Archive-Unzip-Burst
view release on metacpan or search on metacpan
unzip-6.0/inflate.c view on Meta::CPAN
c14k 27 Feb 94 G. Roelofs added some typecasts to avoid warnings.
c14l 9 Apr 94 G. Roelofs fixed split comments on preprocessor lines
to avoid bug in Encore compiler.
c14m 7 Jul 94 P. Kienitz modified to allow assembler version of
inflate_codes() (define ASM_INFLATECODES)
c14n 22 Jul 94 G. Roelofs changed fprintf to macro for DLL versions
c14o 23 Aug 94 C. Spieler added a newline to a debug statement;
G. Roelofs added another typecast to avoid MSC warning
c14p 4 Oct 94 G. Roelofs added (voidp *) cast to free() argument
c14q 30 Oct 94 G. Roelofs changed fprintf macro to MESSAGE()
c14r 1 Nov 94 G. Roelofs fixed possible redefinition of CHECK_EOF
c14s 7 May 95 S. Maxwell OS/2 DLL globals stuff incorporated;
P. Kienitz "fixed" ASM_INFLATECODES macro/prototype
c14t 18 Aug 95 G. Roelofs added UZinflate() to use zlib functions;
changed voidp to zvoid; moved huft_build()
and huft_free() to end of file
c14u 1 Oct 95 G. Roelofs moved G into definition of MESSAGE macro
c14v 8 Nov 95 P. Kienitz changed ASM_INFLATECODES to use a regular
call with __G__ instead of a macro
c15 3 Aug 96 M. Adler fixed bomb-bug on random input data (Adobe)
c15b 24 Aug 96 M. Adler more fixes for random input data
c15c 28 Mar 97 G. Roelofs changed USE_ZLIB fatal exit code from
PK_MEM2 to PK_MEM3
c16 20 Apr 97 J. Altman added memzero(v[]) in huft_build()
c16b 29 Mar 98 C. Spieler modified DLL code for slide redirection
c16c 04 Apr 99 C. Spieler fixed memory leaks when processing gets
stopped because of input data errors
c16d 05 Jul 99 C. Spieler take care of FLUSH() return values and
stop processing in case of errors
c17 31 Dec 00 C. Spieler added preliminary support for Deflate64
c17a 04 Feb 01 C. Spieler complete integration of Deflate64 support
c17b 16 Feb 02 C. Spieler changed type of "extra bits" arrays and
corresponding huft_build() parameter e from
ush into uch, to save space
c17c 9 Mar 02 C. Spieler fixed NEEDBITS() "read beyond EOF" problem
with CHECK_EOF enabled
c17d 23 Jul 05 C. Spieler fixed memory leaks in inflate_dynamic()
when processing invalid compressed literal/
distance table data
c17e 30 Mar 07 C. Spieler in inflate_dynamic(), initialize tl and td
to prevent freeing unallocated huft tables
when processing invalid compressed data and
hitting premature EOF, do not reuse td as
temp work ptr during tables decoding
*/
/*
Inflate deflated (PKZIP's method 8 compressed) data. The compression
method searches for as much of the current string of bytes (up to a
length of 258) in the previous 32K bytes. If it doesn't find any
matches (of at least length 3), it codes the next byte. Otherwise, it
codes the length of the matched string and its distance backwards from
the current position. There is a single Huffman code that codes both
single bytes (called "literals") and match lengths. A second Huffman
code codes the distance information, which follows a length code. Each
length or distance code actually represents a base value and a number
of "extra" (sometimes zero) bits to get to add to the base value. At
the end of each deflated block is a special end-of-block (EOB) literal/
length code. The decoding process is basically: get a literal/length
code; if EOB then done; if a literal, emit the decoded byte; if a
length then get the distance and emit the referred-to bytes from the
sliding window of previously emitted data.
There are (currently) three kinds of inflate blocks: stored, fixed, and
dynamic. The compressor outputs a chunk of data at a time and decides
which method to use on a chunk-by-chunk basis. A chunk might typically
be 32K to 64K, uncompressed. If the chunk is uncompressible, then the
"stored" method is used. In this case, the bytes are simply stored as
is, eight bits per byte, with none of the above coding. The bytes are
preceded by a count, since there is no longer an EOB code.
