/* $OpenBSD: zlib.c,v 1.11 2002/01/23 16:12:17 millert Exp $ */ /* $NetBSD: zlib.c,v 1.2 1996/03/16 23:55:40 christos Exp $ */ /* * This file is derived from various .h and .c files from the zlib-0.95 * distribution by Jean-loup Gailly and Mark Adler, with some additions * by Paul Mackerras to aid in implementing Deflate compression and * decompression for PPP packets. See zlib.h for conditions of * distribution and use. * * Changes that have been made include: * - changed functions not used outside this file to "local" * - added minCompression parameter to deflateInit2 * - added Z_PACKET_FLUSH (see zlib.h for details) * - added inflateIncomp */ /*+++++*/ /* zutil.h -- internal interface and configuration of the compression library * Copyright (C) 1995 Jean-loup Gailly. * For conditions of distribution and use, see copyright notice in zlib.h */ /* WARNING: this file should *not* be used by applications. It is part of the implementation of the compression library and is subject to change. Applications should only use zlib.h. */ /* From: zutil.h,v 1.9 1995/05/03 17:27:12 jloup Exp */ #define _Z_UTIL_H #include "zlib.h" #include #include #ifdef _STANDALONE #include #else #include #endif #ifndef local # define local static #endif /* compile with -Dlocal if your debugger can't find static symbols */ #define FAR typedef unsigned char uch; typedef uch FAR uchf; typedef unsigned short ush; typedef ush FAR ushf; typedef unsigned long ulg; extern char *z_errmsg[]; /* indexed by 1-zlib_error */ #define ERR_RETURN(strm,err) return (strm->msg=z_errmsg[1-err], err) /* To be used only when the state is known to be valid */ #ifndef NULL #define NULL ((void *) 0) #endif /* common constants */ #define DEFLATED 8 #ifndef DEF_WBITS # define DEF_WBITS MAX_WBITS #endif /* default windowBits for decompression. MAX_WBITS is for compression only */ #if MAX_MEM_LEVEL >= 8 # define DEF_MEM_LEVEL 8 #else # define DEF_MEM_LEVEL MAX_MEM_LEVEL #endif /* default memLevel */ #define STORED_BLOCK 0 #define STATIC_TREES 1 #define DYN_TREES 2 /* The three kinds of block type */ #define MIN_MATCH 3 #define MAX_MATCH 258 /* The minimum and maximum match lengths */ /* functions */ #if defined(KERNEL) || defined(_KERNEL) # define zmemcpy(d, s, n) bcopy((s), (d), (n)) # define zmemzero bzero #else #if defined(STDC) && !defined(HAVE_MEMCPY) && !defined(NO_MEMCPY) # define HAVE_MEMCPY #endif #ifdef HAVE_MEMCPY # define zmemcpy memcpy # define zmemzero(dest, len) memset(dest, 0, len) #else extern void zmemcpy OF((Bytef* dest, Bytef* source, uInt len)); extern void zmemzero OF((Bytef* dest, uInt len)); #endif #endif /* Diagnostic functions */ #ifdef DEBUG_ZLIB # include # ifndef verbose # define verbose 0 # endif # define Assert(cond,msg) {if(!(cond)) z_error(msg);} # define Trace(x) fprintf x # define Tracev(x) {if (verbose) fprintf x ;} # define Tracevv(x) {if (verbose>1) fprintf x ;} # define Tracec(c,x) {if (verbose && (c)) fprintf x ;} # define Tracecv(c,x) {if (verbose>1 && (c)) fprintf x ;} #else # define Assert(cond,msg) # define Trace(x) # define Tracev(x) # define Tracevv(x) # define Tracec(c,x) # define Tracecv(c,x) #endif typedef uLong (*check_func) OF((uLong check, Bytef *buf, uInt len)); /* voidpf zcalloc OF((voidpf opaque, unsigned items, unsigned size)); */ /* void zcfree OF((voidpf opaque, voidpf ptr)); */ #define ZALLOC(strm, items, size) \ (*((strm)->zalloc))((strm)->opaque, (items), (size)) #define ZFREE(strm, addr, size) \ (*((strm)->zfree))((strm)->opaque, (voidpf)(addr), (size)) #define TRY_FREE(s, p, n) {if (p) ZFREE(s, p, n);} #ifndef NO_DEFLATE /* deflate.h -- internal compression state * Copyright (C) 1995 Jean-loup Gailly * For conditions of distribution and use, see copyright notice in zlib.h */ /* WARNING: this file should *not* be used by applications. It is part of the implementation of the compression library and is subject to change. Applications should only use zlib.h. */ /*+++++*/ /* From: deflate.h,v 1.5 1995/05/03 17:27:09 jloup Exp */ /* =========================================================================== * Internal compression state. */ /* Data type */ #define BINARY 0 #define ASCII 1 #define UNKNOWN 2 #define LENGTH_CODES 29 /* number of length codes, not counting the special END_BLOCK code */ #define LITERALS 256 /* number of literal bytes 0..255 */ #define L_CODES (LITERALS+1+LENGTH_CODES) /* number of Literal or Length codes, including the END_BLOCK code */ #define D_CODES 30 /* number of distance codes */ #define BL_CODES 19 /* number of codes used to transfer the bit lengths */ #define HEAP_SIZE (2*L_CODES+1) /* maximum heap size */ #define MAX_BITS 15 /* All codes must not exceed MAX_BITS bits */ #define INIT_STATE 42 #define BUSY_STATE 113 #define FLUSH_STATE 124 #define FINISH_STATE 666 /* Stream status */ /* Data structure describing a single value and its code string. */ typedef struct ct_data_s { union { ush freq; /* frequency count */ ush code; /* bit string */ } fc; union { ush dad; /* father node in Huffman tree */ ush len; /* length of bit string */ } dl; } FAR ct_data; #define Freq fc.freq #define Code fc.code #define Dad dl.dad #define Len dl.len typedef struct static_tree_desc_s static_tree_desc; typedef struct tree_desc_s { ct_data *dyn_tree; /* the dynamic tree */ int max_code; /* largest code with non zero frequency */ const static_tree_desc *stat_desc; /* the corresponding static tree */ } FAR tree_desc; typedef ush Pos; typedef Pos FAR Posf; typedef unsigned IPos; /* A Pos is an index in the character window. We use short instead of int to * save space in the various tables. IPos is used only for parameter passing. */ typedef struct deflate_state { z_stream *strm; /* pointer back to this zlib stream */ int status; /* as the name implies */ Bytef *pending_buf; /* output still pending */ Bytef *pending_out; /* next pending byte to output to the stream */ int pending; /* nb of bytes in the pending buffer */ uLong adler; /* adler32 of uncompressed data */ int noheader; /* suppress zlib header and adler32 */ Byte data_type; /* UNKNOWN, BINARY or ASCII */ Byte method; /* STORED (for zip only) or DEFLATED */ int minCompr; /* min size decrease for Z_FLUSH_NOSTORE */ /* used by deflate.c: */ uInt w_size; /* LZ77 window size (32K by default) */ uInt w_bits; /* log2(w_size) (8..16) */ uInt w_mask; /* w_size - 1 */ Bytef *window; /* Sliding window. Input bytes are read into the second half of the window, * and move to the first half later to keep a dictionary of at least wSize * bytes. With this organization, matches are limited to a distance of * wSize-MAX_MATCH bytes, but this ensures that IO is always * performed with a length multiple of the block size. Also, it limits * the window size to 64K, which is quite useful on MSDOS. * To do: use the user input buffer as sliding window. */ ulg window_size; /* Actual size of window: 2*wSize, except when the user input buffer * is directly used as sliding window. */ Posf *prev; /* Link to older string with same hash index. To limit the size of this * array to 64K, this link is maintained only for the last 32K strings. * An index in this array is thus a window index modulo 32K. */ Posf *head; /* Heads of the hash chains or NIL. */ uInt ins_h; /* hash index of string to be inserted */ uInt hash_size; /* number of elements in hash table */ uInt hash_bits; /* log2(hash_size) */ uInt hash_mask; /* hash_size-1 */ uInt hash_shift; /* Number of bits by which ins_h must be shifted at each input * step. It must be such that after MIN_MATCH steps, the oldest * byte no longer takes part in the hash key, that is: * hash_shift * MIN_MATCH >= hash_bits */ long block_start; /* Window position at the beginning of the current output block. Gets * negative when the window is moved backwards. */ uInt match_length; /* length of best match */ IPos prev_match; /* previous match */ int match_available; /* set if previous match exists */ uInt strstart; /* start of string to insert */ uInt match_start; /* start of matching string */ uInt lookahead; /* number of valid bytes ahead in window */ uInt 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. */ uInt 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. */ uInt max_lazy_match; /* Attempt to find a better match only when the current match is strictly * smaller than this value. This mechanism is used only for compression * levels >= 4. */ # define max_insert_length max_lazy_match /* Insert new strings in the hash table only if the match length is not * greater than this length. This saves time but degrades compression. * max_insert_length is used only for compression levels <= 3. */ int level; /* compression level (1..9) */ int strategy; /* favor or force Huffman coding*/ uInt good_match; /* Use a faster search when the previous match is longer than this */ int nice_match; /* Stop searching when current match exceeds this */ /* used by trees.c: */ /* Didn't use ct_data typedef below to supress compiler warning */ struct ct_data_s dyn_ltree[HEAP_SIZE]; /* literal and length tree */ struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */ struct ct_data_s bl_tree[2*BL_CODES+1]; /* Huffman tree for bit lengths */ struct tree_desc_s l_desc; /* desc. for literal tree */ struct tree_desc_s d_desc; /* desc. for distance tree */ struct tree_desc_s bl_desc; /* desc. for bit length tree */ ush bl_count[MAX_BITS+1]; /* number of codes at each bit length for an optimal tree */ int heap[2*L_CODES+1]; /* heap used to build the Huffman trees */ int heap_len; /* number of elements in the heap */ int heap_max; /* element of largest frequency */ /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used. * The same heap array is used to build all trees. */ uch depth[2*L_CODES+1]; /* Depth of each subtree used as tie breaker for trees of equal frequency */ uchf *l_buf; /* buffer for literals or lengths */ uInt lit_bufsize; /* Size of match buffer for literals/lengths. There are 4 reasons for * limiting lit_bufsize to 64K: * - frequencies can be kept in 16 bit counters * - if compression is not successful for the first block, all input * data is still in the window so we can still emit a stored block even * when input comes from standard input. (This can also be done for * all blocks if lit_bufsize is not greater than 32K.) * - if compression is not successful for a file smaller than 64K, we can * even emit a stored file instead of a stored block (saving 5 bytes). * This is applicable only for zip (not gzip or zlib). * - creating new Huffman trees less frequently may not provide fast * adaptation to changes in the input data statistics. (Take for * example a binary file with poorly compressible code followed by * a highly compressible string table.) Smaller buffer sizes give * fast adaptation but have of course the overhead of transmitting * trees more frequently. * - I can't count above 4 */ uInt last_lit; /* running index in l_buf */ ushf *d_buf; /* Buffer for distances. To simplify the code, d_buf and l_buf have * the same number of elements. To use different lengths, an extra flag * array would be necessary. */ ulg opt_len; /* bit length of current block with optimal trees */ ulg static_len; /* bit length of current block with static trees */ ulg compressed_len; /* total bit length of compressed file */ uInt matches; /* number of string matches in current block */ int last_eob_len; /* bit length of EOB code for last block */ #ifdef DEBUG_ZLIB ulg bits_sent; /* bit length of the compressed data */ #endif ush bi_buf; /* Output buffer. bits are inserted starting at the bottom (least * significant bits). */ int bi_valid; /* Number of valid bits in bi_buf. All bits above the last valid bit * are always zero. */ uInt blocks_in_packet; /* Number of blocks produced since the last time Z_PACKET_FLUSH * was used. */ } FAR deflate_state; /* Output a byte on the stream. * IN assertion: there is enough room in pending_buf. */ #define put_byte(s, c) {s->pending_buf[s->pending++] = (c);} #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1) /* Minimum amount of lookahead, except at the end of the input file. * See deflate.c for comments about the MIN_MATCH+1. */ #define MAX_DIST(s) ((s)->w_size-MIN_LOOKAHEAD) /* In order to simplify the code, particularly on 16 bit machines, match * distances are limited to MAX_DIST instead of WSIZE. */ /* in trees.c */ local void ct_init OF((deflate_state *s)); local int ct_tally OF((deflate_state *s, int dist, int lc)); local ulg ct_flush_block OF((deflate_state *s, charf *buf, ulg stored_len, int flush)); local void ct_align OF((deflate_state *s)); local void ct_stored_block OF((deflate_state *s, charf *buf, ulg stored_len, int eof)); local void ct_stored_type_only OF((deflate_state *s)); /*+++++*/ /* deflate.c -- compress data using the deflation algorithm * Copyright (C) 1995 Jean-loup Gailly. * For conditions of distribution and use, see copyright notice in zlib.h */ /* * ALGORITHM * * The "deflation" process depends on being able to identify portions * of the input text which are identical to earlier input (within a * sliding window trailing behind the input currently being processed). * * The most straightforward technique turns out to be the fastest for * most input files: try all possible matches and select the longest. * The key feature of this algorithm is that insertions into the string * dictionary are very simple and thus fast, and deletions are avoided * completely. Insertions are performed at each input character, whereas * string matches are performed only when the previous match ends. So it * is preferable to spend more time in matches to allow very fast string * insertions and avoid deletions. The matching algorithm for small * strings is inspired from that of Rabin & Karp. A brute force approach * is used to find longer strings when a small match has been found. * A similar algorithm is used in comic (by Jan-Mark Wams) and freeze * (by Leonid Broukhis). * A previous version of this file used a more sophisticated algorithm * (by Fiala and Greene) which is guaranteed to run in linear amortized * time, but has a larger average cost, uses more memory and is patented. * However the F&G algorithm may be faster for some highly redundant * files if the parameter max_chain_length (described below) is too large. * * ACKNOWLEDGEMENTS * * The idea of lazy evaluation of matches is due to Jan-Mark Wams, and * I found it in 'freeze' written by Leonid Broukhis. * Thanks to many people for bug reports and testing. * * REFERENCES * * Deutsch, L.P.,"'Deflate' Compressed Data Format Specification". * Available in ftp.uu.net:/pub/archiving/zip/doc/deflate-1.1.doc * * A description of the Rabin and Karp algorithm is given in the book * "Algorithms" by R. Sedgewick, Addison-Wesley, p252. * * Fiala,E.R., and Greene,D.H. * Data Compression with Finite Windows, Comm.ACM, 32,4 (1989) 490-595 * */ /* From: deflate.c,v 1.8 1995/05/03 17:27:08 jloup Exp */ #if 0 local char zlib_copyright[] = " deflate Copyright 1995 Jean-loup Gailly "; #endif /* If you use the zlib library in a product, an acknowledgment is welcome in the documentation of your product. If for some reason you cannot include such an acknowledgment, I would appreciate that you keep this copyright string in the executable of your product. */ #define NIL 0 /* Tail of hash chains */ #ifndef TOO_FAR # define TOO_FAR 4096 #endif /* Matches of length 3 are discarded if their distance exceeds TOO_FAR */ #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1) /* Minimum amount of lookahead, except at the end of the input file. * See deflate.c for comments about the MIN_MATCH+1. */ /* Values for max_lazy_match, good_match and max_chain_length, depending on * the desired pack level (0..9). The values given below have been tuned to * exclude worst case performance for pathological files. Better values may be * found for specific files. */ typedef struct config_s { ush good_length; /* reduce lazy search above this match length */ ush max_lazy; /* do not perform lazy search above this match length */ ush nice_length; /* quit search above this match length */ ush max_chain; } config; local config configuration_table[10] = { /* good lazy nice chain */ /* 0 */ {0, 0, 0, 0}, /* store only */ /* 1 */ {4, 4, 8, 4}, /* maximum speed, no lazy matches */ /* 2 */ {4, 5, 16, 8}, /* 3 */ {4, 6, 32, 32}, /* 4 */ {4, 4, 16, 16}, /* lazy matches */ /* 5 */ {8, 16, 32, 32}, /* 6 */ {8, 16, 128, 128}, /* 7 */ {8, 32, 128, 256}, /* 8 */ {32, 128, 258, 1024}, /* 9 */ {32, 258, 258, 4096}}; /* maximum compression */ /* Note: the deflate() code requires max_lazy >= MIN_MATCH and max_chain >= 4 * For deflate_fast() (levels <= 3) good is ignored and lazy has a different * meaning. */ #define EQUAL 0 /* result of memcmp for equal strings */ /* =========================================================================== * Prototypes for local functions. */ local void fill_window OF((deflate_state *s)); local int deflate_fast OF((deflate_state *s, int flush)); local int deflate_slow OF((deflate_state *s, int flush)); local void lm_init OF((deflate_state *s)); local int longest_match OF((deflate_state *s, IPos cur_match)); local void putShortMSB OF((deflate_state *s, uInt b)); local void flush_pending OF((z_stream *strm)); local int read_buf OF((z_stream *strm, charf *buf, unsigned size)); #ifdef ASMV void match_init OF((void)); /* asm code initialization */ #endif #ifdef DEBUG_ZLIB local void check_match OF((deflate_state *s, 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 * input characters, so that a running hash key can be computed from the * previous key instead of complete recalculation each time. */ #define UPDATE_HASH(s,h,c) (h = (((h)<hash_shift) ^ (c)) & s->hash_mask) /* =========================================================================== * Insert string str 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 * input characters and the first MIN_MATCH bytes of str are valid * (except for the last MIN_MATCH-1 bytes of the input file). */ #define INSERT_STRING(s, str, match_head) \ (UPDATE_HASH(s, s->ins_h, s->window[(str) + (MIN_MATCH-1)]), \ s->prev[(str) & s->w_mask] = match_head = s->head[s->ins_h], \ s->head[s->ins_h] = (str)) /* =========================================================================== * Initialize the hash table (avoiding 64K overflow for 16 bit systems). * prev[] will be initialized on the fly. */ #define CLEAR_HASH(s) \ s->head[s->hash_size-1] = NIL; \ zmemzero((charf *)s->head, (unsigned)(s->hash_size-1)*sizeof(*s->head)); /* ========================================================================= */ int deflateInit (strm, level) z_stream *strm; int level; { return deflateInit2 (strm, level, DEFLATED, MAX_WBITS, DEF_MEM_LEVEL, 0, 0); /* To do: ignore strm->next_in if we use it as window */ } /* ========================================================================= */ int deflateInit2 (strm, level, method, windowBits, memLevel, strategy, minCompression) z_stream *strm; int level; int method; int windowBits; int memLevel; int strategy; int minCompression; { deflate_state *s; int noheader = 0; if (strm == Z_NULL) return Z_STREAM_ERROR; strm->msg = Z_NULL; /* if (strm->zalloc == Z_NULL) strm->zalloc = zcalloc; */ /* if (strm->zfree == Z_NULL) strm->zfree = zcfree; */ if (level == Z_DEFAULT_COMPRESSION) level = 6; if (windowBits < 0) { /* undocumented feature: suppress zlib header */ noheader = 1; windowBits = -windowBits; } if (memLevel < 1 || memLevel > MAX_MEM_LEVEL || method != DEFLATED || windowBits < 8 || windowBits > 15 || level < 1 || level > 9) { return Z_STREAM_ERROR; } s = (deflate_state *) ZALLOC(strm, 1, sizeof(deflate_state)); if (s == Z_NULL) return Z_MEM_ERROR; strm->state = (struct internal_state FAR *)s; s->strm = strm; s->noheader = noheader; s->w_bits = windowBits; s->w_size = 1 << s->w_bits; s->w_mask = s->w_size - 1; s->hash_bits = memLevel + 7; s->hash_size = 1 << s->hash_bits; s->hash_mask = s->hash_size - 1; s->hash_shift = ((s->hash_bits+MIN_MATCH-1)/MIN_MATCH); s->window = (Bytef *) ZALLOC(strm, s->w_size, 2*sizeof(Byte)); s->prev = (Posf *) ZALLOC(strm, s->w_size, sizeof(Pos)); s->head = (Posf *) ZALLOC(strm, s->hash_size, sizeof(Pos)); s->lit_bufsize = 1 << (memLevel + 6); /* 16K elements by default */ s->pending_buf = (uchf *) ZALLOC(strm, s->lit_bufsize, 2*sizeof(ush)); if (s->window == Z_NULL || s->prev == Z_NULL || s->head == Z_NULL || s->pending_buf == Z_NULL) { strm->msg = z_errmsg[1-Z_MEM_ERROR]; deflateEnd (strm); return Z_MEM_ERROR; } s->d_buf = (ushf *) &(s->pending_buf[s->lit_bufsize]); s->l_buf = (uchf *) &(s->pending_buf[3*s->lit_bufsize]); /* We overlay pending_buf and d_buf+l_buf. This works since the average * output size for (length,distance) codes is <= 32 bits (worst case * is 15+15+13=33). */ s->level = level; s->strategy = strategy; s->method = (Byte)method; s->minCompr = minCompression; s->blocks_in_packet = 0; return deflateReset(strm); } /* ========================================================================= */ int deflateReset (strm) z_stream *strm; { deflate_state *s; if (strm == Z_NULL || strm->state == Z_NULL || strm->zalloc == Z_NULL || strm->zfree == Z_NULL) return Z_STREAM_ERROR; strm->total_in = strm->total_out = 0; strm->msg = Z_NULL; /* use zfree if we ever allocate msg dynamically */ strm->data_type = Z_UNKNOWN; s = (deflate_state *)strm->state; s->pending = 0; s->pending_out = s->pending_buf; if (s->noheader < 0) { s->noheader = 0; /* was set to -1 by deflate(..., Z_FINISH); */ } s->status = s->noheader ? BUSY_STATE : INIT_STATE; s->adler = 1; ct_init(s); lm_init(s); return Z_OK; } /* ========================================================================= * Put a short in the pending buffer. The 16-bit value is put in MSB order. * IN assertion: the stream state is correct and there is enough room in * pending_buf. */ local void putShortMSB (s, b) deflate_state *s; uInt b; { put_byte(s, (Byte)(b >> 8)); put_byte(s, (Byte)(b & 0xff)); } /* ========================================================================= * Flush as much pending output as possible. */ local void flush_pending(strm) z_stream *strm; { deflate_state *state = (deflate_state *) strm->state; unsigned len = state->pending; if (len > strm->avail_out) len = strm->avail_out; if (len == 0) return; if (strm->next_out != NULL) { zmemcpy(strm->next_out, state->pending_out, len); strm->next_out += len; } state->pending_out += len; strm->total_out += len; strm->avail_out -= len; state->pending -= len; if (state->pending == 0) { state->pending_out = state->pending_buf; } } /* ========================================================================= */ int deflate (strm, flush) z_stream *strm; int flush; { deflate_state *state = (deflate_state *) strm->state; if (strm == Z_NULL || state == Z_NULL) return Z_STREAM_ERROR; if (strm->next_in == Z_NULL && strm->avail_in != 0) { ERR_RETURN(strm, Z_STREAM_ERROR); } if (strm->avail_out == 0) ERR_RETURN(strm, Z_BUF_ERROR); state->strm = strm; /* just in case */ /* Write the zlib header */ if (state->status == INIT_STATE) { uInt header = (DEFLATED + ((state->w_bits-8)<<4)) << 8; uInt level_flags = (state->level-1) >> 1; if (level_flags > 3) level_flags = 3; header |= (level_flags << 6); header += 31 - (header % 31); state->status = BUSY_STATE; putShortMSB(state, header); } /* Flush as much pending output as possible */ if (state->pending != 0) { flush_pending(strm); if (strm->avail_out == 0) return Z_OK; } /* If we came back in here to get the last output from * a previous flush, we're done for now. */ if (state->status == FLUSH_STATE) { state->status = BUSY_STATE; if (flush != Z_NO_FLUSH && flush != Z_FINISH) return Z_OK; } /* User must not provide more input after the first FINISH: */ if (state->status == FINISH_STATE && strm->avail_in != 0) { ERR_RETURN(strm, Z_BUF_ERROR); } /* Start a new block or continue the current one. */ if (strm->avail_in != 0 || state->lookahead != 0 || (flush == Z_FINISH && state->status != FINISH_STATE)) { int quit; if (flush == Z_FINISH) { state->status = FINISH_STATE; } if (state->level <= 3) { quit = deflate_fast(state, flush); } else { quit = deflate_slow(state, flush); } if (quit || strm->avail_out == 0) return Z_OK; /* If flush != Z_NO_FLUSH && avail_out == 0, the next call * of deflate should use the same flush parameter to make sure * that the flush is complete. So we don't have to output an * empty block here, this will be done at next call. This also * ensures that for a very small output buffer, we emit at most * one empty block. */ } /* If a flush was requested, we have a little more to output now. */ if (flush != Z_NO_FLUSH && flush != Z_FINISH && state->status != FINISH_STATE) { switch (flush) { case Z_PARTIAL_FLUSH: ct_align(state); break; case Z_PACKET_FLUSH: /* Output just the 3-bit `stored' block type value, but not a zero length. */ ct_stored_type_only(state); break; default: ct_stored_block(state, (char*)0, 0L, 0); /* For a full flush, this empty block will be recognized * as a special marker by inflate_sync(). */ if (flush == Z_FULL_FLUSH) { CLEAR_HASH(state); /* forget history */ } } flush_pending(strm); if (strm->avail_out == 0) { /* We'll have to come back to get the rest of the output; * this ensures we don't output a second zero-length stored * block (or whatever). */ state->status = FLUSH_STATE; return Z_OK; } } Assert(strm->avail_out > 0, "bug2"); if (flush != Z_FINISH) return Z_OK; if (state->noheader) return Z_STREAM_END; /* Write the zlib trailer (adler32) */ putShortMSB(state, (uInt)(state->adler >> 16)); putShortMSB(state, (uInt)(state->adler & 0xffff)); flush_pending(strm); /* If avail_out is zero, the application will call deflate again * to flush the rest. */ state->noheader = -1; /* write the trailer only once! */ return state->pending != 0 ? Z_OK : Z_STREAM_END; } /* ========================================================================= */ int deflateEnd (strm) z_stream *strm; { deflate_state *state = (deflate_state *) strm->state; if (strm == Z_NULL || state == Z_NULL) return Z_STREAM_ERROR; TRY_FREE(strm, state->window, state->w_size * 2 * sizeof(Byte)); TRY_FREE(strm, state->prev, state->w_size * sizeof(Pos)); TRY_FREE(strm, state->head, state->hash_size * sizeof(Pos)); TRY_FREE(strm, state->pending_buf, state->lit_bufsize * 2 * sizeof(ush)); ZFREE(strm, state, sizeof(deflate_state)); strm->state = Z_NULL; return Z_OK; } /* =========================================================================== * Read a new buffer from the current input stream, update the adler32 * and total number of bytes read. */ local int read_buf(strm, buf, size) z_stream *strm; charf *buf; unsigned size; { unsigned len = strm->avail_in; deflate_state *state = (deflate_state *) strm->state; if (len > size) len = size; if (len == 0) return 0; strm->avail_in -= len; if (!state->noheader) { state->adler = adler32(state->adler, strm->next_in, len); } zmemcpy(buf, strm->next_in, len); strm->next_in += len; strm->total_in += len; return (int)len; } /* =========================================================================== * Initialize the "longest match" routines for a new zlib stream */ local void lm_init (s) deflate_state *s; { s->window_size = (ulg)2L*s->w_size; CLEAR_HASH(s); /* Set the default configuration parameters: */ s->max_lazy_match = configuration_table[s->level].max_lazy; s->good_match = configuration_table[s->level].good_length; s->nice_match = configuration_table[s->level].nice_length; s->max_chain_length = configuration_table[s->level].max_chain; s->strstart = 0; s->block_start = 0L; s->lookahead = 0; s->match_length = MIN_MATCH-1; s->match_available = 0; s->ins_h = 0; #ifdef ASMV match_init(); /* initialize the asm code */ #endif } /* =========================================================================== * Set match_start to the longest match starting at the given string and * return its length. Matches shorter or equal to prev_length are discarded, * in which case the result is equal to prev_length and match_start is * garbage. * IN assertions: cur_match is the head of the hash chain for the current * string (strstart) and its distance is <= MAX_DIST, and prev_length >= 1 */ #ifndef ASMV /* For 80x86 and 680x0, an optimized version will be provided in match.asm or * match.S. The code will be functionally equivalent. */ local int longest_match(s, cur_match) deflate_state *s; IPos cur_match; /* current match */ { unsigned chain_length = s->max_chain_length;/* max hash chain length */ register Bytef *scan = s->window + s->strstart; /* current string */ register Bytef *match; /* matched string */ register int len; /* length of current match */ int best_len = s->prev_length; /* best match length so far */ IPos limit = s->strstart > (IPos)MAX_DIST(s) ? s->strstart - (IPos)MAX_DIST(s) : NIL; /* Stop when cur_match becomes <= limit. To simplify the code, * we prevent matches with the string of window index 0. */ Posf *prev = s->prev; uInt wmask = s->w_mask; #ifdef UNALIGNED_OK /* Compare two bytes at a time. Note: this is not always beneficial. * Try with and without -DUNALIGNED_OK to check. */ register Bytef *strend = s->window + s->strstart + MAX_MATCH - 1; register ush scan_start = *(ushf*)scan; register ush scan_end = *(ushf*)(scan+best_len-1); #else register Bytef *strend = s->window + s->strstart + MAX_MATCH; register Byte scan_end1 = scan[best_len-1]; register Byte scan_end = scan[best_len]; #endif /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16. * It is easy to get rid of this optimization if necessary. */ Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever"); /* Do not waste too much time if we already have a good match: */ if (s->prev_length >= s->good_match) { chain_length >>= 2; } Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead"); do { Assert(cur_match < s->strstart, "no future"); match = s->window + cur_match; /* Skip to next match if the match length cannot increase * or if the match length is less than 2: */ #if (defined(UNALIGNED_OK) && MAX_MATCH == 258) /* This code assumes sizeof(unsigned short) == 2. Do not use * UNALIGNED_OK if your compiler uses a different size. */ if (*(ushf*)(match+best_len-1) != scan_end || *(ushf*)match != scan_start) continue; /* It is not necessary to compare scan[2] and match[2] since they are * always equal when the other bytes match, given that the hash keys * are equal and that HASH_BITS >= 8. Compare 2 bytes at a time at * strstart+3, +5, ... up to strstart+257. We check for insufficient * lookahead only every 4th comparison; the 128th check will be made * at strstart+257. If MAX_MATCH-2 is not a multiple of 8, it is * necessary to put more guard bytes at the end of the window, or * to check more often for insufficient lookahead. */ Assert(scan[2] == match[2], "scan[2]?"); scan++, match++; do { } while (*(ushf*)(scan+=2) == *(ushf*)(match+=2) && *(ushf*)(scan+=2) == *(ushf*)(match+=2) && *(ushf*)(scan+=2) == *(ushf*)(match+=2) && *(ushf*)(scan+=2) == *(ushf*)(match+=2) && scan < strend); /* The funny "do {}" generates better code on most compilers */ /* Here, scan <= window+strstart+257 */ Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan"); if (*scan == *match) scan++; len = (MAX_MATCH - 1) - (int)(strend-scan); scan = strend - (MAX_MATCH-1); #else /* UNALIGNED_OK */ if (match[best_len] != scan_end || match[best_len-1] != scan_end1 || *match != *scan || *++match != scan[1]) continue; /* The check at best_len-1 can be removed because it will be made * again later. (This heuristic is not always a win.) * It is not necessary to compare scan[2] and match[2] since they * are always equal when the other bytes match, given that * the hash keys are equal and that HASH_BITS >= 8. */ scan += 2, match++; Assert(*scan == *match, "match[2]?"); /* We check for insufficient lookahead only every 8th comparison; * the 256th check will be made at strstart+258. */ do { } while (*++scan == *++match && *++scan == *++match && *++scan == *++match && *++scan == *++match && *++scan == *++match && *++scan == *++match && *++scan == *++match && *++scan == *++match && scan < strend); Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan"); len = MAX_MATCH - (int)(strend - scan); scan = strend - MAX_MATCH; #endif /* UNALIGNED_OK */ if (len > best_len) { s->match_start = cur_match; best_len = len; if (len >= s->nice_match) break; #ifdef UNALIGNED_OK scan_end = *(ushf*)(scan+best_len-1); #else scan_end1 = scan[best_len-1]; scan_end = scan[best_len]; #endif } } while ((cur_match = prev[cur_match & wmask]) > limit && --chain_length != 0); return best_len; } #endif /* ASMV */ #ifdef DEBUG_ZLIB /* =========================================================================== * Check that the match at match_start is indeed a match. */ local void check_match(s, start, match, length) deflate_state *s; IPos start, match; int length; { /* check that the match is indeed a match */ if (memcmp((charf *)s->window + match, (charf *)s->window + start, length) != EQUAL) { fprintf(stderr, " start %u, match %u, length %d\n", start, match, length); do { fprintf(stderr, "%c%c", s->window[match++], s->window[start++]); } while (--length != 0); z_error("invalid match"); } if (verbose > 1) { fprintf(stderr,"\\[%d,%d]", start-match, length); do { putc(s->window[start++], stderr); } while (--length != 0); } } #else # define check_match(s, start, match, length) #endif /* =========================================================================== * Fill the window when the lookahead becomes insufficient. * Updates strstart and lookahead. * * IN assertion: lookahead < MIN_LOOKAHEAD * OUT assertions: strstart <= window_size-MIN_LOOKAHEAD * At least one byte has been read, or avail_in == 0; reads are * performed for at least two bytes (required for the zip translate_eol * option -- not supported here). */ local void fill_window(s) deflate_state *s; { register unsigned n, m; register Posf *p; unsigned more; /* Amount of free space at the end of the window. */ uInt wsize = s->w_size; do { more = (unsigned)(s->window_size -(ulg)s->lookahead -(ulg)s->strstart); /* Deal with !