diff options
author | Anton Samokhvalov <pg83@yandex.ru> | 2022-02-10 16:45:15 +0300 |
---|---|---|
committer | Daniil Cherednik <dcherednik@yandex-team.ru> | 2022-02-10 16:45:15 +0300 |
commit | 72cb13b4aff9bc9cf22e49251bc8fd143f82538f (patch) | |
tree | da2c34829458c7d4e74bdfbdf85dff449e9e7fb8 /contrib/libs/libbz2/huffman.c | |
parent | 778e51ba091dc39e7b7fcab2b9cf4dbedfb6f2b5 (diff) | |
download | ydb-72cb13b4aff9bc9cf22e49251bc8fd143f82538f.tar.gz |
Restoring authorship annotation for Anton Samokhvalov <pg83@yandex.ru>. Commit 1 of 2.
Diffstat (limited to 'contrib/libs/libbz2/huffman.c')
-rw-r--r-- | contrib/libs/libbz2/huffman.c | 406 |
1 files changed, 203 insertions, 203 deletions
diff --git a/contrib/libs/libbz2/huffman.c b/contrib/libs/libbz2/huffman.c index 43a1899e46..b4a0d0e4cb 100644 --- a/contrib/libs/libbz2/huffman.c +++ b/contrib/libs/libbz2/huffman.c @@ -1,205 +1,205 @@ - -/*-------------------------------------------------------------*/ -/*--- Huffman coding low-level stuff ---*/ -/*--- huffman.c ---*/ -/*-------------------------------------------------------------*/ - -/* ------------------------------------------------------------------ - This file is part of bzip2/libbzip2, a program and library for - lossless, block-sorting data compression. - + +/*-------------------------------------------------------------*/ +/*--- Huffman coding low-level stuff ---*/ +/*--- huffman.c ---*/ +/*-------------------------------------------------------------*/ + +/* ------------------------------------------------------------------ + This file is part of bzip2/libbzip2, a program and library for + lossless, block-sorting data compression. + bzip2/libbzip2 version 1.0.8 of 13 July 2019 Copyright (C) 1996-2019 Julian Seward <jseward@acm.org> - - Please read the WARNING, DISCLAIMER and PATENTS sections in the - README file. - - This program is released under the terms of the license contained - in the file LICENSE. - ------------------------------------------------------------------ */ - - -#include "bzlib_private.h" - -/*---------------------------------------------------*/ -#define WEIGHTOF(zz0) ((zz0) & 0xffffff00) -#define DEPTHOF(zz1) ((zz1) & 0x000000ff) -#define MYMAX(zz2,zz3) ((zz2) > (zz3) ? (zz2) : (zz3)) - -#define ADDWEIGHTS(zw1,zw2) \ - (WEIGHTOF(zw1)+WEIGHTOF(zw2)) | \ - (1 + MYMAX(DEPTHOF(zw1),DEPTHOF(zw2))) - -#define UPHEAP(z) \ -{ \ - Int32 zz, tmp; \ - zz = z; tmp = heap[zz]; \ - while (weight[tmp] < weight[heap[zz >> 1]]) { \ - heap[zz] = heap[zz >> 1]; \ - zz >>= 1; \ - } \ - heap[zz] = tmp; \ -} - -#define DOWNHEAP(z) \ -{ \ - Int32 zz, yy, tmp; \ - zz = z; tmp = heap[zz]; \ - while (True) { \ - yy = zz << 1; \ - if (yy > nHeap) break; \ - if (yy < nHeap && \ - weight[heap[yy+1]] < weight[heap[yy]]) \ - yy++; \ - if (weight[tmp] < weight[heap[yy]]) break; \ - heap[zz] = heap[yy]; \ - zz = yy; \ - } \ - heap[zz] = tmp; \ -} - - -/*---------------------------------------------------*/ -void BZ2_hbMakeCodeLengths ( UChar *len, - Int32 *freq, - Int32 alphaSize, - Int32 maxLen ) -{ - /*-- - Nodes and heap entries run from 1. Entry 0 - for both the heap and nodes is a sentinel. - --*/ - Int32 nNodes, nHeap, n1, n2, i, j, k; - Bool tooLong; - - Int32 heap [ BZ_MAX_ALPHA_SIZE + 2 ]; - Int32 weight [ BZ_MAX_ALPHA_SIZE * 2 ]; - Int32 parent [ BZ_MAX_ALPHA_SIZE * 2 ]; - - for (i = 0; i < alphaSize; i++) - weight[i+1] = (freq[i] == 0 ? 