diff options
Diffstat (limited to '')
| -rw-r--r-- | ChangeLog | 13 | ||||
| -rw-r--r-- | INSTALL | 4 | ||||
| -rw-r--r-- | Makefile.in | 2 | ||||
| -rw-r--r-- | NEWS | 39 | ||||
| -rw-r--r-- | README | 21 | ||||
| -rw-r--r-- | carg_parser.c | 110 | ||||
| -rw-r--r-- | carg_parser.h | 14 | ||||
| -rwxr-xr-x | configure | 6 | ||||
| -rw-r--r-- | decoder.c | 35 | ||||
| -rw-r--r-- | decoder.h | 116 | ||||
| -rw-r--r-- | doc/clzip.1 | 21 | ||||
| -rw-r--r-- | doc/clzip.info | 344 | ||||
| -rw-r--r-- | doc/clzip.texi | 326 | ||||
| -rw-r--r-- | encoder.c | 131 | ||||
| -rw-r--r-- | encoder.h | 5 | ||||
| -rw-r--r-- | encoder_base.c | 21 | ||||
| -rw-r--r-- | encoder_base.h | 25 | ||||
| -rw-r--r-- | fast_encoder.c | 23 | ||||
| -rw-r--r-- | fast_encoder.h | 2 | ||||
| -rw-r--r-- | list.c | 63 | ||||
| -rw-r--r-- | lzip.h | 9 | ||||
| -rw-r--r-- | lzip_index.c | 43 | ||||
| -rw-r--r-- | lzip_index.h | 2 | ||||
| -rw-r--r-- | main.c | 173 | ||||
| -rwxr-xr-x | testsuite/check.sh | 17 |
25 files changed, 826 insertions, 739 deletions
@@ -1,3 +1,12 @@ +2022-01-24 Antonio Diaz Diaz <antonio@gnu.org> + + * Version 1.13 released. + * Decompression time has been reduced by 5-12% depending on the file. + * main.c (getnum): Show option name and valid range if error. + * Improve several descriptions in manual, '--help', and man page. + * clzip.texi: Change GNU Texinfo category to 'Compression'. + (Reported by Alfred M. Szmidt). + 2021-01-04 Antonio Diaz Diaz <antonio@gnu.org> * Version 1.12 released. @@ -22,7 +31,7 @@ * main.c (main): Check return value of close( infd ). * main.c: Compile on DOS with DJGPP. * clzip.texi: Improve descriptions of '-0..-9', '-m', and '-s'. - * configure: Accept appending to CFLAGS, 'CFLAGS+=OPTIONS'. + * configure: Accept appending to CFLAGS; 'CFLAGS+=OPTIONS'. * INSTALL: Document use of CFLAGS+='-D __USE_MINGW_ANSI_STDIO'. 2018-02-06 Antonio Diaz Diaz <antonio@gnu.org> @@ -151,7 +160,7 @@ * Translated to C from the C++ source of lzip 1.10. -Copyright (C) 2010-2021 Antonio Diaz Diaz. +Copyright (C) 2010-2022 Antonio Diaz Diaz. This file is a collection of facts, and thus it is not copyrightable, but just in case, you have unlimited permission to copy, distribute, and @@ -1,7 +1,7 @@ Requirements ------------ You will need a C99 compiler. (gcc 3.3.6 or newer is recommended). -I use gcc 6.1.0 and 4.1.2, but the code should compile with any standards +I use gcc 6.1.0 and 3.3.6, but the code should compile with any standards compliant compiler. Gcc is available at http://gcc.gnu.org. @@ -70,7 +70,7 @@ After running 'configure', you can run 'make' and 'make install' as explained above. -Copyright (C) 2010-2021 Antonio Diaz Diaz. +Copyright (C) 2010-2022 Antonio Diaz Diaz. This file is free documentation: you have unlimited permission to copy, distribute, and modify it. diff --git a/Makefile.in b/Makefile.in index ec41f91..5f77295 100644 --- a/Makefile.in +++ b/Makefile.in @@ -21,7 +21,7 @@ objs = carg_parser.o lzip_index.o list.o encoder_base.o encoder.o \ all : $(progname) $(progname) : $(objs) - $(CC) $(LDFLAGS) $(CFLAGS) -o $@ $(objs) + $(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(objs) main.o : main.c $(CC) $(CPPFLAGS) $(CFLAGS) -DPROGVERSION=\"$(pkgversion)\" -c -o $@ $< @@ -1,36 +1,11 @@ -Changes in version 1.12: +Changes in version 1.13: -Clzip now reports an error if a file name is empty (clzip -t ""). +Decompression time has been reduced by 5-12% depending on the file. -Option '-o, --output' now behaves like '-c, --stdout', but sending the -output unconditionally to a file instead of to standard output. See the new -description of '-o' in the manual. This change is backwards compatible only -when (de)compressing from standard input alone. Therefore commands like: - clzip -o foo.lz - bar < foo -must now be split into: - clzip -o foo.lz - < foo - clzip bar -or rewritten as: - clzip - bar < foo > foo.lz +In case of error in a numerical argument to a command line option, clzip +now shows the name of the option and the range of valid values. -When using '-c' or '-o', clzip now checks whether the output is a terminal -only once. +Several descriptions have been improved in manual, '--help', and man page. -Clzip now does not even open the output file if the input file is a terminal. - -The words 'decompressed' and 'compressed' have been replaced with the -shorter 'out' and 'in' in the verbose output when decompressing or testing. - -Option '--list' now reports corruption or truncation of the last header in a -multimenber file specifically instead of showing the generic message "Last -member in input file is truncated or corrupt." - -The commands needed to extract files from a tar.lz archive have been -documented in the manual, in the output of '--help', and in the man page. - -Plzip and tarlz are mentioned in the manual as alternatives for -multiprocessors. - -Several fixes and improvements have been made to the manual. - -9 new test files have been added to the testsuite. +The texinfo category of the manual has been changed from 'Data Compression' +to 'Compression' to match that of gzip. (Reported by Alfred M. Szmidt). @@ -7,13 +7,14 @@ C++ compiler. Lzip is a lossless data compressor with a user interface similar to the one of gzip or bzip2. Lzip uses a simplified form of the 'Lempel-Ziv-Markov -chain-Algorithm' (LZMA) stream format, chosen to maximize safety and -interoperability. Lzip can compress about as fast as gzip (lzip -0) or -compress most files more than bzip2 (lzip -9). Decompression speed is -intermediate between gzip and bzip2. Lzip is better than gzip and bzip2 from -a data recovery perspective. Lzip has been designed, written, and tested -with great care to replace gzip and bzip2 as the standard general-purpose -compressed format for unix-like systems. +chain-Algorithm' (LZMA) stream format and provides a 3 factor integrity +checking to maximize interoperability and optimize safety. Lzip can compress +about as fast as gzip (lzip -0) or compress most files more than bzip2 +(lzip -9). Decompression speed is intermediate between gzip and bzip2. +Lzip is better than gzip and bzip2 from a data recovery perspective. Lzip +has been designed, written, and tested with great care to replace gzip and +bzip2 as the standard general-purpose compressed format for unix-like +systems. For compressing/decompressing large files on multiprocessor machines plzip can be much faster than lzip at the cost of a slightly reduced compression @@ -72,7 +73,7 @@ filename.lz becomes filename filename.tlz becomes filename.tar anyothername becomes anyothername.out -(De)compressing a file is much like copying or moving it; therefore clzip +(De)compressing a file is much like copying or moving it. Therefore clzip preserves the access and modification dates, permissions, and, when possible, ownership of the file just as 'cp -p' does. (If the user ID or the group ID can't be duplicated, the file permission bits S_ISUID and @@ -109,7 +110,7 @@ finding coding sequences of minimum size than the one currently used by lzip could be developed, and the resulting sequence could also be coded using the LZMA coding scheme. -Clzip currently implements two variants of the LZMA algorithm; fast +Clzip currently implements two variants of the LZMA algorithm: fast (used by option '-0') and normal (used by all other compression levels). The high compression of LZMA comes from combining two basic, well-proven @@ -129,7 +130,7 @@ been compressed. Decompressed is used to refer to data which have undergone the process of decompression. -Copyright (C) 2010-2021 Antonio Diaz Diaz. +Copyright (C) 2010-2022 Antonio Diaz Diaz. This file is free documentation: you have unlimited permission to copy, distribute, and modify it. diff --git a/carg_parser.c b/carg_parser.c index d0c05d5..181ba23 100644 --- a/carg_parser.c +++ b/carg_parser.c @@ -1,5 +1,5 @@ /* Arg_parser - POSIX/GNU command line argument parser. (C version) - Copyright (C) 2006-2021 Antonio Diaz Diaz. + Copyright (C) 2006-2022 Antonio Diaz Diaz. This library is free software. Redistribution and use in source and binary forms, with or without modification, are permitted provided @@ -32,10 +32,10 @@ static void * ap_resize_buffer( void * buf, const int min_size ) } -static char push_back_record( struct Arg_parser * const ap, - const int code, const char * const argument ) +static char push_back_record( struct Arg_parser * const ap, const int code, + const char * const long_name, + const char * const argument ) { - const int len = strlen( argument ); struct ap_Record * p; void * tmp = ap_resize_buffer( ap->data, ( ap->data_size + 1 ) * sizeof (struct ap_Record) ); @@ -43,11 +43,29 @@ static char push_back_record( struct Arg_parser * const ap, ap->data = (struct ap_Record *)tmp; p = &(ap->data[ap->data_size]); p->code = code; - p->argument = 0; - tmp = ap_resize_buffer( p->argument, len + 1 ); - if( !tmp ) return 0; - p->argument = (char *)tmp; - strncpy( p->argument, argument, len + 1 ); + if( long_name ) + { + const int len = strlen( long_name ); + p->parsed_name = (char *)malloc( len + 2 + 1 ); + if( !p->parsed_name ) return 0; + p->parsed_name[0] = p->parsed_name[1] = '-'; + strncpy( p->parsed_name + 2, long_name, len + 1 ); + } + else if( code > 0 && code < 256 ) + { + p->parsed_name = (char *)malloc( 2 + 1 ); + if( !p->parsed_name ) return 0; + p->parsed_name[0] = '-'; p->parsed_name[1] = code; p->parsed_name[2] = 0; + } + else p->parsed_name = 0; + if( argument ) + { + const int len = strlen( argument ); + p->argument = (char *)malloc( len + 1 ); + if( !p->argument ) { free( p->parsed_name ); return 0; } + strncpy( p->argument, argument, len + 1 ); + } + else p->argument = 0; ++ap->data_size; return 1; } @@ -68,12 +86,14 @@ static char add_error( struct Arg_parser * const ap, const char * const msg ) static void free_data( struct Arg_parser * const ap ) { int i; - for( i = 0; i < ap->data_size; ++i ) free( ap->data[i].argument ); + for( i = 0; i < ap->data_size; ++i ) + { free( ap->data[i].argument ); free( ap->data[i].parsed_name ); } if( ap->data ) { free( ap->data ); ap->data = 0; } ap->data_size = 0; } +/* Return 0 only if out of memory. */ static char parse_long_option( struct Arg_parser * const ap, const char * const opt, const char * const arg, const struct ap_Option options[], @@ -87,9 +107,10 @@ static char parse_long_option( struct Arg_parser * const ap, /* Test all long options for either exact match or abbreviated matches. */ for( i = 0; options[i].code != 0; ++i ) - if( options[i].name && strncmp( options[i].name, &opt[2], len ) == 0 ) + if( options[i].long_name && + strncmp( options[i].long_name, &opt[2], len ) == 0 ) { - if( strlen( options[i].name ) == len ) /* Exact match found */ + if( strlen( options[i].long_name ) == len ) /* Exact match found */ { index = i; exact = 1; break; } else if( index < 0 ) index = i; /* First nonexact match found */ else if( options[index].code != options[i].code || @@ -117,35 +138,39 @@ static char parse_long_option( struct Arg_parser * const ap, { if( options[index].has_arg == ap_no ) { - add_error( ap, "option '--" ); add_error( ap, options[index].name ); + add_error( ap, "option '--" ); add_error( ap, options[index].long_name ); add_error( ap, "' doesn't allow an argument" ); return 1; } if( options[index].has_arg == ap_yes && !opt[len+3] ) { - add_error( ap, "option '--" ); add_error( ap, options[index].name ); + add_error( ap, "option '--" ); add_error( ap, options[index].long_name ); add_error( ap, "' requires an argument" ); return 1; } - return push_back_record( ap, options[index].code, &opt[len+3] ); + return push_back_record( ap, options[index].code, + options[index].long_name, &opt[len+3] ); } if( options[index].has_arg == ap_yes ) { if( !arg || !arg[0] ) { - add_error( ap, "option '--" ); add_error( ap, options[index].name ); + add_error( ap, "option '--" ); add_error( ap, options[index].long_name ); add_error( ap, "' requires an argument" ); return 1; } ++*argindp; - return push_back_record( ap, options[index].code, arg ); + return push_back_record( ap, options[index].code, + options[index].long_name, arg ); } - return push_back_record( ap, options[index].code, "" ); + return push_back_record( ap, options[index].code, + options[index].long_name, 0 ); } +/* Return 0 only if out of memory. */ static char parse_short_option( struct Arg_parser * const ap, const char * const opt, const char * const arg, const struct ap_Option options[], @@ -156,13 +181,13 @@ static char parse_short_option( struct Arg_parser * const ap, while( cind > 0 ) { int index = -1, i; - const unsigned char code = opt[cind]; + const unsigned char c = opt[cind]; char code_str[2]; - code_str[0] = code; code_str[1] = 0; + code_str[0] = c; code_str[1] = 0; - if( code != 0 ) + if( c != 0 ) for( i = 0; options[i].code; ++i ) - if( code == options[i].code ) + if( c == options[i].code ) { index = i; break; } if( index < 0 ) @@ -176,7 +201,7 @@ static char parse_short_option( struct Arg_parser * const ap, if( options[index].has_arg != ap_no && cind > 0 && opt[cind] ) { - if( !push_back_record( ap, code, &opt[cind] ) ) return 0; + if( !push_back_record( ap, c, 0, &opt[cind] ) ) return 0; ++*argindp; cind = 0; } else if( options[index].has_arg == ap_yes ) @@ -188,9 +213,9 @@ static char parse_short_option( struct Arg_parser * const ap, return 1; } ++*argindp; cind = 0; - if( !push_back_record( ap, code, arg ) ) return 0; + if( !push_back_record( ap, c, 0, arg ) ) return 0; } - else if( !push_back_record( ap, code, "" ) ) return 0; + else if( !push_back_record( ap, c, 0, 0 ) ) return 0; } return 1; } @@ -203,7 +228,7 @@ char ap_init( struct Arg_parser * const ap, const char ** non_options = 0; /* skipped non-options */ int non_options_size = 0; /* number of skipped non-options */ int argind = 1; /* index in argv */ - int i; + char done = 0; /* false until success */ ap->data = 0; ap->error = 0; @@ -223,20 +248,20 @@ char ap_init( struct Arg_parser * const ap, if( ch2 == '-' ) { if( !argv[argind][2] ) { ++argind; break; } /* we found "--" */ - else if( !parse_long_option( ap, opt, arg, options, &argind ) ) return 0; + else if( !parse_long_option( ap, opt, arg, options, &argind ) ) goto out; } - else if( !parse_short_option( ap, opt, arg, options, &argind ) ) return 0; + else if( !parse_short_option( ap, opt, arg, options, &argind ) ) goto out; if( ap->error ) break; } else { if( in_order ) - { if( !push_back_record( ap, 0, argv[argind++] ) ) return 0; } + { if( !push_back_record( ap, 0, 0, argv[argind++] ) ) goto out; } else { void * tmp = ap_resize_buffer( non_options, ( non_options_size + 1 ) * sizeof *non_options ); - if( !tmp ) return 0; + if( !tmp ) goto out; non_options = (const char **)tmp; non_options[non_options_size++] = argv[argind++]; } @@ -245,13 +270,15 @@ char ap_init( struct Arg_parser * const ap, if( ap->error ) free_data( ap ); else { + int i; for( i = 0; i < non_options_size; ++i ) - if( !push_back_record( ap, 0, non_options[i] ) ) return 0; + if( !push_back_record( ap, 0, 0, non_options[i] ) ) goto out; while( argind < argc ) - if( !push_back_record( ap, 0, argv[argind++] ) ) return 0; + if( !push_back_record( ap, 0, 0, argv[argind++] ) ) goto out; } - if( non_options ) free( non_options ); - return 1; + done = 1; +out: if( non_options ) free( non_options ); + return done; } @@ -273,13 +300,20 @@ int ap_arguments( const struct Arg_parser * const ap ) int ap_code( const struct Arg_parser * const ap, const int i ) { - if( i >= 0 && i < ap_arguments( ap ) ) return ap->data[i].code; - else return 0; + if( i < 0 || i >= ap_arguments( ap ) ) return 0; + return ap->data[i].code; + } + + +const char * ap_parsed_name( const struct Arg_parser * const ap, const int i ) + { + if( i < 0 || i >= ap_arguments( ap ) || !ap->data[i].parsed_name ) return ""; + return ap->data[i].parsed_name; } const char * ap_argument( const struct Arg_parser * const ap, const int i ) { - if( i >= 0 && i < ap_arguments( ap ) ) return ap->data[i].argument; - else return ""; + if( i < 0 || i >= ap_arguments( ap ) || !ap->data[i].argument ) return ""; + return ap->data[i].argument; } diff --git a/carg_parser.h b/carg_parser.h index c5f2352..0c64861 100644 --- a/carg_parser.h +++ b/carg_parser.h @@ -1,5 +1,5 @@ /* Arg_parser - POSIX/GNU command line argument parser. (C version) - Copyright (C) 2006-2021 Antonio Diaz Diaz. + Copyright (C) 2006-2022 Antonio Diaz Diaz. This library is free software. Redistribution and use in source and binary forms, with or without modification, are permitted provided @@ -24,9 +24,9 @@ message. 