1 /* vi: set sw=4 ts=4: */
3 * Gzip implementation for busybox
5 * Based on GNU gzip Copyright (C) 1992-1993 Jean-loup Gailly.
7 * Originally adjusted for busybox by Charles P. Wright <cpw@unix.asb.com>
8 * "this is a stripped down version of gzip I put into busybox, it does
9 * only standard in to standard out with -9 compression. It also requires
10 * the zcat module for some important functions."
12 * Adjusted further by Erik Andersen <andersen@codepoet.org> to support
13 * files as well as stdin/stdout, and to generally behave itself wrt
14 * command line handling.
16 * Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
26 #include <sys/types.h>
30 #include <sys/types.h>
37 typedef unsigned char uch;
38 typedef unsigned short ush;
39 typedef unsigned long ulg;
41 /* Return codes from gzip */
46 /* Compression methods (see algorithm.doc) */
47 /* Only STORED and DEFLATED are supported by this BusyBox module */
49 /* methods 4 to 7 reserved */
52 /* To save memory for 16 bit systems, some arrays are overlaid between
53 * the various modules:
54 * deflate: prev+head window d_buf l_buf outbuf
55 * unlzw: tab_prefix tab_suffix stack inbuf outbuf
56 * For compression, input is done in window[]. For decompression, output
57 * is done in window except for unlzw.
62 # define INBUFSIZ 0x2000 /* input buffer size */
64 # define INBUFSIZ 0x8000 /* input buffer size */
67 #define INBUF_EXTRA 64 /* required by unlzw() */
71 # define OUTBUFSIZ 8192 /* output buffer size */
73 # define OUTBUFSIZ 16384 /* output buffer size */
76 #define OUTBUF_EXTRA 2048 /* required by unlzw() */
80 # define DIST_BUFSIZE 0x2000 /* buffer for distances, see trees.c */
82 # define DIST_BUFSIZE 0x8000 /* buffer for distances, see trees.c */
86 # define DECLARE(type, array, size) static type * array
87 # define ALLOC(type, array, size) { \
88 array = (type*)xzalloc((size_t)(((size)+1L)/2) * 2*sizeof(type)); \
90 # define FREE(array) {free(array), array=NULL;}
92 #define tab_suffix window
93 #define tab_prefix prev /* hash link (see deflate.c) */
94 #define head (prev+WSIZE) /* hash head (see deflate.c) */
96 static long bytes_in; /* number of input bytes */
98 #define isize bytes_in
99 /* for compatibility with old zip sources (to be cleaned) */
101 typedef int file_t; /* Do not use stdio */
103 #define NO_FILE (-1) /* in memory compression */
106 #define PACK_MAGIC "\037\036" /* Magic header for packed files */
107 #define GZIP_MAGIC "\037\213" /* Magic header for gzip files, 1F 8B */
108 #define OLD_GZIP_MAGIC "\037\236" /* Magic header for gzip 0.5 = freeze 1.x */
109 #define LZH_MAGIC "\037\240" /* Magic header for SCO LZH Compress files */
110 #define PKZIP_MAGIC "\120\113\003\004" /* Magic header for pkzip files */
113 #define ASCII_FLAG 0x01 /* bit 0 set: file probably ascii text */
114 #define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
115 #define EXTRA_FIELD 0x04 /* bit 2 set: extra field present */
116 #define ORIG_NAME 0x08 /* bit 3 set: original file name present */
117 #define COMMENT 0x10 /* bit 4 set: file comment present */
118 #define RESERVED 0xC0 /* bit 6,7: reserved */
120 /* internal file attribute */
121 #define UNKNOWN 0xffff
126 # define WSIZE 0x8000 /* window size--must be a power of two, and */
127 #endif /* at least 32K for zip's deflate method */
130 #define MAX_MATCH 258
131 /* The minimum and maximum match lengths */
133 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
134 /* Minimum amount of lookahead, except at the end of the input file.
135 * See deflate.c for comments about the MIN_MATCH+1.
138 #define MAX_DIST (WSIZE-MIN_LOOKAHEAD)
139 /* In order to simplify the code, particularly on 16 bit machines, match
140 * distances are limited to MAX_DIST instead of WSIZE.
143 /* put_byte is used for the compressed output */
144 #define put_byte(c) {outbuf[outcnt++]=(uch)(c); if (outcnt==OUTBUFSIZ)\
147 #define seekable() 0 /* force sequential output */
148 #define translate_eol 0 /* no option -a yet */
150 /* Diagnostic functions */
152 # define Assert(cond,msg) {if(!(cond)) bb_error_msg(msg);}
153 # define Trace(x) fprintf x
154 # define Tracev(x) {if (verbose) fprintf x ;}
155 # define Tracevv(x) {if (verbose>1) fprintf x ;}
156 # define Tracec(c,x) {if (verbose && (c)) fprintf x ;}
157 # define Tracecv(c,x) {if (verbose>1 && (c)) fprintf x ;}
159 # define Assert(cond,msg)
164 # define Tracecv(c,x)
167 #define WARN(msg) {if (!quiet) fprintf msg ; \
168 if (exit_code == OK) exit_code = WARNING;}
171 # define MAX_PATH_LEN 1024 /* max pathname length */
176 static int zip(int in, int out);
177 static int file_read(char *buf, unsigned size);
180 static void lm_init(ush * flags);
181 static ulg deflate(void);
184 static void ct_init(ush * attr, int *methodp);
185 static int ct_tally(int dist, int lc);
186 static ulg flush_block(char *buf, ulg stored_len, int eof);
189 static void bi_init(file_t zipfile);
190 static void send_bits(int value, int length);
191 static unsigned bi_reverse(unsigned value, int length);
192 static void bi_windup(void);
193 static void copy_block(char *buf, unsigned len, int header);
194 static int (*read_buf) (char *buf, unsigned size);
197 static void flush_outbuf(void);
199 /* lzw.h -- define the lzw functions.
200 * Copyright (C) 1992-1993 Jean-loup Gailly.
201 * This is free software; you can redistribute it and/or modify it under the
202 * terms of the GNU General Public License, see the file COPYING.
208 #define INIT_BITS 9 /* Initial number of bits per code */
210 #define BIT_MASK 0x1f /* Mask for 'number of compression bits' */
211 /* Mask 0x20 is reserved to mean a fourth header byte, and 0x40 is free.
212 * It's a pity that old uncompress does not check bit 0x20. That makes
213 * extension of the format actually undesirable because old compress
214 * would just crash on the new format instead of giving a meaningful
215 * error message. It does check the number of bits, but it's more
216 * helpful to say "unsupported format, get a new version" than
217 * "can only handle 16 bits".
220 /* tailor.h -- target dependent definitions
221 * Copyright (C) 1992-1993 Jean-loup Gailly.
222 * This is free software; you can redistribute it and/or modify it under the
223 * terms of the GNU General Public License, see the file COPYING.
226 /* The target dependent definitions should be defined here only.
227 * The target dependent functions should be defined in tailor.c.
231 /* Common defaults */
234 # define OS_CODE 0x03 /* assume Unix */
238 # define PATH_SEP '/'
242 # define OPTIONS_VAR "GZIP"
246 # define Z_SUFFIX ".gz"
250 # define MAX_SUFFIX MAX_EXT_CHARS
252 # define MAX_SUFFIX 30
257 DECLARE(uch, inbuf, INBUFSIZ + INBUF_EXTRA);
258 DECLARE(uch, outbuf, OUTBUFSIZ + OUTBUF_EXTRA);
259 DECLARE(ush, d_buf, DIST_BUFSIZE);
260 DECLARE(uch, window, 2L * WSIZE);
261 DECLARE(ush, tab_prefix, 1L << BITS);
263 static int foreground; /* set if program run in foreground */
264 static int method = DEFLATED; /* compression method */
265 static int exit_code = OK; /* program exit code */
266 static long time_stamp; /* original time stamp (modification time) */
267 static char z_suffix[MAX_SUFFIX + 1]; /* default suffix (can be set with --suffix) */
269 static int ifd; /* input file descriptor */
270 static int ofd; /* output file descriptor */
272 static unsigned insize; /* valid bytes in inbuf */
274 static unsigned outcnt; /* bytes in output buffer */
276 static uint32_t *crc_32_tab;
278 /* Output a 16 bit value, lsb first */
279 static void put_short(ush w)
281 if (outcnt < OUTBUFSIZ - 2) {
282 outbuf[outcnt++] = (uch) ((w) & 0xff);
283 outbuf[outcnt++] = (uch) ((ush) (w) >> 8);
285 put_byte((uch) ((w) & 0xff));
286 put_byte((uch) ((ush) (w) >> 8));
290 /* ========================================================================
291 * Signal and error handler.
293 static void abort_gzip(int ATTRIBUTE_UNUSED ignored)
298 /* ===========================================================================
299 * Clear input and output buffers
301 static void clear_bufs(void)
310 /* ===========================================================================
311 * Does the same as write(), but also handles partial pipe writes and checks
314 static void write_buf(int fd, void *buf, unsigned cnt)
318 while ((n = write(fd, buf, cnt)) != cnt) {
319 if (n == (unsigned) (-1)) bb_error_msg_and_die(bb_msg_write_error);
321 buf = (void *) ((char *) buf + n);
325 /* ===========================================================================
326 * Run a set of bytes through the crc shift register. If s is a NULL
327 * pointer, then initialize the crc shift register contents instead.
328 * Return the current crc in either case.
330 static uint32_t updcrc(uch * s, unsigned n)
332 static uint32_t crc = ~0; /* shift register contents */
333 uint32_t c; /* temporary variable */
341 c = crc_32_tab[((int) c ^ (*s++)) & 0xff] ^ (c >> 8);
348 /* bits.c -- output variable-length bit strings
349 * Copyright (C) 1992-1993 Jean-loup Gailly
350 * This is free software; you can redistribute it and/or modify it under the
351 * terms of the GNU General Public License, see the file COPYING.
