[3365] | 1 | /**************************************************************************
|
---|
| 2 | *
|
---|
| 3 | * mgpp_fast_comp_dict.cpp -- Program to generate a fast compression dictionary
|
---|
| 4 | * Copyright (C) 1994 Neil Sharman
|
---|
| 5 | *
|
---|
| 6 | * This program is free software; you can redistribute it and/or modify
|
---|
| 7 | * it under the terms of the GNU General Public License as published by
|
---|
| 8 | * the Free Software Foundation; either version 2 of the License, or
|
---|
| 9 | * (at your option) any later version.
|
---|
| 10 | *
|
---|
| 11 | * This program is distributed in the hope that it will be useful,
|
---|
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 14 | * GNU General Public License for more details.
|
---|
| 15 | *
|
---|
| 16 | * You should have received a copy of the GNU General Public License
|
---|
| 17 | * along with this program; if not, write to the Free Software
|
---|
| 18 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 19 | *
|
---|
| 20 | **************************************************************************/
|
---|
| 21 |
|
---|
| 22 | #define _XOPEN_SOURCE 1
|
---|
| 23 | #define _XOPEN_SOURCE_EXTENDED 1
|
---|
| 24 |
|
---|
| 25 | // need this to avoid bizarre compiler problems under VC++ 6.0
|
---|
| 26 | #if defined (__WIN32__) && !defined (GSDL_USE_IOS_H)
|
---|
| 27 | # include <iostream>
|
---|
| 28 | #endif
|
---|
| 29 |
|
---|
| 30 | /* getopt is in posix.2, so cygwin should have it in unistd, but doesn't */
|
---|
| 31 | #if defined (__WIN32__) || defined (__CYGWIN__)
|
---|
| 32 | # include "getopt_old.h"
|
---|
| 33 | #else
|
---|
| 34 | # include <unistd.h>
|
---|
| 35 | #endif
|
---|
| 36 |
|
---|
| 37 | #include "sysfuncs.h"
|
---|
| 38 | #include "huffman.h"
|
---|
| 39 | #include "messages.h"
|
---|
| 40 | #include "memlib.h"
|
---|
| 41 | #include "netorder.h" /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 42 |
|
---|
| 43 | #include "local_strings.h"
|
---|
| 44 | #include "mg.h"
|
---|
| 45 | #include "text.h"
|
---|
| 46 | #include "invf.h"
|
---|
| 47 | #include "mg_files.h"
|
---|
| 48 | #include "locallib.h"
|
---|
| 49 | #include "words.h"
|
---|
| 50 |
|
---|
| 51 |
|
---|
| 52 |
|
---|
| 53 | #define ALIGN_SIZE(p,t) (((p) + (sizeof(t)-1)) & (~(sizeof(t)-1)))
|
---|
| 54 |
|
---|
| 55 | #define WORDNO(p, base) ((((char*)(p))-((char*)(base)))/sizeof(u_char*))
|
---|
| 56 | #define FIXUP(p) (fixup[WORDNO(p,buffer)/8] |= (1<<(WORDNO(p, buffer) & 7)))
|
---|
| 57 |
|
---|
| 58 | #define IS_FIXUP(p) ((fixup[WORDNO(p,buffer)/8] & (1<<(WORDNO(p, buffer) & 7))) != 0)
|
---|
| 59 |
|
---|
| 60 | #define FIXUP_VALS(vals) do { \
|
---|
| 61 | int i; \
|
---|
[8692] | 62 | for (i=0; i < MAX_HUFFCODE_LEN+1; ++i) \
|
---|
[3365] | 63 | FIXUP(&vals[i]); \
|
---|