'options' is an array of 'struct ap_Option' terminated by an element - containing a code which is zero. A null name means a short-only - option. A code value outside the unsigned char range means a - long-only option. + containing a code which is zero. A null long_name means a short-only + option. A code value outside the unsigned char range means a long-only + option. Arg_parser normally makes it appear as if all the option arguments were specified before all the non-option arguments for the purposes @@ -50,7 +50,7 @@ enum ap_Has_arg { ap_no, ap_yes, ap_maybe }; struct ap_Option { int code; /* Short option letter or code ( code != 0 ) */ - const char * name; /* Long option name (maybe null) */ + const char * long_name; /* Long option name (maybe null) */ enum ap_Has_arg has_arg; }; @@ -58,6 +58,7 @@ struct ap_Option struct ap_Record { int code; + char * parsed_name; char * argument; }; @@ -86,6 +87,9 @@ int ap_arguments( const struct Arg_parser * const ap ); Else ap_argument( i ) is the option's argument (or empty). */ int ap_code( const struct Arg_parser * const ap, const int i ); +/* Full name of the option parsed (short or long). */ +const char * ap_parsed_name( const struct Arg_parser * const ap, const int i ); + const char * ap_argument( const struct Arg_parser * const ap, const int i ); #ifdef __cplusplus @@ -1,12 +1,12 @@ #! /bin/sh # configure script for Clzip - LZMA lossless data compressor -# Copyright (C) 2010-2021 Antonio Diaz Diaz. +# Copyright (C) 2010-2022 Antonio Diaz Diaz. # # This configure script is free software: you have unlimited permission # to copy, distribute, and modify it. pkgname=clzip -pkgversion=1.12 +pkgversion=1.13 progname=clzip srctrigger=doc/${pkgname}.texi @@ -167,7 +167,7 @@ echo "LDFLAGS = ${LDFLAGS}" rm -f Makefile cat > Makefile << EOF # Makefile for Clzip - LZMA lossless data compressor -# Copyright (C) 2010-2021 Antonio Diaz Diaz. +# Copyright (C) 2010-2022 Antonio Diaz Diaz. # This file was generated automatically by configure. Don't edit. # # This Makefile is free software: you have unlimited permission @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -29,8 +29,8 @@ #include "decoder.h" -/* Returns the number of bytes really read. - If (returned value < size) and (errno == 0), means EOF was reached. +/* Return the number of bytes really read. + If (value returned < size) and (errno == 0), means EOF was reached. */ int readblock( const int fd, uint8_t * const buf, const int size ) { @@ -48,8 +48,8 @@ int readblock( const int fd, uint8_t * const buf, const int size ) } -/* Returns the number of bytes really written. - If (returned value < size), it is always an error. +/* Return the number of bytes really written. + If (value returned < size), it is always an error. */ int writeblock( const int fd, const uint8_t * const buf, const int size ) { @@ -105,8 +105,6 @@ static bool LZd_verify_trailer( struct LZ_decoder * const d, int size = Rd_read_data( d->rdec, trailer, Lt_size ); const unsigned long long data_size = LZd_data_position( d ); const unsigned long long member_size = Rd_member_position( d->rdec ); - unsigned td_crc; - unsigned long long td_size, tm_size; bool error = false; if( size < Lt_size ) @@ -121,7 +119,7 @@ static bool LZd_verify_trailer( struct LZ_decoder * const d, while( size < Lt_size ) trailer[size++] = 0; } - td_crc = Lt_get_data_crc( trailer ); + const unsigned td_crc = Lt_get_data_crc( trailer ); if( td_crc != LZd_crc( d ) ) { error = true; @@ -132,7 +130,7 @@ static bool LZd_verify_trailer( struct LZ_decoder * const d, td_crc, LZd_crc( d ) ); } } - td_size = Lt_get_data_size( trailer ); + const unsigned long long td_size = Lt_get_data_size( trailer ); if( td_size != data_size ) { error = true; @@ -143,7 +141,7 @@ static bool LZd_verify_trailer( struct LZ_decoder * const d, td_size, td_size, data_size, data_size ); } } - tm_size = Lt_get_member_size( trailer ); + const unsigned long long tm_size = Lt_get_member_size( trailer ); if( tm_size != member_size ) { error = true; @@ -213,25 +211,19 @@ int LZd_decode_member( struct LZ_decoder * const d, Rd_load( rdec ); while( !Rd_finished( rdec ) ) { - int len; const int pos_state = LZd_data_position( d ) & pos_state_mask; if( Rd_decode_bit( rdec, &bm_match[state][pos_state] ) == 0 ) /* 1st bit */ { /* literal byte */ Bit_model * const bm = bm_literal[get_lit_state(LZd_peek_prev( d ))]; - if( St_is_char( state ) ) - { - state -= ( state < 4 ) ? state : 3; + if( ( state = St_set_char( state ) ) < 4 ) LZd_put_byte( d, Rd_decode_tree8( rdec, bm ) ); - } else - { - state -= ( state < 10 ) ? 3 : 6; LZd_put_byte( d, Rd_decode_matched( rdec, bm, LZd_peek( d, rep0 ) ) ); - } continue; } /* match or repeated match */ + int len; if( Rd_decode_bit( rdec, &bm_rep[state] ) != 0 ) /* 2nd bit */ { if( Rd_decode_bit( rdec, &bm_rep0[state] ) == 0 ) /* 3rd bit */ @@ -257,13 +249,12 @@ int LZd_decode_member( struct LZ_decoder * const d, rep0 = distance; } state = St_set_rep( state ); - len = min_match_len + Rd_decode_len( rdec, &rep_len_model, pos_state ); + len = Rd_decode_len( rdec, &rep_len_model, pos_state ); } else /* match */ { - unsigned distance; - len = min_match_len + Rd_decode_len( rdec, &match_len_model, pos_state ); - distance = Rd_decode_tree6( rdec, bm_dis_slot[get_len_state(len)] ); + len = Rd_decode_len( rdec, &match_len_model, pos_state ); + unsigned distance = Rd_decode_tree6( rdec, bm_dis_slot[get_len_state(len)] ); if( distance >= start_dis_model ) { const unsigned dis_slot = distance; @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -101,12 +101,11 @@ static inline unsigned Rd_decode( struct Range_decoder * const rdec, int i; for( i = num_bits; i > 0; --i ) { - bool bit; Rd_normalize( rdec ); rdec->range >>= 1; /* symbol <<= 1; */ /* if( rdec->code >= rdec->range ) { rdec->code -= rdec->range; symbol |= 1; } */ - bit = ( rdec->code >= rdec->range ); + const bool bit = ( rdec->code >= rdec->range ); symbol <<= 1; symbol += bit; rdec->code -= rdec->range & ( 0U - bit ); } @@ -116,42 +115,75 @@ static inline unsigned Rd_decode( struct Range_decoder * const rdec, static inline unsigned Rd_decode_bit( struct Range_decoder * const rdec, Bit_model * const probability ) { - uint32_t bound; Rd_normalize( rdec ); - bound = ( rdec->range >> bit_model_total_bits ) * *probability; + const uint32_t bound = ( rdec->range >> bit_model_total_bits ) * *probability; if( rdec->code < bound ) { rdec->range = bound; - *probability += (bit_model_total - *probability) >> bit_model_move_bits; + *probability += ( bit_model_total - *probability ) >> bit_model_move_bits; return 0; } else { - rdec->range -= bound; rdec->code -= bound; + rdec->range -= bound; *probability -= *probability >> bit_model_move_bits; return 1; } } -static inline unsigned Rd_decode_tree3( struct Range_decoder * const rdec, - Bit_model bm[] ) +static inline void Rd_decode_symbol_bit( struct Range_decoder * const rdec, + Bit_model * const probability, unsigned * symbol ) { - unsigned symbol = 2 | Rd_decode_bit( rdec, &bm[1] ); - symbol = ( symbol << 1 ) | Rd_decode_bit( rdec, &bm[symbol] ); - symbol = ( symbol << 1 ) | Rd_decode_bit( rdec, &bm[symbol] ); - return symbol & 7; + Rd_normalize( rdec ); + *symbol <<= 1; + const uint32_t bound = ( rdec->range >> bit_model_total_bits ) * *probability; + if( rdec->code < bound ) + { + rdec->range = bound; + *probability += ( bit_model_total - *probability ) >> bit_model_move_bits; + } + else + { + rdec->code -= bound; + rdec->range -= bound; + *probability -= *probability >> bit_model_move_bits; + *symbol |= 1; + } + } + +static inline void Rd_decode_symbol_bit_reversed( struct Range_decoder * const rdec, + Bit_model * const probability, unsigned * model, + unsigned * symbol, const int i ) + { + Rd_normalize( rdec ); + *model <<= 1; + const uint32_t bound = ( rdec->range >> bit_model_total_bits ) * *probability; + if( rdec->code < bound ) + { + rdec->range = bound; + *probability += ( bit_model_total - *probability ) >> bit_model_move_bits; + } + else + { + rdec->code -= bound; + rdec->range -= bound; + *probability -= *probability >> bit_model_move_bits; + *model |= 1; + *symbol |= 1 << i; + } } static inline unsigned Rd_decode_tree6( struct Range_decoder * const rdec, Bit_model bm[] ) { - unsigned symbol = 2 | Rd_decode_bit( rdec, &bm[1] ); - symbol = ( symbol << 1 ) | Rd_decode_bit( rdec, &bm[symbol] ); - symbol = ( symbol << 1 ) | Rd_decode_bit( rdec, &bm[symbol] ); - symbol = ( symbol << 1 ) | Rd_decode_bit( rdec, &bm[symbol] ); - symbol = ( symbol << 1 ) | Rd_decode_bit( rdec, &bm[symbol] ); - symbol = ( symbol << 1 ) | Rd_decode_bit( rdec, &bm[symbol] ); + unsigned symbol = 1; + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); return symbol & 0x3F; } @@ -159,9 +191,14 @@ static inline unsigned Rd_decode_tree8( struct Range_decoder * const rdec, Bit_model bm[] ) { unsigned symbol = 1; - int i; - for( i = 0; i < 8; ++i ) - symbol = ( symbol << 1 ) | Rd_decode_bit( rdec, &bm[symbol] ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); return symbol & 0xFF; } @@ -173,21 +210,19 @@ Rd_decode_tree_reversed( struct Range_decoder * const rdec, unsigned symbol = 0; int i; for( i = 0; i < num_bits; ++i ) - { - const unsigned bit = Rd_decode_bit( rdec, &bm[model] ); - model <<= 1; model += bit; - symbol |= ( bit << i ); - } + Rd_decode_symbol_bit_reversed( rdec, &bm[model], &model, &symbol, i ); return symbol; } static inline unsigned Rd_decode_tree_reversed4( struct Range_decoder * const rdec, Bit_model bm[] ) { - unsigned symbol = Rd_decode_bit( rdec, &bm[1] ); - symbol += Rd_decode_bit( rdec, &bm[2+symbol] ) << 1; - symbol += Rd_decode_bit( rdec, &bm[4+symbol] ) << 2; - symbol += Rd_decode_bit( rdec, &bm[8+symbol] ) << 3; + unsigned model = 1; + unsigned symbol = 0; + Rd_decode_symbol_bit_reversed( rdec, &bm[model], &model, &symbol, 0 ); + Rd_decode_symbol_bit_reversed( rdec, &bm[model], &model, &symbol, 1 ); + Rd_decode_symbol_bit_reversed( rdec, &bm[model], &model, &symbol, 2 ); + Rd_decode_symbol_bit_reversed( rdec, &bm[model], &model, &symbol, 3 ); return symbol; } @@ -210,11 +245,24 @@ static inline unsigned Rd_decode_len( struct Range_decoder * const rdec, struct Len_model * const lm, const int pos_state ) { + Bit_model * bm; + unsigned mask, offset, symbol = 1; + if( Rd_decode_bit( rdec, &lm->choice1 ) == 0 ) - return Rd_decode_tree3( rdec, lm->bm_low[pos_state] ); + { bm = lm->bm_low[pos_state]; mask = 7; offset = 0; goto len3; } if( Rd_decode_bit( rdec, &lm->choice2 ) == 0 ) - return len_low_symbols + Rd_decode_tree3( rdec, lm->bm_mid[pos_state] ); - return len_low_symbols + len_mid_symbols + Rd_decode_tree8( rdec, lm->bm_high ); + { bm = lm->bm_mid[pos_state]; mask = 7; offset = len_low_symbols; goto len3; } + bm = lm->bm_high; mask = 0xFF; offset = len_low_symbols + len_mid_symbols; + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); +len3: + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + Rd_decode_symbol_bit( rdec, &bm[symbol], &symbol ); + return ( symbol & mask ) + min_match_len + offset; } diff --git a/doc/clzip.1 b/doc/clzip.1 index 43cec6d..3525cb4 100644 --- a/doc/clzip.1 +++ b/doc/clzip.1 @@ -1,5 +1,5 @@ .\" DO NOT MODIFY THIS FILE! It was generated by help2man 1.47.16. -.TH CLZIP "1" "January 2021" "clzip 1.12" "User Commands" +.TH CLZIP "1" "January 2022" "clzip 1.13" "User Commands" .SH NAME clzip \- reduces the size of files .SH SYNOPSIS @@ -13,13 +13,14 @@ C++ compiler. .PP Lzip is a lossless data compressor with a user interface similar to the one of gzip or bzip2. Lzip uses a simplified form of the 'Lempel\-Ziv\-Markov -chain\-Algorithm' (LZMA) stream format, chosen to maximize safety and -interoperability. Lzip can compress about as fast as gzip (lzip \fB\-0\fR) or -compress most files more than bzip2 (lzip \fB\-9\fR). Decompression speed is -intermediate between gzip and bzip2. Lzip is better than gzip and bzip2 from -a data recovery perspective. Lzip has been designed, written, and tested -with great care to replace gzip and bzip2 as the standard general\-purpose -compressed format for unix\-like systems. +chain\-Algorithm' (LZMA) stream format and provides a 3 factor integrity +checking to maximize interoperability and optimize safety. Lzip can compress +about as fast as gzip (lzip \fB\-0\fR) or compress most files more than bzip2 +(lzip \fB\-9\fR). Decompression speed is intermediate between gzip and bzip2. +Lzip is better than gzip and bzip2 from a data recovery perspective. Lzip +has been designed, written, and tested with great care to replace gzip and +bzip2 as the standard general\-purpose compressed format for unix\-like +systems. .SH OPTIONS .TP \fB\-h\fR, \fB\-\-help\fR @@ -102,7 +103,7 @@ To extract all the files from archive 'foo.tar.lz', use the commands .PP Exit status: 0 for a normal exit, 1 for environmental problems (file not found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or -invalid input file, 3 for an internal consistency error (eg, bug) which +invalid input file, 3 for an internal consistency error (e.g., bug) which caused clzip to panic. .PP The ideas embodied in clzip are due to (at least) the following people: @@ -115,7 +116,7 @@ Report bugs to lzip\-bug@nongnu.org .br Clzip home page: http://www.nongnu.org/lzip/clzip.html .SH COPYRIGHT -Copyright \(co 2021 Antonio Diaz Diaz. +Copyright \(co 2022 Antonio Diaz Diaz. License GPLv2+: GNU GPL version 2 or later <http://gnu.org/licenses/gpl.html> .br This is free software: you are free to change and redistribute it. diff --git a/doc/clzip.info b/doc/clzip.info index d4bed66..c6dd8ea 100644 --- a/doc/clzip.info +++ b/doc/clzip.info @@ -1,6 +1,6 @@ This is clzip.info, produced by makeinfo version 4.13+ from clzip.texi. -INFO-DIR-SECTION Data Compression +INFO-DIR-SECTION Compression START-INFO-DIR-ENTRY * Clzip: (clzip). LZMA lossless data compressor END-INFO-DIR-ENTRY @@ -11,7 +11,7 @@ File: clzip.info, Node: Top, Next: Introduction, Up: (dir) Clzip Manual ************ -This manual is for Clzip (version 1.12, 4 January 2021). +This manual is for Clzip (version 1.13, 24 January 2022). * Menu: @@ -19,8 +19,8 @@ This manual is for Clzip (version 1.12, 4 January 2021). * Output:: Meaning of clzip's output * Invoking clzip:: Command line interface * Quality assurance:: Design, development, and testing of lzip -* File format:: Detailed format of the compressed file * Algorithm:: How clzip compresses the data +* File format:: Detailed format of the compressed file * Stream format:: Format of the LZMA stream in lzip files * Trailing data:: Extra data appended to the file * Examples:: A small tutorial with examples @@ -29,7 +29,7 @@ This manual is for Clzip (version 1.12, 4 January 2021). * Concept index:: Index of concepts - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This manual is free documentation: you have unlimited permission to copy, distribute, and modify it. @@ -47,13 +47,14 @@ C++ compiler. Lzip is a lossless data compressor with a user interface similar to the one of gzip or bzip2. Lzip uses a simplified form of the 'Lempel-Ziv-Markov -chain-Algorithm' (LZMA) stream format, chosen to maximize safety and -interoperability. Lzip can compress about as fast as gzip (lzip -0) or -compress most files more than bzip2 (lzip -9). Decompression speed is -intermediate between gzip and bzip2. Lzip is better than gzip and bzip2 from -a data recovery perspective. Lzip has been designed, written, and tested -with great care to replace gzip and bzip2 as the standard general-purpose -compressed format for unix-like systems. +chain-Algorithm' (LZMA) stream format and provides a 3 factor integrity +checking to maximize interoperability and optimize safety. Lzip can compress +about as fast as gzip (lzip -0) or compress most files more than bzip2 +(lzip -9). Decompression speed is intermediate between gzip and bzip2. Lzip +is better than gzip and bzip2 from a data recovery perspective. Lzip has +been designed, written, and tested with great care to replace gzip and +bzip2 as the standard general-purpose compressed format for unix-like +systems. For compressing/decompressing large files on multiprocessor machines plzip can be much faster than lzip at the cost of a slightly reduced @@ -91,9 +92,9 @@ byte near the beginning is a thing of the past. The member trailer stores the 32-bit CRC of the original data, the size of the original data, and the size of the member. These values, together -with the end-of-stream marker, provide a 3 factor integrity checking which -guarantees that the decompressed version of the data is identical to the -original. This guards against corruption of the compressed data, and +with the "End Of Stream" marker, provide a 3 factor integrity checking +which guarantees that the decompressed version of the data is identical to +the original. This guards against corruption of the compressed data, and against undetected bugs in clzip (hopefully very unlikely). The chances of data corruption going undetected are microscopic. Be aware, though, that the check occurs upon decompression, so it can only tell you that something @@ -124,7 +125,7 @@ filename.lz becomes filename filename.tlz becomes filename.tar anyothername becomes anyothername.out - (De)compressing a file is much like copying or moving it; therefore clzip + (De)compressing a file is much like copying or moving it. Therefore clzip preserves the access and modification dates, permissions, and, when possible, ownership of the file just as 'cp -p' does. (If the user ID or the group ID can't be duplicated, the file permission bits S_ISUID and @@ -252,10 +253,13 @@ once, the first time it appears in the command line. '-d' '--decompress' - Decompress the files specified. If a file does not exist or can't be - opened, clzip continues decompressing the rest of the files. If a file - fails to decompress, or is a terminal, clzip exits immediately without - decompressing the rest of the files. + Decompress the files specified. If a file does not exist, can't be + opened, or the destination file already exists and '--force' has not + been specified, clzip continues decompressing the rest of the files + and exits with error status 1. If a file fails to decompress, or is a + terminal, clzip exits immediately with error status 2 without + decompressing the rest of the files. A terminal is considered an + uncompressed file, and therefore invalid. '-f' '--force' @@ -281,10 +285,12 @@ once, the first time it appears in the command line. positions and sizes of each member in multimember files are also printed. - '-lq' can be used to verify quickly (without decompressing) the - structural integrity of the files specified. (Use '--test' to verify - the data integrity). '-alq' additionally verifies that none of the - files specified contain trailing data. + If any file is damaged, does not exist, can't be opened, or is not + regular, the final exit status will be > 0. '-lq' can be used to verify + quickly (without decompressing) the structural integrity of the files + specified. (Use '--test' to verify the data integrity). '-alq' + additionally verifies that none of the files specified contain + trailing data. '-m BYTES' '--match-length=BYTES' @@ -423,11 +429,11 @@ Y yottabyte (10^24) | Yi yobibyte (2^80) Exit status: 0 for a normal exit, 1 for environmental problems (file not found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or invalid -input file, 3 for an internal consistency error (eg, bug) which caused +input file, 3 for an internal consistency error (e.g., bug) which caused clzip to panic. -File: clzip.info, Node: Quality assurance, Next: File format, Prev: Invoking clzip, Up: Top +File: clzip.info, Node: Quality assurance, Next: Algorithm, Prev: Invoking clzip, Up: Top 4 Design, development, and testing of lzip ****************************************** @@ -575,12 +581,13 @@ extraction of the decompressed data. ============================= 'Accurate and robust error detection' - The lzip format provides 3 factor integrity checking and the - decompressors report mismatches in each factor separately. This way if - just one byte in one factor fails but the other two factors match the - data, it probably means that the data are intact and the corruption - just affects the mismatching factor (CRC or data size) in the check - sequence. + The lzip format provides 3 factor integrity checking, and the + decompressors report mismatches in each factor separately. This method + detects most false positives for corruption. If just one byte in one + factor fails but the other two factors match the data, it probably + means that the data are intact and the corruption just affects the + mismatching factor (CRC, data size, or member size) in the member + trailer. 'Multiple implementations' Just like the lzip format provides 3 factor protection against @@ -614,82 +621,9 @@ extraction of the decompressed data. -File: clzip.info, Node: File format, Next: Algorithm, Prev: Quality assurance, Up: Top - -5 File format -************* - -Perfection is reached, not when there is no longer anything to add, but -when there is no longer anything to take away. --- Antoine de Saint-Exupery - - - In the diagram below, a box like this: - -+---+ -| | <-- the vertical bars might be missing -+---+ - - represents one byte; a box like this: - -+==============+ -| | -+==============+ - - represents a variable number of bytes. - - - A lzip file consists of a series of "members" (compressed data sets). -The members simply appear one after another in the file, with no additional -information before, between, or after them. - - Each member has the following structure: - -+--+--+--+--+----+----+=============+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ -| ID string | VN | DS | LZMA stream | CRC32 | Data size | Member size | -+--+--+--+--+----+----+=============+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - - All multibyte values are stored in little endian order. - -'ID string (the "magic" bytes)' - A four byte string, identifying the lzip format, with the value "LZIP" - (0x4C, 0x5A, 0x49, 0x50). - -'VN (version number, 1 byte)' - Just in case something needs to be modified in the future. 1 for now. - -'DS (coded dictionary size, 1 byte)' - The dictionary size is calculated by taking a power of 2 (the base - size) and subtracting from it a fraction between 0/16 and 7/16 of the - base size. - Bits 4-0 contain the base 2 logarithm of the base size (12 to 29). - Bits 7-5 contain the numerator of the fraction (0 to 7) to subtract - from the base size to obtain the dictionary size. - Example: 0xD3 = 2^19 - 6 * 2^15 = 512 KiB - 6 * 32 KiB = 320 KiB - Valid values for dictionary size range from 4 KiB to 512 MiB. - -'LZMA stream' - The LZMA stream, finished by an end of stream marker. Uses default - values for encoder properties. *Note Stream format::, for a complete - description. - -'CRC32 (4 bytes)' - Cyclic Redundancy Check (CRC) of the uncompressed original data. - -'Data size (8 bytes)' - Size of the uncompressed original data. - -'Member size (8 bytes)' - Total size of the member, including header and trailer. This field acts - as a distributed index, allows the verification of stream integrity, - and facilitates safe recovery of undamaged members from multimember - files. - - - -File: clzip.info, Node: Algorithm, Next: Stream format, Prev: File format, Up: Top +File: clzip.info, Node: Algorithm, Next: File format, Prev: Quality assurance, Up: Top -6 Algorithm +5 Algorithm *********** In spite of its name (Lempel-Ziv-Markov chain-Algorithm), LZMA is not a @@ -704,7 +638,7 @@ of finding coding sequences of minimum size than the one currently used by clzip could be developed, and the resulting sequence could also be coded using the LZMA coding scheme. - Clzip currently implements two variants of the LZMA algorithm; fast + Clzip currently implements two variants of the LZMA algorithm: fast (used by option '-0') and normal (used by all other compression levels). The high compression of LZMA comes from combining two basic, well-proven @@ -716,7 +650,7 @@ contexts according to what the bits are used for. Clzip is a two stage compressor. The first stage is a Lempel-Ziv coder, which reduces redundancy by translating chunks of data to their corresponding distance-length pairs. The second stage is a range encoder -that uses a different probability model for each type of data; distances, +that uses a different probability model for each type of data: distances, lengths, literal bytes, etc. Here is how it works, step by step: @@ -762,17 +696,90 @@ encoding), Igor Pavlov (for putting all the above together in LZMA), and Julian Seward (for bzip2's CLI). -File: clzip.info, Node: Stream format, Next: Trailing data, Prev: Algorithm, Up: Top +File: clzip.info, Node: File format, Next: Stream format, Prev: Algorithm, Up: Top + +6 File format +************* + +Perfection is reached, not when there is no longer anything to add, but +when there is no longer anything to take away. +-- Antoine de Saint-Exupery + + + In the diagram below, a box like this: + ++---+ +| | <-- the vertical bars might be missing ++---+ + + represents one byte; a box like this: + ++==============+ +| | ++==============+ + + represents a variable number of bytes. + + + A lzip file consists of a series of independent "members" (compressed +data sets). The members simply appear one after another in the file, with no +additional information before, between, or after them. Each member can +encode in compressed form up to 16 EiB - 1 byte of uncompressed data. The +size of a multimember file is unlimited. + + Each member has the following structure: + ++--+--+--+--+----+----+=============+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +| ID string | VN | DS | LZMA stream | CRC32 | Data size | Member size | ++--+--+--+--+----+----+=============+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + + All multibyte values are stored in little endian order. + +'ID string (the "magic" bytes)' + A four byte string, identifying the lzip format, with the value "LZIP" + (0x4C, 0x5A, 0x49, 0x50). + +'VN (version number, 1 byte)' + Just in case something needs to be modified in the future. 1 for now. + +'DS (coded dictionary size, 1 byte)' + The dictionary size is calculated by taking a power of 2 (the base + size) and subtracting from it a fraction between 0/16 and 7/16 of the + base size. + Bits 4-0 contain the base 2 logarithm of the base size (12 to 29). + Bits 7-5 contain the numerator of the fraction (0 to 7) to subtract + from the base size to obtain the dictionary size. + Example: 0xD3 = 2^19 - 6 * 2^15 = 512 KiB - 6 * 32 KiB = 320 KiB + Valid values for dictionary size range from 4 KiB to 512 MiB. + +'LZMA stream' + The LZMA stream, finished by an "End Of Stream" marker. Uses default + values for encoder properties. *Note Stream format::, for a complete + description. + +'CRC32 (4 bytes)' + Cyclic Redundancy Check (CRC) of the original uncompressed data. + +'Data size (8 bytes)' + Size of the original uncompressed data. + +'Member size (8 bytes)' + Total size of the member, including header and trailer. This field acts + as a distributed index, allows the verification of stream integrity, + and facilitates the safe recovery of undamaged members from + multimember files. Member size should be limited to 2 PiB to prevent + the data size field from overflowing. + + + +File: clzip.info, Node: Stream format, Next: Trailing data, Prev: File format, Up: Top 7 Format of the LZMA stream in lzip files ***************************************** -Lzip uses a simplified form of the LZMA stream format chosen to maximize -safety and interoperability. - - The LZMA algorithm has three parameters, called "special LZMA -properties", to adjust it for some kinds of binary data. These parameters -are; 'literal_context_bits' (with a default value of 3), +The LZMA algorithm has three parameters, called "special LZMA properties", +to adjust it for some kinds of binary data. These parameters are: +'literal_context_bits' (with a default value of 3), 'literal_pos_state_bits' (with a default value of 0), and 'pos_state_bits' (with a default value of 2). As a general purpose compressor, lzip only uses the default values for these parameters. In particular @@ -782,12 +789,14 @@ in the code. Lzip finishes the LZMA stream with an "End Of Stream" (EOS) marker (the distance-length pair 0xFFFFFFFFU, 2), which in conjunction with the 'member size' field in the member trailer allows the verification of stream -integrity. The LZMA stream in lzip files always has these two features -(default properties and EOS marker) and is referred to in this document as -LZMA-302eos. The EOS marker is the only marker allowed in lzip files. +integrity. The EOS marker is the only marker allowed in lzip files. The +LZMA stream in lzip files always has these two features (default properties +and EOS marker) and is referred to in this document as LZMA-302eos. This +simplified form of the LZMA stream format has been chosen to maximize +interoperability and safety. The second stage of LZMA is a range encoder that uses a different -probability model for each type of symbol; distances, lengths, literal +probability model for each type of symbol: distances, lengths, literal bytes, etc. Range encoding conceptually encodes all the symbols of the message into one number. Unlike Huffman coding, which assigns to each symbol a bit-pattern and concatenates all the bit-patterns together, range @@ -795,16 +804,16 @@ encoding can compress one symbol to less than one bit. Therefore the compressed data produced by a range encoder can't be split in pieces that could be described individually. - It seems that the only way of describing the LZMA-302eos stream is -describing the algorithm that decodes it. And given the many details about + It seems that the only way of describing the LZMA-302eos stream is to +describe the algorithm that decodes it. And given the many details about the range decoder that need to be described accurately, the source code of -a real decoder seems the only appropriate reference to use. +a real decompressor seems the only appropriate reference to use. What follows is a description of the decoding algorithm for LZMA-302eos streams using as reference the source code of "lzd", an educational decompressor for lzip files which can be downloaded from the lzip download -directory. The source code of lzd is included in appendix A. *Note -Reference source code::. +directory. Lzd is written in C++11 and its source code is included in +appendix A. *Note Reference source code::. 7.1 What is coded @@ -840,7 +849,7 @@ Bit sequence Description 1 + 1 + 8 bits lengths from 18 to 273 - The coding of distances is a little more complicated, so I'll begin + The coding of distances is a little more complicated, so I'll begin by explaining a simpler version of the encoding. Imagine you need to encode a number from 0 to 2^32 - 1, and you want to @@ -850,7 +859,7 @@ which you may find by making a bit scan from the left (from the MSB). A position of 0 means that the number is 0 (no bit is set), 1 means the LSB is the first bit set (the number is 1), and 32 means the MSB is set (i.e., the number is >= 0x80000000). Then, if the position is >= 2, you encode the -remaining position - 1 bits. Let's call these bits "direct_bits" because +remaining position - 1 bits. Let's call these bits "direct bits" because they are coded directly by value instead of indirectly by position. The inconvenient of this simple method is that it needs 6 bits to encode @@ -906,9 +915,10 @@ integers representing the probability of the corresponding bit being 0. of 3. The resulting value is in the range 0 to 3. - In the following table, '!literal' is any sequence except a literal -byte. 