358 * Output variable-length bit strings. Compression can be done
359 * to a file or to memory. (The latter is not supported in this version.)
363 * The PKZIP "deflate" file format interprets compressed file data
364 * as a sequence of bits. Multi-bit strings in the file may cross
365 * byte boundaries without restriction.
367 * The first bit of each byte is the low-order bit.
369 * The routines in this file allow a variable-length bit value to
370 * be output right-to-left (useful for literal values). For
371 * left-to-right output (useful for code strings from the tree routines),
372 * the bits must have been reversed first with bi_reverse().
374 * For in-memory compression, the compressed bit stream goes directly
375 * into the requested output buffer. The input data is read in blocks
376 * by the mem_read() function. The buffer is limited to 64K on 16 bit
381 * void bi_init (FILE *zipfile)
382 * Initialize the bit string routines.
384 * void send_bits (int value, int length)
385 * Write out a bit string, taking the source bits right to
388 * int bi_reverse (int value, int length)
389 * Reverse the bits of a bit string, taking the source bits left to
390 * right and emitting them right to left.
392 * void bi_windup (void)
393 * Write out any remaining bits in an incomplete byte.
395 * void copy_block(char *buf, unsigned len, int header)
396 * Copy a stored block to the zip file, storing first the length and
397 * its one's complement if requested.
401 /* ===========================================================================
402 * Local data used by the "bit string" routines.
405 static file_t zfile; /* output gzip file */
407 static unsigned short bi_buf;
409 /* Output buffer. bits are inserted starting at the bottom (least significant
413 #define Buf_size (8 * 2*sizeof(char))
414 /* Number of bits used within bi_buf. (bi_buf might be implemented on
415 * more than 16 bits on some systems.)
420 /* Current input function. Set to mem_read for in-memory compression */
423 ulg bits_sent; /* bit length of the compressed data */
426 /* ===========================================================================
427 * Initialize the bit string routines.
429 static void bi_init(file_t zipfile)
438 /* Set the defaults for file compression. They are set by memcompress
439 * for in-memory compression.
441 if (zfile != NO_FILE) {
442 read_buf = file_read;
446 /* ===========================================================================
447 * Send a value on a given number of bits.
448 * IN assertion: length <= 16 and value fits in length bits.
450 static void send_bits(int value, int length)
453 Tracev((stderr, " l %2d v %4x ", length, value));
454 Assert(length > 0 && length <= 15, "invalid length");
455 bits_sent += (ulg) length;
457 /* If not enough room in bi_buf, use (valid) bits from bi_buf and
458 * (16 - bi_valid) bits from value, leaving (width - (16-bi_valid))
459 * unused bits in value.
461 if (bi_valid > (int) Buf_size - length) {
462 bi_buf |= (value << bi_valid);
464 bi_buf = (ush) value >> (Buf_size - bi_valid);
465 bi_valid += length - Buf_size;
467 bi_buf |= value << bi_valid;
472 /* ===========================================================================
473 * Reverse the first len bits of a code, using straightforward code (a faster
474 * method would use a table)
475 * IN assertion: 1 <= len <= 15
477 static unsigned bi_reverse(unsigned code, int len)
483 code >>= 1, res <<= 1;
488 /* ===========================================================================
489 * Write out any remaining bits in an incomplete byte.
491 static void bi_windup(void)
495 } else if (bi_valid > 0) {
501 bits_sent = (bits_sent + 7) & ~7;
505 /* ===========================================================================
506 * Copy a stored block to the zip file, storing first the length and its
507 * one's complement if requested.
509 static void copy_block(char *buf, unsigned len, int header)
511 bi_windup(); /* align on byte boundary */
514 put_short((ush) len);
515 put_short((ush) ~ len);
521 bits_sent += (ulg) len << 3;
528 /* deflate.c -- compress data using the deflation algorithm
529 * Copyright (C) 1992-1993 Jean-loup Gailly
530 * This is free software; you can redistribute it and/or modify it under the
531 * terms of the GNU General Public License, see the file COPYING.
537 * Identify new text as repetitions of old text within a fixed-
538 * length sliding window trailing behind the new text.
542 * The "deflation" process depends on being able to identify portions
543 * of the input text which are identical to earlier input (within a
544 * sliding window trailing behind the input currently being processed).
546 * The most straightforward technique turns out to be the fastest for
547 * most input files: try all possible matches and select the longest.
548 * The key feature of this algorithm is that insertions into the string
549 * dictionary are very simple and thus fast, and deletions are avoided
550 * completely. Insertions are performed at each input character, whereas
551 * string matches are performed only when the previous match ends. So it
552 * is preferable to spend more time in matches to allow very fast string
553 * insertions and avoid deletions. The matching algorithm for small
554 * strings is inspired from that of Rabin & Karp. A brute force approach
555 * is used to find longer strings when a small match has been found.
556 * A similar algorithm is used in comic (by Jan-Mark Wams) and freeze
557 * (by Leonid Broukhis).
558 * A previous version of this file used a more sophisticated algorithm
559 * (by Fiala and Greene) which is guaranteed to run in linear amortized
560 * time, but has a larger average cost, uses more memory and is patented.
561 * However the F&G algorithm may be faster for some highly redundant
562 * files if the parameter max_chain_length (described below) is too large.
566 * The idea of lazy evaluation of matches is due to Jan-Mark Wams, and
567 * I found it in 'freeze' written by Leonid Broukhis.
568 * Thanks to many info-zippers for bug reports and testing.
572 * APPNOTE.TXT documentation file in PKZIP 1.93a distribution.
574 * A description of the Rabin and Karp algorithm is given in the book
575 * "Algorithms" by R. Sedgewick, Addison-Wesley, p252.
577 * Fiala,E.R., and Greene,D.H.
578 * Data Compression with Finite Windows, Comm.ACM, 32,4 (1989) 490-595
582 * void lm_init (int pack_level, ush *flags)
583 * Initialize the "longest match" routines for a new file
586 * Processes a new input file and return its compressed length. Sets
587 * the compressed length, crc, deflate flags and internal file
592 /* ===========================================================================
593 * Configuration parameters
596 /* Compile with MEDIUM_MEM to reduce the memory requirements or
597 * with SMALL_MEM to use as little memory as possible. Use BIG_MEM if the
598 * entire input file can be held in memory (not possible on 16 bit systems).
599 * Warning: defining these symbols affects HASH_BITS (see below) and thus
600 * affects the compression ratio. The compressed output
601 * is still correct, and might even be smaller in some cases.
605 # define HASH_BITS 13 /* Number of bits used to hash strings */
608 # define HASH_BITS 14
611 # define HASH_BITS 15
612 /* For portability to 16 bit machines, do not use values above 15. */
615 /* To save space (see unlzw.c), we overlay prev+head with tab_prefix and
616 * window with tab_suffix. Check that we can do this:
618 #if (WSIZE<<1) > (1<<BITS)
619 # error cannot overlay window with tab_suffix and prev with tab_prefix0
621 #if HASH_BITS > BITS-1
622 # error cannot overlay head with tab_prefix1
624 #define HASH_SIZE (unsigned)(1<<HASH_BITS)
625 #define HASH_MASK (HASH_SIZE-1)
626 #define WMASK (WSIZE-1)
627 /* HASH_SIZE and WSIZE must be powers of two */
629 /* Tail of hash chains */
632 /* speed options for the general purpose bit flag */
634 # define TOO_FAR 4096
636 /* Matches of length 3 are discarded if their distance exceeds TOO_FAR */
637 /* ===========================================================================
638 * Local data used by the "longest match" routines.
641 typedef unsigned IPos;
643 /* A Pos is an index in the character window. We use short instead of int to
644 * save space in the various tables. IPos is used only for parameter passing.
647 /* DECLARE(uch, window, 2L*WSIZE); */
648 /* Sliding window. Input bytes are read into the second half of the window,
649 * and move to the first half later to keep a dictionary of at least WSIZE
650 * bytes. With this organization, matches are limited to a distance of
651 * WSIZE-MAX_MATCH bytes, but this ensures that IO is always
652 * performed with a length multiple of the block size. Also, it limits
653 * the window size to 64K, which is quite useful on MSDOS.
654 * To do: limit the window size to WSIZE+BSZ if SMALL_MEM (the code would
655 * be less efficient).
658 /* DECLARE(Pos, prev, WSIZE); */
659 /* Link to older string with same hash index. To limit the size of this
660 * array to 64K, this link is maintained only for the last 32K strings.
661 * An index in this array is thus a window index modulo 32K.
664 /* DECLARE(Pos, head, 1<<HASH_BITS); */
665 /* Heads of the hash chains or NIL. */
667 static const ulg window_size = (ulg) 2 * WSIZE;
669 /* window size, 2*WSIZE except for MMAP or BIG_MEM, where it is the
670 * input file length plus MIN_LOOKAHEAD.
673 static long block_start;
675 /* window position at the beginning of the current output block. Gets
676 * negative when the window is moved backwards.
679 static unsigned ins_h; /* hash index of string to be inserted */
681 #define H_SHIFT ((HASH_BITS+MIN_MATCH-1)/MIN_MATCH)
682 /* Number of bits by which ins_h and del_h must be shifted at each
683 * input step. It must be such that after MIN_MATCH steps, the oldest
684 * byte no longer takes part in the hash key, that is:
685 * H_SHIFT * MIN_MATCH >= HASH_BITS
688 static unsigned int prev_length;
690 /* Length of the best match at previous step. Matches not greater than this
691 * are discarded. This is used in the lazy match evaluation.