| 64 | } while(0)
|
---|
| 65 |
|
---|
| 66 |
|
---|
| 67 |
|
---|
| 68 | static u_long mem_for_comp_dict (char *filename);
|
---|
| 69 | static void load_comp_dict (char *filename);
|
---|
| 70 | static void save_fast_dict (char *filename);
|
---|
| 71 | static void unfixup_buffer (void);
|
---|
| 72 |
|
---|
| 73 |
|
---|
| 74 | static void *buffer;
|
---|
| 75 | static void *cur;
|
---|
| 76 | static u_char *fixup;
|
---|
| 77 | static u_long mem, fixup_mem;
|
---|
| 78 |
|
---|
| 79 | int main (int argc, char **argv)
|
---|
| 80 | {
|
---|
| 81 | int ch;
|
---|
| 82 | char *filename = "";
|
---|
| 83 | opterr = 0;
|
---|
| 84 | msg_prefix = argv[0];
|
---|
| 85 | while ((ch = getopt (argc, argv, "f:d:h")) != -1)
|
---|
| 86 | switch (ch)
|
---|
| 87 | {
|
---|
| 88 | case 'f': /* input file */
|
---|
| 89 | filename = optarg;
|
---|
| 90 | break;
|
---|
| 91 | case 'd':
|
---|
| 92 | set_basepath (optarg);
|
---|
| 93 | break;
|
---|
| 94 | case 'h':
|
---|
| 95 | case '?':
|
---|
| 96 | fprintf (stderr, "usage: %s [-f input_file] [-d data directory] [-h]\n",
|
---|
| 97 | argv[0]);
|
---|
| 98 | exit (1);
|
---|
| 99 | }
|
---|
| 100 |
|
---|
| 101 | mem = mem_for_comp_dict (filename);
|
---|
| 102 |
|
---|
| 103 | fixup_mem = (ALIGN_SIZE (mem, u_char *) / sizeof (u_char *) + 7) / 8;
|
---|
| 104 |
|
---|
| 105 | cur = buffer = Xmalloc (mem);
|
---|
| 106 | memset (buffer, 0, mem);
|
---|
| 107 | fixup = (u_char *) Xmalloc (fixup_mem);
|
---|
| 108 | memset (fixup, 0, fixup_mem);
|
---|
| 109 |
|
---|
| 110 | #ifndef SILENT
|
---|
| 111 | Message ("Estimated memory for fast_dict %u", mem);
|
---|
| 112 | Message ("Estimated memory for fixups %u", fixup_mem);
|
---|
| 113 | #endif
|
---|
| 114 |
|
---|
| 115 | load_comp_dict (filename);
|
---|
| 116 |
|
---|
| 117 | #ifndef SILENT
|
---|
| 118 | Message ("Actual memory for fast_dict %u", (char *) cur - (char *) buffer);
|
---|
| 119 | #endif
|
---|
| 120 |
|
---|
| 121 | if ((u_long) cur > (u_long) buffer + mem)
|
---|
| 122 | FatalError (1, "The buffer was not big enough for the dictionary");
|
---|
| 123 |
|
---|
| 124 | {
|
---|
| 125 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 126 | compression_dict *cd = (compression_dict *) buffer;
|
---|
| 127 | int i, which;
|
---|
| 128 |
|
---|
| 129 | /* cfh */
|
---|
[8692] | 130 | for (which = 0; which <= 1; ++which)
|
---|
[3365] | 131 | {
|
---|
| 132 | int j;
|
---|
| 133 |
|
---|
| 134 | HTONSI(cd->cfh[which]->hd.num_codes);
|
---|
| 135 | HTONSI(cd->cfh[which]->hd.mincodelen);
|
---|
| 136 | HTONSI(cd->cfh[which]->hd.maxcodelen);
|
---|
[8692] | 137 | for (j = 0; j < MAX_HUFFCODE_LEN + 1; ++j)
|
---|
[3365] | 138 | {
|
---|
| 139 | HTONSI(cd->cfh[which]->hd.lencount[j]);
|
---|
| 140 | HTONUL(cd->cfh[which]->hd.min_code[j]);
|
---|
| 141 | }
|
---|
