2017-05-03 15:35:00 +02:00
|
|
|
'use strict';
|
|
|
|
|
2017-12-10 21:51:33 +01:00
|
|
|
// (C) 1995-2013 Jean-loup Gailly and Mark Adler
|
|
|
|
// (C) 2014-2017 Vitaly Puzrin and Andrey Tupitsin
|
|
|
|
//
|
|
|
|
// This software is provided 'as-is', without any express or implied
|
|
|
|
// warranty. In no event will the authors be held liable for any damages
|
|
|
|
// arising from the use of this software.
|
|
|
|
//
|
|
|
|
// Permission is granted to anyone to use this software for any purpose,
|
|
|
|
// including commercial applications, and to alter it and redistribute it
|
|
|
|
// freely, subject to the following restrictions:
|
|
|
|
//
|
|
|
|
// 1. The origin of this software must not be misrepresented; you must not
|
|
|
|
// claim that you wrote the original software. If you use this software
|
|
|
|
// in a product, an acknowledgment in the product documentation would be
|
|
|
|
// appreciated but is not required.
|
|
|
|
// 2. Altered source versions must be plainly marked as such, and must not be
|
|
|
|
// misrepresented as being the original software.
|
|
|
|
// 3. This notice may not be removed or altered from any source distribution.
|
2017-05-03 15:35:00 +02:00
|
|
|
|
|
|
|
var utils = require('../utils/common');
|
|
|
|
var adler32 = require('./adler32');
|
|
|
|
var crc32 = require('./crc32');
|
|
|
|
var inflate_fast = require('./inffast');
|
|
|
|
var inflate_table = require('./inftrees');
|
|
|
|
|
|
|
|
var CODES = 0;
|
|
|
|
var LENS = 1;
|
|
|
|
var DISTS = 2;
|
|
|
|
|
|
|
|
/* Public constants ==========================================================*/
|
|
|
|
/* ===========================================================================*/
|
|
|
|
|
|
|
|
|
|
|
|
/* Allowed flush values; see deflate() and inflate() below for details */
|
|
|
|
//var Z_NO_FLUSH = 0;
|
|
|
|
//var Z_PARTIAL_FLUSH = 1;
|
|
|
|
//var Z_SYNC_FLUSH = 2;
|
|
|
|
//var Z_FULL_FLUSH = 3;
|
|
|
|
var Z_FINISH = 4;
|
|
|
|
var Z_BLOCK = 5;
|
|
|
|
var Z_TREES = 6;
|
|
|
|
|
|
|
|
|
|
|
|
/* Return codes for the compression/decompression functions. Negative values
|
|
|
|
* are errors, positive values are used for special but normal events.
|
|
|
|
*/
|
|
|
|
var Z_OK = 0;
|
|
|
|
var Z_STREAM_END = 1;
|
|
|
|
var Z_NEED_DICT = 2;
|
|
|
|
//var Z_ERRNO = -1;
|
|
|
|
var Z_STREAM_ERROR = -2;
|
|
|
|
var Z_DATA_ERROR = -3;
|
|
|
|
var Z_MEM_ERROR = -4;
|
|
|
|
var Z_BUF_ERROR = -5;
|
|
|
|
//var Z_VERSION_ERROR = -6;
|
|
|
|
|
|
|
|
/* The deflate compression method */
|
|
|
|
var Z_DEFLATED = 8;
|
|
|
|
|
|
|
|
|
|
|
|
/* STATES ====================================================================*/
|
|
|
|
/* ===========================================================================*/
|
|
|
|
|
|
|
|
|
|
|
|
var HEAD = 1; /* i: waiting for magic header */
|
|
|
|
var FLAGS = 2; /* i: waiting for method and flags (gzip) */
|
|
|
|
var TIME = 3; /* i: waiting for modification time (gzip) */
|
|
|
|
var OS = 4; /* i: waiting for extra flags and operating system (gzip) */
|
|
|
|
var EXLEN = 5; /* i: waiting for extra length (gzip) */
|
|
|
|
var EXTRA = 6; /* i: waiting for extra bytes (gzip) */
|
|
|
|
var NAME = 7; /* i: waiting for end of file name (gzip) */
|
|
|
|
var COMMENT = 8; /* i: waiting for end of comment (gzip) */
|
|
|
|
var HCRC = 9; /* i: waiting for header crc (gzip) */
|
|
|
|
var DICTID = 10; /* i: waiting for dictionary check value */
|
|
|
|
var DICT = 11; /* waiting for inflateSetDictionary() call */
|
|
|
|
var TYPE = 12; /* i: waiting for type bits, including last-flag bit */
|
|
|
|
var TYPEDO = 13; /* i: same, but skip check to exit inflate on new block */
|
|
|
|
var STORED = 14; /* i: waiting for stored size (length and complement) */
|
|
|
|
var COPY_ = 15; /* i/o: same as COPY below, but only first time in */
|
|
|
|
var COPY = 16; /* i/o: waiting for input or output to copy stored block */
|
|
|
|
var TABLE = 17; /* i: waiting for dynamic block table lengths */
|
|
|
|
var LENLENS = 18; /* i: waiting for code length code lengths */
|
|
|
|
var CODELENS = 19; /* i: waiting for length/lit and distance code lengths */
|
|
|
|
var LEN_ = 20; /* i: same as LEN below, but only first time in */
|
|
|
|
var LEN = 21; /* i: waiting for length/lit/eob code */
|
|
|
|
var LENEXT = 22; /* i: waiting for length extra bits */
|
|
|
|
var DIST = 23; /* i: waiting for distance code */
|
|
|
|
var DISTEXT = 24; /* i: waiting for distance extra bits */
|
|
|
|
var MATCH = 25; /* o: waiting for output space to copy string */
|
|
|
|
var LIT = 26; /* o: waiting for output space to write literal */
|
|
|
|
var CHECK = 27; /* i: waiting for 32-bit check value */
|
|
|
|
var LENGTH = 28; /* i: waiting for 32-bit length (gzip) */
|
|
|
|
var DONE = 29; /* finished check, done -- remain here until reset */
|
|
|
|
var BAD = 30; /* got a data error -- remain here until reset */
|
|
|
|
var MEM = 31; /* got an inflate() memory error -- remain here until reset */
|
|
|
|
var SYNC = 32; /* looking for synchronization bytes to restart inflate() */
|
|
|
|
|
|
|
|
/* ===========================================================================*/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
var ENOUGH_LENS = 852;
|
|
|
|
var ENOUGH_DISTS = 592;
|
|
|
|
//var ENOUGH = (ENOUGH_LENS+ENOUGH_DISTS);
|
|
|
|
|
|
|
|
var MAX_WBITS = 15;
|
|
|
|
/* 32K LZ77 window */
|
|
|
|
var DEF_WBITS = MAX_WBITS;
|
|
|
|
|
|
|
|
|
|
|
|
function zswap32(q) {
|
|
|
|
return (((q >>> 24) & 0xff) +
|
|
|
|
((q >>> 8) & 0xff00) +
|
|
|
|
((q & 0xff00) << 8) +
|
|
|
|
((q & 0xff) << 24));
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
function InflateState() {
|
|
|
|
this.mode = 0; /* current inflate mode */
|
|
|
|
this.last = false; /* true if processing last block */
|
|
|
|
this.wrap = 0; /* bit 0 true for zlib, bit 1 true for gzip */
|
|
|
|
this.havedict = false; /* true if dictionary provided */
|
|
|
|
this.flags = 0; /* gzip header method and flags (0 if zlib) */
|
|
|
|
this.dmax = 0; /* zlib header max distance (INFLATE_STRICT) */
|
|
|
|
this.check = 0; /* protected copy of check value */
|
|
|
|
this.total = 0; /* protected copy of output count */
|
|
|
|
// TODO: may be {}
|
|
|
|
this.head = null; /* where to save gzip header information */
|
|
|
|
|
|
|
|
/* sliding window */
|
|
|
|
this.wbits = 0; /* log base 2 of requested window size */
|
|
|
|
this.wsize = 0; /* window size or zero if not using window */
|
|
|
|
this.whave = 0; /* valid bytes in the window */
|
|
|
|
this.wnext = 0; /* window write index */
|
|
|
|
this.window = null; /* allocated sliding window, if needed */
|
|
|
|
|
|
|
|
/* bit accumulator */
|
|
|
|
this.hold = 0; /* input bit accumulator */
|
|
|
|
this.bits = 0; /* number of bits in "in" */
|
|
|
|
|
|
|
|
/* for string and stored block copying */