'rep' is any one of 'rep0', 'rep1', 'rep2', or 'rep3'. The types of -previous sequences corresponding to each state are: + The types of previous sequences corresponding to each state are shown in +the following table. '!literal' is any sequence except a literal byte. +'rep' is any one of 'rep0', 'rep1', 'rep2', or 'rep3'. The last type in +each line is the most recent. State Types of previous sequences ------------------------------------------------------ @@ -979,9 +989,9 @@ The LZMA stream is consumed one byte at a time by the range decoder. (See of decoded bits, depending on how well these bits agree with their context. (See 'decode_bit' in the source). - The range decoder state consists of two unsigned 32-bit variables; + The range decoder state consists of two unsigned 32-bit variables: 'range' (representing the most significant part of the range size not yet -decoded), and 'code' (representing the current point within 'range'). +decoded) and 'code' (representing the current point within 'range'). 'range' is initialized to 2^32 - 1, and 'code' is initialized to 0. The range encoder produces a first 0 byte that must be ignored by the @@ -993,7 +1003,7 @@ range decoder. This is done by shifting 5 bytes in the initialization of ========================================== After decoding the member header and obtaining the dictionary size, the -range decoder is initialized and then the LZMA decoder enters a loop (See +range decoder is initialized and then the LZMA decoder enters a loop (see 'decode_member' in the source) where it invokes the range decoder with the appropriate contexts to decode the different coding sequences (matches, repeated matches, and literal bytes), until the "End Of Stream" marker is @@ -1001,8 +1011,8 @@ decoded. Once the "End Of Stream" marker has been decoded, the decompressor reads and decodes the member trailer, and verifies that the three integrity -factors (CRC, data size, and member size) match those calculated by the -LZMA decoder. +factors stored there (CRC, data size, and member size) match those computed +from the data. File: clzip.info, Node: Trailing data, Next: Examples, Prev: Stream format, Up: Top @@ -1079,7 +1089,7 @@ show the compression ratio. clzip -v file -Example 3: Like example 1 but the created 'file.lz' is multimember with a +Example 3: Like example 2 but the created 'file.lz' is multimember with a member size of 1 MiB. The compression ratio is not shown. clzip -b 1MiB file @@ -1097,15 +1107,7 @@ status. clzip -tv file.lz -Example 6: Compress a whole device in /dev/sdc and send the output to -'file.lz'. - - clzip -c /dev/sdc > file.lz - or - clzip /dev/sdc -o file.lz - - -Example 7: The right way of concatenating the decompressed output of two or +Example 6: The right way of concatenating the decompressed output of two or more compressed files. *Note Trailing data::. Don't do this @@ -1114,18 +1116,26 @@ more compressed files. *Note Trailing data::. clzip -cd file1.lz file2.lz file3.lz -Example 8: Decompress 'file.lz' partially until 10 KiB of decompressed data +Example 7: Decompress 'file.lz' partially until 10 KiB of decompressed data are produced. clzip -cd file.lz | dd bs=1024 count=10 -Example 9: Decompress 'file.lz' partially from decompressed byte at offset +Example 8: Decompress 'file.lz' partially from decompressed byte at offset 10000 to decompressed byte at offset 14999 (5000 bytes are produced). clzip -cd file.lz | dd bs=1000 skip=10 count=5 +Example 9: Compress a whole device in /dev/sdc and send the output to +'file.lz'. + + clzip -c /dev/sdc > file.lz + or + clzip /dev/sdc -o file.lz + + Example 10: Create a multivolume compressed tar archive with a volume size of 1440 KiB. @@ -1165,7 +1175,7 @@ Appendix A Reference source code ******************************** /* Lzd - Educational decompressor for the lzip format - Copyright (C) 2013-2021 Antonio Diaz Diaz. + Copyright (C) 2013-2022 Antonio Diaz Diaz. This program is free software. Redistribution and use in source and binary forms, with or without modification, are permitted provided @@ -1195,7 +1205,7 @@ Appendix A Reference source code #include <cstring> #include <stdint.h> #include <unistd.h> -#if defined(__MSVCRT__) || defined(__OS2__) || defined(__DJGPP__) +#if defined __MSVCRT__ || defined __OS2__ || defined __DJGPP__ #include <fcntl.h> #include <io.h> #endif @@ -1585,7 +1595,7 @@ int main( const int argc, const char * const argv[] ) "See the lzip manual for an explanation of the code.\n" "\nUsage: %s [-d] < file.lz > file\n" "Lzd decompresses from standard input to standard output.\n" - "\nCopyright (C) 2021 Antonio Diaz Diaz.\n" + "\nCopyright (C) 2022 Antonio Diaz Diaz.\n" "License 2-clause BSD.\n" "This is free software: you are free to change and redistribute it.\n" "There is NO WARRANTY, to the extent permitted by law.\n" @@ -1595,7 +1605,7 @@ int main( const int argc, const char * const argv[] ) return 0; } -#if defined(__MSVCRT__) || defined(__OS2__) || defined(__DJGPP__) +#if defined __MSVCRT__ || defined __OS2__ || defined __DJGPP__ setmode( STDIN_FILENO, O_BINARY ); setmode( STDOUT_FILENO, O_BINARY ); #endif @@ -1677,23 +1687,23 @@ Concept index Tag Table: -Node: Top210 -Node: Introduction1211 -Node: Output7184 -Node: Invoking clzip8787 -Ref: --trailing-error9585 -Node: Quality assurance18586 -Node: File format27545 -Ref: coded-dict-size28836 -Node: Algorithm29972 -Node: Stream format33379 -Ref: what-is-coded35749 -Node: Trailing data44618 -Node: Examples46881 -Ref: concat-example48493 -Node: Problems49563 -Node: Reference source code50099 -Node: Concept index64964 +Node: Top205 +Node: Introduction1207 +Node: Output7226 +Node: Invoking clzip8829 +Ref: --trailing-error9627 +Node: Quality assurance18961 +Node: Algorithm27986 +Node: File format31397 +Ref: coded-dict-size32827 +Node: Stream format34062 +Ref: what-is-coded36459 +Node: Trailing data45387 +Node: Examples47650 +Ref: concat-example49102 +Node: Problems50332 +Node: Reference source code50868 +Node: Concept index65727 End Tag Table diff --git a/doc/clzip.texi b/doc/clzip.texi index caa40fc..c00599c 100644 --- a/doc/clzip.texi +++ b/doc/clzip.texi @@ -6,10 +6,10 @@ @finalout @c %**end of header -@set UPDATED 4 January 2021 -@set VERSION 1.12 +@set UPDATED 24 January 2022 +@set VERSION 1.13 -@dircategory Data Compression +@dircategory Compression @direntry * Clzip: (clzip). LZMA lossless data compressor @end direntry @@ -40,8 +40,8 @@ This manual is for Clzip (version @value{VERSION}, @value{UPDATED}). * Output:: Meaning of clzip's output * Invoking clzip:: Command line interface * Quality assurance:: Design, development, and testing of lzip -* File format:: Detailed format of the compressed file * Algorithm:: How clzip compresses the data +* File format:: Detailed format of the compressed file * Stream format:: Format of the LZMA stream in lzip files * Trailing data:: Extra data appended to the file * Examples:: A small tutorial with examples @@ -51,7 +51,7 @@ This manual is for Clzip (version @value{VERSION}, @value{UPDATED}). @end menu @sp 1 -Copyright @copyright{} 2010-2021 Antonio Diaz Diaz. +Copyright @copyright{} 2010-2022 Antonio Diaz Diaz. This manual is free documentation: you have unlimited permission to copy, distribute, and modify it. @@ -71,13 +71,14 @@ C++ compiler. @uref{http://www.nongnu.org/lzip/lzip.html,,Lzip} is a lossless data compressor with a user interface similar to the one of gzip or bzip2. Lzip uses a simplified form of the 'Lempel-Ziv-Markov -chain-Algorithm' (LZMA) stream format, chosen to maximize safety and -interoperability. Lzip can compress about as fast as gzip @w{(lzip -0)} or -compress most files more than bzip2 @w{(lzip -9)}. Decompression speed is -intermediate between gzip and bzip2. Lzip is better than gzip and bzip2 from -a data recovery perspective. Lzip has been designed, written, and tested -with great care to replace gzip and bzip2 as the standard general-purpose -compressed format for unix-like systems. +chain-Algorithm' (LZMA) stream format and provides a 3 factor integrity +checking to maximize interoperability and optimize safety. Lzip can compress +about as fast as gzip @w{(lzip -0)} or compress most files more than bzip2 +@w{(lzip -9)}. Decompression speed is intermediate between gzip and bzip2. +Lzip is better than gzip and bzip2 from a data recovery perspective. Lzip +has been designed, written, and tested with great care to replace gzip and +bzip2 as the standard general-purpose compressed format for unix-like +systems. For compressing/decompressing large files on multiprocessor machines @uref{http://www.nongnu.org/lzip/manual/plzip_manual.html,,plzip} can be @@ -87,8 +88,8 @@ much faster than lzip at the cost of a slightly reduced compression ratio. @end ifnothtml For creation and manipulation of compressed tar archives -@uref{http://www.nongnu.org/lzip/manual/tarlz_manual.html,,tarlz} can be -more efficient than using tar and plzip because tarlz is able to keep the +@uref{http://www.nongnu.org/lzip/manual/tarlz_manual.html,,tarlz} can be more +efficient than using tar and plzip because tarlz is able to keep the alignment between tar members and lzip members. @ifnothtml @xref{Top,tarlz manual,,tarlz}. @@ -129,7 +130,7 @@ the beginning is a thing of the past. The member trailer stores the 32-bit CRC of the original data, the size of the original data, and the size of the member. These values, together -with the end-of-stream marker, provide a 3 factor integrity checking +with the "End Of Stream" marker, provide a 3 factor integrity checking which guarantees that the decompressed version of the data is identical to the original. This guards against corruption of the compressed data, and against undetected bugs in clzip (hopefully very unlikely). The @@ -165,9 +166,9 @@ file from that of the compressed file as follows: @item anyothername @tab becomes @tab anyothername.out @end multitable -(De)compressing a file is much like copying or moving it; therefore clzip +(De)compressing a file is much like copying or moving it. Therefore clzip preserves the access and modification dates, permissions, and, when -possible, ownership of the file just as @samp{cp -p} does. (If the user ID or +possible, ownership of the file just as @w{@samp{cp -p}} does. (If the user ID or the group ID can't be duplicated, the file permission bits S_ISUID and S_ISGID are cleared). @@ -305,10 +306,12 @@ and @samp{-S}. @samp{-c} has no effect when testing or listing. @item -d @itemx --decompress -Decompress the files specified. If a file does not exist or can't be -opened, clzip continues decompressing the rest of the files. If a file -fails to decompress, or is a terminal, clzip exits immediately without -decompressing the rest of the files. +Decompress the files specified. If a file does not exist, can't be opened, +or the destination file already exists and @samp{--force} has not been +specified, clzip continues decompressing the rest of the files and exits with +error status 1. If a file fails to decompress, or is a terminal, clzip exits +immediately with error status 2 without decompressing the rest of the files. +A terminal is considered an uncompressed file, and therefore invalid. @item -f @itemx --force @@ -333,10 +336,11 @@ size, the number of members in the file, and the amount of trailing data (if any) are also printed. With @samp{-vv}, the positions and sizes of each member in multimember files are also printed. -@samp{-lq} can be used to verify quickly (without decompressing) the -structural integrity of the files specified. (Use @samp{--test} to verify -the data integrity). @samp{-alq} additionally verifies that none of the -files specified contain trailing data. +If any file is damaged, does not exist, can't be opened, or is not regular, +the final exit status will be @w{> 0}. @samp{-lq} can be used to verify +quickly (without decompressing) the structural integrity of the files +specified. (Use @samp{--test} to verify the data integrity). @samp{-alq} +additionally verifies that none of the files specified contain trailing data. @item -m @var{bytes} @itemx --match-length=@var{bytes} @@ -479,9 +483,9 @@ Table of SI and binary prefixes (unit multipliers): @sp 1 Exit status: 0 for a normal exit, 1 for environmental problems (file not -found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or -invalid input file, 3 for an internal consistency error (eg, bug) which -caused clzip to panic. +found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or invalid +input file, 3 for an internal consistency error (e.g., bug) which caused +clzip to panic. @node Quality assurance @@ -635,11 +639,12 @@ and may limit the number of members or the total uncompressed size. @table @samp @item Accurate and robust error detection -The lzip format provides 3 factor integrity checking and the decompressors -report mismatches in each factor separately. This way if just one byte in -one factor fails but the other two factors match the data, it probably means -that the data are intact and the corruption just affects the mismatching -factor (CRC or data size) in the check sequence. +The lzip format provides 3 factor integrity checking, and the decompressors +report mismatches in each factor separately. This method detects most false +positives for corruption. If just one byte in one factor fails but the other +two factors match the data, it probably means that the data are intact and +the corruption just affects the mismatching factor (CRC, data size, or +member size) in the member trailer. @item Multiple implementations @@ -678,84 +683,6 @@ into the design of gzip. Both bzip2 and lzip are free from this flaw. @end table -@node File format -@chapter File format -@cindex file format - -Perfection is reached, not when there is no longer anything to add, but -when there is no longer anything to take away.@* ---- Antoine de Saint-Exupery - -@sp 1 -In the diagram below, a box like this: - -@verbatim -+---+ -| | <-- the vertical bars might be missing -+---+ -@end verbatim - -represents one byte; a box like this: - -@verbatim -+==============+ -| | -+==============+ -@end verbatim - -represents a variable number of bytes. - -@sp 1 -A lzip file consists of a series of "members" (compressed data sets). -The members simply appear one after another in the file, with no -additional information before, between, or after them. - -Each member has the following structure: - -@verbatim -+--+--+--+--+----+----+=============+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ -| ID string | VN | DS | LZMA stream | CRC32 | Data size | Member size | -+--+--+--+--+----+----+=============+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ -@end verbatim - -All multibyte values are stored in little endian order. - -@table @samp -@item ID string (the "magic" bytes) -A four byte string, identifying the lzip format, with the value "LZIP" -(0x4C, 0x5A, 0x49, 0x50). - -@item VN (version number, 1 byte) -Just in case something needs to be modified in the future. 1 for now. - -@anchor{coded-dict-size} -@item DS (coded dictionary size, 1 byte) -The dictionary size is calculated by taking a power of 2 (the base size) -and subtracting from it a fraction between 0/16 and 7/16 of the base size.@* -Bits 4-0 contain the base 2 logarithm of the base size (12 to 29).@* -Bits 7-5 contain the numerator of the fraction (0 to 7) to subtract -from the base size to obtain the dictionary size.