694 static unsigned strstart; /* start of string to insert */
695 static unsigned match_start; /* start of matching string */
696 static int eofile; /* flag set at end of input file */
697 static unsigned lookahead; /* number of valid bytes ahead in window */
700 max_chain_length = 4096,
702 /* To speed up deflation, hash chains are never searched beyond this length.
703 * A higher limit improves compression ratio but degrades the speed.
706 max_lazy_match = 258,
708 /* Attempt to find a better match only when the current match is strictly
709 * smaller than this value. This mechanism is used only for compression
712 max_insert_length = max_lazy_match,
713 /* Insert new strings in the hash table only if the match length
714 * is not greater than this length. This saves time but degrades compression.
715 * max_insert_length is used only for compression levels <= 3.
720 /* Use a faster search when the previous match is longer than this */
723 /* Values for max_lazy_match, good_match and max_chain_length, depending on
724 * the desired pack level (0..9). The values given below have been tuned to
725 * exclude worst case performance for pathological files. Better values may be
726 * found for specific files.
729 nice_match = 258 /* Stop searching when current match exceeds this */
731 /* Note: the deflate() code requires max_lazy >= MIN_MATCH and max_chain >= 4
732 * For deflate_fast() (levels <= 3) good is ignored and lazy has a different
738 /* result of memcmp for equal strings */
740 /* ===========================================================================
741 * Prototypes for local functions.
743 static void fill_window(void);
745 static int longest_match(IPos cur_match);
748 static void check_match(IPos start, IPos match, int length);
751 /* ===========================================================================
752 * Update a hash value with the given input byte
753 * IN assertion: all calls to to UPDATE_HASH are made with consecutive
754 * input characters, so that a running hash key can be computed from the
755 * previous key instead of complete recalculation each time.
757 #define UPDATE_HASH(h,c) (h = (((h)<<H_SHIFT) ^ (c)) & HASH_MASK)
759 /* ===========================================================================
760 * Insert string s in the dictionary and set match_head to the previous head
761 * of the hash chain (the most recent string with same hash key). Return
762 * the previous length of the hash chain.
763 * IN assertion: all calls to to INSERT_STRING are made with consecutive
764 * input characters and the first MIN_MATCH bytes of s are valid
765 * (except for the last MIN_MATCH-1 bytes of the input file).
767 #define INSERT_STRING(s, match_head) \
768 (UPDATE_HASH(ins_h, window[(s) + MIN_MATCH-1]), \
769 prev[(s) & WMASK] = match_head = head[ins_h], \
772 /* ===========================================================================
773 * Initialize the "longest match" routines for a new file
775 static void lm_init(ush * flags)
779 /* Initialize the hash table. */
780 memset(head, 0, HASH_SIZE * sizeof(*head));
781 /* prev will be initialized on the fly */
784 /* ??? reduce max_chain_length for binary files */
789 lookahead = read_buf((char *) window,
790 sizeof(int) <= 2 ? (unsigned) WSIZE : 2 * WSIZE);
792 if (lookahead == 0 || lookahead == (unsigned) EOF) {
793 eofile = 1, lookahead = 0;
797 /* Make sure that we always have enough lookahead. This is important
798 * if input comes from a device such as a tty.
800 while (lookahead < MIN_LOOKAHEAD && !eofile)
804 for (j = 0; j < MIN_MATCH - 1; j++)
805 UPDATE_HASH(ins_h, window[j]);
806 /* If lookahead < MIN_MATCH, ins_h is garbage, but this is
807 * not important since only literal bytes will be emitted.
811 /* ===========================================================================
812 * Set match_start to the longest match starting at the given string and
813 * return its length. Matches shorter or equal to prev_length are discarded,
814 * in which case the result is equal to prev_length and match_start is
816 * IN assertions: cur_match is the head of the hash chain for the current
817 * string (strstart) and its distance is <= MAX_DIST, and prev_length >= 1
820 /* For MSDOS, OS/2 and 386 Unix, an optimized version is in match.asm or
821 * match.s. The code is functionally equivalent, so you can use the C version
824 static int longest_match(IPos cur_match)
826 unsigned chain_length = max_chain_length; /* max hash chain length */
827 uch *scan = window + strstart; /* current string */
828 uch *match; /* matched string */
829 int len; /* length of current match */
830 int best_len = prev_length; /* best match length so far */
832 strstart > (IPos) MAX_DIST ? strstart - (IPos) MAX_DIST : NIL;
833 /* Stop when cur_match becomes <= limit. To simplify the code,
834 * we prevent matches with the string of window index 0.
837 /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16.
838 * It is easy to get rid of this optimization if necessary.
840 #if HASH_BITS < 8 || MAX_MATCH != 258
841 # error Code too clever
843 uch *strend = window + strstart + MAX_MATCH;
844 uch scan_end1 = scan[best_len - 1];
845 uch scan_end = scan[best_len];
847 /* Do not waste too much time if we already have a good match: */
848 if (prev_length >= good_match) {
851 Assert(strstart <= window_size - MIN_LOOKAHEAD, "insufficient lookahead");
854 Assert(cur_match < strstart, "no future");
855 match = window + cur_match;
857 /* Skip to next match if the match length cannot increase
858 * or if the match length is less than 2:
860 if (match[best_len] != scan_end ||
861 match[best_len - 1] != scan_end1 ||
862 *match != *scan || *++match != scan[1])
865 /* The check at best_len-1 can be removed because it will be made
866 * again later. (This heuristic is not always a win.)
867 * It is not necessary to compare scan[2] and match[2] since they
868 * are always equal when the other bytes match, given that
869 * the hash keys are equal and that HASH_BITS >= 8.
873 /* We check for insufficient lookahead only every 8th comparison;
874 * the 256th check will be made at strstart+258.
877 } while (*++scan == *++match && *++scan == *++match &&
878 *++scan == *++match && *++scan == *++match &&
879 *++scan == *++match && *++scan == *++match &&
880 *++scan == *++match && *++scan == *++match && scan < strend);
882 len = MAX_MATCH - (int) (strend - scan);
883 scan = strend - MAX_MATCH;
885 if (len > best_len) {
886 match_start = cur_match;
888 if (len >= nice_match)
890 scan_end1 = scan[best_len - 1];
891 scan_end = scan[best_len];
893 } while ((cur_match = prev[cur_match & WMASK]) > limit
894 && --chain_length != 0);
900 /* ===========================================================================
901 * Check that the match at match_start is indeed a match.
903 static void check_match(IPos start, IPos match, int length)
905 /* check that the match is indeed a match */
906 if (memcmp((char *) window + match,
907 (char *) window + start, length) != EQUAL) {
908 bb_error_msg(" start %d, match %d, length %d", start, match, length);
909 bb_error_msg("invalid match");
912 bb_error_msg("\\[%d,%d]", start - match, length);
914 putc(window[start++], stderr);
915 } while (--length != 0);
919 # define check_match(start, match, length)
922 /* ===========================================================================
923 * Fill the window when the lookahead becomes insufficient.
924 * Updates strstart and lookahead, and sets eofile if end of input file.
925 * IN assertion: lookahead < MIN_LOOKAHEAD && strstart + lookahead > 0
926 * OUT assertions: at least one byte has been read, or eofile is set;
927 * file reads are performed for at least two bytes (required for the
928 * translate_eol option).
930 static void fill_window(void)
934 (unsigned) (window_size - (ulg) lookahead - (ulg) strstart);
935 /* Amount of free space at the end of the window. */
937 /* If the window is almost full and there is insufficient lookahead,
938 * move the upper half to the lower one to make room in the upper half.
940 if (more == (unsigned) EOF) {
941 /* Very unlikely, but possible on 16 bit machine if strstart == 0
942 * and lookahead == 1 (input done one byte at time)
945 } else if (strstart >= WSIZE + MAX_DIST) {
946 /* By the IN assertion, the window is not empty so we can't confuse
947 * more == 0 with more == 64K on a 16 bit machine.
949 Assert(window_size == (ulg) 2 * WSIZE, "no sliding with BIG_MEM");
951 memcpy((char *) window, (char *) window + WSIZE, (unsigned) WSIZE);
952 match_start -= WSIZE;
953 strstart -= WSIZE; /* we now have strstart >= MAX_DIST: */
955 block_start -= (long) WSIZE;
957 for (n = 0; n < HASH_SIZE; n++) {
959 head[n] = (Pos) (m >= WSIZE ? m - WSIZE : NIL);
961 for (n = 0; n < WSIZE; n++) {
963 prev[n] = (Pos) (m >= WSIZE ? m - WSIZE : NIL);
964 /* If n is not on any hash chain, prev[n] is garbage but
965 * its value will never be used.
970 /* At this point, more >= 2 */
972 n = read_buf((char *) window + strstart + lookahead, more);
973 if (n == 0 || n == (unsigned) EOF) {
981 /* ===========================================================================
982 * Flush the current block, with given end-of-file flag.
983 * IN assertion: strstart is set to the end of the current match.
985 #define FLUSH_BLOCK(eof) \
986 flush_block(block_start >= 0L ? (char*)&window[(unsigned)block_start] : \
987 (char*)NULL, (long)strstart - block_start, (eof))
989 /* ===========================================================================
990 * Same as above, but achieves better compression. We use a lazy
991 * evaluation for matches: a match is finally adopted only if there is
992 * no better match at the next window position.