|
|
|
|
this.length = 0; /* literal or length of data to copy */
|
|
|
|
this.offset = 0; /* distance back to copy string from */
|
|
|
|
|
|
|
|
/* for table and code decoding */
|
|
|
|
this.extra = 0; /* extra bits needed */
|
|
|
|
|
|
|
|
/* fixed and dynamic code tables */
|
|
|
|
this.lencode = null; /* starting table for length/literal codes */
|
|
|
|
this.distcode = null; /* starting table for distance codes */
|
|
|
|
this.lenbits = 0; /* index bits for lencode */
|
|
|
|
this.distbits = 0; /* index bits for distcode */
|
|
|
|
|
|
|
|
/* dynamic table building */
|
|
|
|
this.ncode = 0; /* number of code length code lengths */
|
|
|
|
this.nlen = 0; /* number of length code lengths */
|
|
|
|
this.ndist = 0; /* number of distance code lengths */
|
|
|
|
this.have = 0; /* number of code lengths in lens[] */
|
|
|
|
this.next = null; /* next available space in codes[] */
|
|
|
|
|
|
|
|
this.lens = new utils.Buf16(320); /* temporary storage for code lengths */
|
|
|
|
this.work = new utils.Buf16(288); /* work area for code table building */
|
|
|
|
|
|
|
|
/*
|
|
|
|
because we don't have pointers in js, we use lencode and distcode directly
|
|
|
|
as buffers so we don't need codes
|
|
|
|
*/
|
|
|
|
//this.codes = new utils.Buf32(ENOUGH); /* space for code tables */
|
|
|
|
this.lendyn = null; /* dynamic table for length/literal codes (JS specific) */
|
|
|
|
this.distdyn = null; /* dynamic table for distance codes (JS specific) */
|
|
|
|
this.sane = 0; /* if false, allow invalid distance too far */
|
|
|
|
this.back = 0; /* bits back of last unprocessed length/lit */
|
|
|
|
this.was = 0; /* initial length of match */
|
|
|
|
}
|
|
|
|
|
|
|
|
function inflateResetKeep(strm) {
|
|
|
|
var state;
|
|
|
|
|
|
|
|
if (!strm || !strm.state) { return Z_STREAM_ERROR; }
|
|
|
|
state = strm.state;
|
|
|
|
strm.total_in = strm.total_out = state.total = 0;
|
|
|
|
strm.msg = ''; /*Z_NULL*/
|
|
|
|
if (state.wrap) { /* to support ill-conceived Java test suite */
|
|
|
|
strm.adler = state.wrap & 1;
|
|
|
|
}
|
|
|
|
state.mode = HEAD;
|
|
|
|
state.last = 0;
|
|
|
|
state.havedict = 0;
|
|
|
|
state.dmax = 32768;
|
|
|
|
state.head = null/*Z_NULL*/;
|
|
|
|
state.hold = 0;
|
|
|
|
state.bits = 0;
|
|
|
|
//state.lencode = state.distcode = state.next = state.codes;
|
|
|
|
state.lencode = state.lendyn = new utils.Buf32(ENOUGH_LENS);
|
|
|
|
state.distcode = state.distdyn = new utils.Buf32(ENOUGH_DISTS);
|
|
|
|
|
|
|
|
state.sane = 1;
|
|
|
|
state.back = -1;
|
|
|
|
//Tracev((stderr, "inflate: reset\n"));
|
|
|
|
return Z_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
function inflateReset(strm) {
|
|
|
|
var state;
|
|
|
|
|
|
|
|
if (!strm || !strm.state) { return Z_STREAM_ERROR; }
|
|
|
|
state = strm.state;
|
|
|
|
state.wsize = 0;
|
|
|
|
state.whave = 0;
|
|
|
|
state.wnext = 0;
|
|
|
|
return inflateResetKeep(strm);
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
function inflateReset2(strm, windowBits) {
|
|
|
|
var wrap;
|
|
|
|
var state;
|
|
|
|
|
|
|
|
/* get the state */
|
|
|
|
if (!strm || !strm.state) { return Z_STREAM_ERROR; }
|
|
|
|
state = strm.state;
|
|
|
|
|
|
|
|
/* extract wrap request from windowBits parameter */
|
|
|
|
if (windowBits < 0) {
|
|
|
|
wrap = 0;
|
|
|
|
windowBits = -windowBits;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
wrap = (windowBits >> 4) + 1;
|
|
|
|
if (windowBits < 48) {
|
|
|
|
windowBits &= 15;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* set number of window bits, free window if different */
|
|
|
|
if (windowBits && (windowBits < 8 || windowBits > 15)) {
|
|
|
|
return Z_STREAM_ERROR;
|
|
|
|
}
|
|
|
|
if (state.window !== null && state.wbits !== windowBits) {
|
|
|
|
state.window = null;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* update state and reset the rest of it */
|
|
|
|
state.wrap = wrap;
|
|
|
|
state.wbits = windowBits;
|
|
|
|
return inflateReset(strm);
|
|
|
|
}
|
|
|
|
|
|
|
|
function inflateInit2(strm, windowBits) {
|
|
|
|
var ret;
|
|
|
|
var state;
|
|
|
|
|
|
|
|
if (!strm) { return Z_STREAM_ERROR; }
|
|
|
|
//strm.msg = Z_NULL; /* in case we return an error */
|
|
|
|
|
|
|
|
state = new InflateState();
|
|
|
|
|
|
|
|
//if (state === Z_NULL) return Z_MEM_ERROR;
|
|
|
|
//Tracev((stderr, "inflate: allocated\n"));
|
|
|
|
strm.state = state;
|
|
|
|
state.window = null/*Z_NULL*/;
|
|
|
|
ret = inflateReset2(strm, windowBits);
|
|
|
|
if (ret !== Z_OK) {
|
|
|
|
strm.state = null/*Z_NULL*/;
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
function inflateInit(strm) {
|
|
|
|
return inflateInit2(strm, DEF_WBITS);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Return state with length and distance decoding tables and index sizes set to
|
|
|
|
fixed code decoding. Normally this returns fixed tables from inffixed.h.
|
|
|
|
If BUILDFIXED is defined, then instead this routine builds the tables the
|
|
|
|
first time it's called, and returns those tables the first time and
|
|
|
|
thereafter. This reduces the size of the code by about 2K bytes, in
|
|
|
|
exchange for a little execution time. However, BUILDFIXED should not be
|
|
|
|
used for threaded applications, since the rewriting of the tables and virgin
|
|
|
|
may not be thread-safe.
|
|
|
|
*/
|
|
|
|
var virgin = true;
|
|
|
|
|
|
|
|
var lenfix, distfix; // We have no pointers in JS, so keep tables separate
|
|
|
|
|
|
|
|
function fixedtables(state) {
|
|
|
|
/* build fixed huffman tables if first call (may not be thread safe) */
|
|
|
|
if (virgin) {
|
|
|
|
var sym;
|
|
|
|
|
|
|
|
lenfix = new utils.Buf32(512);
|
|
|
|
distfix = new utils.Buf32(32);
|
|
|
|
|
|
|
|
/* literal/length table */
|
|
|
|
sym = 0;
|
|
|
|
while (sym < 144) { state.lens[sym++] = 8; }
|
|
|
|
while (sym < 256) { state.lens[sym++] = 9; }
|
|
|
|
while (sym < 280) { state.lens[sym++] = 7; }
|
|
|
|
while (sym < 288) { state.lens[sym++] = 8; }
|
|
|
|
|
|
|
|
inflate_table(LENS, state.lens, 0, 288, lenfix, 0, state.work, { bits: 9 });
|
|
|
|
|
|
|
|
/* distance table */
|
|
|
|
sym = 0;
|
|
|
|
while (sym < 32) { state.lens[sym++] = 5; }
|
|
|
|
|
|
|
|
inflate_table(DISTS, state.lens, 0, 32, distfix, 0, state.work, { bits: 5 });
|
|
|
|
|
|
|
|
/* do this just once */
|
|
|
|
virgin = false;
|
|
|
|
}
|
|
|
|
|
|
|
|
state.lencode = lenfix;
|
|
|
|
state.lenbits = 9;
|
|
|
|
state.distcode = distfix;
|
|
|
|
state.distbits = 5;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
Update the window with the last wsize (normally 32K) bytes written before
|
|
|
|
returning. If window does not exist yet, create it. This is only called
|
|
|
|
when a window is already in use, or when output has been written during this
|
|
|
|
inflate call, but the end of the deflate stream has not been reached yet.