@* -Example: 0xD3 = 2^19 - 6 * 2^15 = 512 KiB - 6 * 32 KiB = 320 KiB@* -Valid values for dictionary size range from 4 KiB to 512 MiB. - -@item LZMA stream -The LZMA stream, finished by an end of stream marker. Uses default values -for encoder properties. @xref{Stream format}, for a complete description. - -@item CRC32 (4 bytes) -Cyclic Redundancy Check (CRC) of the uncompressed original data. - -@item Data size (8 bytes) -Size of the uncompressed original data. - -@item Member size (8 bytes) -Total size of the member, including header and trailer. This field acts -as a distributed index, allows the verification of stream integrity, and -facilitates safe recovery of undamaged members from multimember files. - -@end table - - @node Algorithm @chapter Algorithm @cindex algorithm @@ -772,7 +699,7 @@ of finding coding sequences of minimum size than the one currently used by clzip could be developed, and the resulting sequence could also be coded using the LZMA coding scheme. -Clzip currently implements two variants of the LZMA algorithm; fast +Clzip currently implements two variants of the LZMA algorithm: fast (used by option @samp{-0}) and normal (used by all other compression levels). The high compression of LZMA comes from combining two basic, well-proven @@ -784,7 +711,7 @@ contexts according to what the bits are used for. Clzip is a two stage compressor. The first stage is a Lempel-Ziv coder, which reduces redundancy by translating chunks of data to their corresponding distance-length pairs. The second stage is a range encoder -that uses a different probability model for each type of data; +that uses a different probability model for each type of data: distances, lengths, literal bytes, etc. Here is how it works, step by step: @@ -831,32 +758,112 @@ encoding), Igor Pavlov (for putting all the above together in LZMA), and Julian Seward (for bzip2's CLI). +@node File format +@chapter File format +@cindex file format + +Perfection is reached, not when there is no longer anything to add, but +when there is no longer anything to take away.@* +--- Antoine de Saint-Exupery + +@sp 1 +In the diagram below, a box like this: + +@verbatim ++---+ +| | <-- the vertical bars might be missing ++---+ +@end verbatim + +represents one byte; a box like this: + +@verbatim ++==============+ +| | ++==============+ +@end verbatim + +represents a variable number of bytes. + +@sp 1 +A lzip file consists of a series of independent "members" (compressed data +sets). The members simply appear one after another in the file, with no +additional information before, between, or after them. Each member can +encode in compressed form up to @w{16 EiB - 1 byte} of uncompressed data. +The size of a multimember file is unlimited. + +Each member has the following structure: + +@verbatim ++--+--+--+--+----+----+=============+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +| ID string | VN | DS | LZMA stream | CRC32 | Data size | Member size | ++--+--+--+--+----+----+=============+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +@end verbatim + +All multibyte values are stored in little endian order. + +@table @samp +@item ID string (the "magic" bytes) +A four byte string, identifying the lzip format, with the value "LZIP" +(0x4C, 0x5A, 0x49, 0x50). + +@item VN (version number, 1 byte) +Just in case something needs to be modified in the future. 1 for now. + +@anchor{coded-dict-size} +@item DS (coded dictionary size, 1 byte) +The dictionary size is calculated by taking a power of 2 (the base size) +and subtracting from it a fraction between 0/16 and 7/16 of the base size.@* +Bits 4-0 contain the base 2 logarithm of the base size (12 to 29).@* +Bits 7-5 contain the numerator of the fraction (0 to 7) to subtract +from the base size to obtain the dictionary size.@* +Example: 0xD3 = 2^19 - 6 * 2^15 = 512 KiB - 6 * 32 KiB = 320 KiB@* +Valid values for dictionary size range from 4 KiB to 512 MiB. + +@item LZMA stream +The LZMA stream, finished by an "End Of Stream" marker. Uses default values +for encoder properties. @xref{Stream format}, for a complete description. + +@item CRC32 (4 bytes) +Cyclic Redundancy Check (CRC) of the original uncompressed data. + +@item Data size (8 bytes) +Size of the original uncompressed data. + +@item Member size (8 bytes) +Total size of the member, including header and trailer. This field acts +as a distributed index, allows the verification of stream integrity, and +facilitates the safe recovery of undamaged members from multimember files. +Member size should be limited to @w{2 PiB} to prevent the data size field +from overflowing. + +@end table + + @node Stream format @chapter Format of the LZMA stream in lzip files @cindex format of the LZMA stream -Lzip uses a simplified form of the LZMA stream format chosen to maximize -safety and interoperability. - The LZMA algorithm has three parameters, called "special LZMA properties", to adjust it for some kinds of binary data. These -parameters are; @samp{literal_context_bits} (with a default value of 3), +parameters are: @samp{literal_context_bits} (with a default value of 3), @samp{literal_pos_state_bits} (with a default value of 0), and @samp{pos_state_bits} (with a default value of 2). As a general purpose compressor, lzip only uses the default values for these parameters. In particular @samp{literal_pos_state_bits} has been optimized away and does not even appear in the code. -Lzip finishes the LZMA stream with an "End Of Stream" (EOS) marker -(the distance-length pair 0xFFFFFFFFU, 2), which in conjunction with the +Lzip finishes the LZMA stream with an "End Of Stream" (EOS) marker (the +distance-length pair @w{0xFFFFFFFFU, 2}), which in conjunction with the @samp{member size} field in the member trailer allows the verification of -stream integrity. The LZMA stream in lzip files always has these two -features (default properties and EOS marker) and is referred to in this -document as LZMA-302eos. The EOS marker is the only marker allowed in -lzip files. +stream integrity. The EOS marker is the only marker allowed in lzip files. +The LZMA stream in lzip files always has these two features (default +properties and EOS marker) and is referred to in this document as +LZMA-302eos. This simplified form of the LZMA stream format has been chosen +to maximize interoperability and safety. The second stage of LZMA is a range encoder that uses a different -probability model for each type of symbol; distances, lengths, literal +probability model for each type of symbol: distances, lengths, literal bytes, etc. Range encoding conceptually encodes all the symbols of the message into one number. Unlike Huffman coding, which assigns to each symbol a bit-pattern and concatenates all the bit-patterns together, @@ -864,16 +871,16 @@ range encoding can compress one symbol to less than one bit. Therefore the compressed data produced by a range encoder can't be split in pieces that could be described individually. -It seems that the only way of describing the LZMA-302eos stream is -describing the algorithm that decodes it. And given the many details +It seems that the only way of describing the LZMA-302eos stream is to +describe the algorithm that decodes it. And given the many details about the range decoder that need to be described accurately, the source -code of a real decoder seems the only appropriate reference to use. +code of a real decompressor seems the only appropriate reference to use. What follows is a description of the decoding algorithm for LZMA-302eos streams using as reference the source code of "lzd", an educational -decompressor for lzip files which can be downloaded from the lzip -download directory. The source code of lzd is included in appendix A. -@xref{Reference source code}. +decompressor for lzip files which can be downloaded from the lzip download +directory. Lzd is written in C++11 and its source code is included in +appendix A. @xref{Reference source code}. @sp 1 @section What is coded @@ -911,7 +918,7 @@ Lengths (the @samp{len} in the table above) are coded as follows: @end multitable @sp 1 -The coding of distances is a little more complicated, so I'll begin +The coding of distances is a little more complicated, so I'll begin by explaining a simpler version of the encoding. Imagine you need to encode a number from 0 to @w{2^32 - 1}, and you want to @@ -921,7 +928,7 @@ which you may find by making a bit scan from the left (from the MSB). A position of 0 means that the number is 0 (no bit is set), 1 means the LSB is the first bit set (the number is 1), and 32 means the MSB is set (i.e., the number is @w{>= 0x80000000}). Then, if the position is @w{>= 2}, you encode -the remaining @w{position - 1} bits. Let's call these bits "direct_bits" +the remaining @w{position - 1} bits. Let's call these bits "direct bits" because they are coded directly by value instead of indirectly by position. The inconvenient of this simple method is that it needs 6 bits to encode the @@ -981,10 +988,10 @@ of 3. The resulting value is in the range 0 to 3. @end table -In the following table, @samp{!literal} is any sequence except a literal -byte. @samp{rep} is any one of @samp{rep0}, @samp{rep1}, @samp{rep2}, or -@samp{rep3}. The types of previous sequences corresponding to each state -are: +The types of previous sequences corresponding to each state are shown in the +following table. @samp{!literal} is any sequence except a literal byte. +@samp{rep} is any one of @samp{rep0}, @samp{rep1}, @samp{rep2}, or +@samp{rep3}. The last type in each line is the most recent. @multitable {State} {rep or (!literal, shortrep), literal, literal} @headitem State @tab Types of previous sequences @@ -1059,9 +1066,9 @@ The LZMA stream is consumed one byte at a time by the range decoder. variable number of decoded bits, depending on how well these bits agree with their context. (See @samp{decode_bit} in the source). -The range decoder state consists of two unsigned 32-bit variables; +The range decoder state consists of two unsigned 32-bit variables: @samp{range} (representing the most significant part of the range size -not yet decoded), and @samp{code} (representing the current point within +not yet decoded) and @samp{code} (representing the current point within @samp{range}). @samp{range} is initialized to @w{2^32 - 1}, and @samp{code} is initialized to 0. @@ -1075,14 +1082,15 @@ the source). After decoding the member header and obtaining the dictionary size, the range decoder is initialized and then the LZMA decoder enters a loop -(See @samp{decode_member} in the source) where it invokes the range +(see @samp{decode_member} in the source) where it invokes the range decoder with the appropriate contexts to decode the different coding sequences (matches, repeated matches, and literal bytes), until the "End Of Stream" marker is decoded. Once the "End Of Stream" marker has been decoded, the decompressor reads and decodes the member trailer, and verifies that the three integrity factors -(CRC, data size, and member size) match those calculated by the LZMA decoder. +stored there (CRC, data size, and member size) match those computed from the +data. @node Trailing data @@ -1171,7 +1179,7 @@ clzip -v file @sp 1 @noindent -Example 3: Like example 1 but the created @samp{file.lz} is multimember with +Example 3: Like example 2 but the created @samp{file.lz} is multimember with a member size of @w{1 MiB}. The compression ratio is not shown. @example @@ -1197,20 +1205,9 @@ clzip -tv file.lz @end example @sp 1 -@noindent -Example 6: Compress a whole device in /dev/sdc and send the output to -@samp{file.lz}. - -@example - clzip -c /dev/sdc > file.lz -or - clzip /dev/sdc -o file.lz -@end example - -@sp 1 @anchor{concat-example} @noindent -Example 7: The right way of concatenating the decompressed output of two or +Example 6: The right way of concatenating the decompressed output of two or more compressed files. @xref{Trailing data}. @example @@ -1222,7 +1219,7 @@ Do this instead @sp 1 @noindent -Example 8: Decompress @samp{file.lz} partially until @w{10 KiB} of +Example 7: Decompress @samp{file.lz} partially until @w{10 KiB} of decompressed data are produced. @example @@ -1231,7 +1228,7 @@ clzip -cd file.lz | dd bs=1024 count=10 @sp 1 @noindent -Example 9: Decompress @samp{file.lz} partially from decompressed byte at +Example 8: Decompress @samp{file.lz} partially from decompressed byte at offset 10000 to decompressed byte at offset 14999 (5000 bytes are produced). @example @@ -1240,6 +1237,17 @@ clzip -cd file.lz | dd bs=1000 skip=10 count=5 @sp 1 @noindent +Example 9: Compress a whole device in /dev/sdc and send the output to +@samp{file.lz}. + +@example + clzip -c /dev/sdc > file.lz +or + clzip /dev/sdc -o file.lz +@end example + +@sp 1 +@noindent Example 10: Create a multivolume compressed tar archive with a volume size of @w{1440 KiB}. @@ -1287,7 +1295,7 @@ find by running @w{@samp{clzip --version}}. @verbatim /* Lzd - Educational decompressor for the lzip format - Copyright (C) 2013-2021 Antonio Diaz Diaz. + Copyright (C) 2013-2022 Antonio Diaz Diaz. This program is free software. Redistribution and use in source and binary forms, with or without modification, are permitted provided @@ -1317,7 +1325,7 @@ find by running @w{@samp{clzip --version}}. #include <cstring> #include <stdint.h> #include <unistd.h> -#if defined(__MSVCRT__) || defined(__OS2__) || defined(__DJGPP__) +#if defined __MSVCRT__ || defined __OS2__ || defined __DJGPP__ #include <fcntl.h> #include <io.h> #endif @@ -1707,7 +1715,7 @@ int main( const int argc, const char * const argv[] ) "See the lzip manual for an explanation of the code.\n" "\nUsage: %s [-d] < file.lz > file\n" "Lzd decompresses from standard input to standard output.\n" - "\nCopyright (C) 2021 Antonio Diaz Diaz.\n" + "\nCopyright (C) 2022 Antonio Diaz Diaz.\n" "License 2-clause BSD.\n" "This is free software: you are free to change and redistribute it.\n" "There is NO WARRANTY, to the extent permitted by law.\n" @@ -1717,7 +1725,7 @@ int main( const int argc, const char * const argv[] ) return 0; } -#if defined(__MSVCRT__) || defined(__OS2__) || defined(__DJGPP__) +#if defined __MSVCRT__ || defined __OS2__ || defined __DJGPP__ setmode( STDIN_FILENO, O_BINARY ); setmode( STDOUT_FILENO, O_BINARY ); #endif @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -33,32 +33,25 @@ CRC32 crc32; int LZe_get_match_pairs( struct LZ_encoder * const e, struct Pair * pairs ) { - int32_t * ptr0 = e->eb.mb.pos_array + ( e->eb.mb.cyclic_pos << 1 ); - int32_t * ptr1 = ptr0 + 1; - int32_t * newptr; - int len = 0, len0 = 0, len1 = 0; - int maxlen = 3; /* only used if pairs != 0 */ - int num_pairs = 0; - const int pos1 = e->eb.mb.pos + 1; - const int min_pos = ( e->eb.mb.pos > e->eb.mb.dictionary_size ) ? - e->eb.mb.pos - e->eb.mb.dictionary_size : 0; - const uint8_t * const data = Mb_ptr_to_current_pos( &e->eb.mb ); - int count, key2, key3, key4, newpos1; - unsigned tmp; int len_limit = e->match_len_limit; - if( len_limit > Mb_available_bytes( &e->eb.mb ) ) { len_limit = Mb_available_bytes( &e->eb.mb ); if( len_limit < 4 ) return 0; } - tmp = crc32[data[0]] ^ data[1]; - key2 = tmp & ( num_prev_positions2 - 1 ); + int maxlen = 3; /* only used if pairs != 0 */ + int num_pairs = 0; + const int min_pos = ( e->eb.mb.pos > e->eb.mb.dictionary_size ) ? + e->eb.mb.pos - e->eb.mb.dictionary_size : 0; + const uint8_t * const data = Mb_ptr_to_current_pos( &e->eb.mb ); + + unsigned tmp = crc32[data[0]] ^ data[1]; + const int key2 = tmp & ( num_prev_positions2 - 1 ); tmp ^= (unsigned)data[2] << 8; - key3 = num_prev_positions2 + ( tmp & ( num_prev_positions3 - 1 ) ); - key4 = num_prev_positions2 + num_prev_positions3 + - ( ( tmp ^ ( crc32[data[3]] << 5 ) ) & e->eb.mb.key4_mask ); + const int key3 = num_prev_positions2 + ( tmp & ( num_prev_positions3 - 1 ) ); + const int key4 = num_prev_positions2 + num_prev_positions3 + + ( ( tmp ^ ( crc32[data[3]] << 5 ) ) & e->eb.mb.key4_mask ); if( pairs ) { @@ -67,7 +60,7 @@ int LZe_get_match_pairs( struct LZ_encoder * const e, struct Pair * pairs ) if( np2 > min_pos && e->eb.mb.buffer[np2-1] == data[0] ) { pairs[0].dis = e->eb.mb.pos - np2; - pairs[0].len = maxlen = 2; + pairs[0].len = maxlen = 2 + ( np2 == np3 ); num_pairs = 1; } if( np2 != np3 && np3 > min_pos && e->eb.mb.buffer[np3-1] == data[0] ) @@ -86,18 +79,23 @@ int LZe_get_match_pairs( struct LZ_encoder * const e, struct Pair * pairs ) } } + const int pos1 = e->eb.mb.pos + 1; e->eb.mb.prev_positions[key2] = pos1; e->eb.mb.prev_positions[key3] = pos1; - newpos1 = e->eb.mb.prev_positions[key4]; + int newpos1 = e->eb.mb.prev_positions[key4]; e->eb.mb.prev_positions[key4] = pos1; + int32_t * ptr0 = e->eb.mb.pos_array + ( e->eb.mb.cyclic_pos << 1 ); + int32_t * ptr1 = ptr0 + 1; + int len = 0, len0 = 0, len1 = 0; + + int count; for( count = e->cycles; ; ) { - int delta; if( newpos1 <= min_pos || --count < 0 ) { *ptr0 = *ptr1 = 0; break; } - delta = pos1 - newpos1; - newptr = e->eb.mb.pos_array + + const int delta = pos1 - newpos1; + int32_t * const newptr = e->eb.mb.pos_array + ( ( e->eb.mb.cyclic_pos - delta + ( (e->eb.mb.cyclic_pos >= delta) ? 