ZipDeflate Class Reference

The deflater for the file compressor and decompressor. Allows the CCDiskFile to offer compression of files. At present just used to compress the native file format. More...

#include <zdeflate.h>

List of all members.

Public Member Functions
	ZipDeflate ()
	ZipDeflate constructor.
	~ZipDeflate ()
	ZipDeflate destructor.
INT32	Init (z_stream *strm, INT32 level)
	Initializes the internal stream state for compression. The fields zalloc, zfree and opaque must be initialized before by the caller. If zalloc and zfree are set to Z_NULL, deflateInit updates them to use default allocation functions.
INT32	deflate (z_stream *strm, INT32 flush)
	Actually go and deflate the stream.
INT32	End (z_stream *strm)
	End the deflation on this stream.
INT32	Init (z_stream *strm, INT32 level, INT32 method, INT32 windowBits, INT32 memLevel, INT32 strategy)
	This is another version of deflate Init with more compression options. The fields next_in, zalloc and zfree must be initialized before by the caller.
INT32	Reset (z_stream *strm)
	This function is equivalent to deflate End followed by deflateInit, but does not free and reallocate all the internal compression state. The stream will keep the same compression level and any other attributes that may have been set by deflate Init.
INT32	SetDictionary (z_stream *strm, const Bytef *dictionary, uInt dictLength)
	Initializes the compression dictionary (history buffer) from the given byte sequence without producing any compressed output. This function must be called immediately after deflateInit or deflateInit2, before any call of deflate. The compressor and decompressor must use exactly the same dictionary (see inflateSetDictionary). The dictionary should consist of strings (byte sequences) that are likely to be encountered later in the data to be compressed, with the most commonly used strings preferably put towards the end of the dictionary. Using a dictionary is most useful when the data to be compressed is short and can be predicted with good accuracy; the data can then be compressed better than with the default empty dictionary. In this version of the library, only the last 32K bytes of the dictionary are used. Upon return of this function, strm->adler is set to the Adler32 value of the dictionary; the decompressor may later use this value to determine which dictionary has been used by the compressor. (The Adler32 value applies to the whole dictionary even if only a subset of the dictionary is actually used by the compressor.).
INT32	Params (z_stream *strm, INT32 level, INT32 strategy)
	New for version 0.99 Dynamically update the compression level and compression strategy. This can be used to switch between compression and straight copy of the input data, or to switch to a different kind of input data requiring a different strategy. If the compression level is changed, the input available so far is compressed with the old level (and may be flushed); the new level will take effect only at the next call of deflate().
Protected Member Functions
void	ct_init (DeflateState *s)
INT32	ct_tally (DeflateState *s, INT32 dist, INT32 lc)
ulg	ct_flush_block (DeflateState *s, char *buf, ulg stored_len, INT32 eof)
void	ct_stored_block (DeflateState *s, char *buf, ulg stored_len, INT32 eof)
void	_tr_init (DeflateState *s)
	Initialize the tree data structures for a new zlib stream.
INT32	_tr_tally (DeflateState *s, unsigned dist, unsigned lc)
	Save the match info and tally the frequency counts. Return true if the current block must be flushed.
ulg	_tr_flush_block (DeflateState *s, charf *buf, ulg stored_len, INT32 eof)
void	_tr_align (DeflateState *s)
void	_tr_stored_block (DeflateState *s, charf *buf, ulg stored_len, INT32 eof)
Private Member Functions
	CC_DECLARE_MEMDUMP (ZipDeflate)
INT32	deflate_stored (DeflateState *s, INT32 flush)
	Copy without compression 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 insert new strings in the dictionary since uncompressible data is probably not useful. This function is used only for the level=0 compression option. NOTE: this function should be optimized to avoid extra copying.