|
|
|
|
It is also called to create a window for dictionary data when a dictionary
|
|
|
|
is loaded.
|
|
|
|
|
|
|
|
Providing output buffers larger than 32K to inflate() should provide a speed
|
|
|
|
advantage, since only the last 32K of output is copied to the sliding window
|
|
|
|
upon return from inflate(), and since all distances after the first 32K of
|
|
|
|
output will fall in the output data, making match copies simpler and faster.
|
|
|
|
The advantage may be dependent on the size of the processor's data caches.
|
|
|
|
*/
|
|
|
|
function updatewindow(strm, src, end, copy) {
|
|
|
|
var dist;
|
|
|
|
var state = strm.state;
|
|
|
|
|
|
|
|
/* if it hasn't been done already, allocate space for the window */
|
|
|
|
if (state.window === null) {
|
|
|
|
state.wsize = 1 << state.wbits;
|
|
|
|
state.wnext = 0;
|
|
|
|
state.whave = 0;
|
|
|
|
|
|
|
|
state.window = new utils.Buf8(state.wsize);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* copy state->wsize or less output bytes into the circular window */
|
|
|
|
if (copy >= state.wsize) {
|
|
|
|
utils.arraySet(state.window, src, end - state.wsize, state.wsize, 0);
|
|
|
|
state.wnext = 0;
|
|
|
|
state.whave = state.wsize;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
dist = state.wsize - state.wnext;
|
|
|
|
if (dist > copy) {
|
|
|
|
dist = copy;
|
|
|
|
}
|
|
|
|
//zmemcpy(state->window + state->wnext, end - copy, dist);
|
|
|
|
utils.arraySet(state.window, src, end - copy, dist, state.wnext);
|
|
|
|
copy -= dist;
|
|
|
|
if (copy) {
|
|
|
|
//zmemcpy(state->window, end - copy, copy);
|
|
|
|
utils.arraySet(state.window, src, end - copy, copy, 0);
|
|
|
|
state.wnext = copy;
|
|
|
|
state.whave = state.wsize;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
state.wnext += dist;
|
|
|
|
if (state.wnext === state.wsize) { state.wnext = 0; }
|
|
|
|
if (state.whave < state.wsize) { state.whave += dist; }
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
function inflate(strm, flush) {
|
|
|
|
var state;
|
|
|
|
var input, output; // input/output buffers
|
|
|
|
var next; /* next input INDEX */
|
|
|
|
var put; /* next output INDEX */
|
|
|
|
var have, left; /* available input and output */
|
|
|
|
var hold; /* bit buffer */
|
|
|
|
var bits; /* bits in bit buffer */
|
|
|
|
var _in, _out; /* save starting available input and output */
|
|
|
|
var copy; /* number of stored or match bytes to copy */
|
|
|
|
var from; /* where to copy match bytes from */
|
|
|
|
var from_source;
|
|
|
|
var here = 0; /* current decoding table entry */
|
|
|
|
var here_bits, here_op, here_val; // paked "here" denormalized (JS specific)
|
|
|
|
//var last; /* parent table entry */
|
|
|
|
var last_bits, last_op, last_val; // paked "last" denormalized (JS specific)
|
|
|
|
var len; /* length to copy for repeats, bits to drop */
|
|
|
|
var ret; /* return code */
|
|
|
|
var hbuf = new utils.Buf8(4); /* buffer for gzip header crc calculation */
|
|
|
|
var opts;
|
|
|
|
|
|
|
|
var n; // temporary var for NEED_BITS
|
|
|
|
|
|
|
|
var order = /* permutation of code lengths */
|
|
|
|
[ 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15 ];
|
|
|
|
|
|
|
|
|
|
|
|
if (!strm || !strm.state || !strm.output ||
|
|
|
|
(!strm.input && strm.avail_in !== 0)) {
|
|
|
|
return Z_STREAM_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
state = strm.state;
|
|
|
|
if (state.mode === TYPE) { state.mode = TYPEDO; } /* skip check */
|
|
|
|
|
|
|
|
|
|
|
|
//--- LOAD() ---
|
|
|
|
put = strm.next_out;
|
|
|
|
output = strm.output;
|
|
|
|
left = strm.avail_out;
|
|
|
|
next = strm.next_in;
|
|
|
|
input = strm.input;
|
|
|
|
have = strm.avail_in;
|
|
|
|
hold = state.hold;
|
|
|
|
bits = state.bits;
|
|
|
|
//---
|
|
|
|
|
|
|
|
_in = have;
|
|
|
|
_out = left;
|
|
|
|
ret = Z_OK;
|
|
|
|
|
|
|
|
inf_leave: // goto emulation
|
|
|
|
for (;;) {
|
|
|
|
switch (state.mode) {
|
2017-12-10 21:51:33 +01:00
|
|
|
case HEAD:
|
|
|
|
if (state.wrap === 0) {
|
|
|
|
state.mode = TYPEDO;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
//=== NEEDBITS(16);
|
|
|
|
while (bits < 16) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
2017-05-03 15:35:00 +02:00
|
|
|
//===//
|
2017-12-10 21:51:33 +01:00
|
|
|
if ((state.wrap & 2) && hold === 0x8b1f) { /* gzip header */
|
|
|
|
state.check = 0/*crc32(0L, Z_NULL, 0)*/;
|
|
|
|
//=== CRC2(state.check, hold);
|
|
|
|
hbuf[0] = hold & 0xff;
|
|
|
|
hbuf[1] = (hold >>> 8) & 0xff;
|
|
|
|
state.check = crc32(state.check, hbuf, 2, 0);
|
|
|
|
//===//
|
2017-05-03 15:35:00 +02:00
|
|
|
|
2017-12-10 21:51:33 +01:00
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
|
|
|
state.mode = FLAGS;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
state.flags = 0; /* expect zlib header */
|
|
|
|
if (state.head) {
|
|
|
|
state.head.done = false;
|
|
|
|
}
|
|
|
|
if (!(state.wrap & 1) || /* check if zlib header allowed */
|
|
|
|
(((hold & 0xff)/*BITS(8)*/ << 8) + (hold >> 8)) % 31) {
|
|
|
|
strm.msg = 'incorrect header check';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if ((hold & 0x0f)/*BITS(4)*/ !== Z_DEFLATED) {
|
|
|
|
strm.msg = 'unknown compression method';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
//--- DROPBITS(4) ---//
|
|
|
|
hold >>>= 4;
|
|
|
|
bits -= 4;
|
|
|
|
//---//
|
|
|
|
len = (hold & 0x0f)/*BITS(4)*/ + 8;
|
|
|
|
if (state.wbits === 0) {
|
|
|
|
state.wbits = len;
|
|
|
|
}
|
|
|
|
else if (len > state.wbits) {
|
|
|
|
strm.msg = 'invalid window size';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
state.dmax = 1 << len;
|
|
|
|
//Tracev((stderr, "inflate: zlib header ok\n"));
|
|
|
|
strm.adler = state.check = 1/*adler32(0L, Z_NULL, 0)*/;
|
|
|
|
state.mode = hold & 0x200 ? DICTID : TYPE;
|
2017-05-03 15:35:00 +02:00
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
|
|
|
break;
|
2017-12-10 21:51:33 +01:00
|
|
|
case FLAGS:
|
|
|
|
//=== NEEDBITS(16); */
|
|
|
|
while (bits < 16) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