If the data are compressible, then either the fixed or dynamic methods
are used. In the dynamic method, the compressed data are preceded by
an encoding of the literal/length and distance Huffman codes that are
to be used to decode this block. The representation is itself Huffman
coded, and so is preceded by a description of that code. These code
descriptions take up a little space, and so for small blocks, there is
a predefined set of codes, called the fixed codes. The fixed method is
used if the block ends up smaller that way (usually for quite small
chunks); otherwise the dynamic method is used. In the latter case, the
codes are customized to the probabilities in the current block and so
can code it much better than the pre-determined fixed codes can.
The Huffman codes themselves are decoded using a multi-level table
lookup, in order to maximize the speed of decoding plus the speed of
building the decoding tables. See the comments below that precede the
lbits and dbits tuning parameters.
GRR: return values(?)
0 OK
1 incomplete table
2 bad input
3 not enough memory
the following return codes are passed through from FLUSH() errors
50 (PK_DISK) "overflow of output space"
80 (IZ_CTRLC) "canceled by user's request"
*/
/*
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.
14. The Deflate64 (PKZIP method 9) variant of the compression algorithm
differs from "classic" deflate in the following 3 aspect:
unzip-6.0/inflate.c view on Meta::CPAN
where NEEDBITS makes sure that b has at least j bits in it, and
DUMPBITS removes the bits from b. The macros use the variable k
for the number of bits in b. Normally, b and k are register
variables for speed and are initialized at the beginning of a
routine that uses these macros from a global bit buffer and count.
In order to not ask for more bits than there are in the compressed
stream, the Huffman tables are constructed to only ask for just
enough bits to make up the end-of-block code (value 256). Then no
bytes need to be "returned" to the buffer at the end of the last
block. See the huft_build() routine.
Actually, the precautions mentioned above are not sufficient to
prevent fetches of bits beyound the end of the last block in every
case. When the last code fetched before the end-of-block code was
a very short distance code (shorter than "distance-prefetch-bits" -
"end-of-block code bits"), this last distance code fetch already
exausts the available data. To prevent failure of extraction in this
case, the "read beyond EOF" check delays the raise of the "invalid
data" error until an actual overflow of "used data" is detected.
This error condition is only fulfilled when the "number of available
bits" counter k is found to be negative in the NEEDBITS() macro.
An alternate fix for that problem adjusts the size of the distance code
base table so that it does not exceed the length of the end-of-block code
plus the minimum length of a distance code. This alternate fix can be
enabled by defining the preprocessor symbol FIX_PAST_EOB_BY_TABLEADJUST.
*/
/* These have been moved to globals.h */
#if 0
ulg bb; /* bit buffer */
unsigned bk; /* bits in bit buffer */
#endif
#ifndef CHECK_EOF
# define CHECK_EOF /* default as of 5.13/5.2 */
#endif
#ifndef CHECK_EOF
# define NEEDBITS(n) {while(k<(n)){b|=((ulg)NEXTBYTE)<<k;k+=8;}}
#else
# ifdef FIX_PAST_EOB_BY_TABLEADJUST
# define NEEDBITS(n) {while(k<(n)){int c=NEXTBYTE;\
if(c==EOF){retval=1;goto cleanup_and_exit;}\
b|=((ulg)c)<<k;k+=8;}}
# else
# define NEEDBITS(n) {while((int)k<(int)(n)){int c=NEXTBYTE;\
if(c==EOF){if((int)k>=0)break;retval=1;goto cleanup_and_exit;}\
b|=((ulg)c)<<k;k+=8;}}
# endif
#endif
#define DUMPBITS(n) {b>>=(n);k-=(n);}
/*
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
are 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.
*/
/* bits in base literal/length lookup table */
static ZCONST unsigned lbits = 9;
/* bits in base distance lookup table */
static ZCONST unsigned dbits = 6;
#ifndef ASM_INFLATECODES
int inflate_codes(__G__ tl, td, bl, bd)
__GDEF
struct huft *tl, *td; /* literal/length and distance decoder tables */
unsigned bl, bd; /* number of bits decoded by tl[] and td[] */
/* inflate (decompress) the codes in a deflated (compressed) block.