994 static ulg deflate(void)
996 IPos hash_head; /* head of hash chain */
997 IPos prev_match; /* previous match */
998 int flush; /* set if current block must be flushed */
999 int match_available = 0; /* set if previous match exists */
1000 unsigned match_length = MIN_MATCH - 1; /* length of best match */
1002 /* Process the input block. */
1003 while (lookahead != 0) {
1004 /* Insert the string window[strstart .. strstart+2] in the
1005 * dictionary, and set hash_head to the head of the hash chain:
1007 INSERT_STRING(strstart, hash_head);
1009 /* Find the longest match, discarding those <= prev_length.
1011 prev_length = match_length, prev_match = match_start;
1012 match_length = MIN_MATCH - 1;
1014 if (hash_head != NIL && prev_length < max_lazy_match &&
1015 strstart - hash_head <= MAX_DIST) {
1016 /* To simplify the code, we prevent matches with the string
1017 * of window index 0 (in particular we have to avoid a match
1018 * of the string with itself at the start of the input file).
1020 match_length = longest_match(hash_head);
1021 /* longest_match() sets match_start */
1022 if (match_length > lookahead)
1023 match_length = lookahead;
1025 /* Ignore a length 3 match if it is too distant: */
1026 if (match_length == MIN_MATCH && strstart - match_start > TOO_FAR) {
1027 /* If prev_match is also MIN_MATCH, match_start is garbage
1028 * but we will ignore the current match anyway.
1033 /* If there was a match at the previous step and the current
1034 * match is not better, output the previous match:
1036 if (prev_length >= MIN_MATCH && match_length <= prev_length) {
1038 check_match(strstart - 1, prev_match, prev_length);
1041 ct_tally(strstart - 1 - prev_match, prev_length - MIN_MATCH);
1043 /* Insert in hash table all strings up to the end of the match.
1044 * strstart-1 and strstart are already inserted.
1046 lookahead -= prev_length - 1;
1050 INSERT_STRING(strstart, hash_head);
1051 /* strstart never exceeds WSIZE-MAX_MATCH, so there are
1052 * always MIN_MATCH bytes ahead. If lookahead < MIN_MATCH
1053 * these bytes are garbage, but it does not matter since the
1054 * next lookahead bytes will always be emitted as literals.
1056 } while (--prev_length != 0);
1057 match_available = 0;
1058 match_length = MIN_MATCH - 1;
1061 FLUSH_BLOCK(0), block_start = strstart;
1063 } else if (match_available) {
1064 /* If there was no match at the previous position, output a
1065 * single literal. If there was a match but the current match
1066 * is longer, truncate the previous match to a single literal.
1068 Tracevv((stderr, "%c", window[strstart - 1]));
1069 if (ct_tally(0, window[strstart - 1])) {
1070 FLUSH_BLOCK(0), block_start = strstart;
1075 /* There is no previous match to compare with, wait for
1076 * the next step to decide.
1078 match_available = 1;
1082 Assert(strstart <= isize && lookahead <= isize, "a bit too far");
1084 /* Make sure that we always have enough lookahead, except
1085 * at the end of the input file. We need MAX_MATCH bytes
1086 * for the next match, plus MIN_MATCH bytes to insert the
1087 * string following the next match.
1089 while (lookahead < MIN_LOOKAHEAD && !eofile)
1092 if (match_available)
1093 ct_tally(0, window[strstart - 1]);
1095 return FLUSH_BLOCK(1); /* eof */
1098 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
1099 * Copyright (C) 1992-1993 Jean-loup Gailly
1100 * The unzip code was written and put in the public domain by Mark Adler.
1101 * Portions of the lzw code are derived from the public domain 'compress'
1102 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
1103 * Ken Turkowski, Dave Mack and Peter Jannesen.
1105 * See the license_msg below and the file COPYING for the software license.
1106 * See the file algorithm.doc for the compression algorithms and file formats.
1109 /* Compress files with zip algorithm and 'compress' interface.
1110 * See usage() and help() functions below for all options.
1112 * file.gz: compressed file with same mode, owner, and utimes
1113 * or stdout with -c option or if stdin used as input.
1114 * If the output file name had to be truncated, the original name is kept
1115 * in the compressed file.
1120 typedef struct dirent dir_type;
1122 /* ======================================================================== */
1123 int gzip_main(int argc, char **argv)
1134 struct stat statBuf;
1137 opt = bb_getopt_ulflags(argc, argv, "cf123456789q" USE_GUNZIP("d"));
1138 //if (opt & 0x1) // -c
1139 //if (opt & 0x2) // -f
1140 /* Ignore 1-9 (compression level) options */
1141 //if (opt & 0x4) // -1
1142 //if (opt & 0x8) // -2
1143 //if (opt & 0x10) // -3
1144 //if (opt & 0x20) // -4
1145 //if (opt & 0x40) // -5
1146 //if (opt & 0x80) // -6
1147 //if (opt & 0x100) // -7
1148 //if (opt & 0x200) // -8
1149 //if (opt & 0x400) // -9
1150 //if (opt & 0x800) // -q
1151 if (ENABLE_GUNZIP && (opt & 0x1000)) { // -d
1152 /* FIXME: bb_getopt_ulflags should not depend on optind */
1154 return gunzip_main(argc, argv);
1157 foreground = signal(SIGINT, SIG_IGN) != SIG_IGN;
1159 (void) signal(SIGINT, abort_gzip);
1162 if (signal(SIGTERM, SIG_IGN) != SIG_IGN) {
1163 (void) signal(SIGTERM, abort_gzip);
1167 if (signal(SIGHUP, SIG_IGN) != SIG_IGN) {
1168 (void) signal(SIGHUP, abort_gzip);
1172 strncpy(z_suffix, Z_SUFFIX, sizeof(z_suffix) - 1);
1174 /* Allocate all global buffers (for DYN_ALLOC option) */
1175 ALLOC(uch, inbuf, INBUFSIZ + INBUF_EXTRA);
1176 ALLOC(uch, outbuf, OUTBUFSIZ + OUTBUF_EXTRA);
1177 ALLOC(ush, d_buf, DIST_BUFSIZE);
1178 ALLOC(uch, window, 2L * WSIZE);
1179 ALLOC(ush, tab_prefix, 1L << BITS);
1181 /* Initialise the CRC32 table */
1182 crc_32_tab = crc32_filltable(0);
1186 if (optind == argc) {
1188 zip(STDIN_FILENO, STDOUT_FILENO);
1192 for (i = optind; i < argc; i++) {
1196 if (strcmp(argv[i], "-") == 0) {
1198 inFileNum = STDIN_FILENO;
1199 outFileNum = STDOUT_FILENO;
1201 inFileNum = xopen(argv[i], O_RDONLY);
1202 if (fstat(inFileNum, &statBuf) < 0)
1203 bb_perror_msg_and_die("%s", argv[i]);
1204 time_stamp = statBuf.st_ctime;
1206 if (!(opt & OPT_tostdout)) {
1207 path = xasprintf("%s.gz", argv[i]);
1209 /* Open output file */
1210 #if (__GLIBC__ >= 2) && (__GLIBC_MINOR__ >= 1) && defined O_NOFOLLOW
1212 open(path, O_RDWR | O_CREAT | O_EXCL | O_NOFOLLOW);
1214 outFileNum = open(path, O_RDWR | O_CREAT | O_EXCL);
1216 if (outFileNum < 0) {
1217 bb_perror_msg("%s", path);
1222 /* Set permissions on the file */
1223 fchmod(outFileNum, statBuf.st_mode);
1225 outFileNum = STDOUT_FILENO;
1228 if (path == NULL && isatty(outFileNum) && !(opt & OPT_force)) {
1230 ("compressed data not written to a terminal. Use -f to force compression.");
1235 result = zip(inFileNum, outFileNum);
1241 /* Delete the original file */
1243 delFileName = argv[i];
1247 if (unlink(delFileName) < 0)
1248 bb_perror_msg("%s", delFileName);
1258 /* trees.c -- output deflated data using Huffman coding
1259 * Copyright (C) 1992-1993 Jean-loup Gailly
1260 * This is free software; you can redistribute it and/or modify it under the
1261 * terms of the GNU General Public License, see the file COPYING.
1267 * Encode various sets of source values using variable-length
1268 * binary code trees.
1272 * The PKZIP "deflation" process uses several Huffman trees. The more
1273 * common source values are represented by shorter bit sequences.
1275 * Each code tree is stored in the ZIP file in a compressed form
1276 * which is itself a Huffman encoding of the lengths of
1277 * all the code strings (in ascending order by source values).
1278 * The actual code strings are reconstructed from the lengths in
1279 * the UNZIP process, as described in the "application note"
1280 * (APPNOTE.TXT) distributed as part of PKWARE's PKZIP program.
1285 * Data Compression: Techniques and Applications, pp. 53-55.
1286 * Lifetime Learning Publications, 1985. ISBN 0-534-03418-7.
1289 * Data Compression: Methods and Theory, pp. 49-50.
1290 * Computer Science Press, 1988. ISBN 0-7167-8156-5.
1294 * Addison-Wesley, 1983. ISBN 0-201-06672-6.
1298 * void ct_init (ush *attr, int *methodp)
1299 * Allocate the match buffer, initialize the various tables and save
1300 * the location of the internal file attribute (ascii/binary) and
1301 * method (DEFLATE/STORE)
1303 * void ct_tally (int dist, int lc);
1304 * Save the match info and tally the frequency counts.
1306 * long flush_block (char *buf, ulg stored_len, int eof)
1307 * Determine the best encoding for the current block: dynamic trees,
1308 * static trees or store, and output the encoded block to the zip
1309 * file. Returns the total compressed length for the file so far.
