => +U+21d4: == +U+2200: FA +U+2202: dP +U+2203: TE +U+2205: /0 +U+2206: DE +U+2207: NB +U+2208: (- +U+220b: -) +U+220f: *P +U+2211: +Z +U+2212: -2 +U+2213: -+ +U+2217: *- +U+2218: Ob +U+2219: Sb +U+221a: RT +U+221d: 0( +U+221e: 00 +U+221f: -L +U+2220: -V +U+2225: PP +U+2227: AN +U+2228: OR +U+2229: (U +U+222a: )U +U+222b: In +U+222c: DI +U+222e: Io +U+2234: .: +U+2235: :. +U+2236: :R +U+2237: :: +U+223c: ?1 +U+223e: CG +U+2243: ?- +U+2245: ?= +U+2248: ?2 +U+224c: =? +U+2253: HI +U+2260: != +U+2261: =3 +U+2264: =< +U+2265: >= +U+226a: <* +U+226b: *> +U+226e: !< +U+226f: !> +U+2282: (C +U+2283: )C +U+2286: (_ +U+2287: )_ +U+2299: 0. +U+229a: 02 +U+22a5: -T +U+22c5: .P +U+22ee: :3 +U+22ef: .3 +U+2302: Eh +U+2308: <7 +U+2309: >7 +U+230a: 7< +U+230b: 7> +U+2310: NI +U+2312: (A +U+2315: TR +U+2320: Iu +U+2321: Il +U+2329: </ +U+232a: /> +U+2423: Vs +U+2440: 1h +U+2441: 3h +U+2442: 2h +U+2443: 4h +U+2446: 1j +U+2447: 2j +U+2448: 3j +U+2449: 4j +U+2460: _1-o_ +U+2461: _2-o_ +U+2462: _3-o_ +U+2463: _4-o_ +U+2464: _5-o_ +U+2465: _6-o_ +U+2466: _7-o_ +U+2467: _8-o_ +U+2468: _9-o_ +U+2469: _10-o_ +U+246a: _11-o_ +U+246b: _12-o_ +U+246c: _13-o_ +U+246d: _14-o_ +U+246e: _15-o_ +U+246f: _16-o_ +U+2470: _17-o_ +U+2471: _18-o_ +U+2472: _19-o_ +U+2473: _20-o_ +U+2474: _(1)_ +U+2475: _(2)_ +U+2476: _(3)_ +U+2477: _(4)_ +U+2478: _(5)_ +U+2479: _(6)_ +U+247a: _(7)_ +U+247b: _(8)_ +U+247c: _(9)_ +U+247d: _(10)_ +U+247e: _(11)_ +U+247f: _(12)_ +U+2480: _(13)_ +U+2481: _(14)_ +U+2482: _(15)_ +U+2483: _(16)_ +U+2484: _(17)_ +U+2485: _(18)_ +U+2486: _(19)_ +U+2487: _(20)_ +U+2488: 1. +U+2489: 2. +U+248a: 3. +U+248b: 4. +U+248c: 5. +U+248d: 6. +U+248e: 7. +U+248f: 8. +U+2490: 9. +U+2491: _10._ +U+2492: _11._ +U+2493: _12._ +U+2494: _13._ +U+2495: _14._ +U+2496: _15._ +U+2497: _16._ +U+2498: _17._ +U+2499: _18._ +U+249a: _19._ +U+249b: _20._ +U+249c: _(a)_ +U+249d: _(b)_ +U+249e: _(c)_ +U+249f: _(d)_ +U+24a0: _(e)_ +U+24a1: _(f)_ +U+24a2: _(g)_ +U+24a3: _(h)_ +U+24a4: _(i)_ +U+24a5: _(j)_ +U+24a6: _(k)_ +U+24a7: _(l)_ +U+24a8: _(m)_ +U+24a9: _(n)_ +U+24aa: _(o)_ +U+24ab: _(p)_ +U+24ac: _(q)_ +U+24ad: _(r)_ +U+24ae: _(s)_ +U+24af: _(t)_ +U+24b0: _(u)_ +U+24b1: _(v)_ +U+24b2: _(w)_ +U+24b3: _(x)_ +U+24b4: _(y)_ +U+24b5: _(z)_ +U+24b6: _A-o_ +U+24b7: _B-o_ +U+24b8: _C-o_ +U+24b9: _D-o_ +U+24ba: _E-o_ +U+24bb: _F-o_ +U+24bc: _G-o_ +U+24bd: _H-o_ +U+24be: _I-o_ +U+24bf: _J-o_ +U+24c0: _K-o_ +U+24c1: _L-o_ +U+24c2: _M-o_ +U+24c3: _N-o_ +U+24c4: _O-o_ +U+24c5: _P-o_ +U+24c6: _Q-o_ +U+24c7: _R-o_ +U+24c8: _S-o_ +U+24c9: _T-o_ +U+24ca: _U-o_ +U+24cb: _V-o_ +U+24cc: _W-o_ +U+24cd: _X-o_ +U+24ce: _Y-o_ +U+24cf: _Z-o_ +U+24d0: _a-o_ +U+24d1: _b-o_ +U+24d2: _c-o_ +U+24d3: _d-o_ +U+24d4: _e-o_ +U+24d5: _f-o_ +U+24d6: _g-o_ +U+24d7: _h-o_ +U+24d8: _i-o_ +U+24d9: _j-o_ +U+24da: _k-o_ +U+24db: _l-o_ +U+24dc: _m-o_ +U+24dd: _n-o_ +U+24de: _o-o_ +U+24df: _p-o_ +U+24e0: _q-o_ +U+24e1: _r-o_ +U+24e2: _s-o_ +U+24e3: _t-o_ +U+24e4: _u-o_ +U+24e5: _v-o_ +U+24e6: _w-o_ +U+24e7: _x-o_ +U+24e8: _y-o_ +U+24e9: _z-o_ +U+24ea: _0-o_ +U+2500: hh +U+2501: HH +U+2502: vv +U+2503: VV +U+2504: 3- +U+2505: 3_ +U+2506: 3! +U+2507: 3/ +U+2508: 4- +U+2509: 4_ +U+250a: 4! +U+250b: 4/ +U+250c: dr +U+250d: dR +U+250e: Dr +U+250f: DR +U+2510: dl +U+2511: dL +U+2512: Dl +U+2513: LD +U+2514: ur +U+2515: uR +U+2516: Ur +U+2517: UR +U+2518: ul +U+2519: uL +U+251a: Ul +U+251b: UL +U+251c: vr +U+251d: vR +U+251e: _Udr_ +U+251f: _uDr_ +U+2520: Vr +U+2521: _UdR_ +U+2522: _uDR_ +U+2523: VR +U+2524: vl +U+2525: vL +U+2526: _Udl_ +U+2527: _uDl_ +U+2528: Vl +U+2529: _UdL_ +U+252a: _uDL_ +U+252b: VL +U+252c: dh +U+252d: _dLr_ +U+252e: _dlR_ +U+252f: dH +U+2530: Dh +U+2531: _DLr_ +U+2532: _DlR_ +U+2533: DH +U+2534: uh +U+2535: _uLr_ +U+2536: _ulR_ +U+2537: uH +U+2538: Uh +U+2539: _ULr_ +U+253a: _UlR_ +U+253b: UH +U+253c: vh +U+253d: _vLr_ +U+253e: _vlR_ +U+253f: vH +U+2540: _Udh_ +U+2541: _uDh_ +U+2542: Vh +U+2543: _UdLr_ +U+2544: _UdlR_ +U+2545: _uDLr_ +U+2546: _uDlR_ +U+2547: _UdH_ +U+2548: _uDH_ +U+2549: _VLr_ +U+254a: _VlR_ +U+254b: VH +U+2571: FD +U+2572: BD +U+2580: TB +U+2584: LB +U+2588: FB +U+258c: lB +U+2590: RB +U+2591: .S +U+2592: :S +U+2593: ?S +U+25a0: fS +U+25a1: OS +U+25a2: RO +U+25a3: Rr +U+25a4: RF +U+25a5: RY +U+25a6: RH +U+25a7: RZ +U+25a8: RK +U+25a9: RX +U+25aa: sB +U+25ac: SR +U+25ad: Or +U+25b2: UT +U+25b3: uT +U+25b6: PR +U+25b7: Tr +U+25bc: Dt +U+25bd: dT +U+25c0: PL +U+25c1: Tl +U+25c6: Db +U+25c7: Dw +U+25ca: LZ +U+25cb: 0m +U+25ce: 0o +U+25cf: 0M +U+25d0: 0L +U+25d1: 0R +U+25d8: Sn +U+25d9: Ic +U+25e2: Fd +U+25e3: Bd +U+2605: *2 +U+2606: *1 +U+260e: _TEL_ +U+260f: _tel_ +U+261c: <H +U+261e: >H +U+263a: 0u +U+263b: 0U +U+263c: SU +U+2640: Fm +U+2642: Ml +U+2660: cS +U+2661: cH +U+2662: cD +U+2663: cC +U+2664: _cS-_ +U+2665: _cH-_ +U+2666: _cD-_ +U+2667: _cC-_ +U+2669: Md +U+266a: M8 +U+266b: M2 +U+266c: _M16_ +U+266d: Mb +U+266e: Mx +U+266f: MX +U+2713: OK +U+2717: XX +U+2720: -X +U+3000: IS +U+3001: ,_ +U+3002: ._ +U+3003: +" +U+3004: +_ +U+3005: *_ +U+3006: ;_ +U+3007: 0_ +U+300a: <+ +U+300b: >+ +U+300c: <' +U+300d: >' +U+300e: <" +U+300f: >" +U+3010: (" +U+3011: )" +U+3012: =T +U+3013: =_ +U+3014: (' +U+3015: )' +U+3016: (I +U+3017: )I +U+301c: -? +U+3020: _=T:)_ +U+3041: A5 +U+3042: a5 +U+3043: I5 +U+3044: i5 +U+3045: U5 +U+3046: u5 +U+3047: E5 +U+3048: e5 +U+3049: O5 +U+304a: o5 +U+304b: ka +U+304c: ga +U+304d: ki +U+304e: gi +U+304f: ku +U+3050: gu +U+3051: ke +U+3052: ge +U+3053: ko +U+3054: go +U+3055: sa +U+3056: za +U+3057: si +U+3058: zi +U+3059: su +U+305a: zu +U+305b: se +U+305c: ze +U+305d: so +U+305e: zo +U+305f: ta +U+3060: da +U+3061: ti +U+3062: di +U+3063: tU +U+3064: tu +U+3065: du +U+3066: te +U+3067: de +U+3068: to +U+3069: do +U+306a: na +U+306b: ni +U+306c: nu +U+306d: ne +U+306e: no +U+306f: ha +U+3070: ba +U+3071: pa +U+3072: hi +U+3073: bi +U+3074: pi +U+3075: hu +U+3076: bu +U+3077: pu +U+3078: he +U+3079: be +U+307a: pe +U+307b: ho +U+307c: bo +U+307d: po +U+307e: ma +U+307f: mi +U+3080: mu +U+3081: me +U+3082: mo +U+3083: yA +U+3084: ya +U+3085: yU +U+3086: yu +U+3087: yO +U+3088: yo +U+3089: ra +U+308a: ri +U+308b: ru +U+308c: re +U+308d: ro +U+308e: wA +U+308f: wa +U+3090: wi +U+3091: we +U+3092: wo +U+3093: n5 +U+3094: vu +U+309b: "5 +U+309c: 05 +U+309d: *5 +U+309e: +5 +U+30a1: a6 +U+30a2: A6 +U+30a3: i6 +U+30a4: I6 +U+30a5: u6 +U+30a6: U6 +U+30a7: e6 +U+30a8: E6 +U+30a9: o6 +U+30aa: O6 +U+30ab: Ka +U+30ac: Ga +U+30ad: Ki +U+30ae: Gi +U+30af: Ku +U+30b0: Gu +U+30b1: Ke +U+30b2: Ge +U+30b3: Ko +U+30b4: Go +U+30b5: Sa +U+30b6: Za +U+30b7: Si +U+30b8: Zi +U+30b9: Su +U+30ba: Zu +U+30bb: Se +U+30bc: Ze +U+30bd: So +U+30be: Zo +U+30bf: Ta +U+30c0: Da +U+30c1: Ti +U+30c2: Di +U+30c3: TU +U+30c4: Tu +U+30c5: Du +U+30c6: Te +U+30c7: De +U+30c8: To +U+30c9: Do +U+30ca: Na +U+30cb: Ni +U+30cc: Nu +U+30cd: Ne +U+30ce: No +U+30cf: Ha +U+30d0: Ba +U+30d1: Pa +U+30d2: Hi +U+30d3: Bi +U+30d4: Pi +U+30d5: Hu +U+30d6: Bu +U+30d7: Pu +U+30d8: He +U+30d9: Be +U+30da: Pe +U+30db: Ho +U+30dc: Bo +U+30dd: Po +U+30de: Ma +U+30df: Mi +U+30e0: Mu +U+30e1: Me +U+30e2: Mo +U+30e3: YA +U+30e4: Ya +U+30e5: YU +U+30e6: Yu +U+30e7: YO +U+30e8: Yo +U+30e9: Ra +U+30ea: Ri +U+30eb: Ru +U+30ec: Re +U+30ed: Ro +U+30ee: WA +U+30ef: Wa +U+30f0: Wi +U+30f1: We +U+30f2: Wo +U+30f3: N6 +U+30f4: Vu +U+30f5: KA +U+30f6: KE +U+30f7: Va +U+30f8: Vi +U+30f9: Ve +U+30fa: Vo +U+30fb: .6 +U+30fc: -6 +U+30fd: *6 +U+30fe: +6 +U+3105: b4 +U+3106: p4 +U+3107: m4 +U+3108: f4 +U+3109: d4 +U+310a: t4 +U+310b: n4 +U+310c: l4 +U+310d: g4 +U+310e: k4 +U+310f: h4 +U+3110: j4 +U+3111: q4 +U+3112: x4 +U+3113: zh +U+3114: ch +U+3115: sh +U+3116: r4 +U+3117: z4 +U+3118: c4 +U+3119: s4 +U+311a: a4 +U+311b: o4 +U+311c: e4 +U+311d: _eh4_ +U+311e: ai +U+311f: ei +U+3120: au +U+3121: ou +U+3122: an +U+3123: en +U+3124: aN +U+3125: eN +U+3126: er +U+3127: i4 +U+3128: u4 +U+3129: iu +U+312a: v4 +U+312b: nG +U+312c: gn +U+321c: _(JU)_ +U+3220: 1c +U+3221: 2c +U+3222: 3c +U+3223: 4c +U+3224: 5c +U+3225: 6c +U+3226: 7c +U+3227: 8c +U+3228: 9c +U+3229: _10c_ +U+327f: _KSC_ +U+fb00: ff +U+fb01: fi +U+fb02: fl +U+fb03: _ffi_ +U+fb04: _ffl_ +U+fb05: ft +U+fb06: st +U+fe7d: _3+;_ +U+fe82: _aM._ +U+fe84: _aH._ +U+fe8d: _a+-_ +U+fe8e: _a+._ +U+fe8f: _b+-_ +U+fe90: _b+,_ +U+fe91: _b+;_ +U+fe92: _b+._ +U+fe93: _tm-_ +U+fe94: _tm._ +U+fe95: _t+-_ +U+fe96: _t+,_ +U+fe97: _t+;_ +U+fe98: _t+._ +U+fe99: _tk-_ +U+fe9a: _tk,_ +U+fe9b: _tk;_ +U+fe9c: _tk._ +U+fe9d: _g+-_ +U+fe9e: _g+,_ +U+fe9f: _g+;_ +U+fea0: _g+._ +U+fea1: _hk-_ +U+fea2: _hk,_ +U+fea3: _hk;_ +U+fea4: _hk._ +U+fea5: _x+-_ +U+fea6: _x+,_ +U+fea7: _x+;_ +U+fea8: _x+._ +U+fea9: _d+-_ +U+feaa: _d+._ +U+feab: _dk-_ +U+feac: _dk._ +U+fead: _r+-_ +U+feae: _r+._ +U+feaf: _z+-_ +U+feb0: _z+._ +U+feb1: _s+-_ +U+feb2: _s+,_ +U+feb3: _s+;_ +U+feb4: _s+._ +U+feb5: _sn-_ +U+feb6: _sn,_ +U+feb7: _sn;_ +U+feb8: _sn._ +U+feb9: _c+-_ +U+feba: _c+,_ +U+febb: _c+;_ +U+febc: _c+._ +U+febd: _dd-_ +U+febe: _dd,_ +U+febf: _dd;_ +U+fec0: _dd._ +U+fec1: _tj-_ +U+fec2: _tj,_ +U+fec3: _tj;_ +U+fec4: _tj._ +U+fec5: _zH-_ +U+fec6: _zH,_ +U+fec7: _zH;_ +U+fec8: _zH._ +U+fec9: _e+-_ +U+feca: _e+,_ +U+fecb: _e+;_ +U+fecc: _e+._ +U+fecd: _i+-_ +U+fece: _i+,_ +U+fecf: _i+;_ +U+fed0: _i+._ +U+fed1: _f+-_ +U+fed2: _f+,_ +U+fed3: _f+;_ +U+fed4: _f+._ +U+fed5: _q+-_ +U+fed6: _q+,_ +U+fed7: _q+;_ +U+fed8: _q+._ +U+fed9: _k+-_ +U+feda: _k+,_ +U+fedb: _k+;_ +U+fedc: _k+._ +U+fedd: _l+-_ +U+fede: _l+,_ +U+fedf: _l+;_ +U+fee0: _l+._ +U+fee1: _m+-_ +U+fee2: _m+,_ +U+fee3: _m+;_ +U+fee4: _m+._ +U+fee5: _n+-_ +U+fee6: _n+,_ +U+fee7: _n+;_ +U+fee8: _n+._ +U+fee9: _h+-_ +U+feea: _h+,_ +U+feeb: _h+;_ +U+feec: _h+._ +U+feed: _w+-_ +U+feee: _w+._ +U+feef: _j+-_ +U+fef0: _j+._ +U+fef1: _y+-_ +U+fef2: _y+,_ +U+fef3: _y+;_ +U+fef4: _y+._ +U+fef5: _lM-_ +U+fef6: _lM._ +U+fef7: _lH-_ +U+fef8: _lH._ +U+fef9: _lh-_ +U+fefa: _lh._ +U+fefb: _la-_ +U+fefc: _la._ +U+0000: NU +U+0001: SH +U+0002: SX +U+0003: EX +U+0004: ET +U+0005: EQ +U+0006: AK +U+0007: BL +U+0008: BS +U+0009: HT +# U+000a: LF +U+000b: VT +U+000c: FF +U+000d: CR +U+000e: SO +U+000f: SI +U+0010: DL +U+0011: D1 +U+0012: D2 +U+0013: D3 +U+0014: D4 +U+0015: NK +U+0016: SY +U+0017: EB +U+0018: CN +U+0019: EM +U+001a: SB +U+001b: EC +U+001c: FS +U+001d: GS +U+001e: RS +U+001f: US +U+007f: DT +U+0080: PA +U+0081: HO +U+0082: BH +U+0083: NH +U+0084: IN +U+0085: NL +U+0086: SA +U+0087: ES +U+0088: HS +U+0089: HJ +U+008a: VS +U+008b: PD +U+008c: PU +U+008d: RI +U+008e: S2 +U+008f: S3 +U+0090: DC +U+0091: P1 +U+0092: P2 +U+0093: TS +U+0094: CC +U+0095: MW +U+0096: SG +U+0097: EG +U+0098: SS +U+0099: GC +U+009a: SC +U+009b: CI +U+009c: ST +U+009d: OC +U+009e: PM +U+009f: AC +# Characters in Private Use Area (e000-f8ff) do not have ussigned numbers +# according Unicode 2.0 diff --git a/src/chrtrans/next_uni.tbl b/src/chrtrans/next_uni.tbl new file mode 100644 index 0000000..a76ae34 --- /dev/null +++ b/src/chrtrans/next_uni.tbl @@ -0,0 +1,185 @@ +# This file has been modified for lynx (see README.tables) + +#The MIME name of this charset. +Mnext + +#Name as a Display Charset (used on Options screen) +ONeXT character set + +# +# Name: NextStep Encoding to Unicode +# Unicode version: 1.1 +# Table version: 0.1 +# Table format: Format A +# Date: 1999 September 23 +# Authors: Rick McGowan +# +# Copyright (c) 1991-1999 Unicode, Inc. All Rights reserved. +# +# This file is provided as-is by Unicode, Inc. (The Unicode Consortium). +# No claims are made as to fitness for any particular purpose. No +# warranties of any kind are expressed or implied. The recipient +# agrees to determine applicability of information provided. If this +# file has been provided on optical media by Unicode, Inc., the sole +# remedy for any claim will be exchange of defective media within 90 +# days of receipt. +# +# Unicode, Inc. hereby grants the right to freely use the information +# supplied in this file in the creation of products supporting the +# Unicode Standard, and to make copies of this file in any form for +# internal or external distribution as long as this notice remains +# attached. +# +# General notes: +# +# This table contains the data the Unicode Consortium has on how +# NextStep Encoding characters map into Unicode. Since the first +# 128 characters (0x0 - 0x7f) are identical to ASCII and Unicode, +# this table only maps the NextStep range from 0x80 - 0xFF. +# +# This file is provided for historical reference only and pertains +# to NextStep and OpenStep products shipped prior to the acquisition +# of NeXT by Apple Computer, Inc. See http://www.apple.com for +# further information. +# +# Format: Three tab-separated columns +# Column #1 is the NextStep code (in hex as 0xXX) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 NextStep name, Unicode name (follows a comment sign, '#') +# +# The entries are in NextStep order +# +# Any comments or problems, contact info@unicode.org +# +# +0x20-0x7f idem +# +0x80 U+00a0 # NO-BREAK SPACE +0x81 U+00c0 # LATIN CAPITAL LETTER A WITH GRAVE +0x82 U+00c1 # LATIN CAPITAL LETTER A WITH ACUTE +0x83 U+00c2 # LATIN CAPITAL LETTER A WITH CIRCUMFLEX +0x84 U+00c3 # LATIN CAPITAL LETTER A WITH TILDE +0x85 U+00c4 # LATIN CAPITAL LETTER A WITH DIAERESIS +0x86 U+00c5 # LATIN CAPITAL LETTER A WITH RING +0x87 U+00c7 # LATIN CAPITAL LETTER C WITH CEDILLA +0x88 U+00c8 # LATIN CAPITAL LETTER E WITH GRAVE +0x89 U+00c9 # LATIN CAPITAL LETTER E WITH ACUTE +0x8a U+00ca # LATIN CAPITAL LETTER E WITH CIRCUMFLEX +0x8b U+00cb # LATIN CAPITAL LETTER E WITH DIAERESIS +0x8c U+00cc # LATIN CAPITAL LETTER I WITH GRAVE +0x8d U+00cd # LATIN CAPITAL LETTER I WITH ACUTE +0x8e U+00ce # LATIN CAPITAL LETTER I WITH CIRCUMFLEX +0x8f U+00cf # LATIN CAPITAL LETTER I WITH DIAERESIS +0x90 U+00d0 # LATIN CAPITAL LETTER ETH +0x91 U+00d1 # LATIN CAPITAL LETTER N WITH TILDE +0x92 U+00d2 # LATIN CAPITAL LETTER O WITH GRAVE +0x93 U+00d3 # LATIN CAPITAL LETTER O WITH ACUTE +0x94 U+00d4 # LATIN CAPITAL LETTER O WITH CIRCUMFLEX +0x95 U+00d5 # LATIN CAPITAL LETTER O WITH TILDE +0x96 U+00d6 # LATIN CAPITAL LETTER O WITH DIAERESIS +0x97 U+00d9 # LATIN CAPITAL LETTER U WITH GRAVE +0x98 U+00da # LATIN CAPITAL LETTER U WITH ACUTE +0x99 U+00db # LATIN CAPITAL LETTER U WITH CIRCUMFLEX +0x9a U+00dc # LATIN CAPITAL LETTER U WITH DIAERESIS +0x9b U+00dd # LATIN CAPITAL LETTER Y WITH ACUTE +0x9c U+00de # LATIN CAPITAL LETTER THORN +0x9d U+00b5 # MICRO SIGN +0x9e U+00d7 # MULTIPLICATION SIGN +0x9f U+00f7 # DIVISION SIGN +0xa0 U+00a9 # COPYRIGHT SIGN +0xa1 U+00a1 # INVERTED EXCLAMATION MARK +0xa2 U+00a2 # CENT SIGN +0xa3 U+00a3 # POUND SIGN +0xa4 U+2044 # FRACTION SLASH +0xa5 U+00a5 # YEN SIGN +0xa6 U+0192 # LATIN SMALL LETTER F WITH HOOK +0xa7 U+00a7 # SECTION SIGN +0xa8 U+00a4 # CURRENCY SIGN +0xa9 U+2019 # RIGHT SINGLE QUOTATION MARK +0xaa U+201c # LEFT DOUBLE QUOTATION MARK +0xab U+00ab # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xac U+2039 # SINGLE LEFT-POINTING ANGLE QUOTATION MARK +0xad U+203a # SINGLE RIGHT-POINTING ANGLE QUOTATION MARK +0xae U+fb01 # LATIN SMALL LIGATURE FI +0xaf U+fb02 # LATIN SMALL LIGATURE FL +0xb0 U+00ae # REGISTERED SIGN +0xb1 U+2013 # EN DASH +0xb2 U+2020 # DAGGER +0xb3 U+2021 # DOUBLE DAGGER +0xb4 U+00b7 # MIDDLE DOT +0xb5 U+00a6 # BROKEN BAR +0xb6 U+00b6 # PILCROW SIGN +0xb7 U+2022 # BULLET +0xb8 U+201a # SINGLE LOW-9 QUOTATION MARK +0xb9 U+201e # DOUBLE LOW-9 QUOTATION MARK +0xba U+201d # RIGHT DOUBLE QUOTATION MARK +0xbb U+00bb # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xbc U+2026 # HORIZONTAL ELLIPSIS +0xbd U+2030 # PER MILLE SIGN +0xbe U+00ac # NOT SIGN +0xbf U+00bf # INVERTED QUESTION MARK +0xc0 U+00b9 # SUPERSCRIPT ONE +0xc1 U+02cb # MODIFIER LETTER GRAVE ACCENT +0xc2 U+00b4 # ACUTE ACCENT +0xc3 U+02c6 # MODIFIER LETTER CIRCUMFLEX ACCENT +0xc4 U+02dc # SMALL TILDE +0xc5 U+00af # MACRON +0xc6 U+02d8 # BREVE +0xc7 U+02d9 # DOT ABOVE +0xc8 U+00a8 # DIAERESIS +0xc9 U+00b2 # SUPERSCRIPT TWO +0xca U+02da # RING ABOVE +0xcb U+00b8 # CEDILLA +0xcc U+00b3 # SUPERSCRIPT THREE +0xcd U+02dd # DOUBLE ACUTE ACCENT +0xce U+02db # OGONEK +0xcf U+02c7 # CARON +0xd0 U+2014 # EM DASH +0xd1 U+00b1 # PLUS-MINUS SIGN +0xd2 U+00bc # VULGAR FRACTION ONE QUARTER +0xd3 U+00bd # VULGAR FRACTION ONE HALF +0xd4 U+00be # VULGAR FRACTION THREE QUARTERS +0xd5 U+00e0 # LATIN SMALL LETTER A WITH GRAVE +0xd6 U+00e1 # LATIN SMALL LETTER A WITH ACUTE +0xd7 U+00e2 # LATIN SMALL LETTER A WITH CIRCUMFLEX +0xd8 U+00e3 # LATIN SMALL LETTER A WITH TILDE +0xd9 U+00e4 # LATIN SMALL LETTER A WITH DIAERESIS +0xda U+00e5 # LATIN SMALL LETTER A WITH RING ABOVE +0xdb U+00e7 # LATIN SMALL LETTER C WITH CEDILLA +0xdc U+00e8 # LATIN SMALL LETTER E WITH GRAVE +0xdd U+00e9 # LATIN SMALL LETTER E WITH ACUTE +0xde U+00ea # LATIN SMALL LETTER E WITH CIRCUMFLEX +0xdf U+00eb # LATIN SMALL LETTER E WITH DIAERESIS +0xe0 U+00ec # LATIN SMALL LETTER I WITH GRAVE +0xe1 U+00c6 # LATIN CAPITAL LETTER AE +0xe2 U+00ed # LATIN SMALL LETTER I WITH ACUTE +0xe3 U+00aa # FEMININE ORDINAL INDICATOR +0xe4 U+00ee # LATIN SMALL LETTER I WITH CIRCUMFLEX +0xe5 U+00ef # LATIN SMALL LETTER I WITH DIAERESIS +0xe6 U+00f0 # LATIN SMALL LETTER ETH +0xe7 U+00f1 # LATIN SMALL LETTER N WITH TILDE +0xe8 U+0141 # LATIN CAPITAL LETTER L WITH STROKE +0xe9 U+00d8 # LATIN CAPITAL LETTER O WITH STROKE +0xea U+0152 # LATIN CAPITAL LIGATURE OE +0xeb U+00ba # MASCULINE ORDINAL INDICATOR +0xec U+00f2 # LATIN SMALL LETTER O WITH GRAVE +0xed U+00f3 # LATIN SMALL LETTER O WITH ACUTE +0xee U+00f4 # LATIN SMALL LETTER O WITH CIRCUMFLEX +0xef U+00f5 # LATIN SMALL LETTER O WITH TILDE +0xf0 U+00f6 # LATIN SMALL LETTER O WITH DIAERESIS +0xf1 U+00e6 # LATIN SMALL LETTER AE +0xf2 U+00f9 # LATIN SMALL LETTER U WITH GRAVE +0xf3 U+00fa # LATIN SMALL LETTER U WITH ACUTE +0xf4 U+00fb # LATIN SMALL LETTER U WITH CIRCUMFLEX +0xf5 U+0131 # LATIN SMALL LETTER DOTLESS I +0xf6 U+00fc # LATIN SMALL LETTER U WITH DIAERESIS +0xf7 U+00fd # LATIN SMALL LETTER Y WITH ACUTE +0xf8 U+0142 # LATIN SMALL LETTER L WITH STROKE +0xf9 U+00f8 # LATIN SMALL LETTER O WITH STROKE +0xfa U+0153 # LATIN SMALL LIGATURE OE +0xfb U+00df # LATIN SMALL LETTER SHARP S +0xfc U+00fe # LATIN SMALL LETTER THORN +0xfd U+00ff # LATIN SMALL LETTER Y WITH DIAERESIS +#0xfe U+fffd # .notdef, REPLACEMENT CHARACTER +#0xff U+fffd # .notdef, REPLACEMENT CHARACTER + diff --git a/src/chrtrans/pt154_uni.tbl b/src/chrtrans/pt154_uni.tbl new file mode 100644 index 0000000..0bacb52 --- /dev/null +++ b/src/chrtrans/pt154_uni.tbl @@ -0,0 +1,174 @@ +Mptcp154 +# +OCyrillic-Asian (PT154) +# +C1540 + +##### +# +# Charset aliases: +# csPTCP154 +# PT154 +# CP154 +# Cyrillic-Asian +# +# Suitability for use in MIME text: +# Yes +# +# ISO 10646 equivalency table: +# Format: Three tab-separated columns +# Column #1 is the Paratype CP154 code (in hex) +# Column #2 is the Unicode (in hex as 0xXXXX) +# Column #3 is the Unicode name (follows a comment sign, '#') +# +# The entries are in Paratype CP154 order +# +##### + +0x20-0x7e idem +# +0x80 U+0496 # CYRILLIC CAPITAL LETTER ZHE WITH DESCENDER +0x81 U+0492 # CYRILLIC CAPITAL LETTER GHE WITH STROKE +0x82 U+04EE # CYRILLIC CAPITAL LETTER U WITH MACRON +0x83 U+0493 # CYRILLIC SMALL LETTER GHE WITH STROKE +0x84 U+201E # DOUBLE LOW-9 QUOTATION MARK +0x85 U+2026 # HORIZONTAL ELLIPSIS +0x86 U+04B6 # CYRILLIC CAPITAL LETTER CHE WITH DESCENDER +0x87 U+04AE # CYRILLIC CAPITAL LETTER STRAIGHT U +0x88 U+04B2 # CYRILLIC CAPITAL LETTER HA WITH DESCENDER +0x89 U+04AF # CYRILLIC SMALL LETTER STRAIGHT U +0x8a U+04A0 # CYRILLIC CAPITAL LETTER BASHKIR KA +0x8b U+04E2 # CYRILLIC CAPITAL LETTER I WITH MACRON +0x8c U+04A2 # CYRILLIC CAPITAL LETTER EN WITH DESCENDER +0x8d U+049A # CYRILLIC CAPITAL LETTER KA WITH DESCENDER +0x8e U+04BA # CYRILLIC CAPITAL LETTER SHHA +0x8f U+04B8 # CYRILLIC CAPITAL LETTER CHE WITH VERTICAL STROKE +0x90 U+0497 # CYRILLIC SMALL LETTER ZHE WITH DESCENDER +0x91 U+2018 # LEFT SINGLE QUOTATION MARK +0x92 U+2019 # RIGHT SINGLE QUOTATION MARK +0x93 U+201C # LEFT DOUBLE QUOTATION MARK +0x94 U+201D # RIGHT DOUBLE QUOTATION MARK +0x95 U+2022 # BULLET +0x96 U+2013 # EN DASH +0x97 U+2014 # EM DASH +0x98 U+04B3 # CYRILLIC SMALL LETTER HA WITH DESCENDER +0x99 U+04B7 # CYRILLIC SMALL LETTER CHE WITH DESCENDER +0x9a U+04A1 # CYRILLIC SMALL LETTER BASHKIR KA +0x9b U+04E3 # CYRILLIC SMALL LETTER I WITH MACRON +0x9c U+04A3 # CYRILLIC SMALL LETTER EN WITH DESCENDER +0x9d U+049B # CYRILLIC SMALL LETTER KA WITH DESCENDER +0x9e U+04BB # CYRILLIC SMALL LETTER SHHA +0x9f U+04B9 # CYRILLIC SMALL LETTER CHE WITH VERTICAL STROKE +0xa0 U+00A0 # NO-BREAK SPACE +0xa1 U+040E # CYRILLIC CAPITAL LETTER SHORT U (Byelorussian) +0xa2 U+045E # CYRILLIC SMALL LETTER SHORT U (Byelorussian) +0xa3 U+0408 # CYRILLIC CAPITAL LETTER JE +0xa4 U+04E8 # CYRILLIC CAPITAL LETTER BARRED O +0xa5 U+0498 # CYRILLIC CAPITAL LETTER ZE WITH DESCENDER +0xa6 U+04B0 # CYRILLIC CAPITAL LETTER STRAIGHT U WITH STROKE +0xa7 U+00A7 # SECTION SIGN +0xa8 U+0401 # CYRILLIC CAPITAL LETTER IO +0xa9 U+00A9 # COPYRIGHT SIGN +0xaa U+04D8 # CYRILLIC CAPITAL LETTER SCHWA +0xab U+00AB # LEFT-POINTING DOUBLE ANGLE QUOTATION MARK +0xac U+00AC # NOT SIGN +0xad U+04EF # CYRILLIC SMALL LETTER U WITH MACRON +0xae U+00AE # REGISTERED SIGN +0xaf U+049C # CYRILLIC CAPITAL LETTER KA WITH VERTICAL STROKE +0xb0 U+00B0 # DEGREE