X-Git-Url: https://git.gag.com/?a=blobdiff_plain;f=deflate.c;h=4a0f3d37b28278a00695d6440c7993d91e6e5845;hb=da809fd25a5d8f8a8d8a297da86fd275522971d0;hp=629e8d7f9bcade3c7aa826aef7f702bda8a13015;hpb=e7921f54c622e3b32e525f345bc34308821e4ae0;p=debian%2Fgzip diff --git a/deflate.c b/deflate.c index 629e8d7..4a0f3d3 100644 --- a/deflate.c +++ b/deflate.c @@ -1,11 +1,11 @@ /* deflate.c -- compress data using the deflation algorithm - Copyright (C) 1999, 2006 Free Software Foundation, Inc. + Copyright (C) 1999, 2006, 2009-2012 Free Software Foundation, Inc. Copyright (C) 1992-1993 Jean-loup Gailly 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 - the Free Software Foundation; either version 2, or (at your option) + the Free Software Foundation; either version 3, or (at your option) any later version. This program is distributed in the hope that it will be useful, @@ -81,10 +81,6 @@ #include "gzip.h" #include "lzw.h" /* just for consistency checking */ -#ifdef RCSID -static char rcsid[] = "$Id: deflate.c,v 1.5 2006/12/07 23:53:00 eggert Exp $"; -#endif - /* =========================================================================== * Configuration parameters */ @@ -136,11 +132,11 @@ static char rcsid[] = "$Id: deflate.c,v 1.5 2006/12/07 23:53:00 eggert Exp $"; /* Matches of length 3 are discarded if their distance exceeds TOO_FAR */ #ifndef RSYNC_WIN -# define RSYNC_WIN 4096 +# define RSYNC_WIN 8192 #endif /* Size of rsync window, must be < MAX_DIST */ -#define RSYNC_SUM_MATCH(sum) ((sum) % RSYNC_WIN == 0) +#define RSYNC_SUM_MATCH(sum) (((sum) & (RSYNC_WIN - 1)) == 0) /* Whether window sum matches magic value */ /* =========================================================================== @@ -173,7 +169,7 @@ typedef unsigned IPos; /* DECLARE(Pos, head, 1<= HASH_BITS */ -unsigned int near prev_length; + unsigned int near prev_length; /* Length of the best match at previous step. Matches not greater than this * are discarded. This is used in the lazy match evaluation. */ @@ -202,7 +198,7 @@ unsigned int near prev_length; local int eofile; /* flag set at end of input file */ local unsigned lookahead; /* number of valid bytes ahead in window */ -unsigned near max_chain_length; + unsigned max_chain_length; /* To speed up deflation, hash chains are never searched beyond this length. * A higher limit improves compression ratio but degrades the speed. */ @@ -221,7 +217,7 @@ local unsigned int max_lazy_match; local int compr_level; /* compression level (1..9) */ -unsigned near good_match; +unsigned good_match; /* Use a faster search when the previous match is longer than this */ local ulg rsync_sum; /* rolling sum of rsync window */ @@ -240,10 +236,17 @@ typedef struct config { ush max_chain; } config; +#ifdef ASMV +# define static_unless_ASMV +#else +# define static_unless_ASMV static +#endif + #ifdef FULL_SEARCH # define nice_match MAX_MATCH #else - int near nice_match; /* Stop searching when current match exceeds this */ + /* Stop searching when current match exceeds this */ + static_unless_ASMV int nice_match; #endif local config configuration_table[10] = { @@ -265,27 +268,24 @@ local config configuration_table[10] = { * meaning. */ -#define EQUAL 0 -/* result of memcmp for equal strings */ - /* =========================================================================== * Prototypes for local functions. */ -local void fill_window OF((void)); -local off_t deflate_fast OF((void)); +local void fill_window (void); +local off_t deflate_fast (void); - int longest_match OF((IPos cur_match)); #ifdef ASMV - void match_init OF((void)); /* asm code initialization */ + int longest_match (IPos cur_match); + void match_init (void); /* asm code initialization */ #endif #ifdef DEBUG -local void check_match OF((IPos start, IPos match, int length)); +local void check_match (IPos start, IPos match, int length); #endif /* =========================================================================== * Update a hash value with the given input byte - * IN assertion: all calls to to UPDATE_HASH are made with consecutive + * IN assertion: all calls to UPDATE_HASH are made with consecutive * input characters, so that a running hash key can be computed from the * previous key instead of complete recalculation each time. */ @@ -295,7 +295,7 @@ local void check_match OF((IPos start, IPos match, int length)); * Insert string s in the dictionary and set match_head to the previous head * of the hash chain (the most recent string with same hash key). Return * the previous length of the hash chain. - * IN assertion: all calls to to INSERT_STRING are made with consecutive + * IN assertion: all calls to INSERT_STRING are made with consecutive * input characters and the first MIN_MATCH bytes of s are valid * (except for the last MIN_MATCH-1 bytes of the input file). */ @@ -345,12 +345,13 @@ void lm_init (pack_level, flags) strstart = 0; block_start = 0L; + rsync_chunk_end = 0xFFFFFFFFUL; #ifdef ASMV match_init(); /* initialize the asm code */ #endif lookahead = read_buf((char*)window, - sizeof(int) <= 2 ? (unsigned)WSIZE : 2*WSIZE); + sizeof(int) <= 2 ? (unsigned)WSIZE : 2*WSIZE); if (lookahead == 0 || lookahead == (unsigned)EOF) { eofile = 1, lookahead = 0; @@ -382,8 +383,8 @@ void lm_init (pack_level, flags) * match.s. The code is functionally equivalent, so you can use the C version * if desired. */ -int longest_match(cur_match) - IPos cur_match; /* current match */ +static int +longest_match(IPos cur_match) { unsigned chain_length = max_chain_length; /* max hash chain length */ register uch *scan = window + strstart; /* current string */ @@ -502,7 +503,7 @@ int longest_match(cur_match) #endif } } while ((cur_match = prev[cur_match & WMASK]) > limit - && --chain_length != 0); + && --chain_length != 0); return best_len; } @@ -518,7 +519,7 @@ local void check_match(start, match, length) { /* check that the match is indeed a match */ if (memcmp((char*)window + match, - (char*)window + start, length) != EQUAL) { + (char*)window + start, length) != 0) { fprintf(stderr, " start %d, match %d, length %d\n", start, match, length); @@ -587,6 +588,8 @@ local void fill_window() n = read_buf((char*)window+strstart+lookahead, more); if (n == 0 || n == (unsigned)EOF) { eofile = 1; + /* Don't let garbage pollute the dictionary. */ + memzero (window + strstart + lookahead, MIN_MATCH - 1); } else { lookahead += n; } @@ -632,7 +635,7 @@ local void rsync_roll(start, num) */ #define FLUSH_BLOCK(eof) \ flush_block(block_start >= 0L ? (char*)&window[(unsigned)block_start] : \ - (char*)NULL, (long)strstart - block_start, flush-1, (eof)) + (char*)NULL, (long)strstart - block_start, (eof)) /* =========================================================================== * Processes a new input file and return its compressed length. This @@ -643,7 +646,7 @@ local void rsync_roll(start, num) local off_t deflate_fast() { IPos hash_head; /* head of the hash chain */ - int flush; /* set if current block must be flushed, 2=>and padded */ + int flush; /* set if current block must be flushed */ unsigned match_length = 0; /* length of best match */ prev_length = MIN_MATCH-1; @@ -657,7 +660,7 @@ local off_t deflate_fast() * At this point we have always match_length < MIN_MATCH */ if (hash_head != NIL && strstart - hash_head <= MAX_DIST - && strstart <= window_size - MIN_LOOKAHEAD) { + && strstart <= window_size - MIN_LOOKAHEAD) { /* To simplify the code, we prevent matches with the string * of window index 0 (in particular we have to avoid a match * of the string with itself at the start of the input file). @@ -688,12 +691,12 @@ local off_t deflate_fast() * the next lookahead bytes will be emitted as literals. */ } while (--match_length != 0); - strstart++; + strstart++; } else { - strstart += match_length; - match_length = 0; - ins_h = window[strstart]; - UPDATE_HASH(ins_h, window[strstart+1]); + strstart += match_length; + match_length = 0; + ins_h = window[strstart]; + UPDATE_HASH(ins_h, window[strstart+1]); #if MIN_MATCH != 3 Call UPDATE_HASH() MIN_MATCH-3 more times #endif @@ -704,12 +707,16 @@ local off_t deflate_fast() flush = ct_tally (0, window[strstart]); RSYNC_ROLL(strstart, 1); lookahead--; - strstart++; + strstart++; + } + if (rsync && strstart > rsync_chunk_end) { + ush attr = 0; /* ascii/binary flag */ + + flush = 1; + /* Reset huffman tree */ + ct_init(&attr, &method); + rsync_chunk_end = 0xFFFFFFFFUL; } - if (rsync && strstart > rsync_chunk_end) { - rsync_chunk_end = 0xFFFFFFFFUL; - flush = 2; - } if (flush) FLUSH_BLOCK(0), block_start = strstart; /* Make sure that we always have enough lookahead, except @@ -798,8 +805,12 @@ off_t deflate() strstart++; if (rsync && strstart > rsync_chunk_end) { + ush attr = 0; /* ascii/binary flag */ + + /* Reset huffman tree */ + ct_init(&attr, &method); rsync_chunk_end = 0xFFFFFFFFUL; - flush = 2; + flush = 1; } if (flush) FLUSH_BLOCK(0), block_start = strstart; } else if (match_available) { @@ -810,8 +821,13 @@ off_t deflate() Tracevv((stderr,"%c",window[strstart-1])); flush = ct_tally (0, window[strstart-1]); if (rsync && strstart > rsync_chunk_end) { + ush attr = 0; /* ascii/binary flag */ + + /* Reset huffman tree */ + ct_init(&attr, &method); rsync_chunk_end = 0xFFFFFFFFUL; - flush = 2; + + flush = 1; } if (flush) FLUSH_BLOCK(0), block_start = strstart; RSYNC_ROLL(strstart, 1); @@ -822,9 +838,12 @@ off_t deflate() * the next step to decide. */ if (rsync && strstart > rsync_chunk_end) { + ush attr = 0; /* ascii/binary flag */ + /* Reset huffman tree */ + ct_init(&attr, &method); rsync_chunk_end = 0xFFFFFFFFUL; - flush = 2; + FLUSH_BLOCK(0), block_start = strstart; } match_available = 1;