1313 /* ===========================================================================
1318 /* All codes must not exceed MAX_BITS bits */
1320 #define MAX_BL_BITS 7
1321 /* Bit length codes must not exceed MAX_BL_BITS bits */
1323 #define LENGTH_CODES 29
1324 /* number of length codes, not counting the special END_BLOCK code */
1326 #define LITERALS 256
1327 /* number of literal bytes 0..255 */
1329 #define END_BLOCK 256
1330 /* end of block literal code */
1332 #define L_CODES (LITERALS+1+LENGTH_CODES)
1333 /* number of Literal or Length codes, including the END_BLOCK code */
1336 /* number of distance codes */
1339 /* number of codes used to transfer the bit lengths */
1341 typedef uch extra_bits_t;
1343 /* extra bits for each length code */
1344 static const extra_bits_t extra_lbits[LENGTH_CODES]
1345 = { 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4,
1349 /* extra bits for each distance code */
1350 static const extra_bits_t extra_dbits[D_CODES]
1351 = { 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9,
1352 10, 10, 11, 11, 12, 12, 13, 13
1355 /* extra bits for each bit length code */
1356 static const extra_bits_t extra_blbits[BL_CODES]
1357 = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 3, 7 };
1359 #define STORED_BLOCK 0
1360 #define STATIC_TREES 1
1362 /* The three kinds of block type */
1366 # define LIT_BUFSIZE 0x2000
1369 # define LIT_BUFSIZE 0x4000
1371 # define LIT_BUFSIZE 0x8000
1375 #ifndef DIST_BUFSIZE
1376 # define DIST_BUFSIZE LIT_BUFSIZE
1378 /* Sizes of match buffers for literals/lengths and distances. There are
1379 * 4 reasons for limiting LIT_BUFSIZE to 64K:
1380 * - frequencies can be kept in 16 bit counters
1381 * - if compression is not successful for the first block, all input data is
1382 * still in the window so we can still emit a stored block even when input
1383 * comes from standard input. (This can also be done for all blocks if
1384 * LIT_BUFSIZE is not greater than 32K.)
1385 * - if compression is not successful for a file smaller than 64K, we can
1386 * even emit a stored file instead of a stored block (saving 5 bytes).
1387 * - creating new Huffman trees less frequently may not provide fast
1388 * adaptation to changes in the input data statistics. (Take for
1389 * example a binary file with poorly compressible code followed by
1390 * a highly compressible string table.) Smaller buffer sizes give
1391 * fast adaptation but have of course the overhead of transmitting trees
1393 * - I can't count above 4
1394 * The current code is general and allows DIST_BUFSIZE < LIT_BUFSIZE (to save
1395 * memory at the expense of compression). Some optimizations would be possible
1396 * if we rely on DIST_BUFSIZE == LIT_BUFSIZE.
1398 #if LIT_BUFSIZE > INBUFSIZ
1399 #error cannot overlay l_buf and inbuf
1402 /* repeat previous bit length 3-6 times (2 bits of repeat count) */
1403 #define REPZ_3_10 17
1404 /* repeat a zero length 3-10 times (3 bits of repeat count) */
1405 #define REPZ_11_138 18
1406 /* repeat a zero length 11-138 times (7 bits of repeat count) */
1408 /* ===========================================================================
1412 /* Data structure describing a single value and its code string. */
1413 typedef struct ct_data {
1415 ush freq; /* frequency count */
1416 ush code; /* bit string */
1419 ush dad; /* father node in Huffman tree */
1420 ush len; /* length of bit string */
1424 #define Freq fc.freq
1425 #define Code fc.code
1429 #define HEAP_SIZE (2*L_CODES+1)
1430 /* maximum heap size */
1432 static ct_data dyn_ltree[HEAP_SIZE]; /* literal and length tree */
1433 static ct_data dyn_dtree[2 * D_CODES + 1]; /* distance tree */
1435 static ct_data static_ltree[L_CODES + 2];
1437 /* The static literal tree. Since the bit lengths are imposed, there is no
1438 * need for the L_CODES extra codes used during heap construction. However
1439 * The codes 286 and 287 are needed to build a canonical tree (see ct_init
1443 static ct_data static_dtree[D_CODES];
1445 /* The static distance tree. (Actually a trivial tree since all codes use
1449 static ct_data bl_tree[2 * BL_CODES + 1];
1451 /* Huffman tree for the bit lengths */
1453 typedef struct tree_desc {
1454 ct_data *dyn_tree; /* the dynamic tree */
1455 ct_data *static_tree; /* corresponding static tree or NULL */
1456 const extra_bits_t *extra_bits; /* extra bits for each code or NULL */
1457 int extra_base; /* base index for extra_bits */
1458 int elems; /* max number of elements in the tree */
1459 int max_length; /* max bit length for the codes */
1460 int max_code; /* largest code with non zero frequency */
1463 static tree_desc l_desc =
1464 { dyn_ltree, static_ltree, extra_lbits, LITERALS + 1, L_CODES,
1468 static tree_desc d_desc =
1469 { dyn_dtree, static_dtree, extra_dbits, 0, D_CODES, MAX_BITS, 0 };
1471 static tree_desc bl_desc =
1472 { bl_tree, (ct_data *) 0, extra_blbits, 0, BL_CODES, MAX_BL_BITS,
1477 static ush bl_count[MAX_BITS + 1];
1479 /* number of codes at each bit length for an optimal tree */
1481 static const uch bl_order[BL_CODES]
1482 = { 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15 };
1484 /* The lengths of the bit length codes are sent in order of decreasing
1485 * probability, to avoid transmitting the lengths for unused bit length codes.
1488 static int heap[2 * L_CODES + 1]; /* heap used to build the Huffman trees */
1489 static int heap_len; /* number of elements in the heap */
1490 static int heap_max; /* element of largest frequency */
1492 /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
1493 * The same heap array is used to build all trees.
1496 static uch depth[2 * L_CODES + 1];
1498 /* Depth of each subtree used as tie breaker for trees of equal frequency */
1500 static uch length_code[MAX_MATCH - MIN_MATCH + 1];
1502 /* length code for each normalized match length (0 == MIN_MATCH) */
1504 static uch dist_code[512];
1506 /* distance codes. The first 256 values correspond to the distances
1507 * 3 .. 258, the last 256 values correspond to the top 8 bits of
1508 * the 15 bit distances.
1511 static int base_length[LENGTH_CODES];
1513 /* First normalized length for each code (0 = MIN_MATCH) */
1515 static int base_dist[D_CODES];
1517 /* First normalized distance for each code (0 = distance of 1) */
1520 /* DECLARE(uch, l_buf, LIT_BUFSIZE); buffer for literals or lengths */
1522 /* DECLARE(ush, d_buf, DIST_BUFSIZE); buffer for distances */
1524 static uch flag_buf[(LIT_BUFSIZE / 8)];
1526 /* flag_buf is a bit array distinguishing literals from lengths in
1527 * l_buf, thus indicating the presence or absence of a distance.
1530 static unsigned last_lit; /* running index in l_buf */
1531 static unsigned last_dist; /* running index in d_buf */
1532 static unsigned last_flags; /* running index in flag_buf */
1533 static uch flags; /* current flags not yet saved in flag_buf */
1534 static uch flag_bit; /* current bit used in flags */
1536 /* bits are filled in flags starting at bit 0 (least significant).
1537 * Note: these flags are overkill in the current code since we don't
1538 * take advantage of DIST_BUFSIZE == LIT_BUFSIZE.
1541 static ulg opt_len; /* bit length of current block with optimal trees */
1542 static ulg static_len; /* bit length of current block with static trees */
1544 static ulg compressed_len; /* total bit length of compressed file */
1547 static ush *file_type; /* pointer to UNKNOWN, BINARY or ASCII */
1548 static int *file_method; /* pointer to DEFLATE or STORE */
1550 /* ===========================================================================
1551 * Local (static) routines in this file.
1554 static void init_block(void);
1555 static void pqdownheap(ct_data * tree, int k);
1556 static void gen_bitlen(tree_desc * desc);
1557 static void gen_codes(ct_data * tree, int max_code);
1558 static void build_tree(tree_desc * desc);
1559 static void scan_tree(ct_data * tree, int max_code);
1560 static void send_tree(ct_data * tree, int max_code);
1561 static int build_bl_tree(void);
1562 static void send_all_trees(int lcodes, int dcodes, int blcodes);
1563 static void compress_block(ct_data * ltree, ct_data * dtree);
1564 static void set_file_type(void);
1568 # define send_code(c, tree) send_bits(tree[c].Code, tree[c].Len)
1569 /* Send a code of the given tree. c and tree must not have side effects */
1572 # define send_code(c, tree) \
1573 { if (verbose>1) bb_error_msg("\ncd %3d ",(c)); \
1574 send_bits(tree[c].Code, tree[c].Len); }
1577 #define d_code(dist) \
1578 ((dist) < 256 ? dist_code[dist] : dist_code[256+((dist)>>7)])
1579 /* Mapping from a distance to a distance code. dist is the distance - 1 and
1580 * must not have side effects. dist_code[256] and dist_code[257] are never
1584 /* the arguments must not have side effects */
1586 /* ===========================================================================
1587 * Allocate the match buffer, initialize the various tables and save the
1588 * location of the internal file attribute (ascii/binary) and method
1591 static void ct_init(ush * attr, int *methodp)
1593 int n; /* iterates over tree elements */
1594 int bits; /* bit counter */
1595 int length; /* length value */
1596 int code; /* code value */
1597 int dist; /* distance index */
1600 file_method = methodp;
1601 compressed_len = 0L;
1603 if (static_dtree[0].Len != 0)
1604 return; /* ct_init already called */
1606 /* Initialize the mapping length (0..255) -> length code (0..28) */
1608 for (code = 0; code < LENGTH_CODES - 1; code++) {
1609 base_length[code] = length;
1610 for (n = 0; n < (1 << extra_lbits[code]); n++) {
1611 length_code[length++] = (uch) code;
1614 Assert(length == 256, "ct_init: length != 256");
1615 /* Note that the length 255 (match length 258) can be represented
1616 * in two different ways: code 284 + 5 bits or code 285, so we
1617 * overwrite length_code[255] to use the best encoding:
1619 length_code[length - 1] = (uch) code;
1621 /* Initialize the mapping dist (0..32K) -> dist code (0..29) */
1623 for (code = 0; code < 16; code++) {
1624 base_dist[code] = dist;
1625 for (n = 0; n < (1 << extra_dbits[code]); n++) {
1626 dist_code[dist++] = (uch) code;
1629 Assert(dist == 256, "ct_init: dist != 256");
1630 dist >>= 7; /* from now on, all distances are divided by 128 */
1631 for (; code < D_CODES; code++) {
1632 base_dist[code] = dist << 7;
1633 for (n = 0; n < (1 << (extra_dbits[code] - 7)); n++) {
1634 dist_code[256 + dist++] = (uch) code;
1637 Assert(dist == 256, "ct_init: 256+dist != 512");
1639 /* Construct the codes of the static literal tree */
1640 for (bits = 0; bits <= MAX_BITS; bits++)
1644 static_ltree[n++].Len = 8, bl_count[8]++;
1646 static_ltree[n++].Len = 9, bl_count[9]++;
1648 static_ltree[n++].Len = 7, bl_count[7]++;
1650 static_ltree[n++].Len = 8, bl_count[8]++;
1651 /* Codes 286 and 287 do not exist, but we must include them in the
1652 * tree construction to get a canonical Huffman tree (longest code
1655 gen_codes((ct_data *) static_ltree, L_CODES + 1);
1657 /* The static distance tree is trivial: */
1658 for (n = 0; n < D_CODES; n++) {
1659 static_dtree[n].Len = 5;
1660 static_dtree[n].Code = bi_reverse(n, 5);
1663 /* Initialize the first block of the first file: */
1667 /* ===========================================================================
1668 * Initialize a new block.