| 142 | HTONUL(cd->cfh[which]->uncompressed_size);
|
---|
[8692] | 143 | for (j = 0; j < MAX_HUFFCODE_LEN + 1; ++j)
|
---|
[3365] | 144 | HTONUL(cd->cfh[which]->huff_words_size[j]);
|
---|
| 145 | }
|
---|
| 146 | HTONUL(cd->MemForCompDict);
|
---|
| 147 | /* ad */
|
---|
| 148 | if (cd->cdh.novel_method == MG_NOVEL_DELTA ||
|
---|
| 149 | cd->cdh.novel_method == MG_NOVEL_HYBRID)
|
---|
[8692] | 150 | for (which = 0; which <= 1; ++which)
|
---|
[3365] | 151 | {
|
---|
| 152 | int j;
|
---|
| 153 |
|
---|
| 154 | HTONUL(cd->ad->afh[which].num_frags);
|
---|
| 155 | HTONUL(cd->ad->afh[which].mem_for_frags);
|
---|
[8692] | 156 | for (j = 0; j < 33; ++j)
|
---|
[3365] | 157 | {
|
---|
| 158 | HTONSI(cd->ad->blk_start[which][j]);
|
---|
| 159 | HTONSI(cd->ad->blk_end[which][j]);
|
---|
| 160 | }
|
---|
| 161 | }
|
---|
| 162 | /* cdh */
|
---|
| 163 | HTONUL(cd->cdh.dict_type);
|
---|
| 164 | HTONUL(cd->cdh.novel_method);
|
---|
[8692] | 165 | for (i = 0; i < TEXT_PARAMS; ++i)
|
---|
[3365] | 166 | HTONUL(cd->cdh.params[which]);
|
---|
| 167 | HTONUL(cd->cdh.num_words[0]);
|
---|
| 168 | HTONUL(cd->cdh.num_words[1]);
|
---|
| 169 | HTONUL(cd->cdh.num_word_chars[0]);
|
---|
| 170 | HTONUL(cd->cdh.num_word_chars[1]);
|
---|
| 171 | HTONUL(cd->cdh.lookback);
|
---|
| 172 | HTONSI(cd->fast_loaded);
|
---|
| 173 | }
|
---|
| 174 |
|
---|
| 175 | unfixup_buffer ();
|
---|
| 176 |
|
---|
| 177 | save_fast_dict (filename);
|
---|
| 178 |
|
---|
| 179 | return (0);
|
---|
| 180 | }
|
---|
| 181 |
|
---|
| 182 |
|
---|
| 183 |
|
---|
| 184 | static void
|
---|
| 185 | unfixup_buffer ()
|
---|
| 186 | {
|
---|
| 187 | u_long *p;
|
---|
[8692] | 188 | for (p = (u_long *) buffer; (u_long) p < (u_long) cur; ++p)
|
---|
[3365] | 189 | {
|
---|
| 190 | if (IS_FIXUP (p))
|
---|
| 191 | *p = *p - (u_long) buffer;
|
---|
| 192 | }
|
---|
| 193 | }
|
---|
| 194 |
|
---|
| 195 |
|
---|
| 196 |
|
---|
| 197 |
|
---|
| 198 | static u_long
|
---|
| 199 | mem_for_aux_dict (compression_dict_header * /*cdh*/, char *filename)
|
---|
| 200 | {
|
---|
| 201 | int i;
|
---|
| 202 | u_long mem = sizeof (auxiliary_dict);
|
---|
| 203 | FILE *aux;
|
---|
| 204 |
|
---|
| 205 | aux = open_file (filename, TEXT_DICT_AUX_SUFFIX, "rb",
|
---|
| 206 | MAGIC_AUX_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 207 |
|
---|
[8692] | 208 | for (i = 0; i <= 1; ++i)
|
---|
[3365] | 209 | {
|
---|
| 210 | aux_frags_header afh;
|
---|
| 211 | fread (&afh, sizeof (afh), 1, aux);
|
---|
| 212 | NTOHUL(afh.num_frags); /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 213 | NTOHUL(afh.mem_for_frags); /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 214 | mem += afh.num_frags * sizeof (u_char *);
|
---|
| 215 | mem = ALIGN_SIZE (mem + afh.mem_for_frags, u_char *);
|
---|
| 216 | fseek (aux, afh.mem_for_frags, SEEK_CUR);