INT32	deflate_fast (DeflateState *s, INT32 flush)
	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.
INT32	deflate_slow (DeflateState *s, INT32 flush)
	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.
void	fill_window (DeflateState *s)
	Fill the window when the lookahead becomes insufficient. Updates strstart and lookahead.
void	lm_init (DeflateState *s)
	Initialize the "longest match" routines for a new zlib stream.
uInt	longest_match (DeflateState *s, IPos cur_match)
	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 OUT assertion: the match length is not greater than s->lookahead.
void	putShortMSB (DeflateState *s, uInt b)
	Put a short the pending_out buffer. The 16-bit value is put in MSB order. IN assertion: the stream state is correct and there is enough room in the pending_out buffer.
void	flush_pending (z_stream *strm)
	Flush as much pending output as possible. All deflate() output goes through this function so some applications may wish to modify it to avoid allocating a large strm->next_out buffer and copying into it. (See also read_buf()).
INT32	read_buf (z_stream *strm, charf *buf, unsigned size)
	Read a new buffer from the current input stream, update the adler32 and total number of bytes read. All deflate() input goes through this function so some applications may wish to modify it to avoid allocating a large strm->next_in buffer and copying from it. (See also flush_pending()).
void	tr_static_init (void)
	Initialize the various 'constant' tables. To do: do this at compile time.
void	init_block (DeflateState *s)
	Initialize a new block.
void	pqdownheap (DeflateState *s, ct_data *tree, INT32 k)
	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).
void	gen_bitlen (DeflateState *s, tree_desc *desc)
	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.
void	gen_codes (ct_data *tree, INT32 max_code, ushf bl_count[])
	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.
void	build_tree (DeflateState *s, tree_desc *desc)
	Construct one Huffman tree and assigns the code bit strings and lengths. Update the total bit length for the current block. IN assertion: the field freq is set for all tree elements. OUT assertions: the fields len and code are set to the optimal bit length and corresponding code. The length opt_len is updated; static_len is also updated if stree is not null. The field max_code is set.
void	scan_tree (DeflateState *s, ct_data *tree, INT32 max_code)
	Scan a literal or distance tree to determine the frequencies of the codes in the bit length tree.
void	send_tree (DeflateState *s, ct_data *tree, INT32 max_code)
	Send a literal or distance tree in compressed form, using the codes in bl_tree.
INT32	build_bl_tree (DeflateState *s)
	Construct the Huffman tree for the bit lengths and return the index in bl_order of the last bit length code to send.
void	send_all_trees (DeflateState *s, INT32 lcodes, INT32 dcodes, INT32 blcodes)
	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.
void	compress_block (DeflateState *s, ct_data *ltree, ct_data *dtree)
	Send the block data compressed using the given Huffman trees.
void	set_data_type (DeflateState *s)
	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 INT32 on 16 bit machines).
unsigned	bi_reverse (unsigned value, INT32 length)
	Send a value on a given number of bits. IN assertion: length <= 16 and value fits in length bits.Reverse the first len bits of a code, using straightforward code (a faster method would use a table) IN assertion: 1 <= len <= 15.
void	bi_windup (DeflateState *s)
	Write out any remaining bits in an incomplete byte. Flush the bit buffer and align the output on a byte boundary.
void	bi_flush (DeflateState *s)
	Flush the bit buffer, keeping at most 7 bits in it. New Version 0.99.
void	copy_block (DeflateState *s, char *buf, unsigned len, INT32 header)
	Copy a stored block, storing first the length and its one's complement if requested.