1670 static void init_block(void)
1672 int n; /* iterates over tree elements */
1674 /* Initialize the trees. */
1675 for (n = 0; n < L_CODES; n++)
1676 dyn_ltree[n].Freq = 0;
1677 for (n = 0; n < D_CODES; n++)
1678 dyn_dtree[n].Freq = 0;
1679 for (n = 0; n < BL_CODES; n++)
1680 bl_tree[n].Freq = 0;
1682 dyn_ltree[END_BLOCK].Freq = 1;
1683 opt_len = static_len = 0L;
1684 last_lit = last_dist = last_flags = 0;
1690 /* Index within the heap array of least frequent node in the Huffman tree */
1693 /* ===========================================================================
1694 * Remove the smallest element from the heap and recreate the heap with
1695 * one less element. Updates heap and heap_len.
1697 #define pqremove(tree, top) \
1699 top = heap[SMALLEST]; \
1700 heap[SMALLEST] = heap[heap_len--]; \
1701 pqdownheap(tree, SMALLEST); \
1704 /* ===========================================================================
1705 * Compares to subtrees, using the tree depth as tie breaker when
1706 * the subtrees have equal frequency. This minimizes the worst case length.
1708 #define smaller(tree, n, m) \
1709 (tree[n].Freq < tree[m].Freq || \
1710 (tree[n].Freq == tree[m].Freq && depth[n] <= depth[m]))
1712 /* ===========================================================================
1713 * Restore the heap property by moving down the tree starting at node k,
1714 * exchanging a node with the smallest of its two sons if necessary, stopping
1715 * when the heap property is re-established (each father smaller than its
1718 static void pqdownheap(ct_data * tree, int k)
1721 int j = k << 1; /* left son of k */
1723 while (j <= heap_len) {
1724 /* Set j to the smallest of the two sons: */
1725 if (j < heap_len && smaller(tree, heap[j + 1], heap[j]))
1728 /* Exit if v is smaller than both sons */
1729 if (smaller(tree, v, heap[j]))
1732 /* Exchange v with the smallest son */
1736 /* And continue down the tree, setting j to the left son of k */
1742 /* ===========================================================================
1743 * Compute the optimal bit lengths for a tree and update the total bit length
1744 * for the current block.
1745 * IN assertion: the fields freq and dad are set, heap[heap_max] and
1746 * above are the tree nodes sorted by increasing frequency.
1747 * OUT assertions: the field len is set to the optimal bit length, the
1748 * array bl_count contains the frequencies for each bit length.
1749 * The length opt_len is updated; static_len is also updated if stree is
1752 static void gen_bitlen(tree_desc * desc)
1754 ct_data *tree = desc->dyn_tree;
1755 const extra_bits_t *extra = desc->extra_bits;
1756 int base = desc->extra_base;
1757 int max_code = desc->max_code;
1758 int max_length = desc->max_length;
1759 ct_data *stree = desc->static_tree;
1760 int h; /* heap index */
1761 int n, m; /* iterate over the tree elements */
1762 int bits; /* bit length */
1763 int xbits; /* extra bits */
1764 ush f; /* frequency */
1765 int overflow = 0; /* number of elements with bit length too large */
1767 for (bits = 0; bits <= MAX_BITS; bits++)
1770 /* In a first pass, compute the optimal bit lengths (which may
1771 * overflow in the case of the bit length tree).
1773 tree[heap[heap_max]].Len = 0; /* root of the heap */
1775 for (h = heap_max + 1; h < HEAP_SIZE; h++) {
1777 bits = tree[tree[n].Dad].Len + 1;
1778 if (bits > max_length)
1779 bits = max_length, overflow++;
1780 tree[n].Len = (ush) bits;
1781 /* We overwrite tree[n].Dad which is no longer needed */
1784 continue; /* not a leaf node */
1789 xbits = extra[n - base];
1791 opt_len += (ulg) f *(bits + xbits);
1794 static_len += (ulg) f *(stree[n].Len + xbits);
1799 Trace((stderr, "\nbit length overflow\n"));
1800 /* This happens for example on obj2 and pic of the Calgary corpus */
1802 /* Find the first bit length which could increase: */
1804 bits = max_length - 1;
1805 while (bl_count[bits] == 0)
1807 bl_count[bits]--; /* move one leaf down the tree */
1808 bl_count[bits + 1] += 2; /* move one overflow item as its brother */
1809 bl_count[max_length]--;
1810 /* The brother of the overflow item also moves one step up,
1811 * but this does not affect bl_count[max_length]
1814 } while (overflow > 0);
1816 /* Now recompute all bit lengths, scanning in increasing frequency.
1817 * h is still equal to HEAP_SIZE. (It is simpler to reconstruct all
1818 * lengths instead of fixing only the wrong ones. This idea is taken
1819 * from 'ar' written by Haruhiko Okumura.)
1821 for (bits = max_length; bits != 0; bits--) {
1827 if (tree[m].Len != (unsigned) bits) {
1828 Trace((stderr, "code %d bits %d->%d\n", m, tree[m].Len,
1831 ((long) bits - (long) tree[m].Len) * (long) tree[m].Freq;
1832 tree[m].Len = (ush) bits;
1839 /* ===========================================================================
1840 * Generate the codes for a given tree and bit counts (which need not be
1842 * IN assertion: the array bl_count contains the bit length statistics for
1843 * the given tree and the field len is set for all tree elements.
1844 * OUT assertion: the field code is set for all tree elements of non
1847 static void gen_codes(ct_data * tree, int max_code)
1849 ush next_code[MAX_BITS + 1]; /* next code value for each bit length */
1850 ush code = 0; /* running code value */
1851 int bits; /* bit index */
1852 int n; /* code index */
1854 /* The distribution counts are first used to generate the code values
1855 * without bit reversal.
1857 for (bits = 1; bits <= MAX_BITS; bits++) {
1858 next_code[bits] = code = (code + bl_count[bits - 1]) << 1;
1860 /* Check that the bit counts in bl_count are consistent. The last code
1863 Assert(code + bl_count[MAX_BITS] - 1 == (1 << MAX_BITS) - 1,
1864 "inconsistent bit counts");
1865 Tracev((stderr, "\ngen_codes: max_code %d ", max_code));
1867 for (n = 0; n <= max_code; n++) {
1868 int len = tree[n].Len;
1872 /* Now reverse the bits */
1873 tree[n].Code = bi_reverse(next_code[len]++, len);
1875 Tracec(tree != static_ltree,
1876 (stderr, "\nn %3d %c l %2d c %4x (%x) ", n,
1877 (isgraph(n) ? n : ' '), len, tree[n].Code,
1878 next_code[len] - 1));
1882 /* ===========================================================================
1883 * Construct one Huffman tree and assigns the code bit strings and lengths.
1884 * Update the total bit length for the current block.
1885 * IN assertion: the field freq is set for all tree elements.
1886 * OUT assertions: the fields len and code are set to the optimal bit length
1887 * and corresponding code. The length opt_len is updated; static_len is
1888 * also updated if stree is not null. The field max_code is set.
1890 static void build_tree(tree_desc * desc)
1892 ct_data *tree = desc->dyn_tree;
1893 ct_data *stree = desc->static_tree;
1894 int elems = desc->elems;
1895 int n, m; /* iterate over heap elements */
1896 int max_code = -1; /* largest code with non zero frequency */
1897 int node = elems; /* next internal node of the tree */
1899 /* Construct the initial heap, with least frequent element in
1900 * heap[SMALLEST]. The sons of heap[n] are heap[2*n] and heap[2*n+1].
1901 * heap[0] is not used.