2017-05-03 15:35:00 +02:00
|
|
|
//===//
|
2017-12-10 21:51:33 +01:00
|
|
|
state.flags = hold;
|
|
|
|
if ((state.flags & 0xff) !== Z_DEFLATED) {
|
|
|
|
strm.msg = 'unknown compression method';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (state.flags & 0xe000) {
|
|
|
|
strm.msg = 'unknown header flags set';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (state.head) {
|
|
|
|
state.head.text = ((hold >> 8) & 1);
|
|
|
|
}
|
|
|
|
if (state.flags & 0x0200) {
|
|
|
|
//=== CRC2(state.check, hold);
|
|
|
|
hbuf[0] = hold & 0xff;
|
|
|
|
hbuf[1] = (hold >>> 8) & 0xff;
|
|
|
|
state.check = crc32(state.check, hbuf, 2, 0);
|
|
|
|
//===//
|
|
|
|
}
|
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
|
|
|
state.mode = TIME;
|
|
|
|
/* falls through */
|
|
|
|
case TIME:
|
|
|
|
//=== NEEDBITS(32); */
|
|
|
|
while (bits < 32) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
2017-05-03 15:35:00 +02:00
|
|
|
//===//
|
2017-12-10 21:51:33 +01:00
|
|
|
if (state.head) {
|
|
|
|
state.head.time = hold;
|
|
|
|
}
|
|
|
|
if (state.flags & 0x0200) {
|
|
|
|
//=== CRC4(state.check, hold)
|
|
|
|
hbuf[0] = hold & 0xff;
|
|
|
|
hbuf[1] = (hold >>> 8) & 0xff;
|
|
|
|
hbuf[2] = (hold >>> 16) & 0xff;
|
|
|
|
hbuf[3] = (hold >>> 24) & 0xff;
|
|
|
|
state.check = crc32(state.check, hbuf, 4, 0);
|
|
|
|
//===
|
|
|
|
}
|
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
|
|
|
state.mode = OS;
|
|
|
|
/* falls through */
|
|
|
|
case OS:
|
2017-05-03 15:35:00 +02:00
|
|
|
//=== NEEDBITS(16); */
|
|
|
|
while (bits < 16) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
if (state.head) {
|
2017-12-10 21:51:33 +01:00
|
|
|
state.head.xflags = (hold & 0xff);
|
|
|
|
state.head.os = (hold >> 8);
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
|
|
|
if (state.flags & 0x0200) {
|
|
|
|
//=== CRC2(state.check, hold);
|
|
|
|
hbuf[0] = hold & 0xff;
|
|
|
|
hbuf[1] = (hold >>> 8) & 0xff;
|
|
|
|
state.check = crc32(state.check, hbuf, 2, 0);
|
|
|
|
//===//
|
|
|
|
}
|
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
2017-12-10 21:51:33 +01:00
|
|
|
state.mode = EXLEN;
|
|
|
|
/* falls through */
|
|
|
|
case EXLEN:
|
|
|
|
if (state.flags & 0x0400) {
|
|
|
|
//=== NEEDBITS(16); */
|
|
|
|
while (bits < 16) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
state.length = hold;
|
2017-05-03 15:35:00 +02:00
|
|
|
if (state.head) {
|
2017-12-10 21:51:33 +01:00
|
|
|
state.head.extra_len = hold;
|
|
|
|
}
|
|
|
|
if (state.flags & 0x0200) {
|
|
|
|
//=== CRC2(state.check, hold);
|
|
|
|
hbuf[0] = hold & 0xff;
|
|
|
|
hbuf[1] = (hold >>> 8) & 0xff;
|
|
|
|
state.check = crc32(state.check, hbuf, 2, 0);
|
|
|
|
//===//
|
|
|
|
}
|
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
|
|
|
}
|
|
|
|
else if (state.head) {
|
|
|
|
state.head.extra = null/*Z_NULL*/;
|
|
|
|
}
|
|
|
|
state.mode = EXTRA;
|
|
|
|
/* falls through */
|
|
|
|
case EXTRA:
|
|
|
|
if (state.flags & 0x0400) {
|
|
|
|
copy = state.length;
|
|
|
|
if (copy > have) { copy = have; }
|
|
|
|
if (copy) {
|
|
|
|
if (state.head) {
|
|
|
|
len = state.head.extra_len - state.length;
|
|
|
|
if (!state.head.extra) {
|
|
|
|
// Use untyped array for more convenient processing later
|
|
|
|
state.head.extra = new Array(state.head.extra_len);
|
|
|
|
}
|
|
|
|
utils.arraySet(
|
|
|
|
state.head.extra,
|
|
|
|
input,
|
|
|
|
next,
|
|
|
|
// extra field is limited to 65536 bytes
|
|
|
|
// - no need for additional size check
|
|
|
|
copy,
|
|
|
|
/*len + copy > state.head.extra_max - len ? state.head.extra_max : copy,*/
|
|
|
|
len
|
|
|
|
);
|
|
|
|
//zmemcpy(state.head.extra + len, next,
|
|
|
|
// len + copy > state.head.extra_max ?
|
|
|
|
// state.head.extra_max - len : copy);
|
|
|
|
}
|
|
|
|
if (state.flags & 0x0200) {
|
|
|
|
state.check = crc32(state.check, input, copy, next);
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
have -= copy;
|
|
|
|
next += copy;
|
|
|
|
state.length -= copy;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
if (state.length) { break inf_leave; }
|
|
|
|
}
|
|
|
|
state.length = 0;
|
|
|
|
state.mode = NAME;
|
|
|
|
/* falls through */
|
|
|
|
case NAME:
|
|
|
|
if (state.flags & 0x0800) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
copy = 0;
|
|
|
|
do {
|
|
|
|
// TODO: 2 or 1 bytes?
|
|
|
|
len = input[next + copy++];
|
|
|
|
/* use constant limit because in js we should not preallocate memory */
|
|
|
|
if (state.head && len &&
|
|
|
|
(state.length < 65536 /*state.head.name_max*/)) {
|
|
|
|
state.head.name += String.fromCharCode(len);
|
|
|
|
}
|
|
|
|
} while (len && copy < have);
|
|
|
|
|
2017-05-03 15:35:00 +02:00
|
|
|
if (state.flags & 0x0200) {
|
|
|
|
state.check = crc32(state.check, input, copy, next);
|
|
|
|
}
|
|
|
|
have -= copy;
|
|
|
|
next += copy;
|
2017-12-10 21:51:33 +01:00
|
|
|
if (len) { break inf_leave; }
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
else if (state.head) {
|
|
|
|
state.head.name = null;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
state.length = 0;
|
|
|
|
state.mode = COMMENT;
|
|
|
|
/* falls through */
|
|
|
|
case COMMENT:
|
|
|
|
if (state.flags & 0x1000) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
copy = 0;
|
|
|
|
do {
|
|
|
|
len = input[next + copy++];
|
|
|
|
/* use constant limit because in js we should not preallocate memory */
|
|
|
|
if (state.head && len &&
|
|
|
|
(state.length < 65536 /*state.head.comm_max*/)) {
|
|
|
|
state.head.comment += String.fromCharCode(len);
|
|
|
|
}
|
|
|
|
} while (len && copy < have);
|
|
|
|
if (state.flags & 0x0200) {
|
|
|
|
state.check = crc32(state.check, input, copy, next);
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
have -= copy;
|
|
|
|
next += copy;
|
|
|
|
if (len) { break inf_leave; }
|
|
|
|
}
|
|
|
|
else if (state.head) {
|
|
|
|
state.head.comment = null;
|
|
|
|
}
|
|
|
|
state.mode = HCRC;
|
|
|
|
/* falls through */
|
|
|
|
case HCRC:
|
2017-05-03 15:35:00 +02:00
|
|
|
if (state.flags & 0x0200) {
|
2017-12-10 21:51:33 +01:00