Return an error code or zero if it all goes ok. */
{
register unsigned e; /* table entry flag/number of extra bits */
unsigned d; /* index for copy */
UINT_D64 n; /* length for copy (deflate64: might be 64k+2) */
UINT_D64 w; /* current window position (deflate64: up to 64k) */
struct huft *t; /* pointer to table entry */
unsigned ml, md; /* masks for bl and bd bits */
register ulg b; /* bit buffer */
register unsigned k; /* number of bits in bit buffer */
int retval = 0; /* error code returned: initialized to "no error" */
/* make local copies of globals */
b = G.bb; /* initialize bit buffer */
k = G.bk;
w = G.wp; /* initialize window position */
/* inflate the coded data */
ml = mask_bits[bl]; /* precompute masks for speed */
md = mask_bits[bd];
while (1) /* do until end of block */
{
NEEDBITS(bl)
t = tl + ((unsigned)b & ml);
while (1) {
DUMPBITS(t->b)
if ((e = t->e) == 32) /* then it's a literal */
{
redirSlide[w++] = (uch)t->v.n;
if (w == wsize)
{
if ((retval = FLUSH(w)) != 0) goto cleanup_and_exit;
w = 0;
}
break;
}
if (e < 31) /* then it's a length */
{
/* get length of block to copy */
NEEDBITS(e)
n = t->v.n + ((unsigned)b & mask_bits[e]);
DUMPBITS(e)
/* decode distance of block to copy */
NEEDBITS(bd)
t = td + ((unsigned)b & md);
while (1) {
DUMPBITS(t->b)
if ((e = t->e) < 32)
break;
if (IS_INVALID_CODE(e))
return 1;
e &= 31;
NEEDBITS(e)
t = t->v.t + ((unsigned)b & mask_bits[e]);
unzip-6.0/inflate.c view on Meta::CPAN
G.fixed_tl64 = G.fixed_tl;
G.fixed_bl64 = G.fixed_bl;
G.fixed_td64 = G.fixed_td;
G.fixed_bd64 = G.fixed_bd;
} else {
G.fixed_tl32 = G.fixed_tl;
G.fixed_bl32 = G.fixed_bl;
G.fixed_td32 = G.fixed_td;
G.fixed_bd32 = G.fixed_bd;
}
#endif
/* flush out redirSlide and return (success, unless final FLUSH failed) */
return (FLUSH(G.wp));
}
int inflate_free(__G)
__GDEF
{
if (G.fixed_tl != (struct huft *)NULL)
{
huft_free(G.fixed_td);
huft_free(G.fixed_tl);
G.fixed_td = G.fixed_tl = (struct huft *)NULL;
}
return 0;
}
#endif /* ?USE_ZLIB */
/*
* GRR: moved huft_build() and huft_free() down here; used by explode()
* and fUnZip regardless of whether USE_ZLIB defined or not
*/
/* If BMAX needs to be larger than 16, then h and x[] should be ulg. */
#define BMAX 16 /* maximum bit length of any code (16 for explode) */
#define N_MAX 288 /* maximum number of codes in any set */
int huft_build(__G__ b, n, s, d, e, t, m)
__GDEF
ZCONST unsigned *b; /* code lengths in bits (all assumed <= BMAX) */
unsigned n; /* number of codes (assumed <= N_MAX) */
unsigned s; /* number of simple-valued codes (0..s-1) */
ZCONST ush *d; /* list of base values for non-simple codes */
ZCONST uch *e; /* list of extra bits for non-simple codes */
struct huft **t; /* result: starting table */
unsigned *m; /* maximum lookup bits, returns actual */
/* Given a list of code lengths and a maximum table size, make a set of
tables to decode that set of codes. Return zero on success, one if
the given code set is incomplete (the tables are still built in this
case), two if the input is invalid (all zero length codes or an
oversubscribed set of lengths), and three if not enough memory.