---

Detailed Description

The deflater for the file compressor and decompressor. Allows the CCDiskFile to offer compression of files. At present just used to compress the native file format.

Author:

Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Humprhys

Date:

23/5/95

Definition at line 409 of file zdeflate.h.

---

Constructor & Destructor Documentation

ZipDeflate::ZipDeflate (   )

ZipDeflate constructor. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 34/05/95 Definition at line 259 of file zdeflate.cpp. { }

ZipDeflate::~ZipDeflate (   )

ZipDeflate destructor. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 34/05/95 Definition at line 273 of file zdeflate.cpp. { }

---

Member Function Documentation

void ZipDeflate::_tr_align (  DeflateState *  s  )  [protected]

Definition at line 988 of file zdftrees.cpp. { 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 last real code (before * the EOB of the previous block) was thus at least one plus the length * of the EOB plus what we have just sent of the empty static block. */ if (1 + 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; }

ulg ZipDeflate::_tr_flush_block (  DeflateState *  s, charf *  buf, ulg  stored_len, INT32  eof )  [protected]

void ZipDeflate::_tr_init (  DeflateState *  s  )  [protected]

Initialize the tree data structures for a new zlib stream. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Definition at line 405 of file zdftrees.cpp. { tr_static_init(); 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 s->bits_sent = 0L; #endif /* Initialize the first block of the first file: */ init_block(s); }

void ZipDeflate::_tr_stored_block (  DeflateState *  s, charf *  buf, ulg  stored_len, INT32  eof )  [protected]

INT32 ZipDeflate::_tr_tally (  DeflateState *  s, unsigned  dist, unsigned  lc )  [protected]

Save the match info and tally the frequency counts. Return true if the current block must be flushed. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Parameters: s  the current deflate state [INPUTS] dist distance of matched string lc match length-MIN_MATCH or unmatched char (if dist==0) Definition at line 1143 of file zdftrees.cpp. { 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, "_tr_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)((INT32)s->strstart - s->block_start); INT32 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. */ }

void ZipDeflate::bi_flush (  DeflateState *  s  )  [private]

Flush the bit buffer, keeping at most 7 bits in it. New Version 0.99. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Parameters: s  the current deflate state [INPUTS] Definition at line 1340 of file zdftrees.cpp. { 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; } }

unsigned ZipDeflate::bi_reverse (  unsigned  code, INT32  len )  [private]

Send a value on a given number of bits. IN assertion: length <= 16 and value fits in length bits.Reverse the first len bits of a code, using straightforward code (a faster method would use a table) IN assertion: 1 <= len <= 15. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Parameters: code  the value to invert [INPUTS] len its bit length Definition at line 1318 of file zdftrees.cpp. { register unsigned res = 0; do { res |= code & 1; code >>= 1, res <<= 1; } while (--len > 0); return res >> 1; }

void ZipDeflate::bi_windup (  DeflateState *  s  )  [private]

Write out any remaining bits in an incomplete byte. Flush the bit buffer and align the output on a byte boundary. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Parameters: s  the current deflate state [INPUTS] Definition at line 1365 of file zdftrees.cpp. { 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 s->bits_sent = (s->bits_sent+7) & ~7; #endif }

INT32 ZipDeflate::build_bl_tree (  DeflateState *  s  )  [private]

Construct the Huffman tree for the bit lengths and return the index in bl_order of the last bit length code to send. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Parameters: s  the current deflate state [INPUTS] Definition at line 886 of file zdftrees.cpp. { INT32 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; }

void ZipDeflate::build_tree (  DeflateState *  s, tree_desc *  desc )  [private]

Construct one Huffman tree and assigns the code bit strings and lengths. Update the total bit length for the current block. IN assertion: the field freq is set for all tree elements. OUT assertions: the fields len and code are set to the optimal bit length and corresponding code. The length opt_len is updated; static_len is also updated if stree is not null. The field max_code is set. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Parameters: s  current deflate state [INPUTS] desc the tree descriptor Definition at line 682 of file zdftrees.cpp. { ct_data *tree = desc->dyn_tree; ct_data *stree = desc->stat_desc->static_tree; INT32 elems = desc->stat_desc->elems; INT32 n, m; /* iterate over heap elements */ INT32 max_code = -1; /* largest code with non zero frequency */ INT32 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) { _ftprintf(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); }

ZipDeflate::CC_DECLARE_MEMDUMP (  ZipDeflate   )  [private]

void ZipDeflate::compress_block (  DeflateState *  s, ct_data *  ltree, ct_data *  dtree )  [private]

Send the block data compressed using the given Huffman trees. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Parameters: s  the current deflate state [INPUTS] ltree literal tree dtree distance tree Definition at line 1198 of file zdftrees.cpp. { unsigned dist; /* distance of matched string */ INT32 lc; /* match length or unmatched char (if dist == 0) */ UINT32 lx = 0; /* running index in l_buf */ unsigned code; /* the code to send */ INT32 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(_istgraph(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 < ( INT32 ) ( 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; }

void ZipDeflate::copy_block (  DeflateState *  s, char *  buf, unsigned  len, INT32  header )  [private]