|
---|
| 217 | }
|
---|
| 218 | fclose (aux);
|
---|
| 219 |
|
---|
| 220 | return mem;
|
---|
| 221 | }
|
---|
| 222 |
|
---|
| 223 |
|
---|
| 224 |
|
---|
| 225 | static u_long
|
---|
| 226 | mem_for_words (FILE * dict, compression_dict_header * cdh,
|
---|
| 227 | comp_frags_header * cfh)
|
---|
| 228 | {
|
---|
| 229 | u_long mem = 0;
|
---|
| 230 | long i, lookback;
|
---|
| 231 | int ptrs_reqd = 0;
|
---|
| 232 | int mem_reqd = 0;
|
---|
| 233 |
|
---|
| 234 | lookback = cdh->lookback;
|
---|
| 235 |
|
---|
[8692] | 236 | for (i = cfh->hd.mincodelen; i <= cfh->hd.maxcodelen; ++i)
|
---|
[3365] | 237 | {
|
---|
| 238 | ptrs_reqd += (cfh->hd.lencount[i] + ((1 << lookback) - 1)) >> lookback;
|
---|
| 239 | mem_reqd += cfh->huff_words_size[i];
|
---|
| 240 | }
|
---|
| 241 |
|
---|
| 242 | mem += ptrs_reqd * sizeof (u_char *);
|
---|
| 243 | mem += (MAX_HUFFCODE_LEN + 1) * sizeof (u_char **);
|
---|
| 244 | mem += mem_reqd;
|
---|
| 245 |
|
---|
[8692] | 246 | for (i = 0; i < cfh->hd.num_codes; ++i)
|
---|
[3365] | 247 | {
|
---|
| 248 | register int val;
|
---|
[8692] | 249 | for (val = getc (dict) & 0xf; val; --val)
|
---|
[3365] | 250 | getc (dict);
|
---|
| 251 | }
|
---|
| 252 | return ALIGN_SIZE (mem, u_char *);
|
---|
| 253 | }
|
---|
| 254 |
|
---|
| 255 | static u_long mem_skip_hd(FILE *dict, u_long mem)
|
---|
| 256 | { huff_data hd;
|
---|
| 257 |
|
---|
| 258 | mem += sizeof (hd);
|
---|
| 259 | Read_Huffman_Data (dict, &hd, NULL, NULL);
|
---|
| 260 | if (hd.clens)
|
---|
[8692] | 261 | delete []hd.clens;
|
---|
[3365] | 262 | mem += hd.num_codes * sizeof (unsigned long);
|
---|
| 263 | mem += (MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *);
|
---|
| 264 | return mem;
|
---|
| 265 | }
|
---|
| 266 |
|
---|
| 267 | static u_long mem_read_cfh(FILE *dict, compression_dict_header *cdh, comp_frags_header *cfh,
|
---|
| 268 | u_long mem)
|
---|
| 269 | {
|
---|
| 270 | mem += sizeof (comp_frags_header);
|
---|
| 271 | Read_cfh (dict, cfh, NULL, NULL);
|
---|
| 272 |
|
---|
| 273 | /* Don't need to count the space for the clens of the huffman data */
|
---|
| 274 |
|
---|
| 275 | mem += mem_for_words (dict, cdh, cfh);
|
---|
| 276 | if (cfh->hd.clens)
|
---|
[8692] | 277 | delete []cfh->hd.clens;
|
---|
[3365] | 278 |
|
---|
| 279 | return mem;
|
---|
| 280 | }
|
---|
| 281 |
|
---|
| 282 |
|
---|
| 283 | static u_long
|
---|
| 284 | mem_for_comp_dict (char *filename)
|
---|
| 285 | {
|
---|
| 286 | u_long mem = sizeof (compression_dict);
|
---|
| 287 | compression_dict_header cdh;
|
---|
| 288 | comp_frags_header cfh;
|
---|
| 289 | FILE *dict;
|
---|
| 290 | int which;
|
---|
| 291 | int i; /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 292 |
|
---|
| 293 | dict = open_file (filename, TEXT_DICT_SUFFIX, "rb",
|
---|