|
|
|
//=== NEEDBITS(16); */
|
|
|
|
while (bits < 16) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
if (hold !== (state.check & 0xffff)) {
|
|
|
|
strm.msg = 'header crc mismatch';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
if (state.head) {
|
|
|
|
state.head.hcrc = ((state.flags >> 9) & 1);
|
|
|
|
state.head.done = true;
|
|
|
|
}
|
|
|
|
strm.adler = state.check = 0;
|
|
|
|
state.mode = TYPE;
|
|
|
|
break;
|
|
|
|
case DICTID:
|
|
|
|
//=== NEEDBITS(32); */
|
|
|
|
while (bits < 32) {
|
2017-05-03 15:35:00 +02:00
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
2017-12-10 21:51:33 +01:00
|
|
|
strm.adler = state.check = zswap32(hold);
|
2017-05-03 15:35:00 +02:00
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
2017-12-10 21:51:33 +01:00
|
|
|
state.mode = DICT;
|
|
|
|
/* falls through */
|
|
|
|
case DICT:
|
|
|
|
if (state.havedict === 0) {
|
|
|
|
//--- RESTORE() ---
|
|
|
|
strm.next_out = put;
|
|
|
|
strm.avail_out = left;
|
|
|
|
strm.next_in = next;
|
|
|
|
strm.avail_in = have;
|
|
|
|
state.hold = hold;
|
|
|
|
state.bits = bits;
|
|
|
|
//---
|
|
|
|
return Z_NEED_DICT;
|
|
|
|
}
|
|
|
|
strm.adler = state.check = 1/*adler32(0L, Z_NULL, 0)*/;
|
|
|
|
state.mode = TYPE;
|
|
|
|
/* falls through */
|
|
|
|
case TYPE:
|
|
|
|
if (flush === Z_BLOCK || flush === Z_TREES) { break inf_leave; }
|
|
|
|
/* falls through */
|
|
|
|
case TYPEDO:
|
|
|
|
if (state.last) {
|
|
|
|
//--- BYTEBITS() ---//
|
|
|
|
hold >>>= bits & 7;
|
|
|
|
bits -= bits & 7;
|
2017-05-03 15:35:00 +02:00
|
|
|
//---//
|
2017-12-10 21:51:33 +01:00
|
|
|
state.mode = CHECK;
|
|
|
|
break;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
//=== NEEDBITS(3); */
|
2017-05-03 15:35:00 +02:00
|
|
|
while (bits < 3) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
2017-12-10 21:51:33 +01:00
|
|
|
state.last = (hold & 0x01)/*BITS(1)*/;
|
|
|
|
//--- DROPBITS(1) ---//
|
|
|
|
hold >>>= 1;
|
|
|
|
bits -= 1;
|
2017-05-03 15:35:00 +02:00
|
|
|
//---//
|
|
|
|
|
2017-12-10 21:51:33 +01:00
|
|
|
switch ((hold & 0x03)/*BITS(2)*/) {
|
|
|
|
case 0: /* stored block */
|
|
|
|
//Tracev((stderr, "inflate: stored block%s\n",
|
|
|
|
// state.last ? " (last)" : ""));
|
|
|
|
state.mode = STORED;
|
|
|
|
break;
|
|
|
|
case 1: /* fixed block */
|
|
|
|
fixedtables(state);
|
|
|
|
//Tracev((stderr, "inflate: fixed codes block%s\n",
|
|
|
|
// state.last ? " (last)" : ""));
|
|
|
|
state.mode = LEN_; /* decode codes */
|
|
|
|
if (flush === Z_TREES) {
|
|
|
|
//--- DROPBITS(2) ---//
|
|
|
|
hold >>>= 2;
|
|
|
|
bits -= 2;
|
|
|
|
//---//
|
|
|
|
break inf_leave;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
case 2: /* dynamic block */
|
|
|
|
//Tracev((stderr, "inflate: dynamic codes block%s\n",
|
|
|
|
// state.last ? " (last)" : ""));
|
|
|
|
state.mode = TABLE;
|
|
|
|
break;
|
|
|
|
case 3:
|
|
|
|
strm.msg = 'invalid block type';
|
|
|
|
state.mode = BAD;
|
|
|
|
}
|
|
|
|
//--- DROPBITS(2) ---//
|
|
|
|
hold >>>= 2;
|
|
|
|
bits -= 2;
|
|
|
|
//---//
|
|
|
|
break;
|
|
|
|
case STORED:
|
|
|
|
//--- BYTEBITS() ---// /* go to byte boundary */
|
|
|
|
hold >>>= bits & 7;
|
|
|
|
bits -= bits & 7;
|
|
|
|
//---//
|
|
|
|
//=== NEEDBITS(32); */
|
|
|
|
while (bits < 32) {
|
2017-05-03 15:35:00 +02:00
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
2017-12-10 21:51:33 +01:00
|
|
|
}
|
|
|
|
//===//
|
|
|
|
if ((hold & 0xffff) !== ((hold >>> 16) ^ 0xffff)) {
|
|
|
|
strm.msg = 'invalid stored block lengths';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
state.length = hold & 0xffff;
|
|
|
|
//Tracev((stderr, "inflate: stored length %u\n",
|
|
|
|
// state.length));
|
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
|
|
|
state.mode = COPY_;
|
|
|
|
if (flush === Z_TREES) { break inf_leave; }
|
|
|
|
/* falls through */
|
|
|
|
case COPY_:
|
|
|
|
state.mode = COPY;
|
|
|
|
/* falls through */
|
|
|
|
case COPY:
|
|
|
|
copy = state.length;
|
|
|
|
if (copy) {
|
|
|
|
if (copy > have) { copy = have; }
|
|
|
|
if (copy > left) { copy = left; }
|
|
|
|
if (copy === 0) { break inf_leave; }
|
|
|
|
//--- zmemcpy(put, next, copy); ---
|
|
|
|
utils.arraySet(output, input, next, copy, put);
|
2017-05-03 15:35:00 +02:00
|
|
|
//---//
|
2017-12-10 21:51:33 +01:00
|
|
|
have -= copy;
|
|
|
|
next += copy;
|
|
|
|
left -= copy;
|
|
|
|
put += copy;
|
|
|
|
state.length -= copy;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
//Tracev((stderr, "inflate: stored end\n"));
|
|
|
|
state.mode = TYPE;
|
|
|
|
break;
|
|
|
|
case TABLE:
|
|
|
|
//=== NEEDBITS(14); */
|
|
|
|
while (bits < 14) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
//===//
|
|
|
|
state.nlen = (hold & 0x1f)/*BITS(5)*/ + 257;
|
|
|
|
//--- DROPBITS(5) ---//
|
|
|
|
hold >>>= 5;
|
|
|
|
bits -= 5;
|
|
|
|
//---//
|
|
|
|
state.ndist = (hold & 0x1f)/*BITS(5)*/ + 1;
|
|
|
|
//--- DROPBITS(5) ---//
|
|
|
|
hold >>>= 5;
|
|
|
|
bits -= 5;
|
|
|
|
//---//
|
|
|
|
state.ncode = (hold & 0x0f)/*BITS(4)*/ + 4;
|
|
|
|
//--- DROPBITS(4) ---//
|
|
|
|
hold >>>= 4;
|
|
|
|
bits -= 4;
|
|
|
|
//---//
|
|
|
|
//#ifndef PKZIP_BUG_WORKAROUND
|
|
|
|
if (state.nlen > 286 || state.ndist > 30) {
|
|
|
|
strm.msg = 'too many length or distance symbols';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
//#endif
|
|
|
|
//Tracev((stderr, "inflate: table sizes ok\n"));
|
|
|
|
state.have = 0;
|
|
|
|
state.mode = LENLENS;
|
|
|
|
/* falls through */
|
|
|
|
case LENLENS:
|
|
|
|
while (state.have < state.ncode) {
|
|
|
|
//=== NEEDBITS(3);
|
|
|
|
while (bits < 3) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
state.lens[order[state.have++]] = (hold & 0x07);//BITS(3);
|
|
|
|
//--- DROPBITS(3) ---//
|
|
|
|
hold >>>= 3;
|
|
|
|
bits -= 3;
|
2017-05-03 15:35:00 +02:00
|
|
|
//---//
|
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
while (state.have < 19) {
|
|
|
|
state.lens[order[state.have++]] = 0;
|
|
|
|
}
|
|
|
|
// We have separate tables & no pointers. 2 commented lines below not needed.