@#$% 64K limit: */ if (more == 0 && s->strstart == 0 && s->lookahead == 0) { more = wsize; } else if (more == (unsigned)(-1)) { /* Very unlikely, but possible on 16 bit machine if strstart == 0 * and lookahead == 1 (input done one byte at time) */ more--; /* If the window is almost full and there is insufficient lookahead, * move the upper half to the lower one to make room in the upper half. */ } else if (s->strstart >= wsize+MAX_DIST(s)) { /* By the IN assertion, the window is not empty so we can't confuse * more == 0 with more == 64K on a 16 bit machine. */ zmemcpy((charf *)s->window, (charf *)s->window+wsize, (unsigned)wsize); s->match_start -= wsize; s->strstart -= wsize; /* we now have strstart >= MAX_DIST */ s->block_start -= (long) wsize; /* Slide the hash table (could be avoided with 32 bit values at the expense of memory usage): */ n = s->hash_size; p = &s->head[n]; do { m = *--p; *p = (Pos)(m >= wsize ? m-wsize : NIL); } while (--n); n = wsize; p = &s->prev[n]; do { m = *--p; *p = (Pos)(m >= wsize ? m-wsize : NIL); /* If n is not on any hash chain, prev[n] is garbage but * its value will never be used. */ } while (--n); more += wsize; } if (s->strm->avail_in == 0) return; /* If there was no sliding: * strstart <= WSIZE+MAX_DIST-1 && lookahead <= MIN_LOOKAHEAD - 1 && * more == window_size - lookahead - strstart * => more >= window_size - (MIN_LOOKAHEAD-1 + WSIZE + MAX_DIST-1) * => more >= window_size - 2*WSIZE + 2 * In the BIG_MEM or MMAP case (not yet supported), * window_size == input_size + MIN_LOOKAHEAD && * strstart + s->lookahead <= input_size => more >= MIN_LOOKAHEAD. * Otherwise, window_size == 2*WSIZE so more >= 2. * If there was sliding, more >= WSIZE. So in all cases, more >= 2. */ Assert(more >= 2, "more < 2"); n = read_buf(s->strm, (charf *)s->window + s->strstart + s->lookahead, more); s->lookahead += n; /* Initialize the hash value now that we have some input: */ if (s->lookahead >= MIN_MATCH) { s->ins_h = s->window[s->strstart]; UPDATE_HASH(s, s->ins_h, s->window[s->strstart+1]); #if MIN_MATCH != 3 Call UPDATE_HASH() MIN_MATCH-3 more times #endif } /* If the whole input has less than MIN_MATCH bytes, ins_h is garbage, * but this is not important since only literal bytes will be emitted. */ } while (s->lookahead < MIN_LOOKAHEAD && s->strm->avail_in != 0); } /* =========================================================================== * Flush the current block, with given end-of-file flag. * IN assertion: strstart is set to the end of the current match. */ #define FLUSH_BLOCK_ONLY(s, flush) { \ ct_flush_block(s, (s->block_start >= 0L ? \ (charf *)&s->window[(unsigned)s->block_start] : \ (charf *)Z_NULL), (long)s->strstart - s->block_start, (flush)); \ s->block_start = s->strstart; \ flush_pending(s->strm); \ Tracev((stderr,"[FLUSH]")); \ } /* Same but force premature exit if necessary. */ #define FLUSH_BLOCK(s, flush) { \ FLUSH_BLOCK_ONLY(s, flush); \ if (s->strm->avail_out == 0) return 1; \ } /* =========================================================================== * Compress as much as possible from the input stream, return true if * processing was terminated prematurely (no more input or output space). * This function does not perform lazy evaluationof matches and inserts * new strings in the dictionary only for unmatched strings or for short * matches. It is used only for the fast compression options. */ local int deflate_fast(s, flush) deflate_state *s; int flush; { IPos hash_head = NIL; /* head of the hash chain */ int bflush; /* set if current block must be flushed */ s->prev_length = MIN_MATCH-1; for (;;) { /* Make sure that we always have enough lookahead, except * at the end of the input file. We need MAX_MATCH bytes * for the next match, plus MIN_MATCH bytes to insert the * string following the next match. */ if (s->lookahead < MIN_LOOKAHEAD) { fill_window(s); if (s->lookahead < MIN_LOOKAHEAD && flush == Z_NO_FLUSH) return 1; if (s->lookahead == 0) break; /* flush the current block */ } /* Insert the string window[strstart .. strstart+2] in the * dictionary, and set hash_head to the head of the hash chain: */ if (s->lookahead >= MIN_MATCH) { INSERT_STRING(s, s->strstart, hash_head); } /* Find the longest match, discarding those <= prev_length. * At this point we have always match_length < MIN_MATCH */ if (hash_head != NIL && s->strstart - hash_head <= MAX_DIST(s)) { /* 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). */ if (s->strategy != Z_HUFFMAN_ONLY) { s->match_length = longest_match (s, hash_head); } /* longest_match() sets match_start */ if (s->match_length > s->lookahead) s->match_length = s->lookahead; } if (s->match_length >= MIN_MATCH) { check_match(s, s->strstart, s->match_start, s->match_length); bflush = ct_tally(s, s->strstart - s->match_start, s->match_length - MIN_MATCH); s->lookahead -= s->match_length; /* Insert new strings in the hash table only if the match length * is not too large. This saves time but degrades compression. */ if (s->match_length <= s->max_insert_length && s->lookahead >= MIN_MATCH) { s->match_length--; /* string at strstart already in hash table */ do { s->strstart++; INSERT_STRING(s, s->strstart, hash_head); /* strstart never exceeds WSIZE-MAX_MATCH, so there are * always MIN_MATCH bytes ahead. */ } while (--s->match_length != 0); s->strstart++; } else { s->strstart += s->match_length; s->match_length = 0; s->ins_h = s->window[s->strstart]; UPDATE_HASH(s, s->ins_h, s->window[s->strstart+1]); #if MIN_MATCH != 3 Call UPDATE_HASH() MIN_MATCH-3 more times #endif /* If lookahead < MIN_MATCH, ins_h is garbage, but it does not * matter since it will be recomputed at next deflate call. */ } } else { /* No match, output a literal byte */ Tracevv((stderr,"%c", s->window[s->strstart])); bflush = ct_tally (s, 0, s->window[s->strstart]); s->lookahead--; s->strstart++; } if (bflush) FLUSH_BLOCK(s, Z_NO_FLUSH); } FLUSH_BLOCK(s, flush); return 0; /* normal exit */ } /* =========================================================================== * Same as above, but achieves better compression. We use a lazy * evaluation for matches: a match is finally adopted only if there is * no better match at the next window position. */ local int deflate_slow(s, flush) deflate_state *s; int flush; { IPos hash_head = NIL; /* head of hash chain */ int bflush; /* set if current block must be flushed */ /* Process the input block. */ for (;;) { /* Make sure that we always have enough lookahead, except * at the end of the input file. We need MAX_MATCH bytes * for the next match, plus MIN_MATCH bytes to insert the * string following the next match. */ if (s->lookahead < MIN_LOOKAHEAD) { fill_window(s); if (s->lookahead < MIN_LOOKAHEAD && flush == Z_NO_FLUSH) return 1; if (s->lookahead == 0) break; /* flush the current block */ } /* Insert the string window[strstart .. strstart+2] in the * dictionary, and set hash_head to the head of the hash chain: */ if (s->lookahead >= MIN_MATCH) { INSERT_STRING(s, s->strstart, hash_head); } /* Find the longest match, discarding those <= prev_length. */ s->prev_length = s->match_length, s->prev_match = s->match_start; s->match_length = MIN_MATCH-1; if (hash_head != NIL && s->prev_length < s->max_lazy_match && s->strstart - hash_head <= MAX_DIST(s)) { /* 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). */ if (s->strategy != Z_HUFFMAN_ONLY) { s->match_length = longest_match (s, hash_head); } /* longest_match() sets match_start */ if (s->match_length > s->lookahead) s->match_length = s->lookahead; if (s->match_length <= 5 && (s->strategy == Z_FILTERED || (s->match_length == MIN_MATCH && s->strstart - s->match_start > TOO_FAR))) { /* If prev_match is also MIN_MATCH, match_start is garbage * but we will ignore the current match anyway. */ s->match_length = MIN_MATCH-1; } } /* If there was a match at the previous step and the current * match is not better, output the previous match: */ if (s->prev_length >= MIN_MATCH && s->match_length <= s->prev_length) { uInt max_insert = s->strstart + s->lookahead - MIN_MATCH; /* Do not insert strings in hash table beyond this. */ check_match(s, s->strstart-1, s->prev_match, s->prev_length); bflush = ct_tally(s, s->strstart -1 - s->prev_match, s->prev_length - MIN_MATCH); /* Insert in hash table all strings up to the end of the match. * strstart-1 and strstart are already inserted. If there is not * enough lookahead, the last two strings are not inserted in * the hash table. */ s->lookahead -= s->prev_length-1; s->prev_length -= 2; do { if (++s->strstart <= max_insert) { INSERT_STRING(s, s->strstart, hash_head); } } while (--s->prev_length != 0); s->match_available = 0; s->match_length = MIN_MATCH-1; s->strstart++; if (bflush) FLUSH_BLOCK(s, Z_NO_FLUSH); } else if (s->match_available) { /* If there was no match at the previous position, output a * single literal. If there was a match but the current match * is longer, truncate the previous match to a single literal. */ Tracevv((stderr,"%c", s->window[s->strstart-1])); if (ct_tally (s, 0, s->window[s->strstart-1])) { FLUSH_BLOCK_ONLY(s, Z_NO_FLUSH); } s->strstart++; s->lookahead--; if (s->strm->avail_out == 0) return 1; } else { /* There is no previous match to compare with, wait for * the next step to decide. */ s->match_available = 1; s->strstart++; s->lookahead--; } } Assert (flush != Z_NO_FLUSH, "no flush?"); if (s->match_available) { Tracevv((stderr,"%c", s->window[s->strstart-1])); ct_tally (s, 0, s->window[s->strstart-1]); s->match_available = 0; } FLUSH_BLOCK(s, flush); return 0; } /*+++++*/ /* trees.c -- output deflated data using Huffman coding * Copyright (C) 1995 Jean-loup Gailly * For conditions of distribution and use, see copyright notice in zlib.h */ /* * ALGORITHM * * The "deflation" process uses several Huffman trees. The more * common source values are represented by shorter bit sequences. * * Each code tree is stored in a compressed form which is itself * a Huffman encoding of the lengths of all the code strings (in * ascending order by source values). The actual code strings are * reconstructed from the lengths in the inflate process, as described * in the deflate specification. * * REFERENCES * * Deutsch, L.P.,"'Deflate' Compressed Data Format Specification". * Available in ftp.uu.net:/pub/archiving/zip/doc/deflate-1.1.doc * * Storer, James A. * Data Compression: Methods and Theory, pp. 49-50. * Computer Science Press, 1988. ISBN 0-7167-8156-5. * * Sedgewick, R. * Algorithms, p290. * Addison-Wesley, 1983. ISBN 0-201-06672-6. */ /* From: trees.c,v 1.5 1995/05/03 17:27:12 jloup Exp */ #ifdef DEBUG_ZLIB # include #endif /* =========================================================================== * Constants */ #define MAX_BL_BITS 7 /* Bit length codes must not exceed MAX_BL_BITS bits */ #define END_BLOCK 256 /* end of block literal code */ #define REP_3_6 16 /* repeat previous bit length 3-6 times (2 bits of repeat count) */ #define REPZ_3_10 17 /* repeat a zero length 3-10 times (3 bits of repeat count) */ #define REPZ_11_138 18 /* repeat a zero length 11-138 times (7 bits of repeat count) */ local const int extra_lbits[LENGTH_CODES] /* extra bits for each length code */ = {0,0,0,0,0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,5,0}; local const int extra_dbits[D_CODES] /* extra bits for each distance code */ = {0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13}; local const int extra_blbits[BL_CODES]/* extra bits for each bit length code */ = {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,3,7}; local const uch bl_order[BL_CODES] = {16,17,18,0,8,7,9,6,10,5,11,4,12,3,13,2,14,1,15}; /* The lengths of the bit length codes are sent in order of decreasing * probability, to avoid transmitting the lengths for unused bit length codes. */ #define Buf_size (8 * 2*sizeof(char)) /* Number of bits used within bi_buf. (bi_buf might be implemented on * more than 16 bits on some systems.) */ /* =========================================================================== * Local data. These are initialized only once. * To do: initialize at compile time to be completely reentrant. ??? */ local ct_data static_ltree[L_CODES+2]; /* The static literal tree. Since the bit lengths are imposed, there is no * need for the L_CODES extra codes used during heap construction. However * The codes 286 and 287 are needed to build a canonical tree (see ct_init * below). */ local ct_data static_dtree[D_CODES]; /* The static distance tree. (Actually a trivial tree since all codes use * 5 bits.) */ local uch dist_code[512]; /* distance codes. The first 256 values correspond to the distances * 3 .. 