| 294 | MAGIC_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 295 |
|
---|
| 296 | fread (&cdh, sizeof (cdh), 1, dict);
|
---|
| 297 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 298 | NTOHUL(cdh.dict_type);
|
---|
| 299 | NTOHUL(cdh.novel_method);
|
---|
[8692] | 300 | for (i = 0; i < TEXT_PARAMS; ++i)
|
---|
[3365] | 301 | NTOHUL(cdh.params[i]);
|
---|
| 302 | NTOHUL(cdh.num_words[0]);
|
---|
| 303 | NTOHUL(cdh.num_words[1]);
|
---|
| 304 | NTOHUL(cdh.num_word_chars[0]);
|
---|
| 305 | NTOHUL(cdh.num_word_chars[1]);
|
---|
| 306 | NTOHUL(cdh.lookback);
|
---|
| 307 |
|
---|
[8692] | 308 | for (which = 0; which < 2; ++which)
|
---|
[3365] | 309 | switch (cdh.dict_type)
|
---|
| 310 | {
|
---|
| 311 | case MG_COMPLETE_DICTIONARY:
|
---|
| 312 | {
|
---|
| 313 | mem = mem_read_cfh(dict, &cdh, &cfh, mem);
|
---|
| 314 | }
|
---|
| 315 | break;
|
---|
| 316 | case MG_PARTIAL_DICTIONARY:
|
---|
| 317 | {
|
---|
| 318 | // huff_data hd;
|
---|
| 319 | if (cdh.num_words[which])
|
---|
| 320 | {
|
---|
| 321 | mem = mem_read_cfh(dict, &cdh, &cfh, mem);
|
---|
| 322 | }
|
---|
| 323 |
|
---|
| 324 | mem = mem_skip_hd(dict, mem);
|
---|
| 325 | mem = mem_skip_hd(dict, mem);
|
---|
| 326 | }
|
---|
| 327 | break;
|
---|
| 328 | case MG_SEED_DICTIONARY:
|
---|
| 329 | {
|
---|
| 330 | // huff_data hd;
|
---|
| 331 | if (cdh.num_words[which])
|
---|
| 332 | {
|
---|
| 333 | mem = mem_read_cfh(dict, &cdh, &cfh, mem);
|
---|
| 334 | }
|
---|
| 335 | switch (cdh.novel_method)
|
---|
| 336 | {
|
---|
| 337 | case MG_NOVEL_HUFFMAN_CHARS:
|
---|
| 338 | mem = mem_skip_hd(dict, mem);
|
---|
| 339 | mem = mem_skip_hd(dict, mem);
|
---|
| 340 | break;
|
---|
| 341 | case MG_NOVEL_DELTA:
|
---|
| 342 | break;
|
---|
| 343 | case MG_NOVEL_HYBRID:
|
---|
| 344 | break;
|
---|
| 345 | }
|
---|
| 346 | break;
|
---|
| 347 | }
|
---|
| 348 | }
|
---|
| 349 | fclose (dict);
|
---|
| 350 |
|
---|
| 351 | if (cdh.novel_method == MG_NOVEL_DELTA ||
|
---|
| 352 | cdh.novel_method == MG_NOVEL_HYBRID)
|
---|
| 353 | mem += mem_for_aux_dict (&cdh, filename);
|
---|
| 354 |
|
---|
| 355 | return ALIGN_SIZE (mem, u_char *);
|
---|
| 356 | }
|
---|
| 357 |
|
---|
| 358 |
|
---|
| 359 |
|
---|
| 360 | void *
|
---|
| 361 | getmem (u_long size, int align)
|
---|
| 362 | {
|
---|
| 363 | void *res;
|
---|
| 364 | cur = (void *) (((u_long) cur + (align - 1)) & (~(align - 1)));
|
---|
| 365 | res = cur;
|
---|
| 366 | cur = (char *) cur + size;
|
---|
| 367 | if ((u_long) cur > (u_long)buffer + mem)
|
---|
| 368 | FatalError (1, "The buffer was not big enough for the dictionary");
|
---|
| 369 | return res;
|
---|
| 370 | }
|
---|
| 371 |
|
---|
| 372 |
|
---|
| 373 |
|
---|
| 374 |
|
---|
| 375 |
|
---|
| 376 |
|
---|
| 377 |
|
---|
| 378 |
|
---|
| 379 |
|
---|
| 380 |
|
---|
| 381 |
|
---|