|
|
|
|
//state.next = state.codes;
|
|
|
|
//state.lencode = state.next;
|
|
|
|
// Switch to use dynamic table
|
|
|
|
state.lencode = state.lendyn;
|
|
|
|
state.lenbits = 7;
|
|
|
|
|
|
|
|
opts = { bits: state.lenbits };
|
|
|
|
ret = inflate_table(CODES, state.lens, 0, 19, state.lencode, 0, state.work, opts);
|
|
|
|
state.lenbits = opts.bits;
|
|
|
|
|
|
|
|
if (ret) {
|
|
|
|
strm.msg = 'invalid code lengths set';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
//Tracev((stderr, "inflate: code lengths ok\n"));
|
|
|
|
state.have = 0;
|
|
|
|
state.mode = CODELENS;
|
|
|
|
/* falls through */
|
|
|
|
case CODELENS:
|
|
|
|
while (state.have < state.nlen + state.ndist) {
|
|
|
|
for (;;) {
|
|
|
|
here = state.lencode[hold & ((1 << state.lenbits) - 1)];/*BITS(state.lenbits)*/
|
|
|
|
here_bits = here >>> 24;
|
|
|
|
here_op = (here >>> 16) & 0xff;
|
|
|
|
here_val = here & 0xffff;
|
|
|
|
|
|
|
|
if ((here_bits) <= bits) { break; }
|
|
|
|
//--- PULLBYTE() ---//
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
2017-05-03 15:35:00 +02:00
|
|
|
//---//
|
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
if (here_val < 16) {
|
2017-05-03 15:35:00 +02:00
|
|
|
//--- DROPBITS(here.bits) ---//
|
|
|
|
hold >>>= here_bits;
|
|
|
|
bits -= here_bits;
|
|
|
|
//---//
|
2017-12-10 21:51:33 +01:00
|
|
|
state.lens[state.have++] = here_val;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
|
|
|
else {
|
2017-12-10 21:51:33 +01:00
|
|
|
if (here_val === 16) {
|
|
|
|
//=== NEEDBITS(here.bits + 2);
|
|
|
|
n = here_bits + 2;
|
|
|
|
while (bits < n) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
//--- DROPBITS(here.bits) ---//
|
|
|
|
hold >>>= here_bits;
|
|
|
|
bits -= here_bits;
|
|
|
|
//---//
|
|
|
|
if (state.have === 0) {
|
|
|
|
strm.msg = 'invalid bit length repeat';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
len = state.lens[state.have - 1];
|
|
|
|
copy = 3 + (hold & 0x03);//BITS(2);
|
|
|
|
//--- DROPBITS(2) ---//
|
|
|
|
hold >>>= 2;
|
|
|
|
bits -= 2;
|
|
|
|
//---//
|
|
|
|
}
|
|
|
|
else if (here_val === 17) {
|
|
|
|
//=== NEEDBITS(here.bits + 3);
|
|
|
|
n = here_bits + 3;
|
|
|
|
while (bits < n) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
//--- DROPBITS(here.bits) ---//
|
|
|
|
hold >>>= here_bits;
|
|
|
|
bits -= here_bits;
|
|
|
|
//---//
|
|
|
|
len = 0;
|
|
|
|
copy = 3 + (hold & 0x07);//BITS(3);
|
|
|
|
//--- DROPBITS(3) ---//
|
|
|
|
hold >>>= 3;
|
|
|
|
bits -= 3;
|
|
|
|
//---//
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
//=== NEEDBITS(here.bits + 7);
|
|
|
|
n = here_bits + 7;
|
|
|
|
while (bits < n) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
//--- DROPBITS(here.bits) ---//
|
|
|
|
hold >>>= here_bits;
|
|
|
|
bits -= here_bits;
|
|
|
|
//---//
|
|
|
|
len = 0;
|
|
|
|
copy = 11 + (hold & 0x7f);//BITS(7);
|
|
|
|
//--- DROPBITS(7) ---//
|
|
|
|
hold >>>= 7;
|
|
|
|
bits -= 7;
|
|
|
|
//---//
|
|
|
|
}
|
|
|
|
if (state.have + copy > state.nlen + state.ndist) {
|
|
|
|
strm.msg = 'invalid bit length repeat';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
while (copy--) {
|
|
|
|
state.lens[state.have++] = len;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-12-10 21:51:33 +01:00
|
|
|
/* handle error breaks in while */
|
|
|
|
if (state.mode === BAD) { break; }
|
2017-05-03 15:35:00 +02:00
|
|
|
|
2017-12-10 21:51:33 +01:00
|
|
|
/* check for end-of-block code (better have one) */
|
|
|
|
if (state.lens[256] === 0) {
|
|
|
|
strm.msg = 'invalid code -- missing end-of-block';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
|
|
|
|
/* build code tables -- note: do not change the lenbits or distbits
|
|
|
|
values here (9 and 6) without reading the comments in inftrees.h
|
|
|
|
concerning the ENOUGH constants, which depend on those values */
|
|
|
|
state.lenbits = 9;
|
|
|
|
|
|
|
|
opts = { bits: state.lenbits };
|
|
|
|
ret = inflate_table(LENS, state.lens, 0, state.nlen, state.lencode, 0, state.work, opts);
|
|
|
|
// We have separate tables & no pointers. 2 commented lines below not needed.
|
|
|
|
// state.next_index = opts.table_index;
|
|
|
|
state.lenbits = opts.bits;
|
|
|
|
// state.lencode = state.next;
|
|
|
|
|
|
|
|
if (ret) {
|
|
|
|
strm.msg = 'invalid literal/lengths set';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
state.distbits = 6;
|
|
|
|
//state.distcode.copy(state.codes);
|
|
|
|
// Switch to use dynamic table
|
|
|
|
state.distcode = state.distdyn;
|
|
|
|
opts = { bits: state.distbits };
|
|
|
|
ret = inflate_table(DISTS, state.lens, state.nlen, state.ndist, state.distcode, 0, state.work, opts);
|
|
|
|
// We have separate tables & no pointers. 2 commented lines below not needed.
|
|
|
|
// state.next_index = opts.table_index;
|
|
|
|
state.distbits = opts.bits;
|
|
|
|
// state.distcode = state.next;
|
|
|
|
|
|
|
|
if (ret) {
|
|
|
|
strm.msg = 'invalid distances set';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
//Tracev((stderr, 'inflate: codes ok\n'));
|
|
|
|
state.mode = LEN_;
|
|
|
|
if (flush === Z_TREES) { break inf_leave; }
|
|
|
|
/* falls through */
|
|
|
|
case LEN_:
|
|
|
|
state.mode = LEN;
|
|
|
|
/* falls through */
|
|
|
|
case LEN:
|
|
|
|
if (have >= 6 && left >= 258) {
|
|
|
|
//--- RESTORE() ---
|
|
|
|
strm.next_out = put;
|
|
|
|
strm.avail_out = left;
|
|
|
|
strm.next_in = next;
|
|
|
|
strm.avail_in = have;
|
|
|
|
state.hold = hold;
|
|
|
|
state.bits = bits;
|
|
|
|
//---
|
|
|
|
inflate_fast(strm, _out);
|
|
|
|
//--- LOAD() ---
|
|
|
|
put = strm.next_out;
|
|
|
|
output = strm.output;
|
|
|
|
left = strm.avail_out;
|
|
|
|
next = strm.next_in;
|
|
|
|
input = strm.input;
|
|
|
|
have = strm.avail_in;
|
|
|
|
hold = state.hold;
|
|
|
|
bits = state.bits;
|
|
|
|
//---
|
|
|
|
|
|
|
|
if (state.mode === TYPE) {
|
|
|
|
state.back = -1;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
state.back = 0;
|
2017-05-03 15:35:00 +02:00
|
|
|
for (;;) {
|
2017-12-10 21:51:33 +01:00
|
|
|
here = state.lencode[hold & ((1 << state.lenbits) - 1)]; /*BITS(state.lenbits)*/
|
2017-05-03 15:35:00 +02:00
|
|
|
here_bits = here >>> 24;
|
|
|
|
here_op = (here >>> 16) & 0xff;
|
|
|
|
here_val = here & 0xffff;
|
|
|
|
|
2017-12-10 21:51:33 +01:00
|
|
|
if (here_bits <= bits) { break; }
|
2017-05-03 15:35:00 +02:00
|
|
|
//--- PULLBYTE() ---//
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
//---//
|
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
if (here_op && (here_op & 0xf0) === 0) {
|
|
|
|
last_bits = here_bits;
|
|
|
|
last_op = here_op;
|
|
|
|
last_val = here_val;
|
|
|
|
for (;;) {
|
|
|
|
here = state.lencode[last_val +
|
|
|
|
((hold & ((1 << (last_bits + last_op)) - 1))/*BITS(last.bits + last.op)*/ >> last_bits)];
|
|
|
|
here_bits = here >>> 24;
|
|
|
|
here_op = (here >>> 16) & 0xff;
|
|
|
|
here_val = here & 0xffff;
|
|
|
|
|
|
|
|
if ((last_bits + here_bits) <= bits) { break; }
|
|
|
|
//--- PULLBYTE() ---//
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
//---//
|
|
|
|
}
|
|
|
|
//--- DROPBITS(last.bits) ---//
|
|
|
|
hold >>>= last_bits;
|
|
|
|
bits -= last_bits;
|
|
|
|
//---//
|
|
|
|
state.back += last_bits;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
//--- DROPBITS(here.bits) ---//
|
|
|
|
hold >>>= here_bits;
|
|
|
|
bits -= here_bits;
|
2017-05-03 15:35:00 +02:00
|
|
|
//---//
|
2017-12-10 21:51:33 +01:00
|
|
|
state.back += here_bits;
|
|
|
|
state.length = here_val;
|
|
|
|
if (here_op === 0) {
|
|
|
|
//Tracevv((stderr, here.val >= 0x20 && here.val < 0x7f ?