SIGN +0xb1 U+04B1 # CYRILLIC SMALL LETTER STRAIGHT U WITH STROKE +0xb2 U+0406 # CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I +0xb3 U+0456 # CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I +0xb4 U+0499 # CYRILLIC SMALL LETTER ZE WITH DESCENDER +0xb5 U+04E9 # CYRILLIC SMALL LETTER BARRED O +0xb6 U+00B6 # PILCROW SIGN +0xb7 U+00B7 # MIDDLE DOT +0xb8 U+0451 # CYRILLIC SMALL LETTER IO +0xb9 U+2116 # NUMERO SIGN +0xba U+04D9 # CYRILLIC SMALL LETTER SCHWA +0xbb U+00BB # RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK +0xbc U+0458 # CYRILLIC SMALL LETTER JE +0xbd U+04AA # CYRILLIC CAPITAL LETTER ES WITH DESCENDER +0xbe U+04AB # CYRILLIC SMALL LETTER ES WITH DESCENDER +0xbf U+049D # CYRILLIC SMALL LETTER KA WITH VERTICAL STROKE +0xc0 U+0410 # CYRILLIC CAPITAL LETTER A +0xc1 U+0411 # CYRILLIC CAPITAL LETTER BE +0xc2 U+0412 # CYRILLIC CAPITAL LETTER VE +0xc3 U+0413 # CYRILLIC CAPITAL LETTER GHE +0xc4 U+0414 # CYRILLIC CAPITAL LETTER DE +0xc5 U+0415 # CYRILLIC CAPITAL LETTER IE +0xc6 U+0416 # CYRILLIC CAPITAL LETTER ZHE +0xc7 U+0417 # CYRILLIC CAPITAL LETTER ZE +0xc8 U+0418 # CYRILLIC CAPITAL LETTER I +0xc9 U+0419 # CYRILLIC CAPITAL LETTER SHORT I +0xca U+041A # CYRILLIC CAPITAL LETTER KA +0xcb U+041B # CYRILLIC CAPITAL LETTER EL +0xcc U+041C # CYRILLIC CAPITAL LETTER EM +0xcd U+041D # CYRILLIC CAPITAL LETTER EN +0xce U+041E # CYRILLIC CAPITAL LETTER O +0xcf U+041F # CYRILLIC CAPITAL LETTER PE +0xd0 U+0420 # CYRILLIC CAPITAL LETTER ER +0xd1 U+0421 # CYRILLIC CAPITAL LETTER ES +0xd2 U+0422 # CYRILLIC CAPITAL LETTER TE +0xd3 U+0423 # CYRILLIC CAPITAL LETTER U +0xd4 U+0424 # CYRILLIC CAPITAL LETTER EF +0xd5 U+0425 # CYRILLIC CAPITAL LETTER HA +0xd6 U+0426 # CYRILLIC CAPITAL LETTER TSE +0xd7 U+0427 # CYRILLIC CAPITAL LETTER CHE +0xd8 U+0428 # CYRILLIC CAPITAL LETTER SHA +0xd9 U+0429 # CYRILLIC CAPITAL LETTER SHCHA +0xda U+042A # CYRILLIC CAPITAL LETTER HARD SIGN +0xdb U+042B # CYRILLIC CAPITAL LETTER YERU +0xdc U+042C # CYRILLIC CAPITAL LETTER SOFT SIGN +0xdd U+042D # CYRILLIC CAPITAL LETTER E +0xde U+042E # CYRILLIC CAPITAL LETTER YU +0xdf U+042F # CYRILLIC CAPITAL LETTER YA +0xe0 U+0430 # CYRILLIC SMALL LETTER A +0xe1 U+0431 # CYRILLIC SMALL LETTER BE +0xe2 U+0432 # CYRILLIC SMALL LETTER VE +0xe3 U+0433 # CYRILLIC SMALL LETTER GHE +0xe4 U+0434 # CYRILLIC SMALL LETTER DE +0xe5 U+0435 # CYRILLIC SMALL LETTER IE +0xe6 U+0436 # CYRILLIC SMALL LETTER ZHE +0xe7 U+0437 # CYRILLIC SMALL LETTER ZE +0xe8 U+0438 # CYRILLIC SMALL LETTER I +0xe9 U+0439 # CYRILLIC SMALL LETTER SHORT I +0xea U+043A # CYRILLIC SMALL LETTER KA +0xeb U+043B # CYRILLIC SMALL LETTER EL +0xec U+043C # CYRILLIC SMALL LETTER EM +0xed U+043D # CYRILLIC SMALL LETTER EN +0xee U+043E # CYRILLIC SMALL LETTER O +0xef U+043F # CYRILLIC SMALL LETTER PE +0xf0 U+0440 # CYRILLIC SMALL LETTER ER +0xf1 U+0441 # CYRILLIC SMALL LETTER ES +0xf2 U+0442 # CYRILLIC SMALL LETTER TE +0xf3 U+0443 # CYRILLIC SMALL LETTER U +0xf4 U+0444 # CYRILLIC SMALL LETTER EF +0xf5 U+0445 # CYRILLIC SMALL LETTER HA +0xf6 U+0446 # CYRILLIC SMALL LETTER TSE +0xf7 U+0447 # CYRILLIC SMALL LETTER CHE +0xf8 U+0448 # CYRILLIC SMALL LETTER SHA +0xf9 U+0449 # CYRILLIC SMALL LETTER SHCHA +0xfa U+044A # CYRILLIC SMALL LETTER HARD SIGN +0xfb U+044B # CYRILLIC SMALL LETTER YERU +0xfc U+044C # CYRILLIC SMALL LETTER SOFT SIGN +0xfd U+044D # CYRILLIC SMALL LETTER E +0xfe U+044E # CYRILLIC SMALL LETTER YU +0xff U+044F # CYRILLIC SMALL LETTER YA + +##### +# +# Additional information: +# This charset based on CP1251 with added asian cyrillic symbols. +# +# Person & email address to contact for further information: +# Alexander Uskov +# InternetDataCenter of KazakhTelecom. +# e-mail: auskov@idc.kz +# +# Intended usage: +# COMMON +# +# (record created 2002-09-27) +# +##### diff --git a/src/chrtrans/rfc_suni.tbl b/src/chrtrans/rfc_suni.tbl new file mode 100644 index 0000000..65fa17e --- /dev/null +++ b/src/chrtrans/rfc_suni.tbl @@ -0,0 +1,1958 @@ +# Generated from the mnemonic file found under ftp://dkuug.dk/i18n/ +# then hand-tweaked +# perl -n -e \ +# 'if (s|<([^ \t]+)>\s+<U([\dA-Z]{4})>\s.*$|U+\L\2\E:\1|) {s|/?(.)|\1|g&&print}'\ +# mnemonic,ds + +#The MIME name of this charset. +Mmnemonic+ascii+0 + +#Name as a Display Charset (used on Options screen) +ORFC 1345 w/o Intro + +# Don't fall back to default table for unicode -> 8bit +Fallback NO + +U+0020:SP +U+0021:! +U+0022:" +U+0023:Nb +U+0024:DO +U+0025:% +U+0026:& +U+0027:' +U+0028:( +U+0029:) +U+002a:* +U+002b:+ +U+002c:, +U+002d:- +U+002e:. +U+002f:/ +U+0030:0 +U+0031:1 +U+0032:2 +U+0033:3 +U+0034:4 +U+0035:5 +U+0036:6 +U+0037:7 +U+0038:8 +U+0039:9 +U+003a:: +U+003b:; +U+003c:< +U+003d:= +U+003e:> +U+003f:? +U+0040:At +U+0041:A +U+0042:B +U+0043:C +U+0044:D +U+0045:E +U+0046:F +U+0047:G +U+0048:H +U+0049:I +U+004a:J +U+004b:K +U+004c:L +U+004d:M +U+004e:N +U+004f:O +U+0050:P +U+0051:Q +U+0052:R +U+0053:S +U+0054:T +U+0055:U +U+0056:V +U+0057:W +U+0058:X +U+0059:Y +U+005a:Z +U+005b:<( +U+005c:// +U+005d:)> +U+005e:'> +U+005f:_ +U+0060:'! +U+0061:a +U+0062:b +U+0063:c +U+0064:d +U+0065:e +U+0066:f +U+0067:g +U+0068:h +U+0069:i +U+006a:j +U+006b:k +U+006c:l +U+006d:m +U+006e:n +U+006f:o +U+0070:p +U+0071:q +U+0072:r +U+0073:s +U+0074:t +U+0075:u +U+0076:v +U+0077:w +U+0078:x +U+0079:y +U+007a:z +U+007b:(! +U+007c:!! +U+007d:!) +U+007e:'? +U+00a0:NS +U+00a1:!I +U+00a2:Ct +U+00a3:Pd +U+00a4:Cu +U+00a5:Ye +U+00a6:BB +U+00a7:SE +U+00a8:': +U+00a9:Co +U+00aa:-a +U+00ab:<< +U+00ac:NO +U+00ad:-- +U+00ae:Rg +U+00af:'m +U+00b0:DG +U+00b1:+- +U+00b2:2S +U+00b3:3S +U+00b4:'' +U+00b5:My +U+00b6:PI +U+00b7:.M +U+00b8:', +U+00b9:1S +U+00ba:-o +U+00bb:>> +U+00bc:14 +U+00bd:12 +U+00be:34 +U+00bf:?I +U+00c0:A! +U+00c1:A' +U+00c2:A> +U+00c3:A? +U+00c4:A: +U+00c5:AA +U+00c6:AE +U+00c7:C, +U+00c8:E! +U+00c9:E' +U+00ca:E> +U+00cb:E: +U+00cc:I! +U+00cd:I' +U+00ce:I> +U+00cf:I: +U+00d0:D- +U+00d1:N? +U+00d2:O! +U+00d3:O' +U+00d4:O> +U+00d5:O? +U+00d6:O: +U+00d7:*X +U+00d8:O/ +U+00d9:U! +U+00da:U' +U+00db:U> +U+00dc:U: +U+00dd:Y' +U+00de:TH +U+00df:ss +U+00e0:a! +U+00e1:a' +U+00e2:a> +U+00e3:a? +U+00e4:a: +U+00e5:aa +U+00e6:ae +U+00e7:c, +U+00e8:e! +U+00e9:e' +U+00ea:e> +U+00eb:e: +U+00ec:i! +U+00ed:i' +U+00ee:i> +U+00ef:i: +U+00f0:d- +U+00f1:n? +U+00f2:o! +U+00f3:o' +U+00f4:o> +U+00f5:o? +U+00f6:o: +U+00f7:-: +U+00f8:o/ +U+00f9:u! +U+00fa:u' +U+00fb:u> +U+00fc:u: +U+00fd:y' +U+00fe:th +U+00ff:y: +U+0100:A- +U+0101:a- +U+0102:A( +U+0103:a( +U+0104:A; +U+0105:a; +U+0106:C' +U+0107:c' +U+0108:C> +U+0109:c> +U+010a:C. +U+010b:c. +U+010c:C< +U+010d:c< +U+010e:D< +U+010f:d< +U+0110:D/ +U+0111:d/ +U+0112:E- +U+0113:e- +U+0114:E( +U+0115:e( +U+0116:E. +U+0117:e. +U+0118:E; +U+0119:e; +U+011a:E< +U+011b:e< +U+011c:G> +U+011d:g> +U+011e:G( +U+011f:g( +U+0120:G. +U+0121:g. +U+0122:G, +U+0123:g, +U+0124:H> +U+0125:h> +U+0126:H/ +U+0127:h/ +U+0128:I? +U+0129:i? +U+012a:I- +U+012b:i- +U+012c:I( +U+012d:i( +U+012e:I; +U+012f:i; +U+0130:I. +U+0131:i. +U+0132:IJ +U+0133:ij +U+0134:J> +U+0135:j> +U+0136:K, +U+0137:k, +U+0138:kk +U+0139:L' +U+013a:l' +U+013b:L, +U+013c:l, +U+013d:L< +U+013e:l< +U+013f:L. +U+0140:l. +U+0141:L/ +U+0142:l/ +U+0143:N' +U+0144:n' +U+0145:N, +U+0146:n, +U+0147:N< +U+0148:n< +U+0149:'n +U+014a:NG +U+014b:ng +U+014c:O- +U+014d:o- +U+014e:O( +U+014f:o( +U+0150:O" +U+0151:o" +U+0152:OE +U+0153:oe +U+0154:R' +U+0155:r' +U+0156:R, +U+0157:r, +U+0158:R< +U+0159:r< +U+015a:S' +U+015b:s' +U+015c:S> +U+015d:s> +U+015e:S, +U+015f:s, +U+0160:S< +U+0161:s< +U+0162:T, +U+0163:t, +U+0164:T< +U+0165:t< +U+0166:T/ +U+0167:t/ +U+0168:U? +U+0169:u? +U+016a:U- +U+016b:u- +U+016c:U( +U+016d:u( +U+016e:U0 +U+016f:u0 +U+0170:U" +U+0171:u" +U+0172:U; +U+0173:u; +U+0174:W> +U+0175:w> +U+0176:Y> +U+0177:y> +U+0178:Y: +U+0179:Z' +U+017a:z' +U+017b:Z. +U+017c:z. +U+017d:Z< +U+017e:z< +U+017f:s1 +U+0187:C2 +U+0188:c2 +U+0191:F2 +U+0192:f2 +U+0198:K2 +U+0199:k2 +U+01a0:O9 +U+01a1:o9 +U+01a2:OI +U+01a3:oi +U+01a6:yr +U+01af:U9 +U+01b0:u9 +U+01b5:Z/ +U+01b6:z/ +U+01b7:ED +U+01cd:A< +U+01ce:a< +U+01cf:I< +U+01d0:i< +U+01d1:O< +U+01d2:o< +U+01d3:U< +U+01d4:u< +U+01d5:U:- +U+01d6:u:- +U+01d7:U:' +U+01d8:u:' +U+01d9:U:< +U+01da:u:< +U+01db:U:! +U+01dc:u:! +U+01de:A1 +U+01df:a1 +U+01e0:A7 +U+01e1:a7 +U+01e2:A3 +U+01e3:a3 +U+01e4:G/ +U+01e5:g/ +U+01e6:G< +U+01e7:g< +U+01e8:K< +U+01e9:k< +U+01ea:O; +U+01eb:o; +U+01ec:O1 +U+01ed:o1 +U+01ee:EZ +U+01ef:ez +U+01f0:j< +U+01f4:G' +U+01f5:g' +U+01fa:AA' +U+01fb:aa' +U+01fc:AE' +U+01fd:ae' +U+01fe:O/' +U+01ff:o/' +U+0200:A!! +U+0201:a!! +U+0202:A) +U+0203:a) +U+0204:E!! +U+0205:e!! +U+0206:E) +U+0207:e) +U+0208:I!! +U+0209:i!! +U+020a:I) +U+020b:i) +U+020c:O!! +U+020d:o!! +U+020e:O) +U+020f:o) +U+0210:R!! +U+0211:r!! +U+0212:R) +U+0213:r) +U+0214:U!! +U+0215:u!! +U+0216:U) +U+0217:u) +U+0292:ed +U+02bb:;S +U+02c6:1> +U+02c7:'< +U+02c9:1- +U+02cb:1! +U+02d8:'( +U+02d9:'. +U+02da:'0 +U+02db:'; +U+02dc:1? +U+02dd:'" +U+0374:'G +U+0375:,G +U+037a:j3 +U+037e:?% +U+0384:'* +U+0385:'% +U+0386:A% +U+0387:.* +U+0388:E% +U+0389:Y% +U+038a:I% +U+038c:O% +U+038e:U% +U+038f:W% +U+0390:i3 +U+0391:A* +U+0392:B* +U+0393:G* +U+0394:D* +U+0395:E* +U+0396:Z* +U+0397:Y* +U+0398:H* +U+0399:I* +U+039a:K* +U+039b:L* +U+039c:M* +U+039d:N* +U+039e:C* +U+039f:O* +U+03a0:P* +U+03a1:R* +U+03a3:S* +U+03a4:T* +U+03a5:U* +U+03a6:F* +U+03a7:X* +U+03a8:Q* +U+03a9:W* +U+03aa:J* +U+03ab:V* +U+03ac:a% +U+03ad:e% +U+03ae:y% +U+03af:i% +U+03b0:u3 +U+03b1:a* +U+03b2:b* +U+03b3:g* +U+03b4:d* +U+03b5:e* +U+03b6:z* +U+03b7:y* +U+03b8:h* +U+03b9:i* +U+03ba:k* +U+03bb:l* +U+03bc:m* +U+03bd:n* +U+03be:c* +U+03bf:o* +U+03c0:p* +U+03c1:r* +U+03c2:*s +U+03c3:s* +U+03c4:t* +U+03c5:u* +U+03c6:f* +U+03c7:x* +U+03c8:q* +U+03c9:w* +U+03ca:j* +U+03cb:v* +U+03cc:o% +U+03cd:u% +U+03ce:w% +U+03d0:b3 +U+03da:T3 +U+03db:t3 +U+03dc:M3 +U+03dd:m3 +U+03de:K3 +U+03df:k3 +U+03e0:P3 +U+03e1:p3 +U+0401:IO +U+0402:D% +U+0403:G% +U+0404:IE +U+0405:DS +U+0406:II +U+0407:YI +U+0408:J% +U+0409:LJ +U+040a:NJ +U+040b:Ts +U+040c:KJ +U+040e:V% +U+040f:DZ +U+0410:A= +U+0411:B= +U+0412:V= +U+0413:G= +U+0414:D= +U+0415:E= +U+0416:Z% +U+0417:Z= +U+0418:I= +U+0419:J= +U+041a:K= +U+041b:L= +U+041c:M= +U+041d:N= +U+041e:O= +U+041f:P= +U+0420:R= +U+0421:S= +U+0422:T= +U+0423:U= +U+0424:F= +U+0425:H= +U+0426:C= +U+0427:C% +U+0428:S% +U+0429:Sc +U+042a:=" +U+042b:Y= +U+042c:%" +U+042d:JE +U+042e:JU +U+042f:JA +U+0430:a= +U+0431:b= +U+0432:v= +U+0433:g= +U+0434:d= +U+0435:e= +U+0436:z% +U+0437:z= +U+0438:i= +U+0439:j= +U+043a:k= +U+043b:l= +U+043c:m= +U+043d:n= +U+043e:o= +U+043f:p= +U+0440:r= +U+0441:s= +U+0442:t= +U+0443:u= +U+0444:f= +U+0445:h= +U+0446:c= +U+0447:c% +U+0448:s% +U+0449:sc +U+044a:=' +U+044b:y= +U+044c:%' +U+044d:je +U+044e:ju +U+044f:ja +U+0451:io +U+0452:d% +U+0453:g% +U+0454:ie +U+0455:ds +U+0456:ii +U+0457:yi +U+0458:j% +U+0459:lj +U+045a:nj +U+045b:ts +U+045c:kj +U+045e:v% +U+045f:dz +U+0462:Y3 +U+0463:y3 +U+046a:O3 +U+046b:o3 +U+0472:F3 +U+0473:f3 +U+0474:V3 +U+0475:v3 +U+0480:C3 +U+0481:c3 +U+0490:G3 +U+0491:g3 +U+05d0:A+ +U+05d1:B+ +U+05d2:G+ +U+05d3:D+ +U+05d4:H+ +U+05d5:W+ +U+05d6:Z+ +U+05d7:X+ +U+05d8:Tj +U+05d9:J+ +U+05da:K% +U+05db:K+ +U+05dc:L+ +U+05dd:M% +U+05de:M+ +U+05df:N% +U+05e0:N+ +U+05e1:S+ +U+05e2:E+ +U+05e3:P% +U+05e4:P+ +U+05e5:Zj +U+05e6:ZJ +U+05e7:Q+ +U+05e8:R+ +U+05e9:Sh +U+05ea:T+ +U+060c:,+ +U+061b:;+ +U+061f:?+ +U+0621:H' +U+0622:aM +U+0623:aH +U+0624:wH +U+0625:ah +U+0626:yH +U+0627:a+ +U+0628:b+ +U+0629:tm +U+062a:t+ +U+062b:tk +U+062c:g+ +U+062d:hk +U+062e:x+ +U+062f:d+ +U+0630:dk +U+0631:r+ +U+0632:z+ +U+0633:s+ +U+0634:sn +U+0635:c+ +U+0636:dd +U+0637:tj +U+0638:zH +U+0639:e+ +U+063a:i+ +U+0640:++ +U+0641:f+ +U+0642:q+ +U+0643:k+ +U+0644:l+ +U+0645:m+ +U+0646:n+ +U+0647:h+ +U+0648:w+ +U+0649:j+ +U+064a:y+ +U+064b::+ +U+064c:"+ +U+064d:=+ +U+064e:/+ +U+064f:'+ +U+0650:1+ +U+0651:3+ +U+0652:0+ +U+0660:0a +U+0661:1a +U+0662:2a +U+0663:3a +U+0664:4a +U+0665:5a +U+0666:6a +U+0667:7a +U+0668:8a +U+0669:9a +U+0670:aS +U+067e:p+ +U+0681:hH +U+0686:tc +U+0698:zj +U+06a4:v+ +U+06af:gf +U+1e00:A-0 +U+1e01:a-0 +U+1e02:B. +U+1e03:b. +U+1e04:B-. +U+1e05:b-. +U+1e06:B_ +U+1e07:b_ +U+1e08:C,' +U+1e09:c,' +U+1e0a:D. +U+1e0b:d. +U+1e0c:D-. +U+1e0d:d-. +U+1e0e:D_ +U+1e0f:d_ +U+1e10:D, +U+1e11:d, +U+1e12:D-> +U+1e13:d-> +U+1e14:E-! +U+1e15:e-! +U+1e16:E-' +U+1e17:e-' +U+1e18:E-> +U+1e19:e-> +U+1e1a:E-? +U+1e1b:e-? +U+1e1c:E,( +U+1e1d:e,( +U+1e1e:F. +U+1e1f:f. +U+1e20:G- +U+1e21:g- +U+1e22:H. +U+1e23:h. +U+1e24:H-. +U+1e25:h-. +U+1e26:H: +U+1e27:h: +U+1e28:H, +U+1e29:h, +U+1e2a:H-( +U+1e2b:h-( +U+1e2c:I-? +U+1e2d:i-? +U+1e2e:I:' +U+1e2f:i:' +U+1e30:K' +U+1e31:k' +U+1e32:K-. +U+1e33:k-. +U+1e34:K_ +U+1e35:k_ +U+1e36:L-. +U+1e37:l-. +U+1e38:L--. +U+1e39:l--. +U+1e3a:L_ +U+1e3b:l_ +U+1e3c:L-> +U+1e3d:l-> +U+1e3e:M' +U+1e3f:m' +U+1e40:M. +U+1e41:m. +U+1e42:M-. +U+1e43:m-. +U+1e44:N. +U+1e45:n. +U+1e46:N-. +U+1e47:n-. +U+1e48:N_ +U+1e49:n_ +U+1e4a:N-> +U+1e4b:n-> +U+1e4c:O?' +U+1e4d:o?' +U+1e4e:O?: +U+1e4f:o?: +U+1e50:O-! +U+1e51:o-! +U+1e52:O-' +U+1e53:o-' +U+1e54:P' +U+1e55:p' +U+1e56:P. +U+1e57:p. +U+1e58:R. +U+1e59:r. +U+1e5a:R-. +U+1e5b:r-. +U+1e5c:R--. +U+1e5d:r--. +U+1e5e:R_ +U+1e5f:r_ +U+1e60:S. +U+1e61:s. +U+1e62:S-. +U+1e63:s-. +U+1e64:S'. +U+1e65:s'. +U+1e66:S<. +U+1e67:s<. +U+1e68:S.-. +U+1e69:s.-. +U+1e6a:T. +U+1e6b:t. +U+1e6c:T-. +U+1e6d:t-. +U+1e6e:T_ +U+1e6f:t_ +U+1e70:T-> +U+1e71:t-> +U+1e72:U--: +U+1e73:u--: +U+1e74:U-? +U+1e75:u-? +U+1e76:U-> +U+1e77:u-> +U+1e78:U?' +U+1e79:u?' +U+1e7a:U-: +U+1e7b:u-: +U+1e7c:V? +U+1e7d:v? +U+1e7e:V-. +U+1e7f:v-. +U+1e80:W! +U+1e81:w! +U+1e82:W' +U+1e83:w' +U+1e84:W: +U+1e85:w: +U+1e86:W. +U+1e87:w. +U+1e88:W-. +U+1e89:w-. +U+1e8a:X. +U+1e8b:x. +U+1e8c:X: +U+1e8d:x: +U+1e8e:Y. +U+1e8f:y. +U+1e90:Z> +U+1e91:z> +U+1e92:Z-. +U+1e93:z-. +U+1e94:Z_ +U+1e95:z_ +U+1e96:h_ +U+1e97:t: +U+1e98:w0 +U+1e99:y0 +U+1ea0:A-. +U+1ea1:a-. +U+1ea2:A2 +U+1ea3:a2 +U+1ea4:A>' +U+1ea5:a>' +U+1ea6:A>! +U+1ea7:a>! +U+1ea8:A>2 +U+1ea9:a>2 +U+1eaa:A>? +U+1eab:a>? +U+1eac:A>-. +U+1ead:a>-. +U+1eae:A(' +U+1eaf:a(' +U+1eb0:A(! +U+1eb1:a(! +U+1eb2:A(2 +U+1eb3:a(2 +U+1eb4:A(? +U+1eb5:a(? +U+1eb6:A(-. +U+1eb7:a(-. +U+1eb8:E-. +U+1eb9:e-. +U+1eba:E2 +U+1ebb:e2 +U+1ebc:E? +U+1ebd:e? +U+1ebe:E>' +U+1ebf:e>' +U+1ec0:E>! +U+1ec1:e>! +U+1ec2:E>2 +U+1ec3:e>2 +U+1ec4:E>? +U+1ec5:e>? +U+1ec6:E>-. +U+1ec7:e>-. +U+1ec8:I2 +U+1ec9:i2 +U+1eca:I-. +U+1ecb:i-. +U+1ecc:O-. +U+1ecd:o-. +U+1ece:O2 +U+1ecf:o2 +U+1ed0:O>' +U+1ed1:o>' +U+1ed2:O>! +U+1ed3:o>! +U+1ed4:O>2 +U+1ed5:o>2 +U+1ed6:O>? +U+1ed7:o>? +U+1ed8:O>-. +U+1ed9:o>-. +U+1eda:O9' +U+1edb:o9' +U+1edc:O9! +U+1edd:o9! +U+1ede:O92 +U+1edf:o92 +U+1ee0:O9? +U+1ee1:o9? +U+1ee2:O9-. +U+1ee3:o9-. +U+1ee4:U-. +U+1ee5:u-. +U+1ee6:U2 +U+1ee7:u2 +U+1ee8:U9' +U+1ee9:u9' +U+1eea:U9! +U+1eeb:u9! +U+1eec:U92 +U+1eed:u92 +U+1eee:U9? +U+1eef:u9? +U+1ef0:U9-. +U+1ef1:u9-. +U+1ef2:Y! +U+1ef3:y! +U+1ef4:Y-. +U+1ef5:y-. +U+1ef6:Y2 +U+1ef7:y2 +U+1ef8:Y? +U+1ef9:y? +U+1fbf:,, +U+1fc0:?