258, the last 256 values correspond to the top 8 bits of * the 15 bit distances. */ local uch length_code[MAX_MATCH-MIN_MATCH+1]; /* length code for each normalized match length (0 == MIN_MATCH) */ local int base_length[LENGTH_CODES]; /* First normalized length for each code (0 = MIN_MATCH) */ local int base_dist[D_CODES]; /* First normalized distance for each code (0 = distance of 1) */ struct static_tree_desc_s { ct_data *static_tree; /* static tree or NULL */ const intf *extra_bits; /* extra bits for each code or NULL */ int extra_base; /* base index for extra_bits */ int elems; /* max number of elements in the tree */ int max_length; /* max bit length for the codes */ }; local const static_tree_desc static_l_desc = {static_ltree, extra_lbits, LITERALS+1, L_CODES, MAX_BITS}; local const static_tree_desc static_d_desc = {static_dtree, extra_dbits, 0, D_CODES, MAX_BITS}; local const static_tree_desc static_bl_desc = {(ct_data *)0, extra_blbits, 0, BL_CODES, MAX_BL_BITS}; /* =========================================================================== * Local (static) routines in this file. */ local void ct_static_init OF((void)); local void init_block OF((deflate_state *s)); local void pqdownheap OF((deflate_state *s, ct_data *tree, int k)); local void gen_bitlen OF((deflate_state *s, tree_desc *desc)); local void gen_codes OF((ct_data *tree, int max_code, ushf *bl_count)); local void build_tree OF((deflate_state *s, tree_desc *desc)); local void scan_tree OF((deflate_state *s, ct_data *tree, int max_code)); local void send_tree OF((deflate_state *s, ct_data *tree, int max_code)); local int build_bl_tree OF((deflate_state *s)); local void send_all_trees OF((deflate_state *s, int lcodes, int dcodes, int blcodes)); local void compress_block OF((deflate_state *s, ct_data *ltree, ct_data *dtree)); local void set_data_type OF((deflate_state *s)); local unsigned bi_reverse OF((unsigned value, int length)); local void bi_windup OF((deflate_state *s)); local void bi_flush OF((deflate_state *s)); local void copy_block OF((deflate_state *s, charf *buf, unsigned len, int header)); #ifndef DEBUG_ZLIB # define send_code(s, c, tree) send_bits(s, tree[c].Code, tree[c].Len) /* Send a code of the given tree. c and tree must not have side effects */ #else /* DEBUG_ZLIB */ # define send_code(s, c, tree) \ { if (verbose>1) fprintf(stderr,"\ncd %3d ",(c)); \ send_bits(s, tree[c].Code, tree[c].Len); } #endif #define d_code(dist) \ ((dist) < 256 ? dist_code[dist] : dist_code[256+((dist)>>7)]) /* Mapping from a distance to a distance code. dist is the distance - 1 and * must not have side effects. dist_code[256] and dist_code[257] are never * used. */ /* =========================================================================== * Output a short LSB first on the stream. * IN assertion: there is enough room in pendingBuf. */ #define put_short(s, w) { \ put_byte(s, (uch)((w) & 0xff)); \ put_byte(s, (uch)((ush)(w) >> 8)); \ } /* =========================================================================== * Send a value on a given number of bits. * IN assertion: length <= 16 and value fits in length bits. */ #ifdef DEBUG_ZLIB local void send_bits OF((deflate_state *s, int value, int length)); local void send_bits(s, value, length) deflate_state *s; int value; /* value to send */ int length; /* number of bits */ { Tracev((stderr," l %2d v %4x ", length, value)); Assert(length > 0 && length <= 15, "invalid length"); s->bits_sent += (ulg)length; /* If not enough room in bi_buf, use (valid) bits from bi_buf and * (16 - bi_valid) bits from value, leaving (width - (16-bi_valid)) * unused bits in value. */ if (s->bi_valid > (int)Buf_size - length) { s->bi_buf |= (value << s->bi_valid); put_short(s, s->bi_buf); s->bi_buf = (ush)value >> (Buf_size - s->bi_valid); s->bi_valid += length - Buf_size; } else { s->bi_buf |= value << s->bi_valid; s->bi_valid += length; } } #else /* !DEBUG_ZLIB */ #define send_bits(s, value, length) \ { int len = length;\ if (s->bi_valid > (int)Buf_size - len) {\ int val = value;\ s->bi_buf |= (val << s->bi_valid);\ put_short(s, s->bi_buf);\ s->bi_buf = (ush)val >> (Buf_size - s->bi_valid);\ s->bi_valid += len - Buf_size;\ } else {\ s->bi_buf |= (value) << s->bi_valid;\ s->bi_valid += len;\ }\ } #endif /* DEBUG_ZLIB */ /* the arguments must not have side effects */ /* =========================================================================== * Initialize the various 'constant' tables. * To do: do this at compile time. */ local void ct_static_init() { int n; /* iterates over tree elements */ int bits; /* bit counter */ int length; /* length value */ int code; /* code value */ int dist; /* distance index */ ush bl_count[MAX_BITS+1]; /* number of codes at each bit length for an optimal tree */ /* Initialize the mapping length (0..255) -> length code (0..28) */ length = 0; for (code = 0; code < LENGTH_CODES-1; code++) { base_length[code] = length; for (n = 0; n < (1< dist code (0..29) */ dist = 0; for (code = 0 ; code < 16; code++) { base_dist[code] = dist; for (n = 0; n < (1<>= 7; /* from now on, all distances are divided by 128 */ for ( ; code < D_CODES; code++) { base_dist[code] = dist << 7; for (n = 0; n < (1<<(extra_dbits[code]-7)); n++) { dist_code[256 + dist++] = (uch)code; } } Assert (dist == 256, "ct_static_init: 256+dist != 512"); /* Construct the codes of the static literal tree */ for (bits = 0; bits <= MAX_BITS; bits++) bl_count[bits] = 0; n = 0; while (n <= 143) static_ltree[n++].Len = 8, bl_count[8]++; while (n <= 255) static_ltree[n++].Len = 9, bl_count[9]++; while (n <= 279) static_ltree[n++].Len = 7, bl_count[7]++; while (n <= 287) static_ltree[n++].Len = 8, bl_count[8]++; /* Codes 286 and 287 do not exist, but we must include them in the * tree construction to get a canonical Huffman tree (longest code * all ones) */ gen_codes((ct_data *)static_ltree, L_CODES+1, bl_count); /* The static distance tree is trivial: */ for (n = 0; n < D_CODES; n++) { static_dtree[n].Len = 5; static_dtree[n].Code = bi_reverse(n, 5); } } /* =========================================================================== * Initialize the tree data structures for a new zlib stream. */ local void ct_init(s) deflate_state *s; { if (static_dtree[0].Len == 0) { ct_static_init(); /* To do: at compile time */ } s->compressed_len = 0L; s->l_desc.dyn_tree = s->dyn_ltree; s->l_desc.stat_desc = &static_l_desc; s->d_desc.dyn_tree = s->dyn_dtree; s->d_desc.stat_desc = &static_d_desc; s->bl_desc.dyn_tree = s->bl_tree; s->bl_desc.stat_desc = &static_bl_desc; s->bi_buf = 0; s->bi_valid = 0; s->last_eob_len = 8; /* enough lookahead for inflate */ #ifdef DEBUG_ZLIB s->bits_sent = 0L; #endif s->blocks_in_packet = 0; /* Initialize the first block of the first file: */ init_block(s); } /* =========================================================================== * Initialize a new block. */ local void init_block(s) deflate_state *s; { int n; /* iterates over tree elements */ /* Initialize the trees. */ for (n = 0; n < L_CODES; n++) s->dyn_ltree[n].Freq = 0; for (n = 0; n < D_CODES; n++) s->dyn_dtree[n].Freq = 0; for (n = 0; n < BL_CODES; n++) s->bl_tree[n].Freq = 0; s->dyn_ltree[END_BLOCK].Freq = 1; s->opt_len = s->static_len = 0L; s->last_lit = s->matches = 0; } #define SMALLEST 1 /* Index within the heap array of least frequent node in the Huffman tree */ /* =========================================================================== * Remove the smallest element from the heap and recreate the heap with * one less element. Updates heap and heap_len. */ #define pqremove(s, tree, top) \ {\ top = s->heap[SMALLEST]; \ s->heap[SMALLEST] = s->heap[s->heap_len--]; \ pqdownheap(s, tree, SMALLEST); \ } /* =========================================================================== * Compares to subtrees, using the tree depth as tie breaker when * the subtrees have equal frequency. This minimizes the worst case length. */ #define smaller(tree, n, m, depth) \ (tree[n].Freq < tree[m].Freq || \ (tree[n].Freq == tree[m].Freq && depth[n] <= depth[m])) /* =========================================================================== * Restore the heap property by moving down the tree starting at node k, * exchanging a node with the smallest of its two sons if necessary, stopping * when the heap property is re-established (each father smaller than its * two sons). */ local void pqdownheap(s, tree, k) deflate_state *s; ct_data *tree; /* the tree to restore */ int k; /* node to move down */ { int v = s->heap[k]; int j = k << 1; /* left son of k */ while (j <= s->heap_len) { /* Set j to the smallest of the two sons: */ if (j < s->heap_len && smaller(tree, s->heap[j+1], s->heap[j], s->depth)) { j++; } /* Exit if v is smaller than both sons */ if (smaller(tree, v, s->heap[j], s->depth)) break; /* Exchange v with the smallest son */ s->heap[k] = s->heap[j]; k = j; /* And continue down the tree, setting j to the left son of k */ j <<= 1; } s->heap[k] = v; } /* =========================================================================== * Compute the optimal bit lengths for a tree and update the total bit length * for the current block. * IN assertion: the fields freq and dad are set, heap[heap_max] and * above are the tree nodes sorted by increasing frequency. * OUT assertions: the field len is set to the optimal bit length, the * array bl_count contains the frequencies for each bit length. * The length opt_len is updated; static_len is also updated if stree is * not null. */ local void gen_bitlen(s, desc) deflate_state *s; tree_desc *desc; /* the tree descriptor */ { ct_data *tree = desc->dyn_tree; int max_code = desc->max_code; ct_data *stree = desc->stat_desc->static_tree; const intf *extra = desc->stat_desc->extra_bits; int base = desc->stat_desc->extra_base; int max_length = desc->stat_desc->max_length; int h; /* heap index */ int n, m; /* iterate over the tree elements */ int bits; /* bit length */ int xbits; /* extra bits */ ush f; /* frequency */ int overflow = 0; /* number of elements with bit length too large */ for (bits = 0; bits <= MAX_BITS; bits++) s->bl_count[bits] = 0; /* In a first pass, compute the optimal bit lengths (which may * overflow in the case of the bit length tree). */ tree[s->heap[s->heap_max]].Len = 0; /* root of the heap */ for (h = s->heap_max+1; h < HEAP_SIZE; h++) { n = s->heap[h]; bits = tree[tree[n].Dad].Len + 1; if (bits > max_length) bits = max_length, overflow++; tree[n].Len = (ush)bits; /* We overwrite tree[n].Dad which is no longer needed */ if (n > max_code) continue; /* not a leaf node */ s->bl_count[bits]++; xbits = 0; if (n >= base) xbits = extra[n-base]; f = tree[n].Freq; s->opt_len += (ulg)f * (bits + xbits); if (stree) s->static_len += (ulg)f * (stree[n].Len + xbits); } if (overflow == 0) return; Trace((stderr,"\nbit length overflow\n")); /* This happens for example on obj2 and pic of the Calgary corpus */ /* Find the first bit length which could increase: */ do { bits = max_length-1; while (s->bl_count[bits] == 0) bits--; s->bl_count[bits]--; /* move one leaf down the tree */ s->bl_count[bits+1] += 2; /* move one overflow item as its brother */ s->bl_count[max_length]--; /* The brother of the overflow item also moves one step up, * but this does not affect bl_count[max_length] */ overflow -= 2; } while (overflow > 0); /* Now recompute all bit lengths, scanning in increasing frequency. * h is still equal to HEAP_SIZE. (It is simpler to reconstruct all * lengths instead of fixing only the wrong ones. This idea is taken * from 'ar' written by Haruhiko Okumura.) */ for (bits = max_length; bits != 0; bits--) { n = s->bl_count[bits]; while (n != 0) { m = s->heap[--h]; if (m > max_code) continue; if (tree[m].Len != (unsigned) bits) { Trace((stderr,"code %d bits %d->%d\n", m, tree[m].Len, bits)); s->opt_len += ((long)bits - (long)tree[m].Len) *(long)tree[m].Freq; tree[m].Len = (ush)bits; } n--; } } } /* =========================================================================== * Generate the codes for a given tree and bit counts (which need not be * optimal). * IN assertion: the array bl_count contains the bit length statistics for * the given tree and the field len is set for all tree elements. * OUT assertion: the field code is set for all tree elements of non * zero code length. */ local void gen_codes (tree, max_code, bl_count) ct_data *tree; /* the tree to decorate */ int max_code; /* largest code with non zero frequency */ ushf *bl_count; /* number of codes at each bit length */ { ush next_code[MAX_BITS+1]; /* next code value for each bit length */ ush code = 0; /* running code value */ int bits; /* bit index */ int n; /* code index */ /* The distribution counts are first used to generate the code values * without bit reversal. */ for (bits = 1; bits <= MAX_BITS; bits++) { next_code[bits] = code = (code + bl_count[bits-1]) << 1; } /* Check that the bit counts in bl_count are consistent. The last code * must be all ones. */ Assert (code + bl_count[MAX_BITS]-1 == (1<dyn_tree; ct_data *stree = desc->stat_desc->static_tree; int elems = desc->stat_desc->elems; int n, m; /* iterate over heap elements */ int max_code = -1; /* largest code with non zero frequency */ int node; /* new node being created */ /* Construct the initial heap, with least frequent element in * heap[SMALLEST]. The sons of heap[n] are heap[2*n] and heap[2*n+1]. * heap[0] is not used. */ s->heap_len = 0, s->heap_max = HEAP_SIZE; for (n = 0; n < elems; n++) { if (tree[n].Freq != 0) { s->heap[++(s->heap_len)] = max_code = n; s->depth[n] = 0; } else { tree[n].Len = 0; } } /* The pkzip format requires that at least one distance code exists, * and that at least one bit should be sent even if there is only one * possible code. So to avoid special checks later on we force at least * two codes of non zero frequency. */ while (s->heap_len < 2) { node = s->heap[++(s->heap_len)] = (max_code < 2 ? ++max_code : 0); tree[node].Freq = 1; s->depth[node] = 0; s->opt_len--; if (stree) s->static_len -= stree[node].Len; /* node is 0 or 1 so it does not have extra bits */ } desc->max_code = max_code; /* The elements heap[heap_len/2+1 .. heap_len] are leaves of the tree, * establish sub-heaps of increasing lengths: */ for (n = s->heap_len/2; n >= 1; n--) pqdownheap(s, tree, n); /* Construct the Huffman tree by repeatedly combining the least two * frequent nodes. */ node = elems; /* next internal node of the tree */ do { pqremove(s, tree, n); /* n = node of least frequency */ m = s->heap[SMALLEST]; /* m = node of next least frequency */ s->heap[--(s->heap_max)] = n; /* keep the nodes sorted by frequency */ s->heap[--(s->heap_max)] = m; /* Create a new node father of n and m */ tree[node].Freq = tree[n].Freq + tree[m].Freq; s->depth[node] = (uch) (MAX(s->depth[n], s->depth[m]) + 1); tree[n].Dad = tree[m].Dad = (ush)node; #ifdef DUMP_BL_TREE if (tree == s->bl_tree) { fprintf(stderr,"\nnode %d(%d), sons %d(%d) %d(%d)", node, tree[node].Freq, n, tree[n].Freq, m, tree[m].Freq); } #endif /* and insert the new node in the heap */ s->heap[SMALLEST] = node++; pqdownheap(s, tree, SMALLEST); } while (s->heap_len >= 2); s->heap[--(s->heap_max)] = s->heap[SMALLEST]; /* At this point, the fields freq and dad are set. We can now * generate the bit lengths. */ gen_bitlen(s, (tree_desc *)desc); /* The field len is now set, we can generate the bit codes */ gen_codes ((ct_data *)tree, max_code, s->bl_count); } /* =========================================================================== * Scan a literal or distance tree to determine the frequencies of the codes * in the bit length tree. */ local void scan_tree (s, tree, max_code) deflate_state *s; ct_data *tree; /* the tree to be scanned */ int max_code; /* and its largest code of non zero frequency */ { int n; /* iterates over all tree elements */ int prevlen = -1; /* last emitted length */ int curlen; /* length of current code */ int nextlen = tree[0].Len; /* length of next code */ int count = 0; /* repeat count of the current code */ int max_count = 7; /* max repeat count */ int min_count = 4; /* min repeat count */ if (nextlen == 0) max_count = 138, min_count = 3; tree[max_code+1].Len = (ush)0xffff; /* guard */ for (n = 0; n <= max_code; n++) { curlen = nextlen; nextlen = tree[n+1].Len; if (++count < max_count && curlen == nextlen) { continue; } else if (count < min_count) { s->bl_tree[curlen].Freq += count; } else if (curlen != 0) { if (curlen != prevlen) s->bl_tree[curlen].Freq++; s->bl_tree[REP_3_6].Freq++; } else if (count <= 10) { s->bl_tree[REPZ_3_10].Freq++; } else { s->bl_tree[REPZ_11_138].Freq++; } count = 0; prevlen = curlen; if (nextlen == 0) { max_count = 138, min_count = 3; } else if (curlen == nextlen) { max_count = 6, min_count = 3; } else { max_count = 7, min_count = 4; } } } /* =========================================================================== * Send a literal or distance tree in compressed form, using the codes in * bl_tree. */ local void send_tree (s, tree, max_code) deflate_state *s; ct_data *tree; /* the tree to be scanned */ int max_code; /* and its largest code of non zero frequency */ { int n; /* iterates over all tree elements */ int prevlen = -1; /* last emitted length */ int curlen; /* length of current code */ int nextlen = tree[0].Len; /* length of next code */ int count = 0; /* repeat count of the current code */ int max_count = 7; /* max repeat count */ int min_count = 4; /* min repeat count */ /* tree[max_code+1].Len = -1; */ /* guard already set */ if (nextlen == 0) max_count = 138, min_count = 3; for (n = 0; n <= max_code; n++) { curlen = nextlen; nextlen = tree[n+1].Len; if (++count < max_count && curlen == nextlen) { continue; } else if (count < min_count) { do { send_code(s, curlen, s->bl_tree); } while (--count != 0); } else if (curlen != 0) { if (curlen != prevlen) { send_code(s, curlen, s->bl_tree); count--; } Assert(count >= 3 && count <= 6, " 3_6?"); send_code(s, REP_3_6, s->bl_tree); send_bits(s, count-3, 2); } else if (count <= 10) { send_code(s, REPZ_3_10, s->bl_tree); send_bits(s, count-3, 3); } else { send_code(s, REPZ_11_138, s->bl_tree); send_bits(s, count-11, 7); } count = 0; prevlen = curlen; if (nextlen == 0) { max_count = 138, min_count = 3; } else if (curlen == nextlen) { max_count = 6, min_count = 3; } else { max_count = 7, min_count = 4; } } } /* =========================================================================== * Construct the Huffman tree for the bit lengths and return the index in * bl_order of the last bit length code to send. */ local int build_bl_tree(s) deflate_state *s; { int max_blindex; /* index of last bit length code of non zero freq */ /* Determine the bit length frequencies for literal and distance trees */ scan_tree(s, (ct_data *)s->dyn_ltree, s->l_desc.max_code); scan_tree(s, (ct_data *)s->dyn_dtree, s->d_desc.max_code); /* Build the bit length tree: */ build_tree(s, (tree_desc *)(&(s->bl_desc))); /* opt_len now includes the length of the tree representations, except * the lengths of the bit lengths codes and the 5+5+4 bits for the counts. */ /* Determine the number of bit length codes to send. The pkzip format * requires that at least 4 bit length codes be sent. (appnote.txt says * 3 but the actual value used is 4.) */ for (max_blindex = BL_CODES-1; max_blindex >= 3; max_blindex--) { if (s->bl_tree[bl_order[max_blindex]].Len != 0) break; } /* Update opt_len to include the bit length tree and counts */ s->opt_len += 3*(max_blindex+1) + 5+5+4; Tracev((stderr, "\ndyn trees: dyn %ld, stat %ld", s->opt_len, s->static_len)); return max_blindex; } /* =========================================================================== * Send the header for a block using dynamic Huffman trees: the counts, the * lengths of the bit length codes, the literal tree and the distance tree. * IN assertion: lcodes >= 257, dcodes >= 1, blcodes >= 4. */ local void send_all_trees(s, lcodes, dcodes, blcodes) deflate_state *s; int lcodes, dcodes, blcodes; /* number of codes for each tree */ { int rank; /* index in bl_order */ Assert (lcodes >= 257 && dcodes >= 1 && blcodes >= 4, "not enough codes"); Assert (lcodes <= L_CODES && dcodes <= D_CODES && blcodes <= BL_CODES, "too many codes"); Tracev((stderr, "\nbl counts: ")); send_bits(s, lcodes-257, 5); /* not +255 as stated in appnote.txt */ send_bits(s, dcodes-1, 5); send_bits(s, blcodes-4, 4); /* not -3 as stated in appnote.txt */ for (rank = 0; rank < blcodes; rank++) { Tracev((stderr, "\nbl code %2d ", bl_order[rank])); send_bits(s, s->bl_tree[bl_order[rank]].Len, 3); } Tracev((stderr, "\nbl tree: sent %ld", s->bits_sent)); send_tree(s, (ct_data *)s->dyn_ltree, lcodes-1); /* literal tree */ Tracev((stderr, "\nlit tree: sent %ld", s->bits_sent)); send_tree(s, (ct_data *)s->dyn_dtree, dcodes-1); /* distance tree */ Tracev((stderr, "\ndist tree: sent %ld", s->bits_sent)); } /* =========================================================================== * Send a stored block */ local void ct_stored_block(s, buf, stored_len, eof) deflate_state *s; charf *buf; /* input block */ ulg stored_len; /* length of input block */ int eof; /* true if this is the last block for a file */ { send_bits(s, (STORED_BLOCK<<1)+eof, 3); /* send block type */ s->compressed_len = (s->compressed_len + 3 + 7) & ~7L; s->compressed_len += (stored_len + 4) << 3; copy_block(s, buf, (unsigned)stored_len, 1); /* with header */ } /* Send just the `stored block' type code without any length bytes or data. */ local void ct_stored_type_only(s) deflate_state *s; { send_bits(s, (STORED_BLOCK << 1), 3); bi_windup(s); s->compressed_len = (s->compressed_len + 3) & ~7L; } /* =========================================================================== * Send one empty static block to give enough lookahead for inflate. * This takes 10 bits, of which 7 may remain in the bit buffer. * The current inflate code requires 9 bits of lookahead. If the EOB * code for the previous block was coded on 5 bits or less, inflate * may have only 5+3 bits of lookahead to decode this EOB. * (There are no problems if the previous block is stored or fixed.) */ local void ct_align(s) deflate_state *s; { send_bits(s, STATIC_TREES<<1, 3); send_code(s, END_BLOCK, static_ltree); s->compressed_len += 10L; /* 3 for block type, 7 for EOB */ bi_flush(s); /* Of the 10 bits for the empty block, we have already sent * (10 - bi_valid) bits. The lookahead for the EOB of the previous * block was thus its length plus what we have just sent. */ if (s->last_eob_len + 10 - s->bi_valid < 9) { send_bits(s, STATIC_TREES<<1, 3); send_code(s, END_BLOCK, static_ltree); s->compressed_len += 10L; bi_flush(s); } s->last_eob_len = 7; } /* =========================================================================== * Determine the best encoding for the current block: dynamic trees, static * trees or store, and output the encoded block to the zip file. This function * returns the total compressed length for the file so far. */ local ulg ct_flush_block(s, buf, stored_len, flush) deflate_state *s; charf *buf; /* input block, or NULL if too old */ ulg stored_len; /* length of input block */ int flush; /* Z_FINISH if this is the last block for a file */ { ulg opt_lenb, static_lenb; /* opt_len and static_len in bytes */ int max_blindex; /* index of last bit length code of non zero freq */ int eof = flush == Z_FINISH; ++s->blocks_in_packet; /* Check if the file is ascii or binary */ if (s->data_type == UNKNOWN) set_data_type(s); /* Construct the literal and distance trees */ build_tree(s, (tree_desc *)(&(s->l_desc))); Tracev((stderr, "\nlit data: dyn %ld, stat %ld", s->opt_len, s->static_len)); build_tree(s, (tree_desc *)(&(s->d_desc))); Tracev((stderr, "\ndist data: dyn %ld, stat %ld", s->opt_len, s->static_len)); /* At this point, opt_len and static_len are the total bit lengths of * the compressed block data, excluding the tree representations. */ /* Build the bit length tree for the above two trees, and get the index * in bl_order of the last bit length code to send. */ max_blindex = build_bl_tree(s); /* Determine the best encoding. Compute first the block length in bytes */ opt_lenb = (s->opt_len+3+7)>>3; static_lenb = (s->static_len+3+7)>>3; Tracev((stderr, "\nopt %lu(%lu) stat %lu(%lu) stored %lu lit %u ", opt_lenb, s->opt_len, static_lenb, s->static_len, stored_len, s->last_lit)); if (static_lenb <= opt_lenb) opt_lenb = static_lenb; /* If compression failed and this is the first and last block, * and if the .zip file can be seeked (to rewrite the local header), * the whole file is transformed into a stored file: */ #ifdef STORED_FILE_OK # ifdef FORCE_STORED_FILE if (eof && compressed_len == 0L) /* force stored file */ # else if (stored_len <= opt_lenb && eof && s->compressed_len==0L && seekable()) # endif { /* Since LIT_BUFSIZE <= 2*WSIZE, the input data must be there: */ if (buf == (charf*)0) error ("block vanished"); copy_block(buf, (unsigned)stored_len, 0); /* without header */ s->compressed_len = stored_len << 3; s->method = STORED; } else #endif /* STORED_FILE_OK */ /* For Z_PACKET_FLUSH, if we don't achieve the required minimum * compression, and this block contains all the data since the last * time we used Z_PACKET_FLUSH, then just omit this block completely * from the output. */ if (flush == Z_PACKET_FLUSH && s->blocks_in_packet == 1 && opt_lenb > stored_len - s->minCompr) { s->blocks_in_packet = 0; /* output nothing */ } else #ifdef FORCE_STORED if (buf != (char*)0) /* force stored block */ #else if (stored_len+4 <= opt_lenb && buf != (char*)0) /* 4: two words for the lengths */ #endif { /* The test buf != NULL is only necessary if LIT_BUFSIZE > WSIZE. * Otherwise we can't have processed more than WSIZE input bytes since * the last block flush, because compression would have been * successful. If LIT_BUFSIZE <= WSIZE, it is never too late to * transform a block into a stored block. */ ct_stored_block(s, buf, stored_len, eof); } else #ifdef FORCE_STATIC if (static_lenb >= 0) /* force static trees */ #else if (static_lenb == opt_lenb) #endif { send_bits(s, (STATIC_TREES<<1)+eof, 3); compress_block(s, (ct_data *)static_ltree, (ct_data *)static_dtree); s->compressed_len += 3 + s->static_len; } else { send_bits(s, (DYN_TREES<<1)+eof, 3); send_all_trees(s, s->l_desc.max_code+1, s->d_desc.max_code+1, max_blindex+1); compress_block(s, (ct_data *)s->dyn_ltree, (ct_data *)s->dyn_dtree); s->compressed_len += 3 + s->opt_len; } Assert (s->compressed_len == s->bits_sent, "bad compressed size"); init_block(s); if (eof) { bi_windup(s); s->compressed_len += 7; /* align on byte boundary */ } Tracev((stderr,"\ncomprlen %lu(%lu) ", s->compressed_len>>3, s->compressed_len-7*eof)); return s->compressed_len >> 3; } /* =========================================================================== * Save the match info and tally the frequency counts. Return true if * the current block must be flushed. */ local int ct_tally (s, dist, lc) deflate_state *s; int dist; /* distance of matched string */ int lc; /* match length-MIN_MATCH or unmatched char (if dist==0) */ { s->d_buf[s->last_lit] = (ush)dist; s->l_buf[s->last_lit++] = (uch)lc; if (dist == 0) { /* lc is the unmatched char */ s->dyn_ltree[lc].Freq++; } else { s->matches++; /* Here, lc is the match length - MIN_MATCH */ dist--; /* dist = match distance - 1 */ Assert((ush)dist < (ush)MAX_DIST(s) && (ush)lc <= (ush)(MAX_MATCH-MIN_MATCH) && (ush)d_code(dist) < (ush)D_CODES, "ct_tally: bad match"); s->dyn_ltree[length_code[lc]+LITERALS+1].Freq++; s->dyn_dtree[d_code(dist)].Freq++; } /* Try to guess if it is profitable to stop the current block here */ if (s->level > 2 && (s->last_lit & 0xfff) == 0) { /* Compute an upper bound for the compressed length */ ulg out_length = (ulg)s->last_lit*8L; ulg in_length = (ulg)s->strstart - s->block_start; int dcode; for (dcode = 0; dcode < D_CODES; dcode++) { out_length += (ulg)s->dyn_dtree[dcode].Freq * (5L+extra_dbits[dcode]); } out_length >>= 3; Tracev((stderr,"\nlast_lit %u, in %ld, out ~%ld(%ld%%) ", s->last_lit, in_length, out_length, 100L - out_length*100L/in_length)); if (s->matches < s->last_lit/2 && out_length < in_length/2) return 1; } return (s->last_lit == s->lit_bufsize-1); /* We avoid equality with lit_bufsize because of wraparound at 64K * on 16 bit machines and because stored blocks are restricted to * 64K-1 bytes. */ } /* =========================================================================== * Send the block data compressed using the given Huffman trees */ local void compress_block(s, ltree, dtree) deflate_state *s; ct_data *ltree; /* literal tree */ ct_data *dtree; /* distance tree */ { unsigned dist; /* distance of matched string */ int lc; /* match length or unmatched char (if dist == 0) */ unsigned lx = 0; /* running index in l_buf */ unsigned code; /* the code to send */ int extra; /* number of extra bits to send */ if (s->last_lit != 0) do { dist = s->d_buf[lx]; lc = s->l_buf[lx++]; if (dist == 0) { send_code(s, lc, ltree); /* send a literal byte */ Tracecv(isgraph(lc), (stderr," '%c' ", lc)); } else { /* Here, lc is the match length - MIN_MATCH */ code = length_code[lc]; send_code(s, code+LITERALS+1, ltree); /* send the length code */ extra = extra_lbits[code]; if (extra != 0) { lc -= base_length[code]; send_bits(s, lc, extra); /* send the extra length bits */ } dist--; /* dist is now the match distance - 1 */ code = d_code(dist); Assert (code < D_CODES, "bad d_code"); send_code(s, code, dtree); /* send the distance code */ extra = extra_dbits[code]; if (extra != 0) { dist -= base_dist[code]; send_bits(s, dist, extra); /* send the extra distance bits */ } } /* literal or match pair ? */ /* Check that the overlay between pending_buf and d_buf+l_buf is ok: */ Assert(s->pending < s->lit_bufsize + 2*lx, "pendingBuf overflow"); } while (lx < s->last_lit); send_code(s, END_BLOCK, ltree); s->last_eob_len = ltree[END_BLOCK].Len; } /* =========================================================================== * Set the data type to ASCII or BINARY, using a crude approximation: * binary if more than 20% of the bytes are <= 6 or >= 128, ascii otherwise. * IN assertion: the fields freq of dyn_ltree are set and the total of all * frequencies does not exceed 64K (to fit in an int on 16 bit machines). */ local void set_data_type(s) deflate_state *s; { int n = 0; unsigned ascii_freq = 0; unsigned bin_freq = 0; while (n < 7) bin_freq += s->dyn_ltree[n++].Freq; while (n < 128) ascii_freq += s->dyn_ltree[n++].Freq; while (n < LITERALS) bin_freq += s->dyn_ltree[n++].Freq; s->data_type = (Byte)(bin_freq > (ascii_freq >> 2) ? BINARY : ASCII); } /* =========================================================================== * Reverse the first len bits of a code, using straightforward code (a faster * method would use a table) * IN assertion: 1 <= len <= 15 */ local unsigned bi_reverse(code, len) unsigned code; /* the value to invert */ int len; /* its bit length */ { register unsigned res = 0; do { res |= code & 1; code >>= 1, res <<= 1; } while (--len > 0); return res >> 1; } /* =========================================================================== * Flush the bit buffer, keeping at most 7 bits in it. */ local void bi_flush(s) deflate_state *s; { if (s->bi_valid == 16) { put_short(s, s->bi_buf); s->bi_buf = 0; s->bi_valid = 0; } else if (s->bi_valid >= 8) { put_byte(s, (Byte)s->bi_buf); s->bi_buf >>= 8; s->bi_valid -= 8; } } /* =========================================================================== * Flush the bit buffer and align the output on a byte boundary */ local void bi_windup(s) deflate_state *s; { if (s->bi_valid > 8) { put_short(s, s->bi_buf); } else if (s->bi_valid > 0) { put_byte(s, (Byte)s->bi_buf); } s->bi_buf = 0; s->bi_valid = 0; #ifdef DEBUG_ZLIB s->bits_sent = (s->bits_sent+7) & ~7; #endif } /* =========================================================================== * Copy a stored block, storing first the length and its * one's complement if requested. */ local void copy_block(s, buf, len, header) deflate_state *s; charf *buf; /* the input data */ unsigned len; /* its length */ int header; /* true if block header must be written */ { bi_windup(s); /* align on byte boundary */ s->last_eob_len = 8; /* enough lookahead for inflate */ if (header) { put_short(s, (ush)len); put_short(s, (ush)~len); #ifdef DEBUG_ZLIB s->bits_sent += 2*16; #endif } #ifdef DEBUG_ZLIB s->bits_sent += (ulg)len<<3; #endif while (len--) { put_byte(s, *buf++); } } #endif /* NO_DEFLATE */ /*+++++*/ /* infblock.h -- header to use infblock.c * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* WARNING: this file should *not* be used by applications. It is part of the implementation of the compression library and is subject to change. Applications should only use zlib.h. */ struct inflate_blocks_state; typedef struct inflate_blocks_state FAR inflate_blocks_statef; local inflate_blocks_statef * inflate_blocks_new OF(( z_stream *z, check_func c, /* check function */ uInt w)); /* window size */ local int inflate_blocks OF(( inflate_blocks_statef *, z_stream *, int)); /* initial return code */ local void inflate_blocks_reset OF(( inflate_blocks_statef *, z_stream *, uLongf *)); /* check value on output */ local int inflate_blocks_free OF(( inflate_blocks_statef *, z_stream *, uLongf *)); /* check value on output */ local int inflate_addhistory OF(( inflate_blocks_statef *, z_stream *)); local int inflate_packet_flush OF(( inflate_blocks_statef *)); /*+++++*/ /* inftrees.h -- header to use inftrees.c * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* WARNING: this file should *not* be used by applications. It is part of the implementation of the compression library and is subject to change. Applications should only use zlib.h. */ /* Huffman code lookup table entry--this entry is four bytes for machines that have 16-bit pointers (e.g. PC's in the small or medium model). */ typedef struct inflate_huft_s FAR inflate_huft; struct inflate_huft_s { union { struct { Byte Exop; /* number of extra bits or operation */ Byte Bits; /* number of bits in this code or subcode */ } what; uInt Nalloc; /* number of these allocated here */ Bytef *pad; /* pad structure to a power of 2 (4 bytes for */ } word; /* 16-bit, 8 bytes for 32-bit machines) */ union { uInt Base; /* literal, length base, or distance base */ inflate_huft *Next; /* pointer to next level of table */ } more; }; #ifdef DEBUG_ZLIB local uInt inflate_hufts; #endif local int inflate_trees_bits