Copy a stored block, storing first the length and its one's complement if requested. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Parameters: s  the current deflate state [INPUTS] buf the input data len its length header true if block header must be written Definition at line 1394 of file zdftrees.cpp. { 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 s->bits_sent += 2*16; #endif } #ifdef DEBUG s->bits_sent += (ulg)len<<3; #endif while (len--) { put_byte(s, *buf++); } }

ulg ZipDeflate::ct_flush_block (  DeflateState *  s, char *  buf, ulg  stored_len, INT32  eof )  [protected]

void ZipDeflate::ct_init (  DeflateState *  s  )  [protected]

void ZipDeflate::ct_stored_block (  DeflateState *  s, char *  buf, ulg  stored_len, INT32  eof )  [protected]

INT32 ZipDeflate::ct_tally (  DeflateState *  s, INT32  dist, INT32  lc )  [protected]

INT32 ZipDeflate::deflate (  z_stream *  strm, INT32  flush )

Actually go and deflate the stream. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Performs one or both of the following actions: Compress more input starting at next_in and update next_in and avail_in accordingly. If not all input can be processed (because there is not enough room in the output buffer), next_in and avail_in are updated and processing will resume at this point for the next call of deflate(). Provide more output starting at next_out and update next_out and avail_out accordingly. This action is forced if the parameter flush is non zero. Forcing flush frequently degrades the compression ratio, so this parameter should be set only when necessary (in interactive applications). Some output may be provided even if flush is not set. Before the call of deflate(), the application should ensure that at least one of the actions is possible, by providing more input and/or consuming more output, and updating avail_in or avail_out accordingly; avail_out should never be zero before the call. The application can consume the compressed output when it wants, for example when the output buffer is full (avail_out == 0), or after each call of deflate(). If deflate returns Z_OK and with zero avail_out, it must be called again after making room in the output buffer because there might be more output pending. If the parameter flush is set to Z_PARTIAL_FLUSH, the current compression block is terminated and flushed to the output buffer so that the decompressor can get all input data available so far. For method 9, a future variant on method 8, the current block will be flushed but not terminated. If flush is set to Z_FULL_FLUSH, the compression block is terminated, a special marker is output and the compression dictionary is discarded; this is useful to allow the decompressor to synchronize if one compressed block has been damaged (see inflateSync below). Flushing degrades compression and so should be used only when necessary. Using Z_FULL_FLUSH too often can seriously degrade the compression. If the parameter flush is set to Z_FINISH, all pending input is processed, all pending output is flushed and deflate returns with ZStream_END if there was enough output space; if deflate returns with Z_OK, this function must be called again with Z_FINISH and more output space (updated avail_out) but no more input data, until it returns with ZStream_END or an error. After deflate has returned ZStream_END, the only possible operations on the stream are deflateReset or deflateEnd. Z_FINISH can be used immediately after deflateInit if all the compression is to be done in a single step. In this case, avail_out must be at least 0.1% larger than avail_in plus 12 bytes. If deflate does not return ZStream_END, then it must be called again as described above. deflate() may update data_type if it can make a good guess about the input data type (Z_ASCII or Z_BINARY). In doubt, the data is considered binary. This field is only for information purposes and does not affect the compression algorithm in any manner. deflate() returns Z_OK if some progress has been made (more input processed or more output produced), ZStream_END if all input has been consumed and all output has been produced (only when flush is set to Z_FINISH), Z_STREAM_ERROR if the stream state was inconsistent (for example if next_in or next_out was NULL), Z_BUF_ERROR if no progress is possible. Definition at line 779 of file zdeflate.cpp. { INT32 old_flush; /* value of flush param for previous deflate call */ DeflateState *s; if (strm == Z_NULL || strm->De_state == Z_NULL || flush > Z_FINISH || flush < 0) { return Z_STREAM_ERROR; } s = strm->De_state; if (strm->next_out == Z_NULL || (strm->next_in == Z_NULL && strm->avail_in != 0) || (s->status == FINISH_STATE && flush != Z_FINISH)) { ERR_RETURN(strm, Z_STREAM_ERROR); } if (strm->avail_out == 0) ERR_RETURN(strm, Z_BUF_ERROR); strm->De_state->strm = strm; /* just in case */ old_flush = s->last_flush; s->last_flush = flush; /* Write the zlib header */ if (s->status == INIT_STATE) { uInt header = (Z_DEFLATED + ((s->w_bits-8)<<4)) << 8; uInt level_flags = (s->level-1) >> 1; if (level_flags > 3) level_flags = 3; header |= (level_flags << 6); if (s->strstart != 0) header |= PRESET_DICT; header += 31 - (header % 31); s->status = BUSY_STATE; putShortMSB(s, header); /* Save the adler32 of the preset dictionary: */ if (s->strstart != 0) { putShortMSB(s, (uInt)(strm->adler >> 16)); putShortMSB(s, (uInt)(strm->adler & 0xffff)); } strm->adler = 1L; } /* Flush as much pending output as possible */ if (s->pending != 0) { flush_pending(strm); if (strm->avail_out == 0) { /* Since avail_out is 0, deflate will be called again with * more output space, but possibly with both pending and * avail_in equal to zero. There won't be anything to do, * but this is not an error situation so make sure we * return OK instead of BUF_ERROR at next call of deflate: */ s->last_flush = -1; return Z_OK; } /* Make sure there is something to do and avoid duplicate consecutive * flushes. For repeated and useless calls with Z_FINISH, we keep * returning Z_STREAM_END instead of Z_BUFF_ERROR. */ } else if (strm->avail_in == 0 && flush <= old_flush && flush != Z_FINISH) { ERR_RETURN(strm, Z_BUF_ERROR); } /* User must not provide more input after the first FINISH: */ if (s->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 || s->lookahead != 0 || (flush != Z_NO_FLUSH && s->status != FINISH_STATE)) { INT32 quit; if (flush == Z_FINISH) { s->status = FINISH_STATE; } //quit = (*(configuration_table[s->level].func))(s, flush); switch (configuration_table[s->level].func) { case Z_deflate_stored : quit = deflate_stored(s, flush); break; case Z_deflate_fast : quit = deflate_fast(s, flush); break; case Z_deflate_slow : quit = deflate_slow(s, flush); break; default: // We have a type we don't understand so error return Z_STREAM_ERROR; } if (strm->avail_out == 0) { s->last_flush = -1; /* avoid BUF_ERROR at next call, see above */ } 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 (flush != Z_NO_FLUSH && flush != Z_FINISH) { if (flush == Z_PARTIAL_FLUSH) { _tr_align(s); } else { /* FULL_FLUSH or SYNC_FLUSH */ _tr_stored_block(s, (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(s); /* forget history */ } } flush_pending(strm); if (strm->avail_out == 0) { s->last_flush = -1; /* avoid BUF_ERROR at next call, see above */ return Z_OK; } } } Assert(strm->avail_out > 0, "bug2"); if (flush != Z_FINISH) return Z_OK; if (s->noheader) return Z_STREAM_END; /* Write the zlib trailer (adler32) */ putShortMSB(s, (uInt)(strm->adler >> 16)); putShortMSB(s, (uInt)(strm->adler & 0xffff)); flush_pending(strm); /* If avail_out is zero, the application will call deflate again * to flush the rest. */ s->noheader = -1; /* write the trailer only once! */ return s->pending != 0 ? Z_OK : Z_STREAM_END; }

INT32 ZipDeflate::deflate_fast (  DeflateState *  s, INT32  flush )  [private]

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. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Definition at line 1494 of file zdeflate.cpp. { IPos hash_head = NIL; /* head of the hash chain */ INT32 bflush; /* set if current block must be flushed */ 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 >= MIN_MATCH) { check_match(s, s->strstart, s->match_start, s->match_length); bflush = _tr_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 = _tr_tally (s, 0, s->window[s->strstart]); s->lookahead--; s->strstart++; } if (bflush) FLUSH_BLOCK(s, 0); } FLUSH_BLOCK(s, flush == Z_FINISH); return 0; /* normal exit */ }

INT32 ZipDeflate::deflate_slow (  DeflateState *  s, INT32  flush )  [private]

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. Author: Neville_Humphrys (Xara Group Ltd) <camelotdev@xara.com> Date: 24/05/95 Definition at line 1591 of file zdeflate.cpp. { IPos hash_head = NIL; /* head of hash chain */ INT32 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->