The code with value 256 is special, and the tables are constructed
so that no bits beyond that code are fetched when that code is
decoded. */
{
unsigned a; /* counter for codes of length k */
unsigned c[BMAX+1]; /* bit length count table */
unsigned el; /* length of EOB code (value 256) */
unsigned f; /* i repeats in table every f entries */
int g; /* maximum code length */
int h; /* table level */
register unsigned i; /* counter, current code */
register unsigned j; /* counter */
register int k; /* number of bits in current code */
int lx[BMAX+1]; /* memory for l[-1..BMAX-1] */
int *l = lx+1; /* stack of bits per table */
register unsigned *p; /* pointer into c[], b[], or v[] */
register struct huft *q; /* points to current table */
struct huft r; /* table entry for structure assignment */
struct huft *u[BMAX]; /* table stack */
unsigned v[N_MAX]; /* values in order of bit length */
register int w; /* bits before this table == (l * h) */
unsigned x[BMAX+1]; /* bit offsets, then code stack */
unsigned *xp; /* pointer into x */
int y; /* number of dummy codes added */
unsigned z; /* number of entries in current table */
/* Generate counts for each bit length */
el = n > 256 ? b[256] : BMAX; /* set length of EOB code, if any */
memzero((char *)c, sizeof(c));
p = (unsigned *)b; i = n;
do {
c[*p]++; p++; /* assume all entries <= BMAX */
} while (--i);
if (c[0] == n) /* null input--all zero length codes */
{
*t = (struct huft *)NULL;
*m = 0;
return 0;
}
/* Find minimum and maximum length, bound *m by those */
for (j = 1; j <= BMAX; j++)
if (c[j])
break;
k = j; /* minimum code length */
if (*m < j)
*m = j;
for (i = BMAX; i; i--)
if (c[i])
break;
g = i; /* maximum code length */
if (*m > i)
*m = i;
/* 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 2; /* bad input: more codes than bits */
if ((y -= c[i]) < 0)
return 2;
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 */
memzero((char *)v, sizeof(v));
p = (unsigned *)b; i = 0;
do {
if ((j = *p++) != 0)
v[x[j]++] = i;
} while (++i < n);
n = x[g]; /* set n to length of v */
/* 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[-1] = 0; /* no bits decoded yet */
u[0] = (struct huft *)NULL; /* just to keep compilers happy */
q = (struct huft *)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[h++]; /* add bits already decoded */
/* compute minimum size table less than or equal to *m bits */
z = (z = g - w) > *m ? *m : z; /* 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;
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 */
}
}
if ((unsigned)w + j > el && (unsigned)w < el)
j = el - w; /* make EOB code end at table */
z = 1 << j; /* table entries for j-bit table */
l[h] = j; /* set table size in stack */
/* allocate and link in new table */
if ((q = (struct huft *)malloc((z + 1)*sizeof(struct huft))) ==
(struct huft *)NULL)
{
if (h)
huft_free(u[0]);
return 3; /* not enough memory */
}
#ifdef DEBUG
G.hufts += z + 1; /* track memory usage */
#endif
*t = q + 1; /* link to list for huft_free() */
*(t = &(q->v.t)) = (struct huft *)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.b = (uch)l[h-1]; /* bits to dump before this table */
r.e = (uch)(32 + j); /* bits in this table */
r.v.t = q; /* pointer to this table */
j = (i & ((1 << w) - 1)) >> (w - l[h-1]);
u[h-1][j] = r; /* connect to last table */
}
}
/* set up table entry in r */
r.b = (uch)(k - w);
if (p >= v + n)
r.e = INVALID_CODE; /* out of values--invalid code */
else if (*p < s)
{
r.e = (uch)(*p < 256 ? 32 : 31); /* 256 is end-of-block code */
r.v.n = (ush)*p++; /* simple code is just the value */
}
else
{
r.e = e[*p - s]; /* non-simple--look up in lists */
r.v.n = d[*p++ - s];
}
( run in 1.473 second using v1.01-cache-2.11-cpan-0bb4e1dffa6 )