|
|
|
|
// "inflate: literal '%c'\n" :
|
|
|
|
// "inflate: literal 0x%02x\n", here.val));
|
|
|
|
state.mode = LIT;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (here_op & 32) {
|
|
|
|
//Tracevv((stderr, "inflate: end of block\n"));
|
|
|
|
state.back = -1;
|
|
|
|
state.mode = TYPE;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (here_op & 64) {
|
|
|
|
strm.msg = 'invalid literal/length code';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
state.extra = here_op & 15;
|
|
|
|
state.mode = LENEXT;
|
|
|
|
/* falls through */
|
|
|
|
case LENEXT:
|
|
|
|
if (state.extra) {
|
|
|
|
//=== NEEDBITS(state.extra);
|
|
|
|
n = state.extra;
|
|
|
|
while (bits < n) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
state.length += hold & ((1 << state.extra) - 1)/*BITS(state.extra)*/;
|
|
|
|
//--- DROPBITS(state.extra) ---//
|
|
|
|
hold >>>= state.extra;
|
|
|
|
bits -= state.extra;
|
|
|
|
//---//
|
|
|
|
state.back += state.extra;
|
|
|
|
}
|
|
|
|
//Tracevv((stderr, "inflate: length %u\n", state.length));
|
|
|
|
state.was = state.length;
|
|
|
|
state.mode = DIST;
|
|
|
|
/* falls through */
|
|
|
|
case DIST:
|
2017-05-03 15:35:00 +02:00
|
|
|
for (;;) {
|
2017-12-10 21:51:33 +01:00
|
|
|
here = state.distcode[hold & ((1 << state.distbits) - 1)];/*BITS(state.distbits)*/
|
2017-05-03 15:35:00 +02:00
|
|
|
here_bits = here >>> 24;
|
|
|
|
here_op = (here >>> 16) & 0xff;
|
|
|
|
here_val = here & 0xffff;
|
|
|
|
|
2017-12-10 21:51:33 +01:00
|
|
|
if ((here_bits) <= bits) { break; }
|
2017-05-03 15:35:00 +02:00
|
|
|
//--- PULLBYTE() ---//
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
//---//
|
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
if ((here_op & 0xf0) === 0) {
|
|
|
|
last_bits = here_bits;
|
|
|
|
last_op = here_op;
|
|
|
|
last_val = here_val;
|
|
|
|
for (;;) {
|
|
|
|
here = state.distcode[last_val +
|
|
|
|
((hold & ((1 << (last_bits + last_op)) - 1))/*BITS(last.bits + last.op)*/ >> last_bits)];
|
|
|
|
here_bits = here >>> 24;
|
|
|
|
here_op = (here >>> 16) & 0xff;
|
|
|
|
here_val = here & 0xffff;
|
|
|
|
|
|
|
|
if ((last_bits + here_bits) <= bits) { break; }
|
|
|
|
//--- PULLBYTE() ---//
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
//---//
|
|
|
|
}
|
|
|
|
//--- DROPBITS(last.bits) ---//
|
|
|
|
hold >>>= last_bits;
|
|
|
|
bits -= last_bits;
|
|
|
|
//---//
|
|
|
|
state.back += last_bits;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
//--- DROPBITS(here.bits) ---//
|
|
|
|
hold >>>= here_bits;
|
|
|
|
bits -= here_bits;
|
2017-05-03 15:35:00 +02:00
|
|
|
//---//
|
2017-12-10 21:51:33 +01:00
|
|
|
state.back += here_bits;
|
|
|
|
if (here_op & 64) {
|
|
|
|
strm.msg = 'invalid distance code';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
state.offset = here_val;
|
|
|
|
state.extra = (here_op) & 15;
|
|
|
|
state.mode = DISTEXT;
|
|
|
|
/* falls through */
|
|
|
|
case DISTEXT:
|
|
|
|
if (state.extra) {
|
|
|
|
//=== NEEDBITS(state.extra);
|
|
|
|
n = state.extra;
|
|
|
|
while (bits < n) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
state.offset += hold & ((1 << state.extra) - 1)/*BITS(state.extra)*/;
|
|
|
|
//--- DROPBITS(state.extra) ---//
|
|
|
|
hold >>>= state.extra;
|
|
|
|
bits -= state.extra;
|
|
|
|
//---//
|
|
|
|
state.back += state.extra;
|
|
|
|
}
|
2017-05-03 15:35:00 +02:00
|
|
|
//#ifdef INFLATE_STRICT
|
2017-12-10 21:51:33 +01:00
|
|
|
if (state.offset > state.dmax) {
|
|
|
|
strm.msg = 'invalid distance too far back';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
2017-05-03 15:35:00 +02:00
|
|
|
//#endif
|
2017-12-10 21:51:33 +01:00
|
|
|
//Tracevv((stderr, "inflate: distance %u\n", state.offset));
|
|
|
|
state.mode = MATCH;
|
|
|
|
/* falls through */
|
|
|
|
case MATCH:
|
|
|
|
if (left === 0) { break inf_leave; }
|
|
|
|
copy = _out - left;
|
|
|
|
if (state.offset > copy) { /* copy from window */
|
|
|
|
copy = state.offset - copy;
|
|
|
|
if (copy > state.whave) {
|
|
|
|
if (state.sane) {
|
|
|
|
strm.msg = 'invalid distance too far back';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
// (!) This block is disabled in zlib defaults,
|
2017-05-03 15:35:00 +02:00
|
|
|
// don't enable it for binary compatibility
|
|
|
|
//#ifdef INFLATE_ALLOW_INVALID_DISTANCE_TOOFAR_ARRR
|
|
|
|
// Trace((stderr, "inflate.c too far\n"));
|
|
|
|
// copy -= state.whave;
|
|
|
|
// if (copy > state.length) { copy = state.length; }
|
|
|
|
// if (copy > left) { copy = left; }
|
|
|
|
// left -= copy;
|
|
|
|
// state.length -= copy;
|
|
|
|
// do {
|
|
|
|
// output[put++] = 0;
|
|
|
|
// } while (--copy);
|
|
|
|
// if (state.length === 0) { state.mode = LEN; }
|
|
|
|
// break;
|
|
|
|
//#endif
|
2017-12-10 21:51:33 +01:00
|
|
|
}
|
|
|
|
if (copy > state.wnext) {
|
|
|
|
copy -= state.wnext;
|
|
|
|
from = state.wsize - copy;
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
from = state.wnext - copy;
|
|
|
|
}
|
|
|
|
if (copy > state.length) { copy = state.length; }
|
|
|
|
from_source = state.window;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
else { /* copy from output */
|
|
|
|
from_source = output;
|
|
|
|
from = put - state.offset;
|
|
|
|
copy = state.length;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
if (copy > left) { copy = left; }
|
|
|
|
left -= copy;
|
|
|
|
state.length -= copy;
|
|
|
|
do {
|
|
|
|
output[put++] = from_source[from++];
|
|
|
|
} while (--copy);
|
|
|
|
if (state.length === 0) { state.mode = LEN; }
|
|
|
|
break;
|
|
|
|
case LIT:
|
|
|
|
if (left === 0) { break inf_leave; }
|
|
|
|
output[put++] = state.length;
|
|
|
|
left--;
|
|
|
|
state.mode = LEN;
|
|
|
|
break;
|
|
|
|
case CHECK:
|
|
|
|
if (state.wrap) {
|
|
|
|
//=== NEEDBITS(32);
|
|
|
|
while (bits < 32) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
// Use '|' instead of '+' to make sure that result is signed
|
|
|
|
hold |= input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
_out -= left;
|
|
|
|
strm.total_out += _out;
|
|
|
|
state.total += _out;
|
|
|
|
if (_out) {
|
|
|
|
strm.adler = state.check =
|
|
|
|
/*UPDATE(state.check, put - _out, _out);*/
|
|
|
|
(state.flags ? crc32(state.check, output, _out, put - _out) : adler32(state.check, output, _out, put - _out));
