Adding upstream version 2.9.0rel.0.upstream/2.9.0rel.0

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

1 files changed, 76 insertions, 0 deletions

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+for some of the RFC 1345 Mnemonic codes are out of date,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.
+