1903 heap_len = 0, heap_max = HEAP_SIZE;
1905 for (n = 0; n < elems; n++) {
1906 if (tree[n].Freq != 0) {
1907 heap[++heap_len] = max_code = n;
1914 /* The pkzip format requires that at least one distance code exists,
1915 * and that at least one bit should be sent even if there is only one
1916 * possible code. So to avoid special checks later on we force at least
1917 * two codes of non zero frequency.
1919 while (heap_len < 2) {
1920 int new = heap[++heap_len] = (max_code < 2 ? ++max_code : 0);
1926 static_len -= stree[new].Len;
1927 /* new is 0 or 1 so it does not have extra bits */
1929 desc->max_code = max_code;
1931 /* The elements heap[heap_len/2+1 .. heap_len] are leaves of the tree,
1932 * establish sub-heaps of increasing lengths:
1934 for (n = heap_len / 2; n >= 1; n--)
1935 pqdownheap(tree, n);
1937 /* Construct the Huffman tree by repeatedly combining the least two
1941 pqremove(tree, n); /* n = node of least frequency */
1942 m = heap[SMALLEST]; /* m = node of next least frequency */
1944 heap[--heap_max] = n; /* keep the nodes sorted by frequency */
1945 heap[--heap_max] = m;
1947 /* Create a new node father of n and m */
1948 tree[node].Freq = tree[n].Freq + tree[m].Freq;
1949 depth[node] = (uch) (MAX(depth[n], depth[m]) + 1);
1950 tree[n].Dad = tree[m].Dad = (ush) node;
1952 if (tree == bl_tree) {
1953 bb_error_msg("\nnode %d(%d), sons %d(%d) %d(%d)",
1954 node, tree[node].Freq, n, tree[n].Freq, m, tree[m].Freq);
1957 /* and insert the new node in the heap */
1958 heap[SMALLEST] = node++;
1959 pqdownheap(tree, SMALLEST);
1961 } while (heap_len >= 2);
1963 heap[--heap_max] = heap[SMALLEST];
1965 /* At this point, the fields freq and dad are set. We can now
1966 * generate the bit lengths.
1968 gen_bitlen((tree_desc *) desc);
1970 /* The field len is now set, we can generate the bit codes */
1971 gen_codes((ct_data *) tree, max_code);
1974 /* ===========================================================================
1975 * Scan a literal or distance tree to determine the frequencies of the codes
1976 * in the bit length tree. Updates opt_len to take into account the repeat
1977 * counts. (The contribution of the bit length codes will be added later
1978 * during the construction of bl_tree.)
1980 static void scan_tree(ct_data * tree, int max_code)
1982 int n; /* iterates over all tree elements */
1983 int prevlen = -1; /* last emitted length */
1984 int curlen; /* length of current code */
1985 int nextlen = tree[0].Len; /* length of next code */
1986 int count = 0; /* repeat count of the current code */
1987 int max_count = 7; /* max repeat count */
1988 int min_count = 4; /* min repeat count */
1991 max_count = 138, min_count = 3;
1992 tree[max_code + 1].Len = (ush) 0xffff; /* guard */
1994 for (n = 0; n <= max_code; n++) {
1996 nextlen = tree[n + 1].Len;
1997 if (++count < max_count && curlen == nextlen) {
1999 } else if (count < min_count) {
2000 bl_tree[curlen].Freq += count;
2001 } else if (curlen != 0) {
2002 if (curlen != prevlen)
2003 bl_tree[curlen].Freq++;
2004 bl_tree[REP_3_6].Freq++;
2005 } else if (count <= 10) {
2006 bl_tree[REPZ_3_10].Freq++;
2008 bl_tree[REPZ_11_138].Freq++;
2013 max_count = 138, min_count = 3;
2014 } else if (curlen == nextlen) {
2015 max_count = 6, min_count = 3;
2017 max_count = 7, min_count = 4;
2022 /* ===========================================================================
2023 * Send a literal or distance tree in compressed form, using the codes in
2026 static void send_tree(ct_data * tree, int max_code)
2028 int n; /* iterates over all tree elements */
2029 int prevlen = -1; /* last emitted length */
2030 int curlen; /* length of current code */
2031 int nextlen = tree[0].Len; /* length of next code */
2032 int count = 0; /* repeat count of the current code */
2033 int max_count = 7; /* max repeat count */
2034 int min_count = 4; /* min repeat count */
2036 /* tree[max_code+1].Len = -1; *//* guard already set */
2038 max_count = 138, min_count = 3;
2040 for (n = 0; n <= max_code; n++) {
2042 nextlen = tree[n + 1].Len;
2043 if (++count < max_count && curlen == nextlen) {
2045 } else if (count < min_count) {
2047 send_code(curlen, bl_tree);
2048 } while (--count != 0);
2050 } else if (curlen != 0) {
2051 if (curlen != prevlen) {
2052 send_code(curlen, bl_tree);
2055 Assert(count >= 3 && count <= 6, " 3_6?");
2056 send_code(REP_3_6, bl_tree);
2057 send_bits(count - 3, 2);
2059 } else if (count <= 10) {
2060 send_code(REPZ_3_10, bl_tree);
2061 send_bits(count - 3, 3);
2064 send_code(REPZ_11_138, bl_tree);
2065 send_bits(count - 11, 7);
2070 max_count = 138, min_count = 3;
2071 } else if (curlen == nextlen) {
2072 max_count = 6, min_count = 3;
2074 max_count = 7, min_count = 4;
2079 /* ===========================================================================
2080 * Construct the Huffman tree for the bit lengths and return the index in
2081 * bl_order of the last bit length code to send.
2083 static int build_bl_tree(void)
2085 int max_blindex; /* index of last bit length code of non zero freq */
2087 /* Determine the bit length frequencies for literal and distance trees */
2088 scan_tree((ct_data *) dyn_ltree, l_desc.max_code);
2089 scan_tree((ct_data *) dyn_dtree, d_desc.max_code);
2091 /* Build the bit length tree: */
2092 build_tree((tree_desc *) (&bl_desc));
2093 /* opt_len now includes the length of the tree representations, except
2094 * the lengths of the bit lengths codes and the 5+5+4 bits for the counts.
2097 /* Determine the number of bit length codes to send. The pkzip format
2098 * requires that at least 4 bit length codes be sent. (appnote.txt says
2099 * 3 but the actual value used is 4.)
2101 for (max_blindex = BL_CODES - 1; max_blindex >= 3; max_blindex--) {
2102 if (bl_tree[bl_order[max_blindex]].Len != 0)
2105 /* Update opt_len to include the bit length tree and counts */
2106 opt_len += 3 * (max_blindex + 1) + 5 + 5 + 4;
2107 Tracev((stderr, "\ndyn trees: dyn %ld, stat %ld", opt_len, static_len));
2112 /* ===========================================================================
2113 * Send the header for a block using dynamic Huffman trees: the counts, the
2114 * lengths of the bit length codes, the literal tree and the distance tree.
2115 * IN assertion: lcodes >= 257, dcodes >= 1, blcodes >= 4.
2117 static void send_all_trees(int lcodes, int dcodes, int blcodes)
2119 int rank; /* index in bl_order */
2121 Assert(lcodes >= 257 && dcodes >= 1 && blcodes >= 4, "not enough codes");
2122 Assert(lcodes <= L_CODES && dcodes <= D_CODES
2123 && blcodes <= BL_CODES, "too many codes");
2124 Tracev((stderr, "\nbl counts: "));
2125 send_bits(lcodes - 257, 5); /* not +255 as stated in appnote.txt */
2126 send_bits(dcodes - 1, 5);
2127 send_bits(blcodes - 4, 4); /* not -3 as stated in appnote.txt */
2128 for (rank = 0; rank < blcodes; rank++) {
2129 Tracev((stderr, "\nbl code %2d ", bl_order[rank]));
2130 send_bits(bl_tree[bl_order[rank]].Len, 3);
2132 Tracev((stderr, "\nbl tree: sent %ld", bits_sent));
2134 send_tree((ct_data *) dyn_ltree, lcodes - 1); /* send the literal tree */
2135 Tracev((stderr, "\nlit tree: sent %ld", bits_sent));
2137 send_tree((ct_data *) dyn_dtree, dcodes - 1); /* send the distance tree */
2138 Tracev((stderr, "\ndist tree: sent %ld", bits_sent));
2141 /* ===========================================================================
2142 * Determine the best encoding for the current block: dynamic trees, static
2143 * trees or store, and output the encoded block to the zip file. This function
2144 * returns the total compressed length for the file so far.
2146 static ulg flush_block(char *buf, ulg stored_len, int eof)
2148 ulg opt_lenb, static_lenb; /* opt_len and static_len in bytes */
2149 int max_blindex; /* index of last bit length code of non zero freq */
2151 flag_buf[last_flags] = flags; /* Save the flags for the last 8 items */
2153 /* Check if the file is ascii or binary */
2154 if (*file_type == (ush) UNKNOWN)
2157 /* Construct the literal and distance trees */
2158 build_tree((tree_desc *) (&l_desc));
2159 Tracev((stderr, "\nlit data: dyn %ld, stat %ld", opt_len, static_len));
2161 build_tree((tree_desc *) (&d_desc));
2162 Tracev((stderr, "\ndist data: dyn %ld, stat %ld", opt_len, static_len));
2163 /* At this point, opt_len and static_len are the total bit lengths of
2164 * the compressed block data, excluding the tree representations.
2167 /* Build the bit length tree for the above two trees, and get the index
2168 * in bl_order of the last bit length code to send.