|
2017-05-03 15:35:00 +02:00
|
|
|
|
2017-12-10 21:51:33 +01:00
|
|
|
}
|
|
|
|
_out = left;
|
|
|
|
// NB: crc32 stored as signed 32-bit int, zswap32 returns signed too
|
|
|
|
if ((state.flags ? hold : zswap32(hold)) !== state.check) {
|
|
|
|
strm.msg = 'incorrect data check';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
|
|
|
//Tracev((stderr, "inflate: check matches trailer\n"));
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
state.mode = LENGTH;
|
|
|
|
/* falls through */
|
|
|
|
case LENGTH:
|
|
|
|
if (state.wrap && state.flags) {
|
|
|
|
//=== NEEDBITS(32);
|
|
|
|
while (bits < 32) {
|
|
|
|
if (have === 0) { break inf_leave; }
|
|
|
|
have--;
|
|
|
|
hold += input[next++] << bits;
|
|
|
|
bits += 8;
|
|
|
|
}
|
|
|
|
//===//
|
|
|
|
if (hold !== (state.total & 0xffffffff)) {
|
|
|
|
strm.msg = 'incorrect length check';
|
|
|
|
state.mode = BAD;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
//=== INITBITS();
|
|
|
|
hold = 0;
|
|
|
|
bits = 0;
|
|
|
|
//===//
|
|
|
|
//Tracev((stderr, "inflate: length matches trailer\n"));
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
2017-12-10 21:51:33 +01:00
|
|
|
state.mode = DONE;
|
|
|
|
/* falls through */
|
|
|
|
case DONE:
|
|
|
|
ret = Z_STREAM_END;
|
|
|
|
break inf_leave;
|
|
|
|
case BAD:
|
|
|
|
ret = Z_DATA_ERROR;
|
|
|
|
break inf_leave;
|
|
|
|
case MEM:
|
|
|
|
return Z_MEM_ERROR;
|
|
|
|
case SYNC:
|
|
|
|
/* falls through */
|
|
|
|
default:
|
|
|
|
return Z_STREAM_ERROR;
|
2017-05-03 15:35:00 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// inf_leave <- here is real place for "goto inf_leave", emulated via "break inf_leave"
|
|
|
|
|
|
|
|
/*
|
|
|
|
Return from inflate(), updating the total counts and the check value.
|
|
|
|
If there was no progress during the inflate() call, return a buffer
|
|
|
|
error. Call updatewindow() to create and/or update the window state.
|
|
|
|
Note: a memory error from inflate() is non-recoverable.
|
|
|
|
*/
|
|
|
|
|
|
|
|
//--- RESTORE() ---
|
|
|
|
strm.next_out = put;
|
|
|
|
strm.avail_out = left;
|
|
|
|
strm.next_in = next;
|
|
|
|
strm.avail_in = have;
|
|
|
|
state.hold = hold;
|
|
|
|
state.bits = bits;
|
|
|
|
//---
|
|
|
|
|
|
|
|
if (state.wsize || (_out !== strm.avail_out && state.mode < BAD &&
|
|
|
|
(state.mode < CHECK || flush !== Z_FINISH))) {
|
|
|
|
if (updatewindow(strm, strm.output, strm.next_out, _out - strm.avail_out)) {
|
|
|
|
state.mode = MEM;
|
|
|
|
return Z_MEM_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
_in -= strm.avail_in;
|
|
|
|
_out -= strm.avail_out;
|
|
|
|
strm.total_in += _in;
|
|
|
|
strm.total_out += _out;
|
|
|
|
state.total += _out;
|
|
|
|
if (state.wrap && _out) {
|
|
|
|
strm.adler = state.check = /*UPDATE(state.check, strm.next_out - _out, _out);*/
|
|
|
|
(state.flags ? crc32(state.check, output, _out, strm.next_out - _out) : adler32(state.check, output, _out, strm.next_out - _out));
|
|
|
|
}
|
|
|
|
strm.data_type = state.bits + (state.last ? 64 : 0) +
|
|
|
|
(state.mode === TYPE ? 128 : 0) +
|
|
|
|
(state.mode === LEN_ || state.mode === COPY_ ? 256 : 0);
|
|
|
|
if (((_in === 0 && _out === 0) || flush === Z_FINISH) && ret === Z_OK) {
|
|
|
|
ret = Z_BUF_ERROR;
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
function inflateEnd(strm) {
|
|
|
|
|
|
|
|
if (!strm || !strm.state /*|| strm->zfree == (free_func)0*/) {
|
|
|
|
return Z_STREAM_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
var state = strm.state;
|
|
|
|
if (state.window) {
|
|
|
|
state.window = null;
|
|
|
|
}
|
|
|
|
strm.state = null;
|
|
|
|
return Z_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
function inflateGetHeader(strm, head) {
|
|
|
|
var state;
|
|
|
|
|
|
|
|
/* check state */
|
|
|
|
if (!strm || !strm.state) { return Z_STREAM_ERROR; }
|
|
|
|
state = strm.state;
|
|
|
|
if ((state.wrap & 2) === 0) { return Z_STREAM_ERROR; }
|
|
|
|
|
|
|
|
/* save header structure */
|
|
|
|
state.head = head;
|
|
|
|
head.done = false;
|
|
|
|
return Z_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
function inflateSetDictionary(strm, dictionary) {
|
|
|
|
var dictLength = dictionary.length;
|
|
|
|
|
|
|
|
var state;
|
|
|
|
var dictid;
|
|
|
|
var ret;
|
|
|
|
|
|
|
|
/* check state */
|
|
|
|
if (!strm /* == Z_NULL */ || !strm.state /* == Z_NULL */) { return Z_STREAM_ERROR; }
|
|
|
|
state = strm.state;
|
|
|
|
|
|
|
|
if (state.wrap !== 0 && state.mode !== DICT) {
|
|
|
|
return Z_STREAM_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* check for correct dictionary identifier */
|
|
|
|
if (state.mode === DICT) {
|
|
|
|
dictid = 1; /* adler32(0, null, 0)*/
|
|
|
|
/* dictid = adler32(dictid, dictionary, dictLength); */
|
|
|
|
dictid = adler32(dictid, dictionary, dictLength, 0);
|
|
|
|
if (dictid !== state.check) {
|
|
|
|
return Z_DATA_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
/* copy dictionary to window using updatewindow(), which will amend the
|
|
|
|
existing dictionary if appropriate */
|
|
|
|
ret = updatewindow(strm, dictionary, dictLength, dictLength);
|
|
|
|
if (ret) {
|
|
|
|
state.mode = MEM;
|
|
|
|
return Z_MEM_ERROR;
|
|
|
|
}
|
|
|
|
state.havedict = 1;
|
|
|
|
// Tracev((stderr, "inflate: dictionary set\n"));
|
|
|
|
return Z_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
exports.inflateReset = inflateReset;
|
|
|
|
exports.inflateReset2 = inflateReset2;
|
|
|
|
exports.inflateResetKeep = inflateResetKeep;
|
|
|
|
exports.inflateInit = inflateInit;
|
|
|
|
exports.inflateInit2 = inflateInit2;
|
|
|
|
exports.inflate = inflate;
|
|
|
|
exports.inflateEnd = inflateEnd;
|
|
|
|
exports.inflateGetHeader = inflateGetHeader;
|
|
|
|
exports.inflateSetDictionary = inflateSetDictionary;
|
|
|
|
exports.inflateInfo = 'pako inflate (from Nodeca project)';
|
|
|
|
|
|
|
|
/* Not implemented
|
|
|
|
exports.inflateCopy = inflateCopy;
|
|
|
|
exports.inflateGetDictionary = inflateGetDictionary;
|
|
|
|
exports.inflateMark = inflateMark;
|
|
|
|
exports.inflatePrime = inflatePrime;
|
|
|
|
exports.inflateSync = inflateSync;
|
|
|
|
exports.inflateSyncPoint = inflateSyncPoint;
|
|
|
|
exports.inflateUndermine = inflateUndermine;
|
|
|
|
*/
|