* +U+1fc1:?: +U+1fcd:,! +U+1fce:,' +U+1fcf:?, +U+1fdd:;! +U+1fde:;' +U+1fdf:?; +U+1fed:!: +U+1fef:!* +U+1ffe:;; +U+2002:1N +U+2003:1M +U+2004:3M +U+2005:4M +U+2006:6M +U+200e:LR +U+200f:RL +U+2009:1T +U+200a:1H +U+2010:-1 +U+2013:-N +U+2014:-M +U+2015:-3 +U+2016:!2 +U+2017:=2 +U+2018:'6 +U+2019:'9 +U+201a:.9 +U+201b:9' +U+201c:"6 +U+201d:"9 +U+201e::9 +U+201f:9" +U+2020:/- +U+2021:/= +U+2022:Sb +U+2025:.. +U+2026:.3 +U+2030:%0 +U+2032:1' +U+2033:2' +U+2034:3' +U+2035:1" +U+2036:2" +U+2037:3" +U+2038:Ca +U+2039:<1 +U+203a:>1 +U+203b::X +U+203c:!*2 +U+203e:'- +U+2044:/f +U+2070:0S +U+2074:4S +U+2075:5S +U+2076:6S +U+2077:7S +U+2078:8S +U+2079:9S +U+207a:+S +U+207b:-S +U+207c:=S +U+207d:(S +U+207e:)S +U+207f:nS +U+2080:0s +U+2081:1s +U+2082:2s +U+2083:3s +U+2084:4s +U+2085:5s +U+2086:6s +U+2087:7s +U+2088:8s +U+2089:9s +U+208a:+s +U+208b:-s +U+208c:=s +U+208d:(s +U+208e:)s +U+20a3:Ff +U+20a4:Li +U+20a7:Pt +U+20a9:W= +U+2103:oC +U+2105:co +U+2109:oF +U+2116:N0 +U+2117:PO +U+211e:Rx +U+2120:SM +U+2122:TM +U+2126:Om +U+212b:AO +U+2153:13 +U+2154:23 +U+2155:15 +U+2156:25 +U+2157:35 +U+2158:45 +U+2159:16 +U+215a:56 +U+215b:18 +U+215c:38 +U+215d:58 +U+215e:78 +U+2160:1R +U+2161:2R +U+2162:3R +U+2163:4R +U+2164:5R +U+2165:6R +U+2166:7R +U+2167:8R +U+2168:9R +U+2169:aR +U+216a:bR +U+216b:cR +U+216c:50R +U+216d:100R +U+216e:500R +U+216f:1000R +U+2170:1r +U+2171:2r +U+2172:3r +U+2173:4r +U+2174:5r +U+2175:6r +U+2176:7r +U+2177:8r +U+2178:9r +U+2179:ar +U+217a:br +U+217b:cr +U+217c:50r +U+217d:100r +U+217e:500r +U+217f:1000r +U+2180:1000RCD +U+2181:5000R +U+2182:10000R +U+2190:<- +U+2191:-! +U+2192:-> +U+2193:-v +U+2194:<> +U+2195:UD +U+2196:<!! +U+2197://> +U+2198:!!> +U+2199:<// +U+21a8:UD- +U+21c0:>V +U+21d0:<= +U+21d2:=> +U+21d4:== +U+2200:FA +U+2202:dP +U+2203:TE +U+2205:/0 +U+2206:DE +U+2207:NB +U+2208:(- +U+220b:-) +U+220f:*P +U+2211:+Z +U+2212:-2 +U+2213:-+ +U+2214:.+ +U+2217:*- +U+2218:Ob +U+2219:sb +U+221a:RT +U+221d:0( +U+221e:00 +U+221f:-L +U+2220:-V +U+2225:PP +U+2227:AN +U+2228:OR +U+2229:(U +U+222a:)U +U+222b:In +U+222c:DI +U+222e:Io +U+2234:.: +U+2235::. +U+2236::R +U+2237::: +U+223c:?1 +U+223e:CG +U+2243:?- +U+2245:?= +U+2248:?2 +U+224c:=? +U+2253:HI +U+2260:!= +U+2261:=3 +U+2264:=< +U+2265:>= +U+226a:<* +U+226b:*> +U+226e:!< +U+226f:!> +U+2282:(C +U+2283:)C +U+2286:(_ +U+2287:)_ +U+2299:0. +U+229a:02 +U+22a5:-T +U+22c5:.P +U+22ee::3 +U+2302:Eh +U+2308:<7 +U+2309:>7 +U+230a:7< +U+230b:7> +U+2310:NI +U+2312:(A +U+2315:TR +U+2318:88 +U+2320:Iu +U+2321:Il +U+2329:</ +U+232a:/> +U+2423:Vs +U+2440:1h +U+2441:3h +U+2442:2h +U+2443:4h +U+2446:1j +U+2447:2j +U+2448:3j +U+2449:4j +U+2460:1-o +U+2461:2-o +U+2462:3-o +U+2463:4-o +U+2464:5-o +U+2465:6-o +U+2466:7-o +U+2467:8-o +U+2468:9-o +U+2469:10-o +U+246a:11-o +U+246b:12-o +U+246c:13-o +U+246d:14-o +U+246e:15-o +U+246f:16-o +U+2470:17-o +U+2471:18-o +U+2472:19-o +U+2473:20-o +U+2474:(1) +U+2475:(2) +U+2476:(3) +U+2477:(4) +U+2478:(5) +U+2479:(6) +U+247a:(7) +U+247b:(8) +U+247c:(9) +U+247d:(10) +U+247e:(11) +U+247f:(12) +U+2480:(13) +U+2481:(14) +U+2482:(15) +U+2483:(16) +U+2484:(17) +U+2485:(18) +U+2486:(19) +U+2487:(20) +U+2488:1. +U+2489:2. +U+248a:3. +U+248b:4. +U+248c:5. +U+248d:6. +U+248e:7. +U+248f:8. +U+2490:9. +U+2491:10. +U+2492:11. +U+2493:12. +U+2494:13. +U+2495:14. +U+2496:15. +U+2497:16. +U+2498:17. +U+2499:18. +U+249a:19. +U+249b:20. +U+249c:(a) +U+249d:(b) +U+249e:(c) +U+249f:(d) +U+24a0:(e) +U+24a1:(f) +U+24a2:(g) +U+24a3:(h) +U+24a4:(i) +U+24a5:(j) +U+24a6:(k) +U+24a7:(l) +U+24a8:(m) +U+24a9:(n) +U+24aa:(o) +U+24ab:(p) +U+24ac:(q) +U+24ad:(r) +U+24ae:(s) +U+24af:(t) +U+24b0:(u) +U+24b1:(v) +U+24b2:(w) +U+24b3:(x) +U+24b4:(y) +U+24b5:(z) +U+24b6:A-o +U+24b7:B-o +U+24b8:C-o +U+24b9:D-o +U+24ba:E-o +U+24bb:F-o +U+24bc:G-o +U+24bd:H-o +U+24be:I-o +U+24bf:J-o +U+24c0:K-o +U+24c1:L-o +U+24c2:M-o +U+24c3:N-o +U+24c4:O-o +U+24c5:P-o +U+24c6:Q-o +U+24c7:R-o +U+24c8:S-o +U+24c9:T-o +U+24ca:U-o +U+24cb:V-o +U+24cc:W-o +U+24cd:X-o +U+24ce:Y-o +U+24cf:Z-o +U+24d0:a-o +U+24d1:b-o +U+24d2:c-o +U+24d3:d-o +U+24d4:e-o +U+24d5:f-o +U+24d6:g-o +U+24d7:h-o +U+24d8:i-o +U+24d9:j-o +U+24da:k-o +U+24db:l-o +U+24dc:m-o +U+24dd:n-o +U+24de:o-o +U+24df:p-o +U+24e0:q-o +U+24e1:r-o +U+24e2:s-o +U+24e3:t-o +U+24e4:u-o +U+24e5:v-o +U+24e6:w-o +U+24e7:x-o +U+24e8:y-o +U+24e9:z-o +U+24ea:0-o +U+2500:hh +U+2501:HH- +U+2502:vv +U+2503:VV- +U+2504:3- +U+2505:3_ +U+2506:3! +U+2507:3/ +U+2508:4- +U+2509:4_ +U+250a:4! +U+250b:4/ +U+250c:dr +U+250d:dR- +U+250e:Dr- +U+250f:DR- +U+2510:dl +U+2511:dL- +U+2512:Dl- +U+2513:LD- +U+2514:ur +U+2515:uR- +U+2516:Ur- +U+2517:UR- +U+2518:ul +U+2519:uL- +U+251a:Ul- +U+251b:UL- +U+251c:vr +U+251d:vR- +U+251e:Udr +U+251f:uDr +U+2520:Vr- +U+2521:UdR +U+2522:uDR +U+2523:VR- +U+2524:vl +U+2525:vL- +U+2526:Udl +U+2527:uDl +U+2528:Vl- +U+2529:UdL +U+252a:uDL +U+252b:VL- +U+252c:dh +U+252d:dLr +U+252e:dlR +U+252f:dH- +U+2530:Dh- +U+2531:DLr +U+2532:DlR +U+2533:DH- +U+2534:uh +U+2535:uLr +U+2536:ulR +U+2537:uH- +U+2538:Uh- +U+2539:ULr +U+253a:UlR +U+253b:UH- +U+253c:vh +U+253d:vLr +U+253e:vlR +U+253f:vH- +U+2540:Udh +U+2541:uDh +U+2542:Vh- +U+2543:UdLr +U+2544:UdlR +U+2545:uDLr +U+2546:uDlR +U+2547:UdH +U+2548:uDH +U+2549:VLr +U+254a:VlR +U+254b:VH- +U+2550:HH +U+2551:VV +U+2552:dR +U+2553:Dr +U+2554:DR +U+2555:dL +U+2556:Dl +U+2557:LD +U+2558:uR +U+2559:Ur +U+255a:UR +U+255b:uL +U+255c:Ul +U+255d:UL +U+255e:vR +U+255f:Vr +U+2560:VR +U+2561:vL +U+2562:Vl +U+2563:VL +U+2564:dH +U+2565:Dh +U+2566:DH +U+2567:uH +U+2568:Uh +U+2569:UH +U+256a:vH +U+256b:Vh +U+256c:VH +U+2571:FD +U+2572:BD +U+2580:TB +U+2584:LB +U+2588:FB +U+258c:lB +U+2590:RB +U+2591:.S +U+2592::S +U+2593:?S +U+25a0:fS +U+25a1:OS +U+25a2:RO +U+25a3:Rr +U+25a4:RF +U+25a5:RY +U+25a6:RH +U+25a7:RZ +U+25a8:RK +U+25a9:RX +U+25aa:sB +U+25ac:SR +U+25ad:Or +U+25b2:UT +U+25b3:uT +U+25b7:Tr +U+25ba:PR +U+25bc:Dt +U+25bd:dT +U+25c1:Tl +U+25c4:PL +U+25c6:Db +U+25c7:Dw +U+25ca:LZ +U+25cb:0m +U+25ce:0o +U+25cf:0M +U+25d0:0L +U+25d1:0R +U+25d8:Sn +U+25d9:Ic +U+25e2:Fd +U+25e3:Bd +U+25ef:Ci +U+2605:*2 +U+2606:*1 +U+260e:TEL +U+260f:tel +U+261c:<H +U+261e:>H +U+263a:0u +U+263b:0U +U+263c:SU +U+2640:Fm +U+2642:Ml +U+2660:cS +U+2661:cH +U+2662:cD +U+2663:cC +U+2664:cS- +U+2665:cH- +U+2666:cD- +U+2667:cC- +U+2669:Md +U+266a:M8 +U+266b:M2 +U+266c:M16 +U+266d:Mb +U+266e:Mx +U+266f:MX +U+2713:OK +U+2717:XX +U+2720:-X +U+3000:IS +U+3001:,_ +U+3002:._ +U+3003:+" +U+3004:JIS +U+3005:*_ +U+3006:;_ +U+3007:0_ +U+300a:<+ +U+300b:>+ +U+300c:<' +U+300d:>' +U+300e:<" +U+300f:>" +U+3010:(" +U+3011:)" +U+3012:=T +U+3013:=_ +U+3014:(' +U+3015:)' +U+3016:(I +U+3017:)I +U+301c:-? +U+3020:=T:) +U+3041:A5 +U+3042:a5 +U+3043:I5 +U+3044:i5 +U+3045:U5 +U+3046:u5 +U+3047:E5 +U+3048:e5 +U+3049:O5 +U+304a:o5 +U+304b:ka +U+304c:ga +U+304d:ki +U+304e:gi +U+304f:ku +U+3050:gu +U+3051:ke +U+3052:ge +U+3053:ko +U+3054:go +U+3055:sa +U+3056:za +U+3057:si +U+3058:zi +U+3059:su +U+305a:zu +U+305b:se +U+305c:ze +U+305d:so +U+305e:zo +U+305f:ta +U+3060:da +U+3061:ti +U+3062:di +U+3063:tU +U+3064:tu +U+3065:du +U+3066:te +U+3067:de +U+3068:to +U+3069:do +U+306a:na +U+306b:ni +U+306c:nu +U+306d:ne +U+306e:no +U+306f:ha +U+3070:ba +U+3071:pa +U+3072:hi +U+3073:bi +U+3074:pi +U+3075:hu +U+3076:bu +U+3077:pu +U+3078:he +U+3079:be +U+307a:pe +U+307b:ho +U+307c:bo +U+307d:po +U+307e:ma +U+307f:mi +U+3080:mu +U+3081:me +U+3082:mo +U+3083:yA +U+3084:ya +U+3085:yU +U+3086:yu +U+3087:yO +U+3088:yo +U+3089:ra +U+308a:ri +U+308b:ru +U+308c:re +U+308d:ro +U+308e:wA +U+308f:wa +U+3090:wi +U+3091:we +U+3092:wo +U+3093:n5 +U+3094:vu +U+309b:"5 +U+309c:05 +U+309d:*5 +U+309e:+5 +U+30a1:a6 +U+30a2:A6 +U+30a3:i6 +U+30a4:I6 +U+30a5:u6 +U+30a6:U6 +U+30a7:e6 +U+30a8:E6 +U+30a9:o6 +U+30aa:O6 +U+30ab:Ka +U+30ac:Ga +U+30ad:Ki +U+30ae:Gi +U+30af:Ku +U+30b0:Gu +U+30b1:Ke +U+30b2:Ge +U+30b3:Ko +U+30b4:Go +U+30b5:Sa +U+30b6:Za +U+30b7:Si +U+30b8:Zi +U+30b9:Su +U+30ba:Zu +U+30bb:Se +U+30bc:Ze +U+30bd:So +U+30be:Zo +U+30bf:Ta +U+30c0:Da +U+30c1:Ti +U+30c2:Di +U+30c3:TU +U+30c4:Tu +U+30c5:Du +U+30c6:Te +U+30c7:De +U+30c8:To +U+30c9:Do +U+30ca:Na +U+30cb:Ni +U+30cc:Nu +U+30cd:Ne +U+30ce:No +U+30cf:Ha +U+30d0:Ba +U+30d1:Pa +U+30d2:Hi +U+30d3:Bi +U+30d4:Pi +U+30d5:Hu +U+30d6:Bu +U+30d7:Pu +U+30d8:He +U+30d9:Be +U+30da:Pe +U+30db:Ho +U+30dc:Bo +U+30dd:Po +U+30de:Ma +U+30df:Mi +U+30e0:Mu +U+30e1:Me +U+30e2:Mo +U+30e3:YA +U+30e4:Ya +U+30e5:YU +U+30e6:Yu +U+30e7:YO +U+30e8:Yo +U+30e9:Ra +U+30ea:Ri +U+30eb:Ru +U+30ec:Re +U+30ed:Ro +U+30ee:WA +U+30ef:Wa +U+30f0:Wi +U+30f1:We +U+30f2:Wo +U+30f3:N6 +U+30f4:Vu +U+30f5:KA +U+30f6:KE +U+30f7:Va +U+30f8:Vi +U+30f9:Ve +U+30fa:Vo +U+30fb:.6 +U+30fc:-6 +U+30fd:*6 +U+30fe:+6 +U+3105:b4 +U+3106:p4 +U+3107:m4 +U+3108:f4 +U+3109:d4 +U+310a:t4 +U+310b:n4 +U+310c:l4 +U+310d:g4 +U+310e:k4 +U+310f:h4 +U+3110:j4 +U+3111:q4 +U+3112:x4 +U+3113:zh +U+3114:ch +U+3115:sh +U+3116:r4 +U+3117:z4 +U+3118:c4 +U+3119:s4 +U+311a:a4 +U+311b:o4 +U+311c:e4 +U+311d:eh4 +U+311e:ai +U+311f:ei +U+3120:au +U+3121:ou +U+3122:an +U+3123:en +U+3124:aN +U+3125:eN +U+3126:er +U+3127:i4 +U+3128:u4 +U+3129:iu +U+312a:v4 +U+312b:nG +U+312c:gn +U+321c:(JU) +U+3220:1c +U+3221:2c +U+3222:3c +U+3223:4c +U+3224:5c +U+3225:6c +U+3226:7c +U+3227:8c +U+3228:9c +U+3229:10c +U+327f:KSC +U+33c2:am +U+33d8:pm +U+fb00:ff +U+fb01:fi +U+fb02:fl +U+fb03:ffi +U+fb04:ffl +U+fb05:St +U+fb06:st +U+fe7d:3+; +U+fe82:aM. +U+fe84:aH. +U+fe88:ah. +U+fe8d:a+- +U+fe8e:a+. +U+fe8f:b+- +U+fe90:b+. +U+fe91:b+, +U+fe92:b+; +U+fe93:tm- +U+fe94:tm. +U+fe95:t+- +U+fe96:t+. +U+fe97:t+, +U+fe98:t+; +U+fe99:tk- +U+fe9a:tk. +U+fe9b:tk, +U+fe9c:tk; +U+fe9d:g+- +U+fe9e:g+. +U+fe9f:g+, +U+fea0:g+; +U+fea1:hk- +U+fea2:hk. +U+fea3:hk, +U+fea4:hk; +U+fea5:x+- +U+fea6:x+. +U+fea7:x+, +U+fea8:x+; +U+fea9:d+- +U+feaa:d+. +U+feab:dk- +U+feac:dk. +U+fead:r+- +U+feae:r+. +U+feaf:z+- +U+feb0:z+. +U+feb1:s+- +U+feb2:s+. +U+feb3:s+, +U+feb4:s+; +U+feb5:sn- +U+feb6:sn. +U+feb7:sn, +U+feb8:sn; +U+feb9:c+- +U+feba:c+. +U+febb:c+, +U+febc:c+; +U+febd:dd- +U+febe:dd. +U+febf:dd, +U+fec0:dd; +U+fec1:tj- +U+fec2:tj. +U+fec3:tj, +U+fec4:tj; +U+fec5:zH- +U+fec6:zH. +U+fec7:zH, +U+fec8:zH; +U+fec9:e+- +U+feca:e+. +U+fecb:e+, +U+fecc:e+; +U+fecd:i+- +U+fece:i+. +U+fecf:i+, +U+fed0:i+; +U+fed1:f+- +U+fed2:f+. +U+fed3:f+, +U+fed4:f+; +U+fed5:q+- +U+fed6:q+. +U+fed7:q+, +U+fed8:q+; +U+fed9:k+- +U+feda:k+. +U+fedb:k+, +U+fedc:k+; +U+fedd:l+- +U+fede:l+. +U+fedf:l+, +U+fee0:l+; +U+fee1:m+- +U+fee2:m+. +U+fee3:m+, +U+fee4:m+; +U+fee5:n+- +U+fee6:n+. +U+fee7:n+, +U+fee8:n+; +U+fee9:h+- +U+feea:h+. +U+feeb:h+, +U+feec:h+; +U+feed:w+- +U+feee:w+. +U+feef:j+- +U+fef0:j+. +U+fef1:y+- +U+fef2:y+. +U+fef3:y+, +U+fef4:y+; +U+fef5:lM- +U+fef6:lM. +U+fef7:lH- +U+fef8:lH. +U+fef9:lh- +U+fefa:lh. +U+fefb:la- +U+fefc:la. +U+0000:NU +U+0001:SH +U+0002:SX +U+0003:EX +U+0004:ET +U+0005:EQ +U+0006:AK +U+0007:BL +U+0008:BS +U+0009:HT +U+000a:LF +U+000b:VT +U+000c:FF +U+000d:CR +U+000e:SO +U+000f:SI +U+0010:DL +U+0011:D1 +U+0012:D2 +U+0013:D3 +U+0014:D4 +U+0015:NK +U+0016:SY +U+0017:EB +U+0018:CN +U+0019:EM +U+001a:SB +U+001b:EC +U+001c:FS +U+001d:GS +U+001e:RS +U+001f:US +U+007f:DT +U+0080:PA +U+0081:HO +U+0082:BH +U+0083:NH +U+0084:IN +U+0085:NL +U+0086:SA +U+0087:ES +U+0088:HS +U+0089:HJ +U+008a:VS +U+008b:PD +U+008c:PU +U+008d:RI +U+008e:S2 +U+008f:S3 +U+0090:DC +U+0091:P1 +U+0092:P2 +U+0093:TS +U+0094:CC +U+0095:MW +U+0096:SG +U+0097:EG +U+0098:SS +U+0099:GC +U+009a:SC +U+009b:CI +U+009c:ST +U+009d:OC +U+009e:PM +U+009f:AC +# Characters in Private Use Area (e000-f8ff) do not have ussigned numbers +# according Unicode 2.0 diff --git a/src/chrtrans/rot13_kb.h b/src/chrtrans/rot13_kb.h new file mode 100644 index 0000000..1df5f32 --- /dev/null +++ b/src/chrtrans/rot13_kb.h @@ -0,0 +1,22 @@ +static LYKbLayout_t kb_layout_rot13[128] = +{ + 0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007, /* 00..07 */ + 0x0008, 0x0009, 0x000A, 0x000B, 0x000C, 0x000D, 0x000E, 0x000F, /* 08..0F */ + 0x0010, 0x0011, 0x0012, 0x0013, 0x0014, 0x0015, 0x0016, 0x0017, /* 10..17 */ + 0x0018, 0x0019, 0x001A, 0x001B, 0x001C, 0x001D, 0x001E, 0x001F, /* 18..1F */ + + 0x0020, 0x0021, 0x0022, 0x0023, 0x0024, 0x0025, 0x0026, 0x0027, /* 20..27 */ + 0x0028, 0x0029, 0x002A, 0x002B, 0x002C, 0x002D, 0x002E, 0x002F, /* 28..2F */ + 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 0x0037, /* 30..37 */ + 0x0038, 0x0039, 0x003A, 0x003B, 0x003C, 0x003D, 0x003E, 0x003F, /* 38..3F */ + + 0x0000, 0x004e, 0x004f, 0x0050, 0x0051, 0x0052, 0x0053, 0x0054, /* 40..48 */ + 0x0055, 0x0056, 0x0057, 0x0058, 0x0059, 0x005a, 0x0041, 0x0042, /* 40..4F */ + 0x0043, 0x0044, 0x0045, 0x0046, 0x0047, 0x0048, 0x0049, 0x004a, /* 50..58 */ + 0x004b, 0x004c, 0x004d, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /* 50..5F */ + + 0x0000, 0x006e, 0x006f, 0x0070, 0x0071, 0x0072, 0x0073, 0x0074, /* 60..68 */ + 0x0075, 0x0076, 0x0077, 0x0078, 0x0079, 0x007a, 0x0061, 0x0062, /* 60..6F */ + 0x0063, 0x0064, 0x0065, 0x0066, 0x0067, 0x0068, 0x0069, 0x006a, /* 70..78 */ + 0x006b, 0x006c, 0x006d, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000 /* 70..7F */ +}; diff --git a/src/chrtrans/utf8_uni.tbl b/src/chrtrans/utf8_uni.tbl new file mode 100644 index 0000000..88ad492 --- /dev/null +++ b/src/chrtrans/utf8_uni.tbl @@ -0,0 +1,35 @@ +# +# This one is not really much of a "translation table", it mostly just +# tells Lynx that "utf-8" is Unicode/UCS2 encoded in UTF8. Note that +# "unicode-1-1-utf-8" and "utf8" are treated as synonyms. +# +#The MIME name of this charset. +Mutf-8 + +#Name as a Display Charset (used on Options screen) +OUNICODE (UTF-8) + +# Some kind of raw Unicode? +# Use 6 for for really "raw" 16bit UCS-2, 7 for UTF-8, ... + +# most of these codes currently don't make much sense in a *.tbl file, +# but for completeness (from UCDefs.h): +# #define UCT_ENC_7BIT 0 +# #define UCT_ENC_8BIT 1 +# #define UCT_ENC_8859 2 +# #define UCT_ENC_8BIT_C0 3 +# #define UCT_ENC_MAYBE2022 4 +# #define UCT_ENC_CJK 5 +# #define UCT_ENC_16BIT 6 +# #define UCT_ENC_UTF8 7 + +R 7 + +#Shall this become the "default" translation? +#There has to be exactly one table marked as "default". +Default NO + +# Don't fall back to default table for unicode -> 8bit +Fallback NO + +0x20-0x7f idem diff --git a/src/chrtrans/viscii_uni.tbl b/src/chrtrans/viscii_uni.tbl new file mode 100644 index 0000000..617f1e1 --- /dev/null +++ b/src/chrtrans/viscii_uni.tbl @@ -0,0 +1,300 @@ +# +# Unicode mapping table for VISCII 1.1 fonts and charset=viscii, +# described in RFC 1456. +# See also <URL:http://www.trichlor.org/vietstd/report/rep92.htm>, +# also for testing. +# The 6 characters encoded in the C0 control region should not +# be passed through to the terminal but be mapped to VIQR strings. +# THe two changed mappings of MacVISCII are recognized in documents. + +# [convert with makeuctb] +# +#The MIME name of this charset. +Mviscii + +#Name as a Display Charset (used on Options screen). +OVietnamese (VISCII) + +# Special 'enc' flag to signal that some C0 characters are used. +# Tables with R3 should properly map the allowed C0 control chars! +# +# most of these codes currently don't make much sense in a *.tbl file, +# but for completeness (from UCDefs.h): +# #define UCT_ENC_7BIT 0 +# #define UCT_ENC_8BIT 1 +# #define UCT_ENC_8859 2 +# #define UCT_ENC_8BIT_C0 3 +# #define UCT_ENC_MAYBE2022 4 +# #define UCT_ENC_CJK 5 +# #define UCT_ENC_16BIT 6 +# #define UCT_ENC_UTF8 7 + +R 3 + +#0x00 U+0000 +#0x01 U+0001 +#0x03 U+0003 +#0x04 U+0004 +#0x07 U+0007 +#0x08 U+0008 +0x09 U+0009 +0x0a U+000a +#0x0b U+000b +0x0c U+000c +U+000c " " +0x0d U+000d +#0x0e U+000e +#0x0f U+000f +#0x10 U+0010 +#0x11 U+0011 +#0x12 U+0012 +#0x13 U+0013 +#0x15 U+0015 +#0x16 U+0016 +#0x17 U+0017 +#0x18 U+0018 +0x1a U+001a +U+001a:^Z +#0x1b U+001b +#0x1c U+001c +#0x1d U+001d +#0x1f U+001f +#0x20 U+0020 +#0x21 U+0021 +#0x22 U+0022 +#0x23 U+0023 +#0x24 U+0024 +#0x25 U+0025 +#0x26 U+0026 +#0x27 U+0027 +#0x28 U+0028 +#0x29 U+0029 +#0x2a U+002a +#0x2b U+002b +#0x2c U+002c +#0x2d U+002d +#0x2e U+002e +#0x2f U+002f +#0x30 U+0030 +#0x31 U+0031 +#0x32 U+0032 +#0x33 U+0033 +#0x34 U+0034 +#0x35 U+0035 +#0x36 U+0036 +#0x37 U+0037 +#0x38 U+0038 +#0x39 U+0039 +#0x3a U+003a +#0x3b U+003b +#0x3c U+003c +#0x3d U+003d +#0x3e U+003e +#0x3f U+003f +#0x40 U+0040 +#0x41 U+0041 +#0x42 U+0042 +#0x43 U+0043 +#0x44 U+0044 +#0x45 U+0045 +#0x46 U+0046 +#0x47 U+0047 +#0x48 U+0048 +#0x49 U+0049 +#0x4a U+004a +#0x4b U+004b +#0x4c U+004c +#0x4d U+004d +#0x4e U+004e +#0x4f U+004f +#0x50 U+0050 +#0x51 U+0051 +#0x52 U+0052 +#0x53 U+0053 +#0x54 U+0054 +#0x55 U+0055 +#0x56 U+0056 +#0x57 U+0057 +#0x58 U+0058 +#0x59 U+0059 +#0x5a U+005a +#0x5b U+005b +#0x5c U+005c +#0x5d U+005d +#0x5e U+005e +#0x5f U+005f +#0x60 U+0060 +#0x61 U+0061 +#0x62 U+0062 +#0x63 U+0063 +#0x64 U+0064 +#0x65 U+0065 +#0x66 U+0066 +#0x67 U+0067 +#0x68 U+0068 +#0x69 U+0069 +#0x6a U+006a +#0x6b U+006b +#0x6c U+006c +#0x6d U+006d +#0x6e U+006e +#0x6f U+006f +#0x70 U+0070 +#0x71 U+0071 +#0x72 U+0072 +#0x73 U+0073 +#0x74 U+0074 +#0x75 U+0075 +#0x76 U+0076 +#0x77 U+0077 +#0x78 U+0078 +#0x79 U+0079 +#0x7a U+007a +#0x7b U+007b +#0x7c U+007c +#0x7d U+007d +#0x7e U+007e +#0x7f U+007f +0xc0 U+00c0 +0xc1 U+00c1 +0xc2 U+00c2 +0xc3 U+00c3 +0xc8 U+00c8 +0xc9 U+00c9 +0xca U+00ca +0xcc U+00cc +0xcd U+00cd +0xd2 U+00d2 +0xd3 U+00d3 +0xd4 U+00d4 +0xa0 U+00d5 +0xd9 U+00d9 +0xda U+00da +0xdd U+00dd +0xe0 U+00e0 +0xe1 U+00e1 +0xe2 U+00e2 +0xe3 U+00e3 +0xe8 U+00e8 +0xe9 U+00e9 +0xea U+00ea +0xec U+00ec +0xed U+00ed +0xf2 U+00f2 +0xf3 U+00f3 +0xf4 U+00f4 +0xf5 U+00f5 +0xf9 U+00f9 +0xfa U+00fa +0xfd U+00fd +0xc5 U+0102 +0xe5 U+0103 +0xd0 U+0110 +0xf0 U+0111 U+00f0 # "edh" is similar enough to map it here +0xce U+0128 +0xee U+0129 +0x9d U+0168 +0xfb U+0169 +0xb4 U+01a0 +0xbd U+01a1 +0xbf U+01af +0xdf U+01b0 +0x80 U+1ea0 +0xd5 U+1ea1 +0xc4 U+1ea2 +0xe4 U+1ea3 +0x84 U+1ea4 +0xa4 U+1ea5 +0x85 U+1ea6 +0xa5 U+1ea7 +0x86 U+1ea8 +0xa6 U+1ea9 +0x06 U+1eaa +U+1eaa "\302~" # A with circumflex (same code as in iso-8859-1) and tilde +0xe7 U+1eab +0x87 U+1eac +0xa7 U+1ead +0x81 U+1eae +0xa1 U+1eaf +0x82 U+1eb0 +0xa2 U+1eb1 +0x02 U+1eb2 +U+1eb2:A(? +0xc6 U+1eb3 +0x05 U+1eb4 +U+1eb4:A(~ +0xc7 U+1eb5 +0x83 U+1eb6 +0xa3 U+1eb7 +0x89 U+1eb8 +0xa9 U+1eb9 +0xcb U+1eba +0xeb U+1ebb +0x88 U+1ebc +0xa8 U+1ebd +0x8a U+1ebe +0xaa U+1ebf +0x8b U+1ec0 +0xab U+1ec1 +0x8c U+1ec2 +0xac U+1ec3 +0x8d U+1ec4 +0xad U+1ec5 +0x8e U+1ec6 +0xae U+1ec7 +0x9b U+1ec8 +0xef U+1ec9 +0x98 U+1eca +0xb8 U+1ecb +0x9a U+1ecc +0xf7 U+1ecd +0x99 U+1ece +0xf6 U+1ecf +0x8f U+1ed0 +0xaf U+1ed1 +0x90 U+1ed2 +0xb0 U+1ed3 +0x91 U+1ed4 +0xb1 U+1ed5 +0x92 U+1ed6 +0xb2 U+1ed7 +0x93 U+1ed8 +0xb5 U+1ed9 +0x95 U+1eda +0xbe U+1edb +0x96 U+1edc +0xb6 U+1edd +0x97 U+1ede +0xb7 U+1edf +0xb3 U+1ee0 +0xde U+1ee1 +0x94 U+1ee2 +0xfe U+1ee3 +0x9e U+1ee4 +0xf8 U+1ee5 +0x9c U+1ee6 +0xfc U+1ee7 +0xba U+1ee8 +0xd1 U+1ee9 +0xbb U+1eea +0xd7 U+1eeb +0xbc U+1eec +0xd8 U+1eed +0xff U+1eee +0xe6 U+1eef +0xb9 U+1ef0 +0xf1 U+1ef1 +0x9f U+1ef2 +0xcf U+1ef3 +0x18 U+1ef4 # MacVISCII +0x1e U+1ef4 +U+1ef4:Y. +0xdc U+1ef5 +0x17 U+1ef6 # MacVISCII +0x14 U+1ef6 +U+1ef6:Y? +0xd6 U+1ef7 +0x19 U+1ef8 +U+1ef8:Y~ +0xdb U+1ef9 + diff --git a/src/chrtrans/yawerty_kb.h b/src/chrtrans/yawerty_kb.h new file mode 100644 index 0000000..8301c81 --- /dev/null +++ b/src/chrtrans/yawerty_kb.h @@ -0,0 +1,22 @@ +static LYKbLayout_t kb_layout_yawerty[128] = +{ + 0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007, /* 00..07 */ + 0x0008, 0x0009, 0x000A, 0x000B, 0x000C, 0x000D, 0x000E, 0x000F, /* 08..0F */ + 0x0010, 0x0011, 0x0012, 0x0013, 0x0014, 0x0015, 0x0016, 0x0017, /* 10..17 */ + 0x0018, 0x0019, 0x001A, 0x001B, 0x001C, 0x001D, 0x001E, 0x001F, /* 18..1F */ + + 0x0020, 0x0021, 0x0022, 0x0023, 0x0024, 0x0025, 0x0026, 0x0027, /* 20..27 */ + 0x0028, 0x0029, 0x002A, 0x002B, 0x002C, 0x002D, 0x002E, 0x002F, /* 28..2F */ + 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 0x0037, /* 30..37 */ + 0x0038, 0x0039, 0x003A, 0x003B, 0x003C, 0x003D, 0x003E, 0x003F, /* 38..3F */ + + 0x042e, 0x0410, 0x0411, 0x0426, 0x0414, 0x0415, 0x0424, 0x0413, /* 40..47 */ + 0x0425, 0x0418, 0x0419, 0x041a, 0x041b, 0x041c, 0x041d, 0x041e, /* 48..4F */ + 0x041f, 0x042f, 0x0420, 0x0421, 0x0422, 0x0423, 0x0416, 0x0412, /* 50..57 */ + 0x042c, 0x042b, 0x0417, 0x0448, 0x044d, 0x0449, 0x0447, 0x044a, /* 58..5F */ + + 0x044e, 0x0430, 0x0431, 0x0446, 0x0434, 0x0435, 0x0444, 0x0433, /* 60..67 */ + 0x0445, 0x0438, 0x0439, 0x043a, 0x043b, 0x043c, 0x043d, 0x043e, /* 68..6F */ + 0x043f, 0x044f, 0x0440, 0x0441, 0x0442, 0x0443, 0x0436, 0x0432, /* 70..77 */ + 0x044c, 0x044b, 0x0437, 0x0428, 0x042d, 0x0429, 0x0427, 0x0000 /* 78..7F */ +}; |