modules/libbz2/src/huffman.c@6474c204b198
modules/libbz2/src/huffman.c
Wed, 31 Dec 2014 06:09:35 +0100
- author
- Michael Schloh von Bennewitz <michael@schloh.com>
- date
- Wed, 31 Dec 2014 06:09:35 +0100
- changeset 0
- 6474c204b198
- permissions
- -rw-r--r--
Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.
2 /*-------------------------------------------------------------*/
3 /*--- Huffman coding low-level stuff ---*/
4 /*--- huffman.c ---*/
5 /*-------------------------------------------------------------*/
7 /* ------------------------------------------------------------------
8 This file is part of bzip2/libbzip2, a program and library for
9 lossless, block-sorting data compression.
11 bzip2/libbzip2 version 1.0.4 of 20 December 2006
12 Copyright (C) 1996-2006 Julian Seward <jseward@bzip.org>
14 Please read the WARNING, DISCLAIMER and PATENTS sections in the
15 README file.
17 This program is released under the terms of the license contained
18 in the file LICENSE.
19 ------------------------------------------------------------------ */
22 #include "bzlib_private.h"
24 /*---------------------------------------------------*/
25 #define WEIGHTOF(zz0) ((zz0) & 0xffffff00)
26 #define DEPTHOF(zz1) ((zz1) & 0x000000ff)
27 #define MYMAX(zz2,zz3) ((zz2) > (zz3) ? (zz2) : (zz3))
29 #define ADDWEIGHTS(zw1,zw2) \
30 (WEIGHTOF(zw1)+WEIGHTOF(zw2)) | \
31 (1 + MYMAX(DEPTHOF(zw1),DEPTHOF(zw2)))
33 #define UPHEAP(z) \
34 { \
35 Int32 zz, tmp; \
36 zz = z; tmp = heap[zz]; \
37 while (weight[tmp] < weight[heap[zz >> 1]]) { \
38 heap[zz] = heap[zz >> 1]; \
39 zz >>= 1; \
40 } \
41 heap[zz] = tmp; \
42 }
44 #define DOWNHEAP(z) \
45 { \
46 Int32 zz, yy, tmp; \
47 zz = z; tmp = heap[zz]; \
48 while (True) { \
49 yy = zz << 1; \
50 if (yy > nHeap) break; \
51 if (yy < nHeap && \
52 weight[heap[yy+1]] < weight[heap[yy]]) \
53 yy++; \
54 if (weight[tmp] < weight[heap[yy]]) break; \
55 heap[zz] = heap[yy]; \
56 zz = yy; \
57 } \
58 heap[zz] = tmp; \
59 }
62 /*---------------------------------------------------*/
63 void BZ2_hbMakeCodeLengths ( UChar *len,
64 Int32 *freq,
65 Int32 alphaSize,
66 Int32 maxLen )
67 {
68 /*--
69 Nodes and heap entries run from 1. Entry 0
70 for both the heap and nodes is a sentinel.