0 : e->eb.mb.dictionary_size + 1 ) ) << 1 ); if( data[len-delta] == data[len] ) @@ -152,7 +150,6 @@ static void LZe_update_distance_prices( struct LZ_encoder * const e ) for( len_state = 0; len_state < len_states; ++len_state ) { int * const dsp = e->dis_slot_prices[len_state]; - int * const dp = e->dis_prices[len_state]; const Bit_model * const bmds = e->eb.bm_dis_slot[len_state]; int slot = 0; for( ; slot < end_dis_model; ++slot ) @@ -161,6 +158,7 @@ static void LZe_update_distance_prices( struct LZ_encoder * const e ) dsp[slot] = price_symbol6( bmds, slot ) + (((( slot >> 1 ) - 1 ) - dis_align_bits ) << price_shift_bits ); + int * const dp = e->dis_prices[len_state]; for( dis = 0; dis < start_dis_model; ++dis ) dp[dis] = dsp[dis]; for( ; dis < modeled_distances; ++dis ) @@ -169,7 +167,7 @@ static void LZe_update_distance_prices( struct LZ_encoder * const e ) } -/* Returns the number of bytes advanced (ahead). +/* Return the number of bytes advanced (ahead). trials[0]..trials[ahead-1] contain the steps to encode. ( trials[0].dis4 == -1 ) means literal. A match/rep longer or equal than match_len_limit finishes the sequence. @@ -178,9 +176,8 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, const int reps[num_rep_distances], const State state ) { - int main_len, num_pairs, i, rep, num_trials, len; - int rep_index = 0, cur = 0; - int replens[num_rep_distances]; + int num_pairs, num_trials; + int i, rep, len; if( e->pending_num_pairs > 0 ) /* from previous call */ { @@ -189,8 +186,10 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, } else num_pairs = LZe_read_match_distances( e ); - main_len = ( num_pairs > 0 ) ? e->pairs[num_pairs-1].len : 0; + const int main_len = ( num_pairs > 0 ) ? e->pairs[num_pairs-1].len : 0; + int replens[num_rep_distances]; + int rep_index = 0; for( i = 0; i < num_rep_distances; ++i ) { replens[i] = Mb_true_match_len( &e->eb.mb, 0, reps[i] + 1 ); @@ -212,10 +211,7 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, return main_len; } - { const int pos_state = Mb_data_position( &e->eb.mb ) & pos_state_mask; - const int match_price = price1( e->eb.bm_match[state][pos_state] ); - const int rep_match_price = match_price + price1( e->eb.bm_rep[state] ); const uint8_t prev_byte = Mb_peek( &e->eb.mb, 1 ); const uint8_t cur_byte = Mb_peek( &e->eb.mb, 0 ); const uint8_t match_byte = Mb_peek( &e->eb.mb, reps[0] + 1 ); @@ -227,6 +223,9 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, e->trials[1].price += LZeb_price_matched( &e->eb, prev_byte, cur_byte, match_byte ); e->trials[1].dis4 = -1; /* literal */ + const int match_price = price1( e->eb.bm_match[state][pos_state] ); + const int rep_match_price = match_price + price1( e->eb.bm_rep[state] ); + if( match_byte == cur_byte ) Tr_update( &e->trials[1], rep_match_price + LZeb_price_shortrep( &e->eb, state, pos_state ), 0, 0 ); @@ -250,9 +249,8 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, for( rep = 0; rep < num_rep_distances; ++rep ) { - int price; if( replens[rep] < min_match_len ) continue; - price = rep_match_price + LZeb_price_rep( &e->eb, rep, state, pos_state ); + const int price = rep_match_price + LZeb_price_rep( &e->eb, rep, state, pos_state ); for( len = min_match_len; len <= replens[rep]; ++len ) Tr_update( &e->trials[len], price + Lp_price( &e->rep_len_prices, len, pos_state ), rep, 0 ); @@ -272,17 +270,10 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, if( ++len > e->pairs[i].len && ++i >= num_pairs ) break; } } - } + int cur = 0; while( true ) /* price optimization loop */ { - struct Trial *cur_trial, *next_trial; - int newlen, pos_state, triable_bytes, len_limit; - int start_len = min_match_len; - int next_price, match_price, rep_match_price; - State cur_state; - uint8_t prev_byte, cur_byte, match_byte; - Mb_move_pos( &e->eb.mb ); if( ++cur >= num_trials ) /* no more initialized trials */ { @@ -290,8 +281,8 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, return cur; } - num_pairs = LZe_read_match_distances( e ); - newlen = ( num_pairs > 0 ) ? e->pairs[num_pairs-1].len : 0; + const int num_pairs = LZe_read_match_distances( e ); + const int newlen = ( num_pairs > 0 ) ? e->pairs[num_pairs-1].len : 0; if( newlen >= e->match_len_limit ) { e->pending_num_pairs = num_pairs; @@ -300,7 +291,8 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, } /* give final values to current trial */ - cur_trial = &e->trials[cur]; + struct Trial * cur_trial = &e->trials[cur]; + State cur_state; { const int dis4 = cur_trial->dis4; int prev_index = cur_trial->prev_index; @@ -331,25 +323,25 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, mtf_reps( dis4, cur_trial->reps ); /* literal is ignored */ } - pos_state = Mb_data_position( &e->eb.mb ) & pos_state_mask; - prev_byte = Mb_peek( &e->eb.mb, 1 ); - cur_byte = Mb_peek( &e->eb.mb, 0 ); - match_byte = Mb_peek( &e->eb.mb, cur_trial->reps[0] + 1 ); + const int pos_state = Mb_data_position( &e->eb.mb ) & pos_state_mask; + const uint8_t prev_byte = Mb_peek( &e->eb.mb, 1 ); + const uint8_t cur_byte = Mb_peek( &e->eb.mb, 0 ); + const uint8_t match_byte = Mb_peek( &e->eb.mb, cur_trial->reps[0] + 1 ); - next_price = cur_trial->price + - price0( e->eb.bm_match[cur_state][pos_state] ); + int next_price = cur_trial->price + + price0( e->eb.bm_match[cur_state][pos_state] ); if( St_is_char( cur_state ) ) next_price += LZeb_price_literal( &e->eb, prev_byte, cur_byte ); else next_price += LZeb_price_matched( &e->eb, prev_byte, cur_byte, match_byte ); /* try last updates to next trial */ - next_trial = &e->trials[cur+1]; + struct Trial * next_trial = &e->trials[cur+1]; Tr_update( next_trial, next_price, -1, cur ); /* literal */ - match_price = cur_trial->price + price1( e->eb.bm_match[cur_state][pos_state] ); - rep_match_price = match_price + price1( e->eb.bm_rep[cur_state] ); + const int match_price = cur_trial->price + price1( e->eb.bm_match[cur_state][pos_state] ); + const int rep_match_price = match_price + price1( e->eb.bm_rep[cur_state] ); if( match_byte == cur_byte && next_trial->dis4 != 0 && next_trial->prev_index2 == single_step_trial ) @@ -364,11 +356,11 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, } } - triable_bytes = + const int triable_bytes = min( Mb_available_bytes( &e->eb.mb ), max_num_trials - 1 - cur ); if( triable_bytes < min_match_len ) continue; - len_limit = min( e->match_len_limit, triable_bytes ); + const int len_limit = min( e->match_len_limit, triable_bytes ); /* try literal + rep0 */ if( match_byte != cur_byte && next_trial->prev_index != cur ) @@ -392,19 +384,20 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, } } + int start_len = min_match_len; + /* try rep distances */ for( rep = 0; rep < num_rep_distances; ++rep ) { const uint8_t * const data = Mb_ptr_to_current_pos( &e->eb.mb ); const int dis = cur_trial->reps[rep] + 1; - int price; if( data[0-dis] != data[0] || data[1-dis] != data[1] ) continue; for( len = min_match_len; len < len_limit; ++len ) if( data[len-dis] != data[len] ) break; while( num_trials < cur + len ) e->trials[++num_trials].price = infinite_price; - price = rep_match_price + LZeb_price_rep( &e->eb, rep, cur_state, pos_state ); + int price = rep_match_price + LZeb_price_rep( &e->eb, rep, cur_state, pos_state ); for( i = min_match_len; i <= len; ++i ) Tr_update( &e->trials[cur+i], price + Lp_price( &e->rep_len_prices, i, pos_state ), rep, cur ); @@ -412,17 +405,14 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, if( rep == 0 ) start_len = len + 1; /* discard shorter matches */ /* try rep + literal + rep0 */ - { int len2 = len + 1; const int limit = min( e->match_len_limit + len2, triable_bytes ); - int pos_state2; - State state2; while( len2 < limit && data[len2-dis] == data[len2] ) ++len2; len2 -= len + 1; if( len2 < min_match_len ) continue; - pos_state2 = ( pos_state + len ) & pos_state_mask; - state2 = St_set_rep( cur_state ); + int pos_state2 = ( pos_state + len ) & pos_state_mask; + State state2 = St_set_rep( cur_state ); price += Lp_price( &e->rep_len_prices, len, pos_state ) + price0( e->eb.bm_match[state2][pos_state2] ) + LZeb_price_matched( &e->eb, data[len-1], data[len], data[len-dis] ); @@ -435,21 +425,19 @@ static int LZe_sequence_optimizer( struct LZ_encoder * const e, e->trials[++num_trials].price = infinite_price; Tr_update3( &e->trials[cur+len+1+len2], price, rep, cur + len + 1, cur ); } - } /* try matches */ if( newlen >= start_len && newlen <= len_limit ) { - int dis; const int normal_match_price = match_price + price0( e->eb.bm_rep[cur_state] ); while( num_trials < cur + newlen ) e->trials[++num_trials].price = infinite_price; - i = 0; + int i = 0; while( e->pairs[i].len < start_len ) ++i; - dis = e->pairs[i].dis; + int dis = e->pairs[i].dis; for( len = start_len; ; ++len ) { int price = normal_match_price + LZe_price_pair( e, dis, len, pos_state ); @@ -502,7 +490,7 @@ bool LZe_encode_member( struct LZ_encoder * const e, int price_counter = 0; /* counters may decrement below 0 */ int dis_price_counter = 0; int align_price_counter = 0; - int ahead, i; + int i; int reps[num_rep_distances]; State state = 0; for( i = 0; i < num_rep_distances; ++i ) reps[i] = 0; @@ -539,7 +527,7 @@ bool LZe_encode_member( struct LZ_encoder * const e, Lp_update_prices( &e->rep_len_prices ); } - ahead = LZe_sequence_optimizer( e, reps, state ); + int ahead = LZe_sequence_optimizer( e, reps, state ); price_counter -= ahead; for( i = 0; ahead > 0; ) @@ -556,14 +544,13 @@ bool LZe_encode_member( struct LZ_encoder * const e, const uint8_t prev_byte = Mb_peek( &e->eb.mb, ahead + 1 ); const uint8_t cur_byte = Mb_peek( &e->eb.mb, ahead ); CRC32_update_byte( &e->eb.crc, cur_byte ); - if( St_is_char( state ) ) + if( ( state = St_set_char( state ) ) < 4 ) LZeb_encode_literal( &e->eb, prev_byte, cur_byte ); else { const uint8_t match_byte = Mb_peek( &e->eb.mb, ahead + reps[0] + 1 ); LZeb_encode_matched( &e->eb, prev_byte, cur_byte, match_byte ); } - state = St_set_char( state ); } else /* match or repeated match */ { @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -188,10 +188,9 @@ static inline int LZeb_price_rep( const struct LZ_encoder_base * const eb, const int rep, const State state, const int pos_state ) { - int price; if( rep == 0 ) return price0( eb->bm_rep0[state] ) + price1( eb->bm_len[state][pos_state] ); - price = price1( eb->bm_rep0[state] ); + int price = price1( eb->bm_rep0[state] ); if( rep == 1 ) price += price0( eb->bm_rep1[state] ); else diff --git a/encoder_base.c b/encoder_base.c index 7dd7eca..91c4623 100644 --- a/encoder_base.c +++ b/encoder_base.c @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -40,8 +40,7 @@ bool Mb_read_block( struct Matchfinder_base * const mb ) mb->stream_pos += rd; if( rd != size && errno ) { show_error( "Read error", errno, false ); cleanup_and_fail( 1 ); } - if( rd < size ) - { mb->at_stream_end = true; mb->pos_limit = mb->buffer_size; } + if( rd < size ) { mb->at_stream_end = true; mb->pos_limit = mb->buffer_size; } } return mb->pos < mb->stream_pos; } @@ -77,7 +76,6 @@ bool Mb_init( struct Matchfinder_base * const mb, const int before_size, { const int buffer_size_limit = ( dict_factor * dict_size ) + before_size + after_size; - unsigned size; int i; mb->partial_data_pos = 0; @@ -107,9 +105,8 @@ bool Mb_init( struct Matchfinder_base * const mb, const int before_size, mb->dictionary_size = dict_size; mb->pos_limit = mb->buffer_size; if( !mb->at_stream_end ) mb->pos_limit -= after_size; - size = 1 << max( 16, real_bits( mb->dictionary_size - 1 ) - 2 ); - if( mb->dictionary_size > 1 << 26 ) /* 64 MiB */ - size >>= 1; + unsigned size = 1 << max( 16, real_bits( mb->dictionary_size - 1 ) - 2 ); + if( mb->dictionary_size > 1 << 26 ) size >>= 1; /* 64 MiB */ mb->key4_mask = size - 1; /* increases with dictionary size */ size += num_prev_positions23; mb->num_prev_positions = size; @@ -138,11 +135,9 @@ void Mb_reset( struct Matchfinder_base * const mb ) Mb_read_block( mb ); if( mb->at_stream_end && mb->stream_pos < mb->dictionary_size ) { - int size; mb->dictionary_size = max( min_dictionary_size, mb->stream_pos ); - size = 1 << max( 16, real_bits( mb->dictionary_size - 1 ) - 2 ); - if( mb->dictionary_size > 1 << 26 ) /* 64 MiB */ - size >>= 1; + int size = 1 << max( 16, real_bits( mb->dictionary_size - 1 ) - 2 ); + if( mb->dictionary_size > 1 << 26 ) size >>= 1; /* 64 MiB */ mb->key4_mask = size - 1; size += mb->num_prev_positions23; mb->num_prev_positions = size; @@ -169,16 +164,16 @@ void Re_flush_data( struct Range_encoder * const renc ) /* End Of Stream marker => (dis == 0xFFFFFFFFU, len == min_match_len) */ void LZeb_full_flush( struct LZ_encoder_base * const eb, const State state ) { - int i; const int pos_state = Mb_data_position( &eb->mb ) & pos_state_mask; - Lzip_trailer trailer; Re_encode_bit( &eb->renc, &eb->bm_match[state][pos_state], 1 ); Re_encode_bit( &eb->renc, &eb->bm_rep[state], 0 ); LZeb_encode_pair( eb, 0xFFFFFFFFU, min_match_len, pos_state ); Re_flush( &eb->renc ); + Lzip_trailer trailer; Lt_set_data_crc( trailer, LZeb_crc( eb ) ); Lt_set_data_size( trailer, Mb_data_position( &eb->mb ) ); Lt_set_member_size( trailer, Re_member_position( &eb->renc ) + Lt_size ); + int i; for( i = 0; i < Lt_size; ++i ) Re_put_byte( &eb->renc, trailer[i] ); Re_flush_data( &eb->renc ); diff --git a/encoder_base.h b/encoder_base.h index 64f147b..f47ec48 100644 --- a/encoder_base.h +++ b/encoder_base.h @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -83,10 +83,9 @@ static inline int price_bit( const Bit_model bm, const bool bit ) static inline int price_symbol3( const Bit_model bm[], int symbol ) { - int price; bool bit = symbol & 1; symbol |= 8; symbol >>= 1; - price = price_bit( bm[symbol], bit ); + int price = price_bit( bm[symbol], bit ); bit = symbol & 1; symbol >>= 1; price += price_bit( bm[symbol], bit ); return price + price_bit( bm[1], symbol & 1 ); } @@ -94,10 +93,9 @@ static inline int price_symbol3( const Bit_model bm[], int symbol ) static inline int price_symbol6( const Bit_model bm[], unsigned symbol ) { - int price; bool bit = symbol & 1; symbol |= 64; symbol >>= 1; - price = price_bit( bm[symbol], bit ); + int price = price_bit( bm[symbol], bit ); bit = symbol & 1; symbol >>= 1; price += price_bit( bm[symbol], bit ); bit = symbol & 1; symbol >>= 1; price += price_bit( bm[symbol], bit ); bit = symbol & 1; symbol >>= 1; price += price_bit( bm[symbol], bit ); @@ -108,10 +106,9 @@ static inline int price_symbol6( const Bit_model bm[], unsigned symbol ) static inline int price_symbol8( const Bit_model bm[], int symbol ) { - int price; bool bit = symbol & 1; symbol |= 0x100; symbol >>= 1; - price = price_bit( bm[symbol], bit ); + int price = price_bit( bm[symbol], bit ); bit = symbol & 1; symbol >>= 1; price += price_bit( bm[symbol], bit ); bit = symbol & 1; symbol >>= 1; price += price_bit( bm[symbol], bit ); bit = symbol & 1; symbol >>= 1; price += price_bit( bm[symbol], bit ); @@ -307,8 +304,7 @@ static inline void Re_encode( struct Range_encoder * const renc, { renc->range >>= 1; if( symbol & mask ) renc->low += renc->range; - if( renc->range <= 0x00FFFFFFU ) - { renc->range <<= 8; Re_shift_low( renc ); } + if( renc->range <= 0x00FFFFFFU ) { renc->range <<= 8; Re_shift_low( renc ); } } } @@ -333,10 +329,9 @@ static inline