OF(( uIntf *, /* 19 code lengths */ uIntf *, /* bits tree desired/actual depth */ inflate_huft * FAR *, /* bits tree result */ z_stream *)); /* for zalloc, zfree functions */ local int inflate_trees_dynamic OF(( uInt, /* number of literal/length codes */ uInt, /* number of distance codes */ uIntf *, /* that many (total) code lengths */ uIntf *, /* literal desired/actual bit depth */ uIntf *, /* distance desired/actual bit depth */ inflate_huft * FAR *, /* literal/length tree result */ inflate_huft * FAR *, /* distance tree result */ z_stream *)); /* for zalloc, zfree functions */ local int inflate_trees_fixed OF(( uIntf *, /* literal desired/actual bit depth */ uIntf *, /* distance desired/actual bit depth */ inflate_huft * FAR *, /* literal/length tree result */ inflate_huft * FAR *)); /* distance tree result */ local int inflate_trees_free OF(( inflate_huft *, /* tables to free */ z_stream *)); /* for zfree function */ /*+++++*/ /* infcodes.h -- header to use infcodes.c * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* WARNING: this file should *not* be used by applications. It is part of the implementation of the compression library and is subject to change. Applications should only use zlib.h. */ struct inflate_codes_state; typedef struct inflate_codes_state FAR inflate_codes_statef; local inflate_codes_statef *inflate_codes_new OF(( uInt, uInt, inflate_huft *, inflate_huft *, z_stream *)); local int inflate_codes OF(( inflate_blocks_statef *, z_stream *, int)); local void inflate_codes_free OF(( inflate_codes_statef *, z_stream *)); /*+++++*/ /* inflate.c -- zlib interface to inflate modules * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* inflate private state */ struct internal_state { /* mode */ enum { METHOD, /* waiting for method byte */ FLAG, /* waiting for flag byte */ BLOCKS, /* decompressing blocks */ CHECK4, /* four check bytes to go */ CHECK3, /* three check bytes to go */ CHECK2, /* two check bytes to go */ CHECK1, /* one check byte to go */ DONE, /* finished check, done */ BAD} /* got an error--stay here */ mode; /* current inflate mode */ /* mode dependent information */ union { uInt method; /* if FLAGS, method byte */ struct { uLong was; /* computed check value */ uLong need; /* stream check value */ } check; /* if CHECK, check values to compare */ uInt marker; /* if BAD, inflateSync's marker bytes count */ } sub; /* submode */ /* mode independent information */ int nowrap; /* flag for no wrapper */ uInt wbits; /* log2(window size) (8..15, defaults to 15) */ inflate_blocks_statef *blocks; /* current inflate_blocks state */ }; int inflateReset(z) z_stream *z; { uLong c; if (z == Z_NULL || z->state == Z_NULL) return Z_STREAM_ERROR; z->total_in = z->total_out = 0; z->msg = Z_NULL; z->state->mode = z->state->nowrap ? BLOCKS : METHOD; inflate_blocks_reset(z->state->blocks, z, &c); Trace((stderr, "inflate: reset\n")); return Z_OK; } int inflateEnd(z) z_stream *z; { uLong c; if (z == Z_NULL || z->state == Z_NULL || z->zfree == Z_NULL) return Z_STREAM_ERROR; if (z->state->blocks != Z_NULL) inflate_blocks_free(z->state->blocks, z, &c); ZFREE(z, z->state, sizeof(struct internal_state)); z->state = Z_NULL; Trace((stderr, "inflate: end\n")); return Z_OK; } int inflateInit2(z, w) z_stream *z; int w; { /* initialize state */ if (z == Z_NULL) return Z_STREAM_ERROR; /* if (z->zalloc == Z_NULL) z->zalloc = zcalloc; */ /* if (z->zfree == Z_NULL) z->zfree = zcfree; */ if ((z->state = (struct internal_state FAR *) ZALLOC(z,1,sizeof(struct internal_state))) == Z_NULL) return Z_MEM_ERROR; z->state->blocks = Z_NULL; /* handle undocumented nowrap option (no zlib header or check) */ z->state->nowrap = 0; if (w < 0) { w = - w; z->state->nowrap = 1; } /* set window size */ if (w < 8 || w > 15) { inflateEnd(z); return Z_STREAM_ERROR; } z->state->wbits = (uInt)w; /* create inflate_blocks state */ if ((z->state->blocks = inflate_blocks_new(z, z->state->nowrap ? Z_NULL : adler32, 1 << w)) == Z_NULL) { inflateEnd(z); return Z_MEM_ERROR; } Trace((stderr, "inflate: allocated\n")); /* reset state */ inflateReset(z); return Z_OK; } int inflateInit(z) z_stream *z; { return inflateInit2(z, DEF_WBITS); } #define NEEDBYTE {if(z->avail_in==0)goto empty;r=Z_OK;} #define NEXTBYTE (z->avail_in--,z->total_in++,*z->next_in++) int inflate(z, f) z_stream *z; int f; { int r; uInt b; if (z == Z_NULL || z->next_in == Z_NULL) return Z_STREAM_ERROR; r = Z_BUF_ERROR; while (1) switch (z->state->mode) { case METHOD: NEEDBYTE if (((z->state->sub.method = NEXTBYTE) & 0xf) != DEFLATED) { z->state->mode = BAD; z->msg = "unknown compression method"; z->state->sub.marker = 5; /* can't try inflateSync */ break; } if ((z->state->sub.method >> 4) + 8 > z->state->wbits) { z->state->mode = BAD; z->msg = "invalid window size"; z->state->sub.marker = 5; /* can't try inflateSync */ break; } z->state->mode = FLAG; case FLAG: NEEDBYTE if ((b = NEXTBYTE) & 0x20) { z->state->mode = BAD; z->msg = "invalid reserved bit"; z->state->sub.marker = 5; /* can't try inflateSync */ break; } if (((z->state->sub.method << 8) + b) % 31) { z->state->mode = BAD; z->msg = "incorrect header check"; z->state->sub.marker = 5; /* can't try inflateSync */ break; } Trace((stderr, "inflate: zlib header ok\n")); z->state->mode = BLOCKS; case BLOCKS: r = inflate_blocks(z->state->blocks, z, r); if (f == Z_PACKET_FLUSH && z->avail_in == 0 && z->avail_out != 0) r = inflate_packet_flush(z->state->blocks); if (r == Z_DATA_ERROR) { z->state->mode = BAD; z->state->sub.marker = 0; /* can try inflateSync */ break; } if (r != Z_STREAM_END) return r; r = Z_OK; inflate_blocks_reset(z->state->blocks, z, &z->state->sub.check.was); if (z->state->nowrap) { z->state->mode = DONE; break; } z->state->mode = CHECK4; case CHECK4: NEEDBYTE z->state->sub.check.need = (uLong)NEXTBYTE << 24; z->state->mode = CHECK3; case CHECK3: NEEDBYTE z->state->sub.check.need += (uLong)NEXTBYTE << 16; z->state->mode = CHECK2; case CHECK2: NEEDBYTE z->state->sub.check.need += (uLong)NEXTBYTE << 8; z->state->mode = CHECK1; case CHECK1: NEEDBYTE z->state->sub.check.need += (uLong)NEXTBYTE; if (z->state->sub.check.was != z->state->sub.check.need) { z->state->mode = BAD; z->msg = "incorrect data check"; z->state->sub.marker = 5; /* can't try inflateSync */ break; } Trace((stderr, "inflate: zlib check ok\n")); z->state->mode = DONE; case DONE: return Z_STREAM_END; case BAD: return Z_DATA_ERROR; default: return Z_STREAM_ERROR; } empty: if (f != Z_PACKET_FLUSH) return r; z->state->mode = BAD; z->state->sub.marker = 0; /* can try inflateSync */ return Z_DATA_ERROR; } /* * This subroutine adds the data at next_in/avail_in to the output history * without performing any output. The output buffer must be "caught up"; * i.e. no pending output (hence s->read equals s->write), and the state must * be BLOCKS (i.e. we should be willing to see the start of a series of * BLOCKS). On exit, the output will also be caught up, and the checksum * will have been updated if need be. */ int inflateIncomp(z) z_stream *z; { if (z->state->mode != BLOCKS) return Z_DATA_ERROR; return inflate_addhistory(z->state->blocks, z); } int inflateSync(z) z_stream *z; { uInt n; /* number of bytes to look at */ Bytef *p; /* pointer to bytes */ uInt m; /* number of marker bytes found in a row */ uLong r, w; /* temporaries to save total_in and total_out */ /* set up */ if (z == Z_NULL || z->state == Z_NULL) return Z_STREAM_ERROR; if (z->state->mode != BAD) { z->state->mode = BAD; z->state->sub.marker = 0; } if ((n = z->avail_in) == 0) return Z_BUF_ERROR; p = z->next_in; m = z->state->sub.marker; /* search */ while (n && m < 4) { if (*p == (Byte)(m < 2 ? 0 : 0xff)) m++; else if (*p) m = 0; else m = 4 - m; p++, n--; } /* restore */ z->total_in += p - z->next_in; z->next_in = p; z->avail_in = n; z->state->sub.marker = m; /* return no joy or set up to restart on a new block */ if (m != 4) return Z_DATA_ERROR; r = z->total_in; w = z->total_out; inflateReset(z); z->total_in = r; z->total_out = w; z->state->mode = BLOCKS; return Z_OK; } #undef NEEDBYTE #undef NEXTBYTE /*+++++*/ /* infutil.h -- types and macros common to blocks and codes * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* WARNING: this file should *not* be used by applications. It is part of the implementation of the compression library and is subject to change. Applications should only use zlib.h. */ /* inflate blocks semi-private state */ struct inflate_blocks_state { /* mode */ enum { TYPE, /* get type bits (3, including end bit) */ LENS, /* get lengths for stored */ STORED, /* processing stored block */ TABLE, /* get table lengths */ BTREE, /* get bit lengths tree for a dynamic block */ DTREE, /* get length, distance trees for a dynamic block */ CODES, /* processing fixed or dynamic block */ DRY, /* output remaining window bytes */ DONEB, /* finished last block, done */ BADB} /* got a data error--stuck here */ mode; /* current inflate_block mode */ /* mode dependent information */ union { uInt left; /* if STORED, bytes left to copy */ struct { uInt table; /* table lengths (14 bits) */ uInt index; /* index into blens (or border) */ uIntf *blens; /* bit lengths of codes */ uInt bb; /* bit length tree depth */ inflate_huft *tb; /* bit length decoding tree */ int nblens; /* # elements allocated at blens */ } trees; /* if DTREE, decoding info for trees */ struct { inflate_huft *tl, *td; /* trees to free */ inflate_codes_statef *codes; } decode; /* if CODES, current state */ } sub; /* submode */ uInt last; /* true if this block is the last block */ /* mode independent information */ uInt bitk; /* bits in bit buffer */ uLong bitb; /* bit buffer */ Bytef *window; /* sliding window */ Bytef *end; /* one byte after sliding window */ Bytef *read; /* window read pointer */ Bytef *write; /* window write pointer */ check_func checkfn; /* check function */ uLong check; /* check on output */ }; /* defines for inflate input/output */ /* update pointers and return */ #define UPDBITS {s->bitb=b;s->bitk=k;} #define UPDIN {z->avail_in=n;z->total_in+=p-z->next_in;z->next_in=p;} #define UPDOUT {s->write=q;} #define UPDATE {UPDBITS UPDIN UPDOUT} #define LEAVE {UPDATE return inflate_flush(s,z,r);} /* get bytes and bits */ #define LOADIN {p=z->next_in;n=z->avail_in;b=s->bitb;k=s->bitk;} #define NEEDBYTE {if(n)r=Z_OK;else LEAVE} #define NEXTBYTE (n--,*p++) #define NEEDBITS(j) {while(k<(j)){NEEDBYTE;b|=((uLong)NEXTBYTE)<>=(j);k-=(j);} /* output bytes */ #define WAVAIL (qread?s->read-q-1:s->end-q) #define LOADOUT {q=s->write;m=WAVAIL;} #define WRAP {if(q==s->end&&s->read!=s->window){q=s->window;m=WAVAIL;}} #define FLUSH {UPDOUT r=inflate_flush(s,z,r); LOADOUT} #define NEEDOUT {if(m==0){WRAP if(m==0){FLUSH WRAP if(m==0) LEAVE}}r=Z_OK;} #define OUTBYTE(a) {*q++=(Byte)(a);m--;} /* load local pointers */ #define LOAD {LOADIN LOADOUT} /* And'ing with mask[n] masks the lower n bits */ local const uInt inflate_mask[] = { 0x0000, 0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff, 0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff }; /* copy as much as possible from the sliding window to the output area */ local int inflate_flush OF(( inflate_blocks_statef *, z_stream *, int)); /*+++++*/ /* inffast.h -- header to use inffast.c * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* WARNING: this file should *not* be used by applications. It is part of the implementation of the compression library and is subject to change. Applications should only use zlib.h. */ local int inflate_fast OF(( uInt, uInt, inflate_huft *, inflate_huft *, inflate_blocks_statef *, z_stream *)); /*+++++*/ /* infblock.c -- interpret and process block types to last block * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* Table for deflate from PKZIP's appnote.txt. */ local uInt border[] = { /* Order of the bit length code lengths */ 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15}; /* Notes beyond the 1.93a appnote.txt: 1. Distance pointers never point before the beginning of the output stream. 2. Distance pointers can point back across blocks, up to 32k away. 3. There is an implied maximum of 7 bits for the bit length table and 15 bits for the actual data. 4. If only one code exists, then it is encoded using one bit. (Zero would be more efficient, but perhaps a little confusing.) If two codes exist, they are coded using one bit each (0 and 1). 5. There is no way of sending zero distance codes--a dummy must be sent if there are none. (History: a pre 2.0 version of PKZIP would store blocks with no distance codes, but this was discovered to be too harsh a criterion.) Valid only for 1.93a. 2.04c does allow zero distance codes, which is sent as one code of zero bits in length. 6. There are up to 286 literal/length codes. Code 256 represents the end-of-block. Note however that the static length tree defines 288 codes just to fill out the Huffman codes. Codes 286 and 287 cannot be used though, since there is no length base or extra bits defined for them. Similarily, there are up to 30 distance codes. However, static trees define 32 codes (all 5 bits) to fill out the Huffman codes, but the last two had better not show up in the data. 7. Unzip can check dynamic Huffman blocks for complete code sets. The exception is that a single code would not be complete (see #4). 8. The five bits following the block type is really the number of literal codes sent minus 257. 9. Length codes 8,16,16 are interpreted as 13 length codes of 8 bits (1+6+6). Therefore, to output three times the length, you output three codes (1+1+1), whereas to output four times the same length, you only need two codes (1+3). Hmm. 10. In the tree reconstruction algorithm, Code = Code + Increment only if BitLength(i) is not zero. (Pretty obvious.) 11. Correction: 4 Bits: # of Bit Length codes - 4 (4 - 19) 12. Note: length code 284 can represent 227-258, but length code 285 really is 258. The last length deserves its own, short code since it gets used a lot in very redundant files. The length 258 is special since 258 - 3 (the min match length) is 255. 