2170 max_blindex = build_bl_tree();
2172 /* Determine the best encoding. Compute first the block length in bytes */
2173 opt_lenb = (opt_len + 3 + 7) >> 3;
2174 static_lenb = (static_len + 3 + 7) >> 3;
2177 "\nopt %lu(%lu) stat %lu(%lu) stored %lu lit %u dist %u ",
2178 opt_lenb, opt_len, static_lenb, static_len, stored_len,
2179 last_lit, last_dist));
2181 if (static_lenb <= opt_lenb)
2182 opt_lenb = static_lenb;
2184 /* If compression failed and this is the first and last block,
2185 * and if the zip file can be seeked (to rewrite the local header),
2186 * the whole file is transformed into a stored file:
2188 if (stored_len <= opt_lenb && eof && compressed_len == 0L && seekable()) {
2189 /* Since LIT_BUFSIZE <= 2*WSIZE, the input data must be there: */
2190 if (buf == (char *) 0)
2191 bb_error_msg("block vanished");
2193 copy_block(buf, (unsigned) stored_len, 0); /* without header */
2194 compressed_len = stored_len << 3;
2195 *file_method = STORED;
2197 } else if (stored_len + 4 <= opt_lenb && buf != (char *) 0) {
2198 /* 4: two words for the lengths */
2199 /* The test buf != NULL is only necessary if LIT_BUFSIZE > WSIZE.
2200 * Otherwise we can't have processed more than WSIZE input bytes since
2201 * the last block flush, because compression would have been
2202 * successful. If LIT_BUFSIZE <= WSIZE, it is never too late to
2203 * transform a block into a stored block.
2205 send_bits((STORED_BLOCK << 1) + eof, 3); /* send block type */
2206 compressed_len = (compressed_len + 3 + 7) & ~7L;
2207 compressed_len += (stored_len + 4) << 3;
2209 copy_block(buf, (unsigned) stored_len, 1); /* with header */
2211 } else if (static_lenb == opt_lenb) {
2212 send_bits((STATIC_TREES << 1) + eof, 3);
2213 compress_block((ct_data *) static_ltree, (ct_data *) static_dtree);
2214 compressed_len += 3 + static_len;
2216 send_bits((DYN_TREES << 1) + eof, 3);
2217 send_all_trees(l_desc.max_code + 1, d_desc.max_code + 1,
2219 compress_block((ct_data *) dyn_ltree, (ct_data *) dyn_dtree);
2220 compressed_len += 3 + opt_len;
2222 Assert(compressed_len == bits_sent, "bad compressed size");
2227 compressed_len += 7; /* align on byte boundary */
2229 Tracev((stderr, "\ncomprlen %lu(%lu) ", compressed_len >> 3,
2230 compressed_len - 7 * eof));
2232 return compressed_len >> 3;
2235 /* ===========================================================================
2236 * Save the match info and tally the frequency counts. Return true if
2237 * the current block must be flushed.
2239 static int ct_tally(int dist, int lc)
2241 l_buf[last_lit++] = (uch) lc;
2243 /* lc is the unmatched char */
2244 dyn_ltree[lc].Freq++;
2246 /* Here, lc is the match length - MIN_MATCH */
2247 dist--; /* dist = match distance - 1 */
2248 Assert((ush) dist < (ush) MAX_DIST &&
2249 (ush) lc <= (ush) (MAX_MATCH - MIN_MATCH) &&
2250 (ush) d_code(dist) < (ush) D_CODES, "ct_tally: bad match");
2252 dyn_ltree[length_code[lc] + LITERALS + 1].Freq++;
2253 dyn_dtree[d_code(dist)].Freq++;
2255 d_buf[last_dist++] = (ush) dist;
2260 /* Output the flags if they fill a byte: */
2261 if ((last_lit & 7) == 0) {
2262 flag_buf[last_flags++] = flags;
2263 flags = 0, flag_bit = 1;
2265 /* Try to guess if it is profitable to stop the current block here */
2266 if ((last_lit & 0xfff) == 0) {
2267 /* Compute an upper bound for the compressed length */
2268 ulg out_length = (ulg) last_lit * 8L;
2269 ulg in_length = (ulg) strstart - block_start;
2272 for (dcode = 0; dcode < D_CODES; dcode++) {
2274 (ulg) dyn_dtree[dcode].Freq * (5L + extra_dbits[dcode]);
2278 "\nlast_lit %u, last_dist %u, in %ld, out ~%ld(%ld%%) ",
2279 last_lit, last_dist, in_length, out_length,
2280 100L - out_length * 100L / in_length));
2281 if (last_dist < last_lit / 2 && out_length < in_length / 2)
2284 return (last_lit == LIT_BUFSIZE - 1 || last_dist == DIST_BUFSIZE);
2285 /* We avoid equality with LIT_BUFSIZE because of wraparound at 64K
2286 * on 16 bit machines and because stored blocks are restricted to
2291 /* ===========================================================================
2292 * Send the block data compressed using the given Huffman trees
2294 static void compress_block(ct_data * ltree, ct_data * dtree)
2296 unsigned dist; /* distance of matched string */
2297 int lc; /* match length or unmatched char (if dist == 0) */
2298 unsigned lx = 0; /* running index in l_buf */
2299 unsigned dx = 0; /* running index in d_buf */
2300 unsigned fx = 0; /* running index in flag_buf */
2301 uch flag = 0; /* current flags */
2302 unsigned code; /* the code to send */
2303 int extra; /* number of extra bits to send */
2308 flag = flag_buf[fx++];
2310 if ((flag & 1) == 0) {
2311 send_code(lc, ltree); /* send a literal byte */
2312 Tracecv(isgraph(lc), (stderr, " '%c' ", lc));
2314 /* Here, lc is the match length - MIN_MATCH */
2315 code = length_code[lc];
2316 send_code(code + LITERALS + 1, ltree); /* send the length code */
2317 extra = extra_lbits[code];
2319 lc -= base_length[code];
2320 send_bits(lc, extra); /* send the extra length bits */
2323 /* Here, dist is the match distance - 1 */
2324 code = d_code(dist);
2325 Assert(code < D_CODES, "bad d_code");
2327 send_code(code, dtree); /* send the distance code */
2328 extra = extra_dbits[code];
2330 dist -= base_dist[code];
2331 send_bits(dist, extra); /* send the extra distance bits */
2333 } /* literal or match pair ? */
2335 } while (lx < last_lit);
2337 send_code(END_BLOCK, ltree);
2340 /* ===========================================================================
2341 * Set the file type to ASCII or BINARY, using a crude approximation:
2342 * binary if more than 20% of the bytes are <= 6 or >= 128, ascii otherwise.
2343 * IN assertion: the fields freq of dyn_ltree are set and the total of all
2344 * frequencies does not exceed 64K (to fit in an int on 16 bit machines).
2346 static void set_file_type(void)
2349 unsigned ascii_freq = 0;
2350 unsigned bin_freq = 0;
2353 bin_freq += dyn_ltree[n++].Freq;
2355 ascii_freq += dyn_ltree[n++].Freq;
2356 while (n < LITERALS)
2357 bin_freq += dyn_ltree[n++].Freq;
2358 *file_type = bin_freq > (ascii_freq >> 2) ? BINARY : ASCII;
2359 if (*file_type == BINARY && translate_eol) {
2360 bb_error_msg("-l used on binary file");
2364 /* zip.c -- compress files to the gzip or pkzip format
2365 * Copyright (C) 1992-1993 Jean-loup Gailly
2366 * This is free software; you can redistribute it and/or modify it under the
2367 * terms of the GNU General Public License, see the file COPYING.
2371 static uint32_t crc; /* crc on uncompressed file data */
2372 static long header_bytes; /* number of bytes in gzip header */
2374 static void put_long(ulg n)
2376 put_short((n) & 0xffff);
2377 put_short(((ulg) (n)) >> 16);
2380 /* put_header_byte is used for the compressed output
2381 * - for the initial 4 bytes that can't overflow the buffer.
2383 #define put_header_byte(c) {outbuf[outcnt++]=(uch)(c);}
2385 /* ===========================================================================
2386 * Deflate in to out.
2387 * IN assertions: the input and output buffers are cleared.
2388 * The variables time_stamp and save_orig_name are initialized.
2390 static int zip(int in, int out)
2392 uch my_flags = 0; /* general purpose bit flags */
2393 ush attr = 0; /* ascii/binary flag */
2394 ush deflate_flags = 0; /* pkzip -es, -en or -ex equivalent */
2400 /* Write the header to the gzip file. See algorithm.doc for the format */
2404 put_header_byte(GZIP_MAGIC[0]); /* magic header */
2405 put_header_byte(GZIP_MAGIC[1]);
2406 put_header_byte(DEFLATED); /* compression method */
2408 put_header_byte(my_flags); /* general flags */
2409 put_long(time_stamp);
2411 /* Write deflated file to zip file */
2415 ct_init(&attr, &method);
2416 lm_init(&deflate_flags);
2418 put_byte((uch) deflate_flags); /* extra flags */
2419 put_byte(OS_CODE); /* OS identifier */
2421 header_bytes = (long) outcnt;
2425 /* Write the crc and uncompressed size */
2428 header_bytes += 2 * sizeof(long);
2435 /* ===========================================================================
2436 * Read a new buffer from the current input file, perform end-of-line
2437 * translation, and update the crc and input file size.
2438 * IN assertion: size >= 2 (for end-of-line translation)
2440 static int file_read(char *buf, unsigned size)
2444 Assert(insize == 0, "inbuf not empty");
2446 len = read(ifd, buf, size);
2447 if (len == (unsigned) (-1) || len == 0)
2450 crc = updcrc((uch *) buf, len);
2455 /* ===========================================================================
2456 * Write the output buffer outbuf[0..outcnt-1] and update bytes_out.
2457 * (used for the compressed data only)
2459 static void flush_outbuf(void)
2464 write_buf(ofd, (char *) outbuf, outcnt);