71 --*/
72 Int32 nNodes, nHeap, n1, n2, i, j, k;
73 Bool tooLong;
75 Int32 heap [ BZ_MAX_ALPHA_SIZE + 2 ];
76 Int32 weight [ BZ_MAX_ALPHA_SIZE * 2 ];
77 Int32 parent [ BZ_MAX_ALPHA_SIZE * 2 ];
79 for (i = 0; i < alphaSize; i++)
80 weight[i+1] = (freq[i] == 0 ? 1 : freq[i]) << 8;
82 while (True) {
84 nNodes = alphaSize;
85 nHeap = 0;
87 heap[0] = 0;
88 weight[0] = 0;
89 parent[0] = -2;
91 for (i = 1; i <= alphaSize; i++) {
92 parent[i] = -1;
93 nHeap++;
94 heap[nHeap] = i;
95 UPHEAP(nHeap);
96 }
98 AssertH( nHeap < (BZ_MAX_ALPHA_SIZE+2), 2001 );
100 while (nHeap > 1) {
101 n1 = heap[1]; heap[1] = heap[nHeap]; nHeap--; DOWNHEAP(1);
102 n2 = heap[1]; heap[1] = heap[nHeap]; nHeap--; DOWNHEAP(1);
103 nNodes++;
104 parent[n1] = parent[n2] = nNodes;
105 weight[nNodes] = ADDWEIGHTS(weight[n1], weight[n2]);
106 parent[nNodes] = -1;
107 nHeap++;
108 heap[nHeap] = nNodes;
109 UPHEAP(nHeap);
110 }
112 AssertH( nNodes < (BZ_MAX_ALPHA_SIZE * 2), 2002 );
114 tooLong = False;
115 for (i = 1; i <= alphaSize; i++) {
116 j = 0;
117 k = i;
118 while (parent[k] >= 0) { k = parent[k]; j++; }
119 len[i-1] = j;
120 if (j > maxLen) tooLong = True;
121 }
123 if (! tooLong) break;
125 /* 17 Oct 04: keep-going condition for the following loop used
126 to be 'i < alphaSize', which missed the last element,
127 theoretically leading to the possibility of the compressor
128 looping. However, this count-scaling step is only needed if
129 one of the generated Huffman code words is longer than
130 maxLen, which up to and including version 1.0.2 was 20 bits,
131 which is extremely unlikely. In version 1.0.3 maxLen was
132 changed to 17 bits, which has minimal effect on compression
133 ratio, but does mean this scaling step is used from time to
134 time, enough to verify that it works.
136 This means that bzip2-1.0.3 and later will only produce
137 Huffman codes with a maximum length of 17 bits. However, in
138 order to preserve backwards compatibility with bitstreams
139 produced by versions pre-1.0.3, the decompressor must still
140 handle lengths of up to 20. */
142 for (i = 1; i <= alphaSize; i++) {
143 j = weight[i] >> 8;
144 j = 1 + (j / 2);
145 weight[i] = j << 8;
146 }
147 }
148 }
151 /*---------------------------------------------------*/
152 void BZ2_hbAssignCodes ( Int32 *code,
153 UChar *length,
154 Int32 minLen,
155 Int32 maxLen,
156 Int32 alphaSize )
157 {
158 Int32 n, vec, i;
160 vec = 0;
161 for (n = minLen; n <= maxLen; n++) {
162 for (i = 0; i < alphaSize; i++)
163 if (length[i] == n) { code[i] = vec; vec++; };
164 vec <<= 1;
165 }
166 }
169 /*---------------------------------------------------*/
170 void BZ2_hbCreateDecodeTables ( Int32 *limit,
171 Int32 *base,
172 Int32 *perm,
173 UChar *length,
174 Int32 minLen,
175 Int32 maxLen,
176 Int32 alphaSize )
177 {
178 Int32 pp, i, j, vec;
180 pp = 0;
181 for (i = minLen; i <= maxLen; i++)
182 for (j = 0; j < alphaSize; j++)
183 if (length[j] == i) { perm[pp] = j; pp++; };
185 for (i = 0; i < BZ_MAX_CODE_LEN; i++) base[i] = 0;
186 for (i = 0; i < alphaSize; i++) base[length[i]+1]++;
188 for (i = 1; i < BZ_MAX_CODE_LEN; i++) base[i] += base[i-1];
190 for (i = 0; i < BZ_MAX_CODE_LEN; i++) limit[i] = 0;
191 vec = 0;
193 for (i = minLen; i <= maxLen; i++) {
194 vec += (base[i+1] - base[i]);
195 limit[i] = vec-1;
196 vec <<= 1;
197 }
198 for (i = minLen + 1; i <= maxLen; i++)
199 base[i] = ((limit[i-1] + 1) << 1) - base[i];
200 }
203 /*-------------------------------------------------------------*/
204 /*--- end huffman.c ---*/
205 /*-------------------------------------------------------------*/