13. The literal/length and distance code bit lengths are read as a single stream of lengths. It is possible (and advantageous) for a repeat code (16, 17, or 18) to go across the boundary between the two sets of lengths. */ local void inflate_blocks_reset(s, z, c) inflate_blocks_statef *s; z_stream *z; uLongf *c; { if (s->checkfn != Z_NULL) *c = s->check; if (s->mode == BTREE || s->mode == DTREE) ZFREE(z, s->sub.trees.blens, s->sub.trees.nblens * sizeof(uInt)); if (s->mode == CODES) { inflate_codes_free(s->sub.decode.codes, z); inflate_trees_free(s->sub.decode.td, z); inflate_trees_free(s->sub.decode.tl, z); } s->mode = TYPE; s->bitk = 0; s->bitb = 0; s->read = s->write = s->window; if (s->checkfn != Z_NULL) s->check = (*s->checkfn)(0L, Z_NULL, 0); Trace((stderr, "inflate: blocks reset\n")); } local inflate_blocks_statef *inflate_blocks_new(z, c, w) z_stream *z; check_func c; uInt w; { inflate_blocks_statef *s; if ((s = (inflate_blocks_statef *)ZALLOC (z,1,sizeof(struct inflate_blocks_state))) == Z_NULL) return s; if ((s->window = (Bytef *)ZALLOC(z, 1, w)) == Z_NULL) { ZFREE(z, s, sizeof(struct inflate_blocks_state)); return Z_NULL; } s->end = s->window + w; s->checkfn = c; s->mode = TYPE; Trace((stderr, "inflate: blocks allocated\n")); inflate_blocks_reset(s, z, &s->check); return s; } local int inflate_blocks(s, z, r) inflate_blocks_statef *s; z_stream *z; int r; { uInt t; /* temporary storage */ uLong b; /* bit buffer */ uInt k; /* bits in bit buffer */ Bytef *p; /* input data pointer */ uInt n; /* bytes available there */ Bytef *q; /* output window write pointer */ uInt m; /* bytes to end of window or read pointer */ /* copy input/output information to locals (UPDATE macro restores) */ LOAD /* process input based on current state */ while (1) switch (s->mode) { case TYPE: NEEDBITS(3) t = (uInt)b & 7; s->last = t & 1; switch (t >> 1) { case 0: /* stored */ Trace((stderr, "inflate: stored block%s\n", s->last ? " (last)" : "")); DUMPBITS(3) t = k & 7; /* go to byte boundary */ DUMPBITS(t) s->mode = LENS; /* get length of stored block */ break; case 1: /* fixed */ Trace((stderr, "inflate: fixed codes block%s\n", s->last ? " (last)" : "")); { uInt bl, bd; inflate_huft *tl, *td; inflate_trees_fixed(&bl, &bd, &tl, &td); s->sub.decode.codes = inflate_codes_new(bl, bd, tl, td, z); if (s->sub.decode.codes == Z_NULL) { r = Z_MEM_ERROR; LEAVE } s->sub.decode.tl = Z_NULL; /* don't try to free these */ s->sub.decode.td = Z_NULL; } DUMPBITS(3) s->mode = CODES; break; case 2: /* dynamic */ Trace((stderr, "inflate: dynamic codes block%s\n", s->last ? " (last)" : "")); DUMPBITS(3) s->mode = TABLE; break; case 3: /* illegal */ DUMPBITS(3) s->mode = BADB; z->msg = "invalid block type"; r = Z_DATA_ERROR; LEAVE } break; case LENS: NEEDBITS(32) if (((~b) >> 16) != (b & 0xffff)) { s->mode = BADB; z->msg = "invalid stored block lengths"; r = Z_DATA_ERROR; LEAVE } s->sub.left = (uInt)b & 0xffff; b = k = 0; /* dump bits */ Tracev((stderr, "inflate: stored length %u\n", s->sub.left)); s->mode = s->sub.left ? STORED : TYPE; break; case STORED: if (n == 0) LEAVE NEEDOUT t = s->sub.left; if (t > n) t = n; if (t > m) t = m; zmemcpy(q, p, t); p += t; n -= t; q += t; m -= t; if ((s->sub.left -= t) != 0) break; Tracev((stderr, "inflate: stored end, %lu total out\n", z->total_out + (q >= s->read ? q - s->read : (s->end - s->read) + (q - s->window)))); s->mode = s->last ? DRY : TYPE; break; case TABLE: NEEDBITS(14) s->sub.trees.table = t = (uInt)b & 0x3fff; #ifndef PKZIP_BUG_WORKAROUND if ((t & 0x1f) > 29 || ((t >> 5) & 0x1f) > 29) { s->mode = BADB; z->msg = "too many length or distance symbols"; r = Z_DATA_ERROR; LEAVE } #endif t = 258 + (t & 0x1f) + ((t >> 5) & 0x1f); if (t < 19) t = 19; if ((s->sub.trees.blens = (uIntf*)ZALLOC(z, t, sizeof(uInt))) == Z_NULL) { r = Z_MEM_ERROR; LEAVE } s->sub.trees.nblens = t; DUMPBITS(14) s->sub.trees.index = 0; Tracev((stderr, "inflate: table sizes ok\n")); s->mode = BTREE; case BTREE: while (s->sub.trees.index < 4 + (s->sub.trees.table >> 10)) { NEEDBITS(3) s->sub.trees.blens[border[s->sub.trees.index++]] = (uInt)b & 7; DUMPBITS(3) } while (s->sub.trees.index < 19) s->sub.trees.blens[border[s->sub.trees.index++]] = 0; s->sub.trees.bb = 7; t = inflate_trees_bits(s->sub.trees.blens, &s->sub.trees.bb, &s->sub.trees.tb, z); if (t != Z_OK) { r = t; if (r == Z_DATA_ERROR) { ZFREE(z, s->sub.trees.blens, s->sub.trees.nblens * sizeof(uInt)); s->mode = BADB; } LEAVE } s->sub.trees.index = 0; Tracev((stderr, "inflate: bits tree ok\n")); s->mode = DTREE; case DTREE: while (t = s->sub.trees.table, s->sub.trees.index < 258 + (t & 0x1f) + ((t >> 5) & 0x1f)) { inflate_huft *h; uInt i, j, c; t = s->sub.trees.bb; NEEDBITS(t) h = s->sub.trees.tb + ((uInt)b & inflate_mask[t]); t = h->word.what.Bits; c = h->more.Base; if (c < 16) { DUMPBITS(t) s->sub.trees.blens[s->sub.trees.index++] = c; } else /* c == 16..18 */ { i = c == 18 ? 7 : c - 14; j = c == 18 ? 11 : 3; NEEDBITS(t + i) DUMPBITS(t) j += (uInt)b & inflate_mask[i]; DUMPBITS(i) i = s->sub.trees.index; t = s->sub.trees.table; if (i + j > 258 + (t & 0x1f) + ((t >> 5) & 0x1f) || (c == 16 && i < 1)) { s->mode = BADB; z->msg = "invalid bit length repeat"; r = Z_DATA_ERROR; LEAVE } c = c == 16 ? s->sub.trees.blens[i - 1] : 0; do { s->sub.trees.blens[i++] = c; } while (--j); s->sub.trees.index = i; } } inflate_trees_free(s->sub.trees.tb, z); s->sub.trees.tb = Z_NULL; { uInt bl, bd; inflate_huft *tl, *td; inflate_codes_statef *c; bl = 9; /* must be <= 9 for lookahead assumptions */ bd = 6; /* must be <= 9 for lookahead assumptions */ t = s->sub.trees.table; t = inflate_trees_dynamic(257 + (t & 0x1f), 1 + ((t >> 5) & 0x1f), s->sub.trees.blens, &bl, &bd, &tl, &td, z); if (t != Z_OK) { if (t == (uInt)Z_DATA_ERROR) { ZFREE(z, s->sub.trees.blens, s->sub.trees.nblens * sizeof(uInt)); s->mode = BADB; } r = t; LEAVE } Tracev((stderr, "inflate: trees ok\n")); if ((c = inflate_codes_new(bl, bd, tl, td, z)) == Z_NULL) { inflate_trees_free(td, z); inflate_trees_free(tl, z); r = Z_MEM_ERROR; LEAVE } ZFREE(z, s->sub.trees.blens, s->sub.trees.nblens * sizeof(uInt)); s->sub.decode.codes = c; s->sub.decode.tl = tl; s->sub.decode.td = td; } s->mode = CODES; case CODES: UPDATE if ((r = inflate_codes(s, z, r)) != Z_STREAM_END) return inflate_flush(s, z, r); r = Z_OK; inflate_codes_free(s->sub.decode.codes, z); inflate_trees_free(s->sub.decode.td, z); inflate_trees_free(s->sub.decode.tl, z); LOAD Tracev((stderr, "inflate: codes end, %lu total out\n", z->total_out + (q >= s->read ? q - s->read : (s->end - s->read) + (q - s->window)))); if (!s->last) { s->mode = TYPE; break; } if (k > 7) /* return unused byte, if any */ { Assert(k < 16, "inflate_codes grabbed too many bytes") k -= 8; n++; p--; /* can always return one */ } s->mode = DRY; case DRY: FLUSH if (s->read != s->write) LEAVE s->mode = DONEB; case DONEB: r = Z_STREAM_END; LEAVE case BADB: r = Z_DATA_ERROR; LEAVE default: r = Z_STREAM_ERROR; LEAVE } } local int inflate_blocks_free(s, z, c) inflate_blocks_statef *s; z_stream *z; uLongf *c; { inflate_blocks_reset(s, z, c); ZFREE(z, s->window, s->end - s->window); ZFREE(z, s, sizeof(struct inflate_blocks_state)); Trace((stderr, "inflate: blocks freed\n")); return Z_OK; } /* * This subroutine adds the data at next_in/avail_in to the output history * without performing any output. The output buffer must be "caught up"; * i.e. no pending output (hence s->read equals s->write), and the state must * be BLOCKS (i.e. we should be willing to see the start of a series of * BLOCKS). On exit, the output will also be caught up, and the checksum * will have been updated if need be. */ local int inflate_addhistory(s, z) inflate_blocks_statef *s; z_stream *z; { uLong b; /* bit buffer */ /* NOT USED HERE */ uInt k; /* bits in bit buffer */ /* NOT USED HERE */ uInt t; /* temporary storage */ Bytef *p; /* input data pointer */ uInt n; /* bytes available there */ Bytef *q; /* output window write pointer */ uInt m; /* bytes to end of window or read pointer */ if (s->read != s->write) return Z_STREAM_ERROR; if (s->mode != TYPE) return Z_DATA_ERROR; /* we're ready to rock */ LOAD /* while there is input ready, copy to output buffer, moving * pointers as needed. */ while (n) { t = n; /* how many to do */ /* is there room until end of buffer? */ if (t > m) t = m; /* update check information */ if (s->checkfn != Z_NULL) s->check = (*s->checkfn)(s->check, q, t); zmemcpy(q, p, t); q += t; p += t; n -= t; z->total_out += t; s->read = q; /* drag read pointer forward */ /* WRAP */ /* expand WRAP macro by hand to handle s->read */ if (q == s->end) { s->read = q = s->window; m = WAVAIL; } } UPDATE return Z_OK; } /* * At the end of a Deflate-compressed PPP packet, we expect to have seen * a `stored' block type value but not the (zero) length bytes. */ local int inflate_packet_flush(s) inflate_blocks_statef *s; { if (s->mode != LENS) return Z_DATA_ERROR; s->mode = TYPE; return Z_OK; } /*+++++*/ /* inftrees.c -- generate Huffman trees for efficient decoding * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* simplify the use of the inflate_huft type with some defines */ #define base more.Base #define next more.Next #define exop word.what.Exop #define bits word.what.Bits local int huft_build OF(( uIntf *, /* code lengths in bits */ uInt, /* number of codes */ uInt, /* number of "simple" codes */ const uIntf *, /* list of base values for non-simple codes */ const uIntf *, /* list of extra bits for non-simple codes */ inflate_huft * FAR*,/* result: starting table */ uIntf *, /* maximum lookup bits (returns actual) */ z_stream *)); /* for zalloc function */ local voidpf falloc OF(( voidpf, /* opaque pointer (not used) */ uInt, /* number of items */ uInt)); /* size of item */ local void ffree OF(( voidpf q, /* opaque pointer (not used) */ voidpf p, /* what to free (not used) */ uInt n)); /* number of bytes (not used) */ /* Tables for deflate from PKZIP's appnote.txt. */ local const uInt cplens[] = { /* Copy lengths for literal codes 257..285 */ 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31, 35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0}; /* actually lengths - 2; also see note #13 above about 258 */ local const uInt cplext[] = { /* Extra bits for literal codes 257..285 */ 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 192, 192}; /* 192==invalid */ local const uInt cpdist[] = { /* Copy offsets for distance codes 0..29 */ 1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193, 257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145, 8193, 12289, 16385, 24577}; local const uInt cpdext[] = { /* Extra bits for distance codes */ 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13}; /* Huffman code decoding is performed using a multi-level table lookup. The fastest way to decode is to simply build a lookup table whose size is determined by the longest code. However, the time it takes to build this table can also be a factor if the data being decoded is not very long. The most common codes are necessarily the shortest codes, so those codes dominate the decoding time, and hence the speed. The idea is you can have a shorter table that decodes the shorter, more probable codes, and then point to subsidiary tables for the longer codes. The time it costs to decode the longer codes is then traded against the time it takes to make longer tables. This results of this trade are in the variables lbits and dbits below. lbits is the number of bits the first level table for literal/ length codes can decode in one step, and dbits is the same thing for the distance codes. Subsequent tables are also less than or equal to those sizes. These values may be adjusted either when all of the codes are shorter than that, in which case the longest code length in bits is used, or when the shortest code is *longer* than the requested table size, in which case the length of the shortest code in bits is used. There are two different values for the two tables, since they code a different number of possibilities each. The literal/length table codes 286 possible values, or in a flat code, a little over eight bits. The distance table codes 30 possible values, or a little less than five bits, flat. The optimum values for speed end up being about one bit more than those, so lbits is 8+1 and dbits is 5+1. The optimum values may differ though from machine to machine, and possibly even between compilers. Your mileage may vary. */ /* If BMAX needs to be larger than 16, then h and x[] should be uLong. */ #define BMAX 15 /* maximum bit length of any code */ #define N_MAX 288 /* maximum number of codes in any set */ #ifdef DEBUG_ZLIB uInt inflate_hufts; #endif local int huft_build(b, n, s, d, e, t, m, zs) uIntf *b; /* code lengths in bits (all assumed <= BMAX) */ uInt n; /* number of codes (assumed <= N_MAX) */ uInt s; /* number of simple-valued codes (0..s-1) */ const uIntf *d; /* list of base values for non-simple codes */ const uIntf *e; /* list of extra bits for non-simple codes */ inflate_huft * FAR *t; /* result: starting table */ uIntf *m; /* maximum lookup bits, returns actual */ z_stream *zs; /* for zalloc function */ /* Given a list of code lengths and a maximum table size, make a set of tables to decode that set of codes. Return Z_OK on success, Z_BUF_ERROR if the given code set is incomplete (the tables are still built in this case), Z_DATA_ERROR if the input is invalid (all zero length codes or an over-subscribed set of lengths), or Z_MEM_ERROR if not enough memory. */ { uInt a; /* counter for codes of length k */ uInt c[BMAX+1]; /* bit length count table */ uInt f; /* i repeats in table every f entries */ int g; /* maximum code length */ int h; /* table level */ register uInt i; /* counter, current code */ register uInt j; /* counter */ register int k; /* number of bits in current code */ int l; /* bits per table (returned in m) */ register uIntf *p; /* pointer into c[], b[], or v[] */ inflate_huft *q; /* points to current table */ struct inflate_huft_s r; /* table entry for structure assignment */ inflate_huft *u[BMAX]; /* table stack */ uInt v[N_MAX]; /* values in order of bit length */ register int w; /* bits before this table == (l * h) */ uInt x[BMAX+1]; /* bit offsets, then code stack */ uIntf *xp; /* pointer into x */ int y; /* number of dummy codes added */ uInt z; /* number of entries in current table */ /* Generate counts for each bit length */ p = c; #define C0 *p++ = 0; #define C2 C0 C0 C0 C0 #define C4 C2 C2 C2 C2 C4 /* clear c[]--assume BMAX+1 is 16 */ p = b; i = n; do { c[*p++]++; /* assume all entries <= BMAX */ } while (--i); if (c[0] == n) /* null input--all zero length codes */ { *t = (inflate_huft *)Z_NULL; *m = 0; return Z_OK; } /* Find minimum and maximum length, bound *m by those */ l = *m; for (j = 1; j <= BMAX; j++) if (c[j]) break; k = j; /* minimum code length */ if ((uInt)l < j) l = j; for (i = BMAX; i; i--) if (c[i]) break; g = i; /* maximum code length */ if ((uInt)l > i) l = i; *m = l; /* Adjust last length count to fill out codes, if needed */ for (y = 1 << j; j < i; j++, y <<= 1) if ((y -= c[j]) < 0) return Z_DATA_ERROR; if ((y -= c[i]) < 0) return Z_DATA_ERROR; c[i] += y; /* Generate starting offsets into the value table for each length */ x[1] = j = 0; p = c + 1; xp = x + 2; while (--i) { /* note that i == g from above */ *xp++ = (j += *p++); } /* Make a table of values in order of bit lengths */ p = b; i = 0; do { if ((j = *p++) != 0) v[x[j]++] = i; } while (++i < n); /* Generate the Huffman codes and for each, make the table entries */ x[0] = i = 0; /* first Huffman code is zero */ p = v; /* grab values in bit order */ h = -1; /* no tables yet--level -1 */ w = -l; /* bits decoded == (l * h) */ u[0] = (inflate_huft *)Z_NULL; /* just to keep compilers happy */ q = (inflate_huft *)Z_NULL; /* ditto */ z = 0; /* ditto */ /* go through the bit lengths (k already is bits in shortest code) */ for (; k <= g; k++) { a = c[k]; while (a--) { /* here i is the Huffman code of length k bits for value *p */ /* make tables up to required level */ while (k > w + l) { h++; w += l; /* previous table always l bits */ /* compute minimum size table less than or equal to l bits */ z = (z = g - w) > (uInt)l ? l : z; /* table size upper limit */ if ((f = 1 << (j = k - w)) > a + 1) /* try a k-w bit table */ { /* too few codes for k-w bit table */ f -= a + 1; /* deduct codes from patterns left */ xp = c + k; if (j < z) while (++j < z) /* try smaller tables up to z bits */ { if ((f <<= 1) <= *++xp) break; /* enough codes to use up j bits */ f -= *xp; /* else deduct codes from patterns */ } } z = 1 << j; /* table entries for j-bit table */ /* allocate and link in new table */ if ((q = (inflate_huft *)ZALLOC (zs,z + 1,sizeof(inflate_huft))) == Z_NULL) { if (h) inflate_trees_free(u[0], zs); return Z_MEM_ERROR; /* not enough memory */ } q->word.Nalloc = z + 1; #ifdef DEBUG_ZLIB inflate_hufts += z + 1; #endif *t = q + 1; /* link to list for huft_free() */ *(t = &(q->next)) = Z_NULL; u[h] = ++q; /* table starts after link */ /* connect to last table, if there is one */ if (h) { x[h] = i; /* save pattern for backing up */ r.bits = (Byte)l; /* bits to dump before this table */ r.exop = (Byte)j; /* bits in this table */ r.next = q; /* pointer to this table */ j = i >> (w - l); /* (get around Turbo C bug) */ u[h-1][j] = r; /* connect to last table */ } } /* set up table entry in r */ r.bits = (Byte)(k - w); if (p >= v + n) r.exop = 128 + 64; /* out of values--invalid code */ else if (*p < s) { r.exop = (Byte)(*p < 256 ? 