void Re_encode_bit( struct Range_encoder * const renc, static inline void Re_encode_tree3( struct Range_encoder * const renc, Bit_model bm[], const int symbol ) { - int model; bool bit = ( symbol >> 2 ) & 1; Re_encode_bit( renc, &bm[1], bit ); - model = 2 | bit; + int model = 2 | bit; bit = ( symbol >> 1 ) & 1; Re_encode_bit( renc, &bm[model], bit ); model <<= 1; model |= bit; Re_encode_bit( renc, &bm[model], symbol & 1 ); @@ -345,10 +340,9 @@ static inline void Re_encode_tree3( struct Range_encoder * const renc, static inline void Re_encode_tree6( struct Range_encoder * const renc, Bit_model bm[], const unsigned symbol ) { - int model; bool bit = ( symbol >> 5 ) & 1; Re_encode_bit( renc, &bm[1], bit ); - model = 2 | bit; + int model = 2 | bit; bit = ( symbol >> 4 ) & 1; Re_encode_bit( renc, &bm[model], bit ); model <<= 1; model |= bit; bit = ( symbol >> 3 ) & 1; @@ -479,8 +473,7 @@ static inline int LZeb_price_matched( const struct LZ_encoder_base * const eb, static inline void LZeb_encode_literal( struct LZ_encoder_base * const eb, const uint8_t prev_byte, const uint8_t symbol ) - { Re_encode_tree8( &eb->renc, eb->bm_literal[get_lit_state(prev_byte)], - symbol ); } + { Re_encode_tree8( &eb->renc, eb->bm_literal[get_lit_state(prev_byte)], symbol ); } static inline void LZeb_encode_matched( struct LZ_encoder_base * const eb, const uint8_t prev_byte, const uint8_t symbol, const uint8_t match_byte ) @@ -491,8 +484,8 @@ static inline void LZeb_encode_pair( struct LZ_encoder_base * const eb, const unsigned dis, const int len, const int pos_state ) { - const unsigned dis_slot = get_slot( dis ); Re_encode_len( &eb->renc, &eb->match_len_model, len, pos_state ); + const unsigned dis_slot = get_slot( dis ); Re_encode_tree6( &eb->renc, eb->bm_dis_slot[get_len_state(len)], dis_slot ); if( dis_slot >= start_dis_model ) diff --git a/fast_encoder.c b/fast_encoder.c index 1904428..f69b999 100644 --- a/fast_encoder.c +++ b/fast_encoder.c @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -31,25 +31,24 @@ int FLZe_longest_match_len( struct FLZ_encoder * const fe, int * const distance ) { enum { len_limit = 16 }; - const uint8_t * const data = Mb_ptr_to_current_pos( &fe->eb.mb ); - int32_t * ptr0 = fe->eb.mb.pos_array + fe->eb.mb.cyclic_pos; - const int pos1 = fe->eb.mb.pos + 1; - int maxlen = 0, newpos1, count; const int available = min( Mb_available_bytes( &fe->eb.mb ), max_match_len ); if( available < len_limit ) return 0; + const uint8_t * const data = Mb_ptr_to_current_pos( &fe->eb.mb ); fe->key4 = ( ( fe->key4 << 4 ) ^ data[3] ) & fe->eb.mb.key4_mask; - newpos1 = fe->eb.mb.prev_positions[fe->key4]; + const int pos1 = fe->eb.mb.pos + 1; + int newpos1 = fe->eb.mb.prev_positions[fe->key4]; fe->eb.mb.prev_positions[fe->key4] = pos1; + int32_t * ptr0 = fe->eb.mb.pos_array + fe->eb.mb.cyclic_pos; + int maxlen = 0, count; for( count = 4; ; ) { - int32_t * newptr; int delta; if( newpos1 <= 0 || --count < 0 || ( delta = pos1 - newpos1 ) > fe->eb.mb.dictionary_size ) { *ptr0 = 0; break; } - newptr = fe->eb.mb.pos_array + + int32_t * const newptr = fe->eb.mb.pos_array + ( fe->eb.mb.cyclic_pos - delta + ( ( fe->eb.mb.cyclic_pos >= delta ) ? 0 : fe->eb.mb.dictionary_size + 1 ) ); @@ -118,11 +117,10 @@ bool FLZe_encode_member( struct FLZ_encoder * const fe, Re_encode_bit( &fe->eb.renc, &fe->eb.bm_len[state][pos_state], 1 ); else { - int distance; Re_encode_bit( &fe->eb.renc, &fe->eb.bm_rep1[state], rep > 1 ); if( rep > 1 ) Re_encode_bit( &fe->eb.renc, &fe->eb.bm_rep2[state], rep > 2 ); - distance = reps[rep]; + const int distance = reps[rep]; for( i = rep; i > 0; --i ) reps[i] = reps[i-1]; reps[0] = distance; } @@ -147,7 +145,6 @@ bool FLZe_encode_member( struct FLZ_encoder * const fe, continue; } - { const uint8_t prev_byte = Mb_peek( &fe->eb.mb, 1 ); const uint8_t cur_byte = Mb_peek( &fe->eb.mb, 0 ); const uint8_t match_byte = Mb_peek( &fe->eb.mb, reps[0] + 1 ); @@ -178,12 +175,10 @@ bool FLZe_encode_member( struct FLZ_encoder * const fe, /* literal byte */ Re_encode_bit( &fe->eb.renc, &fe->eb.bm_match[state][pos_state], 0 ); - if( St_is_char( state ) ) + if( ( state = St_set_char( state ) ) < 4 ) LZeb_encode_literal( &fe->eb, prev_byte, cur_byte ); else LZeb_encode_matched( &fe->eb, prev_byte, cur_byte, match_byte ); - state = St_set_char( state ); - } } LZeb_full_flush( &fe->eb, state ); diff --git a/fast_encoder.h b/fast_encoder.h index 868128c..64ea2cc 100644 --- a/fast_encoder.h +++ b/fast_encoder.h @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -52,17 +52,15 @@ int list_files( const char * const filenames[], const int num_filenames, bool stdin_used = false; for( i = 0; i < num_filenames; ++i ) { - const char * input_filename; - struct Lzip_index lzip_index; - struct stat in_stats; /* not used */ - int infd; const bool from_stdin = ( strcmp( filenames[i], "-" ) == 0 ); if( from_stdin ) { if( stdin_used ) continue; else stdin_used = true; } - input_filename = from_stdin ? "(stdin)" : filenames[i]; - infd = from_stdin ? STDIN_FILENO : + const char * const input_filename = from_stdin ? "(stdin)" : filenames[i]; + struct stat in_stats; /* not used */ + const int infd = from_stdin ? STDIN_FILENO : open_instream( input_filename, &in_stats, false, true ); if( infd < 0 ) { set_retval( &retval, 1 ); continue; } + struct Lzip_index lzip_index; Li_init( &lzip_index, infd, ignore_trailing, loose_trailing ); close( infd ); if( lzip_index.retval != 0 ) @@ -71,37 +69,36 @@ int list_files( const char * const filenames[], const int num_filenames, set_retval( &retval, lzip_index.retval ); Li_free( &lzip_index ); continue; } - if( verbosity >= 0 ) + if( verbosity < 0 ) { Li_free( &lzip_index ); continue; } + const unsigned long long udata_size = Li_udata_size( &lzip_index ); + const unsigned long long cdata_size = Li_cdata_size( &lzip_index ); + total_comp += cdata_size; total_uncomp += udata_size; ++files; + const long members = lzip_index.members; + if( first_post ) { - const unsigned long long udata_size = Li_udata_size( &lzip_index ); - const unsigned long long cdata_size = Li_cdata_size( &lzip_index ); - total_comp += cdata_size; total_uncomp += udata_size; ++files; - if( first_post ) - { - first_post = false; - if( verbosity >= 1 ) fputs( " dict memb trail ", stdout ); - fputs( " uncompressed compressed saved name\n", stdout ); - } - if( verbosity >= 1 ) - printf( "%s %5ld %6lld ", format_ds( lzip_index.dictionary_size ), - lzip_index.members, Li_file_size( &lzip_index ) - cdata_size ); - list_line( udata_size, cdata_size, input_filename ); + first_post = false; + if( verbosity >= 1 ) fputs( " dict memb trail ", stdout ); + fputs( " uncompressed compressed saved name\n", stdout ); + } + if( verbosity >= 1 ) + printf( "%s %5ld %6lld ", format_ds( lzip_index.dictionary_size ), + members, Li_file_size( &lzip_index ) - cdata_size ); + list_line( udata_size, cdata_size, input_filename ); - if( verbosity >= 2 && lzip_index.members > 1 ) + if( verbosity >= 2 && members > 1 ) + { + long i; + fputs( " member data_pos data_size member_pos member_size\n", stdout ); + for( i = 0; i < members; ++i ) { - long i; - fputs( " member data_pos data_size member_pos member_size\n", stdout ); - for( i = 0; i < lzip_index.members; ++i ) - { - const struct Block * db = Li_dblock( &lzip_index, i ); - const struct Block * mb = Li_mblock( &lzip_index, i ); - printf( "%6ld %14llu %14llu %14llu %14llu\n", - i + 1, db->pos, db->size, mb->pos, mb->size ); - } - first_post = true; /* reprint heading after list of members */ + const struct Block * db = Li_dblock( &lzip_index, i ); + const struct Block * mb = Li_mblock( &lzip_index, i ); + printf( "%6ld %14llu %14llu %14llu %14llu\n", + i + 1, db->pos, db->size, mb->pos, mb->size ); } - fflush( stdout ); + first_post = true; /* reprint heading after list of members */ } + fflush( stdout ); Li_free( &lzip_index ); } if( verbosity >= 0 && files > 1 ) @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -136,6 +136,7 @@ static inline void CRC32_init( void ) static inline void CRC32_update_byte( uint32_t * const crc, const uint8_t byte ) { *crc = crc32[(*crc^byte)&0xFF] ^ ( *crc >> 8 ); } +/* about as fast as it is possible without messing with endianness */ static inline void CRC32_update_buf( uint32_t * const crc, const uint8_t * const buffer, const int size ) @@ -269,12 +270,12 @@ static inline bool Lt_verify_consistency( const Lzip_trailer data ) { const unsigned crc = Lt_get_data_crc( data ); const unsigned long long dsize = Lt_get_data_size( data ); - const unsigned long long msize = Lt_get_member_size( data ); - const unsigned long long mlimit = ( 9 * dsize + 7 ) / 8 + min_member_size; - const unsigned long long dlimit = 7090 * ( msize - 26 ) - 1; if( ( crc == 0 ) != ( dsize == 0 ) ) return false; + const unsigned long long msize = Lt_get_member_size( data ); if( msize < min_member_size ) return false; + const unsigned long long mlimit = ( 9 * dsize + 7 ) / 8 + min_member_size; if( mlimit > dsize && msize > mlimit ) return false; + const unsigned long long dlimit = 7090 * ( msize - 26 ) - 1; if( dlimit > msize && dsize > dlimit ) return false; return true; } diff --git a/lzip_index.c b/lzip_index.c index 5db5022..41d0e13 100644 --- a/lzip_index.c +++ b/lzip_index.c @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -133,44 +133,38 @@ static bool Li_skip_trailing_data( struct Lzip_index * const li, const int fd, const bool ignore_trailing, const bool loose_trailing ) { + if( *pos < min_member_size ) return false; enum { block_size = 16384, buffer_size = block_size + Lt_size - 1 + Lh_size }; uint8_t buffer[buffer_size]; int bsize = *pos % block_size; /* total bytes in buffer */ - int search_size, rd_size; - unsigned long long ipos; - int i; - if( *pos < min_member_size ) return false; if( bsize <= buffer_size - block_size ) bsize += block_size; - search_size = bsize; /* bytes to search for trailer */ - rd_size = bsize; /* bytes to read from file */ - ipos = *pos - rd_size; /* aligned to block_size */ + int search_size = bsize; /* bytes to search for trailer */ + int rd_size = bsize; /* bytes to read from file */ + unsigned long long ipos = *pos - rd_size; /* aligned to block_size */ while( true ) { - const uint8_t max_msb = ( ipos + search_size ) >> 56; if( seek_read( fd, buffer, rd_size, ipos ) != rd_size ) - { Li_set_errno_error( li, "Error seeking member trailer: " ); - return false; } + { Li_set_errno_error( li, "Error seeking member trailer: " ); return false; } + const uint8_t max_msb = ( ipos + search_size ) >> 56; + int i; for( i = search_size; i >= Lt_size; --i ) if( buffer[i-1] <= max_msb ) /* most significant byte of member_size */ { - Lzip_header header; - const Lzip_header * header2; const Lzip_trailer * const trailer = (const Lzip_trailer *)( buffer + i - Lt_size ); const unsigned long long member_size = Lt_get_member_size( *trailer ); - unsigned dictionary_size; - bool full_h2; if( member_size == 0 ) /* skip trailing zeros */ { while( i > Lt_size && buffer[i-9] == 0 ) --i; continue; } if( member_size > ipos + i || !Lt_verify_consistency( *trailer ) ) continue; + Lzip_header header; if( !Li_read_header( li, fd, header, ipos + i - member_size ) ) return false; if( !Lh_verify( header ) ) continue; - header2 = (const Lzip_header *)( buffer + i ); - full_h2 = bsize - i >= Lh_size; + const Lzip_header * header2 = (const Lzip_header *)( buffer + i ); + const bool full_h2 = bsize - i >= Lh_size; if( Lh_verify_prefix( *header2, bsize - i ) ) /* last member */ { if( !full_h2 ) add_error( li, "Last member in input file is truncated." ); @@ -183,7 +177,7 @@ static bool Li_skip_trailing_data( struct Lzip_index * const li, const int fd, if( !ignore_trailing ) { add_error( li, trailing_msg ); li->retval = 2; return false; } *pos = ipos + i - member_size; - dictionary_size = Lh_get_dictionary_size( header ); + const unsigned dictionary_size = Lh_get_dictionary_size( header ); if( li->dictionary_size < dictionary_size ) li->dictionary_size = dictionary_size; return push_back_member( li, 0, Lt_get_data_size( *trailer ), *pos, @@ -204,9 +198,6 @@ static bool Li_skip_trailing_data( struct Lzip_index * const li, const int fd, bool Li_init( struct Lzip_index * const li, const int infd, const bool ignore_trailing, const bool loose_trailing ) { - Lzip_header header; - unsigned long long pos; - long i; li->member_vector = 0; li->error = 0; li->insize = lseek( infd, 0, SEEK_END ); @@ -223,18 +214,17 @@ bool Li_init( struct Lzip_index * const li, const int infd, { add_error( li, "Input file is too long (2^63 bytes or more)." ); li->retval = 2; return false; } + Lzip_header header; if( !Li_read_header( li, infd, header, 0 ) ) return false; if( Li_check_header_error( li, header ) ) return false; - pos = li->insize; /* always points to a header or to EOF */ + unsigned long long pos = li->insize; /* always points to a header or to EOF */ while( pos >= min_member_size ) { Lzip_trailer trailer; - unsigned long long member_size; - unsigned dictionary_size; if( seek_read( infd, trailer, Lt_size, pos - Lt_size ) != Lt_size ) { Li_set_errno_error( li, "Error reading member trailer: " ); break; } - member_size = Lt_get_member_size( trailer ); + const unsigned long long member_size = Lt_get_member_size( trailer ); if( member_size > pos || !Lt_verify_consistency( trailer ) ) { /* bad trailer */ if( li->members <= 0 ) @@ -253,7 +243,7 @@ bool Li_init( struct Lzip_index * const li, const int infd, break; } pos -= member_size; - dictionary_size = Lh_get_dictionary_size( header ); + const unsigned dictionary_size = Lh_get_dictionary_size( header ); if( li->dictionary_size < dictionary_size ) li->dictionary_size = dictionary_size; if( !push_back_member( li, 0, Lt_get_data_size( trailer ), pos, @@ -268,6 +258,7 @@ bool Li_init( struct Lzip_index * const li, const int infd, return false; } Li_reverse_member_vector( li ); + long i; for( i = 0; ; ++i ) { const long long end = block_end( li->member_vector[i].dblock ); diff --git a/lzip_index.h b/lzip_index.h index d7db038..b76560e 100644 --- a/lzip_index.h +++ b/lzip_index.h @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -1,5 +1,5 @@ /* Clzip - LZMA lossless data compressor - Copyright (C) 2010-2021 Antonio Diaz Diaz. + Copyright (C) 2010-2022 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -18,7 +18,7 @@ Exit status: 0 for a normal exit, 1 for environmental problems (file not found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or invalid input file, 3 for an internal consistency error - (eg, bug) which caused clzip to panic. + (e.g., bug) which caused clzip to panic. */ #define _FILE_OFFSET_BITS 64 @@ -36,9 +36,9 @@ #include <unistd.h> #include <utime.h> #include <sys/stat.h> -#if defined(__MSVCRT__) || defined(__OS2__) || defined(__DJGPP__) +#if defined __MSVCRT__ || defined __OS2__ || defined __DJGPP__ #include <io.h> -#if defined(__MSVCRT__) +#if defined __MSVCRT__ #define fchmod(x,y) 0 #define fchown(x,y,z) 0 #define strtoull strtoul @@ -51,7 +51,7 @@ #define S_IWOTH 0 #endif #endif -#if defined(__DJGPP__) +#if defined __DJGPP__ #define S_ISSOCK(x) 0 #define S_ISVTX 0 #endif @@ -72,10 +72,15 @@ #error "Environments where CHAR_BIT != 8 are not supported." #endif +#if ( defined SIZE_MAX && SIZE_MAX < UINT_MAX ) || \ + ( defined SSIZE_MAX && SSIZE_MAX < INT_MAX ) +#error "Environments where 'size_t' is narrower than 'int' are not supported." +#endif + int verbosity = 0; static const char * const program_name = "clzip"; -static const char * const program_year = "2021"; +static const char * const program_year = "2022"; static const char * invocation_name = "clzip"; /* default value */ static const struct { const char * from; const char * to; } known_extensions[] = { @@ -106,13 +111,14 @@ static void show_help( void ) "C++ compiler.