1 : freq[i]) << 8; - - while (True) { - - nNodes = alphaSize; - nHeap = 0; - - heap[0] = 0; - weight[0] = 0; - parent[0] = -2; - - for (i = 1; i <= alphaSize; i++) { - parent[i] = -1; - nHeap++; - heap[nHeap] = i; - UPHEAP(nHeap); - } - - AssertH( nHeap < (BZ_MAX_ALPHA_SIZE+2), 2001 ); - - while (nHeap > 1) { - n1 = heap[1]; heap[1] = heap[nHeap]; nHeap--; DOWNHEAP(1); - n2 = heap[1]; heap[1] = heap[nHeap]; nHeap--; DOWNHEAP(1); - nNodes++; - parent[n1] = parent[n2] = nNodes; - weight[nNodes] = ADDWEIGHTS(weight[n1], weight[n2]); - parent[nNodes] = -1; - nHeap++; - heap[nHeap] = nNodes; - UPHEAP(nHeap); - } - - AssertH( nNodes < (BZ_MAX_ALPHA_SIZE * 2), 2002 ); - - tooLong = False; - for (i = 1; i <= alphaSize; i++) { - j = 0; - k = i; - while (parent[k] >= 0) { k = parent[k]; j++; } - len[i-1] = j; - if (j > maxLen) tooLong = True; - } - - if (! tooLong) break; - - /* 17 Oct 04: keep-going condition for the following loop used - to be 'i < alphaSize', which missed the last element, - theoretically leading to the possibility of the compressor - looping. However, this count-scaling step is only needed if - one of the generated Huffman code words is longer than - maxLen, which up to and including version 1.0.2 was 20 bits, - which is extremely unlikely. In version 1.0.3 maxLen was - changed to 17 bits, which has minimal effect on compression - ratio, but does mean this scaling step is used from time to - time, enough to verify that it works. - - This means that bzip2-1.0.3 and later will only produce - Huffman codes with a maximum length of 17 bits. However, in - order to preserve backwards compatibility with bitstreams - produced by versions pre-1.0.3, the decompressor must still - handle lengths of up to 20. */ - - for (i = 1; i <= alphaSize; i++) { - j = weight[i] >> 8; - j = 1 + (j / 2); - weight[i] = j << 8; - } - } -} - - -/*---------------------------------------------------*/ -void BZ2_hbAssignCodes ( Int32 *code, - UChar *length, - Int32 minLen, - Int32 maxLen, - Int32 alphaSize ) -{ - Int32 n, vec, i; - - vec = 0; - for (n = minLen; n <= maxLen; n++) { - for (i = 0; i < alphaSize; i++) - if (length[i] == n) { code[i] = vec; vec++; }; - vec <<= 1; - } -} - - -/*---------------------------------------------------*/ -void BZ2_hbCreateDecodeTables ( Int32 *limit, - Int32 *base, - Int32 *perm, - UChar *length, - Int32 minLen, - Int32 maxLen, - Int32 alphaSize ) -{ - Int32 pp, i, j, vec; - - pp = 0; - for (i = minLen; i <= maxLen; i++) - for (j = 0; j < alphaSize; j++) - if (length[j] == i) { perm[pp] = j; pp++; }; - - for (i = 0; i < BZ_MAX_CODE_LEN; i++) base[i] = 0; - for (i = 0; i < alphaSize; i++) base[length[i]+1]++; - - for (i = 1; i < BZ_MAX_CODE_LEN; i++) base[i] += base[i-1]; - - for (i = 0; i < BZ_MAX_CODE_LEN; i++) limit[i] = 0; - vec = 0; - - for (i = minLen; i <= maxLen; i++) { - vec += (base[i+1] - base[i]); - limit[i] = vec-1; - vec <<= 1; - } - for (i = minLen + 1; i <= maxLen; i++) - base[i] = ((limit[i-1] + 1) << 1) - base[i]; -} - - -/*-------------------------------------------------------------*/ -/*--- end huffman.c ---*/ -/*-------------------------------------------------------------*/ + + Please read the WARNING, DISCLAIMER and PATENTS sections in the + README file. + + This program is released under the terms of the license contained + in the file LICENSE. + ------------------------------------------------------------------ */ + + +#include "bzlib_private.h" + +/*---------------------------------------------------*/ +#define WEIGHTOF(zz0) ((zz0) & 0xffffff00) +#define DEPTHOF(zz1) ((zz1) & 0x000000ff) +#define MYMAX(zz2,zz3) ((zz2) > (zz3) ? (zz2) : (zz3)) + +#define ADDWEIGHTS(zw1,zw2) \ + (WEIGHTOF(zw1)+WEIGHTOF(zw2)) | \ + (1 + MYMAX(DEPTHOF(zw1),DEPTHOF(zw2))) + +#define UPHEAP(z) \ +{ \ + Int32 zz, tmp; \ + zz = z; tmp = heap[zz]; \ + while (weight[tmp] < weight[heap[zz >> 1]]) { \ + heap[zz] = heap[zz >> 1]; \ + zz >>= 1; \ + } \ + heap[zz] = tmp; \ +} + +#define DOWNHEAP(z) \ +{ \ + Int32 zz, yy, tmp; \ + zz = z; tmp = heap[zz]; \ + while (True) { \ + yy = zz << 1; \ + if (yy > nHeap) break; \ + if (yy < nHeap && \ + weight[heap[yy+1]] < weight[heap[yy]]) \ + yy++; \ + if (weight[tmp] < weight[heap[yy]]) break; \ + heap[zz] = heap[yy]; \ + zz = yy; \ + } \ + heap[zz] = tmp; \ +} + + +/*---------------------------------------------------*/ +void BZ2_hbMakeCodeLengths ( UChar *len, + Int32 *freq, + Int32 alphaSize, + Int32 maxLen ) +{ + /*-- + Nodes and heap entries run from 1. Entry 0 + for both the heap and nodes is a sentinel. + --*/ + Int32 nNodes, nHeap, n1, n2, i, j, k; + Bool tooLong; + + Int32 heap [ BZ_MAX_ALPHA_SIZE + 2 ]; + Int32 weight [ BZ_MAX_ALPHA_SIZE * 2 ]; + Int32 parent [ BZ_MAX_ALPHA_SIZE * 2 ]; + + for (i = 0; i < alphaSize; i++) + weight[i+1] = (freq[i] == 0 ? 1 : freq[i]) << 8; + + while (True) { + + nNodes = alphaSize; + nHeap = 0; + + heap[0] = 0; + weight[0] = 0; + parent[0] = -2; + + for (i = 1; i <= alphaSize; i++) { + parent[i] = -1; + nHeap++; + heap[nHeap] = i; + UPHEAP(nHeap); + } + + AssertH( nHeap < (BZ_MAX_ALPHA_SIZE+2), 2001 ); + + while (nHeap > 1) { + n1 = heap[1]; heap[1] = heap[nHeap]; nHeap--; DOWNHEAP(1); + n2 = heap[1]; heap[1] = heap[nHeap]; nHeap--; DOWNHEAP(1); + nNodes++; + parent[n1] = parent[n2] = nNodes; + weight[nNodes] = ADDWEIGHTS(weight[n1], weight[n2]); + parent[nNodes] = -1; + nHeap++; + heap[nHeap] = nNodes; + UPHEAP(nHeap); + } + + AssertH( nNodes < (BZ_MAX_ALPHA_SIZE * 2), 2002 ); + + tooLong = False; + for (i = 1; i <= alphaSize; i++) { + j = 0; + k = i; + while (parent[k] >= 0) { k = parent[k]; j++; } + len[i-1] = j; + if (j > maxLen) tooLong = True; + } + + if (! tooLong) break; + + /* 17 Oct 04: keep-going condition for the following loop used + to be 'i < alphaSize', which missed the last element, + theoretically leading to the possibility of the compressor + looping. However, this count-scaling step is only needed if + one of the generated Huffman code words is longer than + maxLen, which up to and including version 1.0.2 was 20 bits, + which is extremely unlikely. In version 1.0.3 maxLen was + changed to 17 bits, which has minimal effect on compression + ratio, but does mean this scaling step is used from time to + time, enough to verify that it works. + + This means that bzip2-1.0.3 and later will only produce + Huffman codes with a maximum length of 17 bits. However, in + order to preserve backwards compatibility with bitstreams + produced by versions pre-1.0.3, the decompressor must still + handle lengths of up to 20. */ + + for (i = 1; i <= alphaSize; i++) { + j = weight[i] >> 8; + j = 1 + (j / 2); + weight[i] = j << 8; + } + } +} + + +/*---------------------------------------------------*/ +void BZ2_hbAssignCodes ( Int32 *code, + UChar *length, + Int32 minLen, + Int32 maxLen, + Int32 alphaSize ) +{ + Int32 n, vec, i; + + vec = 0; + for (n = minLen; n <= maxLen; n++) { + for (i = 0; i < alphaSize; i++) + if (length[i] == n) { code[i] = vec; vec++; }; + vec <<= 1; + } +} + + +/*---------------------------------------------------*/ +void BZ2_hbCreateDecodeTables ( Int32 *limit, + Int32 *base, + Int32 *perm, + UChar *length, + Int32 minLen, + Int32 maxLen, + Int32 alphaSize ) +{ + Int32 pp, i, j, vec; + + pp = 0; + for (i = minLen; i <= maxLen; i++) + for (j = 0; j < alphaSize; j++) + if (length[j] == i) { perm[pp] = j; pp++; }; + + for (i = 0; i < BZ_MAX_CODE_LEN; i++) base[i] = 0; + for (i = 0; i < alphaSize; i++) base[length[i]+1]++; + + for (i = 1; i < BZ_MAX_CODE_LEN; i++) base[i] += base[i-1]; + + for (i = 0; i < BZ_MAX_CODE_LEN; i++) limit[i] = 0; + vec = 0; + + for (i = minLen; i <= maxLen; i++) { + vec += (base[i+1] - base[i]); + limit[i] = vec-1; + vec <<= 1; + } + for (i = minLen + 1; i <= maxLen; i++) + base[i] = ((limit[i-1] + 1) << 1) - base[i]; +} + + +/*-------------------------------------------------------------*/ +/*--- end huffman.c ---*/ +/*-------------------------------------------------------------*/ |