0 : 32 + 64); /* 256 is end-of-block */ r.base = *p++; /* simple code is just the value */ } else { r.exop = (Byte)e[*p - s] + 16 + 64; /* non-simple--look up in lists */ r.base = d[*p++ - s]; } /* fill code-like entries with r */ f = 1 << (k - w); for (j = i >> w; j < z; j += f) q[j] = r; /* backwards increment the k-bit code i */ for (j = 1 << (k - 1); i & j; j >>= 1) i ^= j; i ^= j; /* backup over finished tables */ while ((i & ((1 << w) - 1)) != x[h]) { h--; /* don't need to update q */ w -= l; } } } /* Return Z_BUF_ERROR if we were given an incomplete table */ return y != 0 && g != 1 ? Z_BUF_ERROR : Z_OK; } local int inflate_trees_bits(c, bb, tb, z) uIntf *c; /* 19 code lengths */ uIntf *bb; /* bits tree desired/actual depth */ inflate_huft * FAR *tb; /* bits tree result */ z_stream *z; /* for zfree function */ { int r; r = huft_build(c, 19, 19, (uIntf*)Z_NULL, (uIntf*)Z_NULL, tb, bb, z); if (r == Z_DATA_ERROR) z->msg = "oversubscribed dynamic bit lengths tree"; else if (r == Z_BUF_ERROR) { inflate_trees_free(*tb, z); z->msg = "incomplete dynamic bit lengths tree"; r = Z_DATA_ERROR; } return r; } local int inflate_trees_dynamic(nl, nd, c, bl, bd, tl, td, z) uInt nl; /* number of literal/length codes */ uInt nd; /* number of distance codes */ uIntf *c; /* that many (total) code lengths */ uIntf *bl; /* literal desired/actual bit depth */ uIntf *bd; /* distance desired/actual bit depth */ inflate_huft * FAR *tl; /* literal/length tree result */ inflate_huft * FAR *td; /* distance tree result */ z_stream *z; /* for zfree function */ { int r; /* build literal/length tree */ if ((r = huft_build(c, nl, 257, cplens, cplext, tl, bl, z)) != Z_OK) { if (r == Z_DATA_ERROR) z->msg = "oversubscribed literal/length tree"; else if (r == Z_BUF_ERROR) { inflate_trees_free(*tl, z); z->msg = "incomplete literal/length tree"; r = Z_DATA_ERROR; } return r; } /* build distance tree */ if ((r = huft_build(c + nl, nd, 0, cpdist, cpdext, td, bd, z)) != Z_OK) { if (r == Z_DATA_ERROR) z->msg = "oversubscribed literal/length tree"; else if (r == Z_BUF_ERROR) { #ifdef PKZIP_BUG_WORKAROUND r = Z_OK; } #else inflate_trees_free(*td, z); z->msg = "incomplete literal/length tree"; r = Z_DATA_ERROR; } inflate_trees_free(*tl, z); return r; #endif } /* done */ return Z_OK; } /* build fixed tables only once--keep them here */ local int fixed_lock = 0; local int fixed_built = 0; #define FIXEDH 530 /* number of hufts used by fixed tables */ local uInt fixed_left = FIXEDH; local inflate_huft fixed_mem[FIXEDH]; local uInt fixed_bl; local uInt fixed_bd; local inflate_huft *fixed_tl; local inflate_huft *fixed_td; local voidpf falloc(q, n, s) voidpf q; /* opaque pointer (not used) */ uInt n; /* number of items */ uInt s; /* size of item */ { Assert(s == sizeof(inflate_huft) && n <= fixed_left, "inflate_trees falloc overflow"); if (q) s++; /* to make some compilers happy */ fixed_left -= n; return (voidpf)(fixed_mem + fixed_left); } local void ffree(q, p, n) voidpf q; voidpf p; uInt n; { Assert(0, "inflate_trees ffree called!"); if (q) q = p; /* to make some compilers happy */ } local int inflate_trees_fixed(bl, bd, tl, td) uIntf *bl; /* literal desired/actual bit depth */ uIntf *bd; /* distance desired/actual bit depth */ inflate_huft * FAR *tl; /* literal/length tree result */ inflate_huft * FAR *td; /* distance tree result */ { /* build fixed tables if not built already--lock out other instances */ while (++fixed_lock > 1) fixed_lock--; if (!fixed_built) { int k; /* temporary variable */ unsigned c[288]; /* length list for huft_build */ z_stream z; /* for falloc function */ /* set up fake z_stream for memory routines */ z.zalloc = falloc; z.zfree = ffree; z.opaque = Z_NULL; /* literal table */ for (k = 0; k < 144; k++) c[k] = 8; for (; k < 256; k++) c[k] = 9; for (; k < 280; k++) c[k] = 7; for (; k < 288; k++) c[k] = 8; fixed_bl = 7; huft_build(c, 288, 257, cplens, cplext, &fixed_tl, &fixed_bl, &z); /* distance table */ for (k = 0; k < 30; k++) c[k] = 5; fixed_bd = 5; huft_build(c, 30, 0, cpdist, cpdext, &fixed_td, &fixed_bd, &z); /* done */ fixed_built = 1; } fixed_lock--; *bl = fixed_bl; *bd = fixed_bd; *tl = fixed_tl; *td = fixed_td; return Z_OK; } local int inflate_trees_free(t, z) inflate_huft *t; /* table to free */ z_stream *z; /* for zfree function */ /* Free the malloc'ed tables built by huft_build(), which makes a linked list of the tables it made, with the links in a dummy first entry of each table. */ { register inflate_huft *p, *q; /* Go through linked list, freeing from the malloced (t[-1]) address. */ p = t; while (p != Z_NULL) { q = (--p)->next; ZFREE(z, p, p->word.Nalloc * sizeof(inflate_huft)); p = q; } return Z_OK; } /*+++++*/ /* infcodes.c -- process literals and length/distance pairs * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* simplify the use of the inflate_huft type with some defines */ #define base more.Base #define next more.Next #define exop word.what.Exop #define bits word.what.Bits /* inflate codes private state */ struct inflate_codes_state { /* mode */ enum { /* waiting for "i:"=input, "o:"=output, "x:"=nothing */ START, /* x: set up for LEN */ LEN, /* i: get length/literal/eob next */ LENEXT, /* i: getting length extra (have base) */ DIST, /* i: get distance next */ DISTEXT, /* i: getting distance extra */ COPY, /* o: copying bytes in window, waiting for space */ LIT, /* o: got literal, waiting for output space */ WASH, /* o: got eob, possibly still output waiting */ END, /* x: got eob and all data flushed */ BADCODE} /* x: got error */ mode; /* current inflate_codes mode */ /* mode dependent information */ uInt len; union { struct { inflate_huft *tree; /* pointer into tree */ uInt need; /* bits needed */ } code; /* if LEN or DIST, where in tree */ uInt lit; /* if LIT, literal */ struct { uInt get; /* bits to get for extra */ uInt dist; /* distance back to copy from */ } copy; /* if EXT or COPY, where and how much */ } sub; /* submode */ /* mode independent information */ Byte lbits; /* ltree bits decoded per branch */ Byte dbits; /* dtree bits decoder per branch */ inflate_huft *ltree; /* literal/length/eob tree */ inflate_huft *dtree; /* distance tree */ }; local inflate_codes_statef *inflate_codes_new(bl, bd, tl, td, z) uInt bl, bd; inflate_huft *tl, *td; z_stream *z; { inflate_codes_statef *c; if ((c = (inflate_codes_statef *) ZALLOC(z,1,sizeof(struct inflate_codes_state))) != Z_NULL) { c->mode = START; c->lbits = (Byte)bl; c->dbits = (Byte)bd; c->ltree = tl; c->dtree = td; Tracev((stderr, "inflate: codes new\n")); } return c; } local int inflate_codes(s, z, r) inflate_blocks_statef *s; z_stream *z; int r; { uInt j; /* temporary storage */ inflate_huft *t; /* temporary pointer */ uInt e; /* extra bits or operation */ uLong b; /* bit buffer */ uInt k; /* bits in bit buffer */ Bytef *p; /* input data pointer */ uInt n; /* bytes available there */ Bytef *q; /* output window write pointer */ uInt m; /* bytes to end of window or read pointer */ Bytef *f; /* pointer to copy strings from */ inflate_codes_statef *c = s->sub.decode.codes; /* codes state */ /* copy input/output information to locals (UPDATE macro restores) */ LOAD /* process input and output based on current state */ while (1) switch (c->mode) { /* waiting for "i:"=input, "o:"=output, "x:"=nothing */ case START: /* x: set up for LEN */ #ifndef SLOW if (m >= 258 && n >= 10) { UPDATE r = inflate_fast(c->lbits, c->dbits, c->ltree, c->dtree, s, z); LOAD if (r != Z_OK) { c->mode = r == Z_STREAM_END ? WASH : BADCODE; break; } } #endif /* !SLOW */ c->sub.code.need = c->lbits; c->sub.code.tree = c->ltree; c->mode = LEN; case LEN: /* i: get length/literal/eob next */ j = c->sub.code.need; NEEDBITS(j) t = c->sub.code.tree + ((uInt)b & inflate_mask[j]); DUMPBITS(t->bits) e = (uInt)(t->exop); if (e == 0) /* literal */ { c->sub.lit = t->base; Tracevv((stderr, t->base >= 0x20 && t->base < 0x7f ? "inflate: literal '%c'\n" : "inflate: literal 0x%02x\n", t->base)); c->mode = LIT; break; } if (e & 16) /* length */ { c->sub.copy.get = e & 15; c->len = t->base; c->mode = LENEXT; break; } if ((e & 64) == 0) /* next table */ { c->sub.code.need = e; c->sub.code.tree = t->next; break; } if (e & 32) /* end of block */ { Tracevv((stderr, "inflate: end of block\n")); c->mode = WASH; break; } c->mode = BADCODE; /* invalid code */ z->msg = "invalid literal/length code"; r = Z_DATA_ERROR; LEAVE case LENEXT: /* i: getting length extra (have base) */ j = c->sub.copy.get; NEEDBITS(j) c->len += (uInt)b & inflate_mask[j]; DUMPBITS(j) c->sub.code.need = c->dbits; c->sub.code.tree = c->dtree; Tracevv((stderr, "inflate: length %u\n", c->len)); c->mode = DIST; case DIST: /* i: get distance next */ j = c->sub.code.need; NEEDBITS(j) t = c->sub.code.tree + ((uInt)b & inflate_mask[j]); DUMPBITS(t->bits) e = (uInt)(t->exop); if (e & 16) /* distance */ { c->sub.copy.get = e & 15; c->sub.copy.dist = t->base; c->mode = DISTEXT; break; } if ((e & 64) == 0) /* next table */ { c->sub.code.need = e; c->sub.code.tree = t->next; break; } c->mode = BADCODE; /* invalid code */ z->msg = "invalid distance code"; r = Z_DATA_ERROR; LEAVE case DISTEXT: /* i: getting distance extra */ j = c->sub.copy.get; NEEDBITS(j) c->sub.copy.dist += (uInt)b & inflate_mask[j]; DUMPBITS(j) Tracevv((stderr, "inflate: distance %u\n", c->sub.copy.dist)); c->mode = COPY; case COPY: /* o: copying bytes in window, waiting for space */ #ifndef __TURBOC__ /* Turbo C bug for following expression */ f = (uInt)(q - s->window) < c->sub.copy.dist ? s->end - (c->sub.copy.dist - (q - s->window)) : q - c->sub.copy.dist; #else f = q - c->sub.copy.dist; if ((uInt)(q - s->window) < c->sub.copy.dist) f = s->end - (c->sub.copy.dist - (q - s->window)); #endif while (c->len) { NEEDOUT OUTBYTE(*f++) if (f == s->end) f = s->window; c->len--; } c->mode = START; break; case LIT: /* o: got literal, waiting for output space */ NEEDOUT OUTBYTE(c->sub.lit) c->mode = START; break; case WASH: /* o: got eob, possibly more output */ FLUSH if (s->read != s->write) LEAVE c->mode = END; case END: r = Z_STREAM_END; LEAVE case BADCODE: /* x: got error */ r = Z_DATA_ERROR; LEAVE default: r = Z_STREAM_ERROR; LEAVE } } local void inflate_codes_free(c, z) inflate_codes_statef *c; z_stream *z; { ZFREE(z, c, sizeof(struct inflate_codes_state)); Tracev((stderr, "inflate: codes free\n")); } /*+++++*/ /* inflate_util.c -- data and routines common to blocks and codes * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* copy as much as possible from the sliding window to the output area */ local int inflate_flush(s, z, r) inflate_blocks_statef *s; z_stream *z; int r; { uInt n; Bytef *p, *q; /* local copies of source and destination pointers */ p = z->next_out; q = s->read; /* compute number of bytes to copy as far as end of window */ n = (uInt)((q <= s->write ? s->write : s->end) - q); if (n > z->avail_out) n = z->avail_out; if (n && r == Z_BUF_ERROR) r = Z_OK; /* update counters */ z->avail_out -= n; z->total_out += n; /* update check information */ if (s->checkfn != Z_NULL) s->check = (*s->checkfn)(s->check, q, n); /* copy as far as end of window */ if (p != NULL) { zmemcpy(p, q, n); p += n; } q += n; /* see if more to copy at beginning of window */ if (q == s->end) { /* wrap pointers */ q = s->window; if (s->write == s->end) s->write = s->window; /* compute bytes to copy */ n = (uInt)(s->write - q); if (n > z->avail_out) n = z->avail_out; if (n && r == Z_BUF_ERROR) r = Z_OK; /* update counters */ z->avail_out -= n; z->total_out += n; /* update check information */ if (s->checkfn != Z_NULL) s->check = (*s->checkfn)(s->check, q, n); /* copy */ if (p != NULL) { zmemcpy(p, q, n); p += n; } q += n; } /* update pointers */ z->next_out = p; s->read = q; /* done */ return r; } /*+++++*/ /* inffast.c -- process literals and length/distance pairs fast * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* simplify the use of the inflate_huft type with some defines */ #define base more.Base #define next more.Next #define exop word.what.Exop #define bits word.what.Bits /* macros for bit input with no checking and for returning unused bytes */ #define GRABBITS(j) {while(k<(j)){b|=((uLong)NEXTBYTE)<>3);p-=c;k&=7;} /* Called with number of bytes left to write in window at least 258 (the maximum string length) and number of input bytes available at least ten. The ten bytes are six bytes for the longest length/ distance pair plus four bytes for overloading the bit buffer. */ local int inflate_fast(bl, bd, tl, td, s, z) uInt bl, bd; inflate_huft *tl, *td; inflate_blocks_statef *s; z_stream *z; { inflate_huft *t; /* temporary pointer */ uInt e; /* extra bits or operation */ uLong b; /* bit buffer */ uInt k; /* bits in bit buffer */ Bytef *p; /* input data pointer */ uInt n; /* bytes available there */ Bytef *q; /* output window write pointer */ uInt m; /* bytes to end of window or read pointer */ uInt ml; /* mask for literal/length tree */ uInt md; /* mask for distance tree */ uInt c; /* bytes to copy */ uInt d; /* distance back to copy from */ Bytef *r; /* copy source pointer */ /* load input, output, bit values */ LOAD /* initialize masks */ ml = inflate_mask[bl]; md = inflate_mask[bd]; /* do until not enough input or output space for fast loop */ do { /* assume called with m >= 258 && n >= 10 */ /* get literal/length code */ GRABBITS(20) /* max bits for literal/length code */ if ((e = (t = tl + ((uInt)b & ml))->exop) == 0) { DUMPBITS(t->bits) Tracevv((stderr, t->base >= 0x20 && t->base < 0x7f ? "inflate: * literal '%c'\n" : "inflate: * literal 0x%02x\n", t->base)); *q++ = (Byte)t->base; m--; continue; } do { DUMPBITS(t->bits) if (e & 16) { /* get extra bits for length */ e &= 15; c = t->base + ((uInt)b & inflate_mask[e]); DUMPBITS(e) Tracevv((stderr, "inflate: * length %u\n", c)); /* decode distance base of block to copy */ GRABBITS(15); /* max bits for distance code */ e = (t = td + ((uInt)b & md))->exop; do { DUMPBITS(t->bits) if (e & 16) { /* get extra bits to add to distance base */ e &= 15; GRABBITS(e) /* get extra bits (up to 13) */ d = t->base + ((uInt)b & inflate_mask[e]); DUMPBITS(e) Tracevv((stderr, "inflate: * distance %u\n", d)); /* do the copy */ m -= c; if ((uInt)(q - s->window) >= d) /* offset before dest */ { /* just copy */ r = q - d; *q++ = *r++; c--; /* minimum count is three, */ *q++ = *r++; c--; /* so unroll loop a little */ } else /* else offset after destination */ { e = d - (q - s->window); /* bytes from offset to end */ r = s->end - e; /* pointer to offset */ if (c > e) /* if source crosses, */ { c -= e; /* copy to end of window */ do { *q++ = *r++; } while (--e); r = s->window; /* copy rest from start of window */ } } do { /* copy all or what's left */ *q++ = *r++; } while (--c); break; } else if ((e & 64) == 0) e = (t = t->next + ((uInt)b & inflate_mask[e]))->exop; else { z->msg = "invalid distance code"; UNGRAB UPDATE return Z_DATA_ERROR; } } while (1); break; } if ((e & 64) == 0) { if ((e = (t = t->next + ((uInt)b & inflate_mask[e]))->exop) == 0) { DUMPBITS(t->bits) Tracevv((stderr, t->base >= 0x20 && t->base < 0x7f ? "inflate: * literal '%c'\n" : "inflate: * literal 0x%02x\n", t->base)); *q++ = (Byte)t->base; m--; break; } } else if (e & 32) { Tracevv((stderr, "inflate: * end of block\n")); UNGRAB UPDATE return Z_STREAM_END; } else { z->msg = "invalid literal/length code"; UNGRAB UPDATE return Z_DATA_ERROR; } } while (1); } while (m >= 258 && n >= 10); /* not enough input or output--restore pointers and return */ UNGRAB UPDATE return Z_OK; } /*+++++*/ /* zutil.c -- target dependent utility functions for the compression library * Copyright (C) 1995 Jean-loup Gailly. * For conditions of distribution and use, see copyright notice in zlib.h */ /* From: zutil.c,v 1.8 1995/05/03 17:27:12 jloup Exp */ char *zlib_version = ZLIB_VERSION; #ifndef NO_DEFLATE char *z_errmsg[] = { "stream end", /* Z_STREAM_END 1 */ "", /* Z_OK 0 */ "file error", /* Z_ERRNO (-1) */ "stream error", /* Z_STREAM_ERROR (-2) */ "data error", /* Z_DATA_ERROR (-3) */ "insufficient memory", /* Z_MEM_ERROR (-4) */ "buffer error", /* Z_BUF_ERROR (-5) */ ""}; #endif /* NO_DEFLATE */ /*+++++*/ /* adler32.c -- compute the Adler-32 checksum of a data stream * Copyright (C) 1995 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h */ /* From: adler32.c,v 1.6 1995/05/03 17:27:08 jloup Exp */ #define BASE 65521L /* largest prime smaller than 65536 */ #define NMAX 5552 /* NMAX is the largest n such that 255n(n+1)/2 + (n+1)(BASE-1) <= 2^32-1 */ #define DO1(buf) {s1 += *buf++; s2 += s1;} #define DO2(buf) DO1(buf); DO1(buf); #define DO4(buf) DO2(buf); DO2(buf); #define DO8(buf) DO4(buf); DO4(buf); #define DO16(buf) DO8(buf); DO8(buf); /* ========================================================================= */ uLong adler32(adler, buf, len) uLong adler; Bytef *buf; uInt len; { unsigned long s1 = adler & 0xffff; unsigned long s2 = (adler >> 16) & 0xffff; int k; if (buf == Z_NULL) return 1L; while (len > 0) { k = len < NMAX ? len : NMAX; len -= k; while (k >= 16) { DO16(buf); k -= 16; } if (k != 0) do { DO1(buf); } while (--k); s1 %= BASE; s2 %= BASE; } return (s2 << 16) | s1; }