\n" "\nLzip is a lossless data compressor with a user interface similar to the one\n" "of gzip or bzip2. Lzip uses a simplified form of the 'Lempel-Ziv-Markov\n" - "chain-Algorithm' (LZMA) stream format, chosen to maximize safety and\n" - "interoperability. Lzip can compress about as fast as gzip (lzip -0) or\n" - "compress most files more than bzip2 (lzip -9). Decompression speed is\n" - "intermediate between gzip and bzip2. Lzip is better than gzip and bzip2 from\n" - "a data recovery perspective. Lzip has been designed, written, and tested\n" - "with great care to replace gzip and bzip2 as the standard general-purpose\n" - "compressed format for unix-like systems.\n" + "chain-Algorithm' (LZMA) stream format and provides a 3 factor integrity\n" + "checking to maximize interoperability and optimize safety. Lzip can compress\n" + "about as fast as gzip (lzip -0) or compress most files more than bzip2\n" + "(lzip -9). Decompression speed is intermediate between gzip and bzip2.\n" + "Lzip is better than gzip and bzip2 from a data recovery perspective. Lzip\n" + "has been designed, written, and tested with great care to replace gzip and\n" + "bzip2 as the standard general-purpose compressed format for unix-like\n" + "systems.\n" "\nUsage: %s [options] [files]\n", invocation_name ); printf( "\nOptions:\n" " -h, --help display this help and exit\n" @@ -150,7 +156,7 @@ static void show_help( void ) "'tar -xf foo.tar.lz' or 'clzip -cd foo.tar.lz | tar -xf -'.\n" "\nExit status: 0 for a normal exit, 1 for environmental problems (file\n" "not found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or\n" - "invalid input file, 3 for an internal consistency error (eg, bug) which\n" + "invalid input file, 3 for an internal consistency error (e.g., bug) which\n" "caused clzip to panic.\n" "\nThe ideas embodied in clzip are due to (at least) the following people:\n" "Abraham Lempel and Jacob Ziv (for the LZ algorithm), Andrey Markov (for the\n" @@ -194,8 +200,6 @@ struct Pretty_print static void Pp_init( struct Pretty_print * const pp, const char * const filenames[], const int num_filenames ) { - unsigned stdin_name_len; - int i; pp->name = 0; pp->padded_name = 0; pp->stdin_name = "(stdin)"; @@ -203,7 +207,8 @@ static void Pp_init( struct Pretty_print * const pp, pp->first_post = false; if( verbosity <= 0 ) return; - stdin_name_len = strlen( pp->stdin_name ); + const unsigned stdin_name_len = strlen( pp->stdin_name ); + int i; for( i = 0; i < num_filenames; ++i ) { const char * const s = filenames[i]; @@ -237,16 +242,14 @@ static void Pp_reset( struct Pretty_print * const pp ) void Pp_show_msg( struct Pretty_print * const pp, const char * const msg ) { - if( verbosity >= 0 ) + if( verbosity < 0 ) return; + if( pp->first_post ) { - if( pp->first_post ) - { - pp->first_post = false; - fputs( pp->padded_name, stderr ); - if( !msg ) fflush( stderr ); - } - if( msg ) fprintf( stderr, "%s\n", msg ); + pp->first_post = false; + fputs( pp->padded_name, stderr ); + if( !msg ) fflush( stderr ); } + if( msg ) fprintf( stderr, "%s\n", msg ); } @@ -284,17 +287,53 @@ void show_header( const unsigned dictionary_size ) } -static unsigned long long getnum( const char * const ptr, +/* separate large numbers >= 100_000 in groups of 3 digits using '_' */ +static const char * format_num3( unsigned long long num ) + { + const char * const si_prefix = "kMGTPEZY"; + const char * const binary_prefix = "KMGTPEZY"; + enum { buffers = 8, bufsize = 4 * sizeof (long long) }; + static char buffer[buffers][bufsize]; /* circle of static buffers for printf */ + static int current = 0; + int i; + char * const buf = buffer[current++]; current %= buffers; + char * p = buf + bufsize - 1; /* fill the buffer backwards */ + *p = 0; /* terminator */ + if( num > 1024 ) + { + char prefix = 0; /* try binary first, then si */ + for( i = 0; i < 8 && num >= 1024 && num % 1024 == 0; ++i ) + { num /= 1024; prefix = binary_prefix[i]; } + if( prefix ) *(--p) = 'i'; + else + for( i = 0; i < 8 && num >= 1000 && num % 1000 == 0; ++i ) + { num /= 1000; prefix = si_prefix[i]; } + if( prefix ) *(--p) = prefix; + } + const bool split = num >= 100000; + + for( i = 0; ; ) + { + *(--p) = num % 10 + '0'; num /= 10; if( num == 0 ) break; + if( split && ++i >= 3 ) { i = 0; *(--p) = '_'; } + } + return p; + } + + +static unsigned long long getnum( const char * const arg, + const char * const option_name, const unsigned long long llimit, const unsigned long long ulimit ) { - unsigned long long result; char * tail; errno = 0; - result = strtoull( ptr, &tail, 0 ); - if( tail == ptr ) + unsigned long long result = strtoull( arg, &tail, 0 ); + if( tail == arg ) { - show_error( "Bad or missing numerical argument.", 0, true ); + if( verbosity >= 0 ) + fprintf( stderr, "%s: Bad or missing numerical argument in " + "option '%s'.\n", program_name, option_name ); exit( 1 ); } @@ -317,7 +356,9 @@ static unsigned long long getnum( const char * const ptr, } if( exponent <= 0 ) { - show_error( "Bad multiplier in numerical argument.", 0, true ); + if( verbosity >= 0 ) + fprintf( stderr, "%s: Bad multiplier in numerical argument of " + "option '%s'.\n", program_name, option_name ); exit( 1 ); } for( i = 0; i < exponent; ++i ) @@ -329,21 +370,24 @@ static unsigned long long getnum( const char * const ptr, if( !errno && ( result < llimit || result > ulimit ) ) errno = ERANGE; if( errno ) { - show_error( "Numerical argument out of limits.", 0, false ); + if( verbosity >= 0 ) + fprintf( stderr, "%s: Numerical argument out of limits [%s,%s] " + "in option '%s'.\n", program_name, format_num3( llimit ), + format_num3( ulimit ), option_name ); exit( 1 ); } return result; } -static int get_dict_size( const char * const arg ) +static int get_dict_size( const char * const arg, const char * const option_name ) { char * tail; const long bits = strtol( arg, &tail, 0 ); if( bits >= min_dictionary_bits && bits <= max_dictionary_bits && *tail == 0 ) return 1 << bits; - return getnum( arg, min_dictionary_size, max_dictionary_size ); + return getnum( arg, option_name, min_dictionary_size, max_dictionary_size ); } @@ -519,7 +563,7 @@ static bool check_tty_in( const char * const input_filename, const int infd, isatty( infd ) ) /* for example /dev/tty */ { show_file_error( input_filename, "I won't read compressed data from a terminal.", 0 ); - close( infd ); set_retval( retval, 1 ); + close( infd ); set_retval( retval, 2 ); if( program_mode != m_test ) cleanup_and_fail( *retval ); return false; } return true; @@ -600,7 +644,6 @@ static int compress( const unsigned long long cfile_size, struct Pretty_print * const pp, const struct stat * const in_statsp, const bool zero ) { - unsigned long long in_size = 0, out_size = 0, partial_volume_size = 0; int retval = 0; struct Poly_encoder encoder = { 0, 0, 0 }; /* polymorphic encoder */ if( verbosity >= 1 ) Pp_show_msg( pp, 0 ); @@ -633,6 +676,7 @@ static int compress( const unsigned long long cfile_size, } } + unsigned long long in_size = 0, out_size = 0, partial_volume_size = 0; while( true ) /* encode one member per iteration */ { const unsigned long long size = ( volume_size > 0 ) ? @@ -729,12 +773,9 @@ static int decompress( const unsigned long long cfile_size, const int infd, for( first_member = true; ; first_member = false ) { - int result, size; - unsigned dictionary_size; Lzip_header header; - struct LZ_decoder decoder; Rd_reset_member_position( &rdec ); - size = Rd_read_data( &rdec, header, Lh_size ); + const int size = Rd_read_data( &rdec, header, Lh_size ); if( Rd_finished( &rdec ) ) /* End Of File */ { if( first_member ) @@ -764,17 +805,18 @@ static int decompress( const unsigned long long cfile_size, const int infd, if( !Lh_verify_version( header ) ) { Pp_show_msg( pp, bad_version( Lh_version( header ) ) ); retval = 2; break; } - dictionary_size = Lh_get_dictionary_size( header ); + const unsigned dictionary_size = Lh_get_dictionary_size( header ); if( !isvalid_ds( dictionary_size ) ) { Pp_show_msg( pp, bad_dict_msg ); retval = 2; break; } if( verbosity >= 2 || ( verbosity == 1 && first_member ) ) Pp_show_msg( pp, 0 ); + struct LZ_decoder decoder; if( !LZd_init( &decoder, &rdec, dictionary_size, outfd ) ) { Pp_show_msg( pp, mem_msg ); retval = 1; break; } show_dprogress( cfile_size, partial_file_pos, &rdec, pp ); /* init */ - result = LZd_decode_member( &decoder, pp ); + const int result = LZd_decode_member( &decoder, pp ); partial_file_pos += Rd_member_position( &rdec ); LZd_free( &decoder ); if( result != 0 ) @@ -911,24 +953,16 @@ int main( const int argc, const char * const argv[] ) unsigned long long member_size = max_member_size; unsigned long long volume_size = 0; const char * default_output_filename = ""; - static struct Arg_parser parser; /* static because valgrind complains */ - static struct Pretty_print pp; /* and memory management in C sucks */ - static const char ** filenames = 0; - int num_filenames = 0; enum Mode program_mode = m_compress; - int argind = 0; - int failed_tests = 0; - int retval = 0; int i; - bool filenames_given = false; bool force = false; bool ignore_trailing = true; bool keep_input_files = false; bool loose_trailing = false; bool recompress = false; - bool stdin_used = false; bool to_stdout = false; bool zero = false; + if( argc > 0 ) invocation_name = argv[0]; enum { opt_lt = 256 }; const struct ap_Option options[] = @@ -964,19 +998,22 @@ int main( const int argc, const char * const argv[] ) { opt_lt, "loose-trailing", ap_no }, { 0, 0, ap_no } }; - if( argc > 0 ) invocation_name = argv[0]; CRC32_init(); + /* static because valgrind complains and memory management in C sucks */ + static struct Arg_parser parser; if( !ap_init( &parser, argc, argv, options, 0 ) ) { show_error( mem_msg, 0, false ); return 1; } if( ap_error( &parser ) ) /* bad option */ { show_error( ap_error( &parser ), 0, true ); return 1; } + int argind = 0; for( ; argind < ap_arguments( &parser ); ++argind ) { const int code = ap_code( &parser, argind ); - const char * const arg = ap_argument( &parser, argind ); if( !code ) break; /* no more options */ + const char * const pn = ap_parsed_name( &parser, argind ); + const char * const arg = ap_argument( &parser, argind ); switch( code ) { case '0': case '1': case '2': case '3': case '4': @@ -984,7 +1021,7 @@ int main( const int argc, const char * const argv[] ) zero = ( code == '0' ); encoder_options = option_mapping[code-'0']; break; case 'a': ignore_trailing = false; break; - case 'b': member_size = getnum( arg, 100000, max_member_size ); break; + case 'b': member_size = getnum( arg, pn, 100000, max_member_size ); break; case 'c': to_stdout = true; break; case 'd': set_mode( &program_mode, m_decompress ); break; case 'f': force = true; break; @@ -993,15 +1030,15 @@ int main( const int argc, const char * const argv[] ) case 'k': keep_input_files = true; break; case 'l': set_mode( &program_mode, m_list ); break; case 'm': encoder_options.match_len_limit = - getnum( arg, min_match_len_limit, max_match_len ); + getnum( arg, pn, min_match_len_limit, max_match_len ); zero = false; break; case 'n': break; case 'o': if( strcmp( arg, "-" ) == 0 ) to_stdout = true; else { default_output_filename = arg; } break; case 'q': verbosity = -1; break; - case 's': encoder_options.dictionary_size = get_dict_size( arg ); + case 's': encoder_options.dictionary_size = get_dict_size( arg, pn ); zero = false; break; - case 'S': volume_size = getnum( arg, 100000, max_volume_size ); break; + case 'S': volume_size = getnum( arg, pn, 100000, max_volume_size ); break; case 't': set_mode( &program_mode, m_test ); break; case 'v': if( verbosity < 4 ) ++verbosity; break; case 'V': show_version(); return 0; @@ -1010,15 +1047,17 @@ int main( const int argc, const char * const argv[] ) } } /* end process options */ -#if defined(__MSVCRT__) || defined(__OS2__) || defined(__DJGPP__) +#if defined __MSVCRT__ || defined __OS2__ || defined __DJGPP__ setmode( STDIN_FILENO, O_BINARY ); setmode( STDOUT_FILENO, O_BINARY ); #endif - num_filenames = max( 1, ap_arguments( &parser ) - argind ); + static const char ** filenames = 0; + int num_filenames = max( 1, ap_arguments( &parser ) - argind ); filenames = resize_buffer( filenames, num_filenames * sizeof filenames[0] ); filenames[0] = "-"; + bool filenames_given = false; for( i = 0; argind + i < ap_arguments( &parser ); ++i ) { filenames[i] = ap_argument( &parser, argind + i ); @@ -1052,17 +1091,18 @@ int main( const int argc, const char * const argv[] ) if( !to_stdout && program_mode != m_test && ( filenames_given || to_file ) ) set_signals( signal_handler ); + static struct Pretty_print pp; Pp_init( &pp, filenames, num_filenames ); + int failed_tests = 0; + int retval = 0; const bool one_to_one = !to_stdout && program_mode != m_test && !to_file; + bool stdin_used = false; for( i = 0; i < num_filenames; ++i ) { - unsigned long long cfile_size; const char * input_filename = ""; int infd; - int tmp; struct stat in_stats; - const struct stat * in_statsp; Pp_set_name( &pp, filenames[i] ); if( strcmp( filenames[i], "-" ) == 0 ) @@ -1104,9 +1144,12 @@ int main( const int argc, const char * const argv[] ) return 1; /* check tty only once and don't try to delete a tty */ } - in_statsp = ( input_filename[0] && one_to_one ) ? &in_stats : 0; - cfile_size = ( input_filename[0] && S_ISREG( in_stats.st_mode ) ) ? - ( in_stats.st_size + 99 ) / 100 : 0; + const struct stat * const in_statsp = + ( input_filename[0] && one_to_one ) ? &in_stats : 0; + const unsigned long long cfile_size = + ( input_filename[0] && S_ISREG( in_stats.st_mode ) ) ? + ( in_stats.st_size + 99 ) / 100 : 0; + int tmp; if( program_mode == m_compress ) tmp = compress( cfile_size, member_size, volume_size, infd, &encoder_options, &pp, in_statsp, zero ); diff --git a/testsuite/check.sh b/testsuite/check.sh index adda5f0..0b80dd1 100755 --- a/testsuite/check.sh +++ b/testsuite/check.sh @@ -1,6 +1,6 @@ #! /bin/sh # check script for Clzip - LZMA lossless data compressor -# Copyright (C) 2010-2021 Antonio Diaz Diaz. +# Copyright (C) 2010-2022 Antonio Diaz Diaz. # # This script is free software: you have unlimited permission # to copy, distribute, and modify it. @@ -100,6 +100,7 @@ done printf "LZIP\001-.............................." | "${LZIP}" -t 2> /dev/null printf "LZIP\002-.............................." | "${LZIP}" -t 2> /dev/null printf "LZIP\001+.............................." | "${LZIP}" -t 2> /dev/null +rm -f out || framework_failure printf "\ntesting decompression..." @@ -123,17 +124,22 @@ lines=$("${LZIP}" -tvv "${in_em}" 2>&1 | wc -l) || test_failed $LINENO lines=$("${LZIP}" -lvv "${in_em}" | wc -l) || test_failed $LINENO [ "${lines}" -eq 11 ] || test_failed $LINENO "${lines}" +"${LZIP}" -cd "${fox_lz}" > fox || test_failed $LINENO cat "${in_lz}" > copy.lz || framework_failure "${LZIP}" -dk copy.lz || test_failed $LINENO cmp in copy || test_failed $LINENO -printf "to be overwritten" > copy || framework_failure -"${LZIP}" -d copy.lz 2> /dev/null +cat fox > copy || framework_failure +cat "${in_lz}" > out.lz || framework_failure +rm -f out || framework_failure +"${LZIP}" -d copy.lz out.lz 2> /dev/null # skip copy, decompress out [ $? = 1 ] || test_failed $LINENO +cmp fox copy || test_failed $LINENO +cmp in out || test_failed $LINENO "${LZIP}" -df copy.lz || test_failed $LINENO [ ! -e copy.lz ] || test_failed $LINENO cmp in copy || test_failed $LINENO +rm -f copy out || framework_failure -rm -f copy || framework_failure cat "${in_lz}" > copy.lz || framework_failure "${LZIP}" -d -S100k copy.lz || test_failed $LINENO # ignore -S [ ! -e copy.lz ] || test_failed $LINENO @@ -167,7 +173,7 @@ rm -f copy anyothername.out || framework_failure [ $? = 1 ] || test_failed $LINENO "${LZIP}" -cdq in "${in_lz}" > copy [ $? = 2 ] || test_failed $LINENO -cat copy in | cmp in - || test_failed $LINENO +cat copy in | cmp in - || test_failed $LINENO # copy must be empty "${LZIP}" -cdq nx_file.lz "${in_lz}" > copy [ $? = 1 ] || test_failed $LINENO cmp in copy || test_failed $LINENO @@ -375,7 +381,6 @@ for i in fox_v2.lz fox_s11.lz fox_de20.lz \ [ $? = 2 ] || test_failed $LINENO $i done -"${LZIP}" -cd "${fox_lz}" > fox || test_failed $LINENO for i in fox_bcrc.lz fox_crc0.lz fox_das46.lz fox_mes81.lz ; do "${LZIP}" -cdq "${testdir}"/$i > out [ $? = 2 ] || test_failed $LINENO $i |