| 382 | static auxiliary_dict *
|
---|
| 383 | LoadAuxDict (compression_dict_header * cdh,
|
---|
| 384 | char *filename)
|
---|
| 385 | {
|
---|
| 386 | auxiliary_dict *ad;
|
---|
| 387 | int i;
|
---|
| 388 | FILE *aux;
|
---|
| 389 |
|
---|
| 390 | aux = open_file (filename, TEXT_DICT_AUX_SUFFIX, "rb",
|
---|
| 391 | MAGIC_AUX_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 392 |
|
---|
| 393 | ad = (auxiliary_dict *) getmem (sizeof (auxiliary_dict), sizeof (u_char *));
|
---|
| 394 |
|
---|
[8692] | 395 | for (i = 0; i <= 1; ++i)
|
---|
[3365] | 396 | {
|
---|
| 397 | unsigned int j;
|
---|
| 398 | u_char *pos;
|
---|
| 399 |
|
---|
| 400 | fread (&ad->afh[i], sizeof (aux_frags_header), 1, aux);
|
---|
| 401 |
|
---|
| 402 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 403 | NTOHUL(ad->afh[i].num_frags);
|
---|
| 404 | NTOHUL(ad->afh[i].mem_for_frags);
|
---|
| 405 |
|
---|
| 406 | ad->word_data[i] = (u_char *) getmem (ad->afh[i].mem_for_frags, 1);
|
---|
| 407 | FIXUP (&ad->word_data[i]);
|
---|
| 408 |
|
---|
| 409 | ad->words[i] = (u_char **) getmem (ad->afh[i].num_frags * sizeof (u_char *),
|
---|
| 410 | sizeof (u_char *));
|
---|
| 411 | FIXUP (&ad->words[i]);
|
---|
| 412 |
|
---|
| 413 | fread (ad->word_data[i], ad->afh[i].mem_for_frags, sizeof (u_char), aux);
|
---|
| 414 |
|
---|
| 415 | pos = ad->word_data[i];
|
---|
[8692] | 416 | for (j = 0; j < ad->afh[i].num_frags; ++j)
|
---|
[3365] | 417 | {
|
---|
| 418 | ad->words[i][j] = pos;
|
---|
| 419 | FIXUP (&ad->words[i][j]);
|
---|
| 420 | pos += *pos + 1;
|
---|
| 421 | }
|
---|
| 422 | if (cdh->novel_method == MG_NOVEL_HYBRID)
|
---|
| 423 | {
|
---|
| 424 | int num;
|
---|
| 425 | num = 1;
|
---|
| 426 | ad->blk_start[i][0] = 0;
|
---|
| 427 | ad->blk_end[i][0] = cdh->num_words[i] - 1;
|
---|
| 428 | while (num < 33)
|
---|
| 429 | {
|
---|
| 430 | ad->blk_start[i][num] = ad->blk_end[i][num - 1] + 1;
|
---|
| 431 | ad->blk_end[i][num] = ad->blk_start[i][num] +
|
---|
| 432 | (ad->blk_end[i][num - 1] - ad->blk_start[i][num - 1]) * 2;
|
---|
[8692] | 433 | ++num;
|
---|
[3365] | 434 | }
|
---|
| 435 | }
|
---|
| 436 | }
|
---|
| 437 | fclose (aux);
|
---|
| 438 | return (ad);
|
---|
| 439 | }
|
---|
| 440 |
|
---|
| 441 |
|
---|
| 442 |
|
---|
| 443 |
|
---|
| 444 |
|
---|
| 445 | static u_char ***
|
---|
| 446 | ReadInWords (FILE * dict, compression_dict * cd,
|
---|
| 447 | comp_frags_header * cfh, u_char ** escape)
|
---|
| 448 | {
|
---|
| 449 | int i, lookback;
|
---|
| 450 | int ptrs_reqd = 0;
|
---|
| 451 | int mem_reqd = 0;
|
---|
| 452 | int num_set[MAX_HUFFCODE_LEN + 1];
|
---|
| 453 | u_char *next_word[MAX_HUFFCODE_LEN + 1];
|
---|
| 454 | u_char **vals;
|
---|
| 455 | u_char ***values;
|
---|
| 456 | u_char word[MAXWORDLEN + 1];
|
---|
| 457 | u_char last_word[MAX_HUFFCODE_LEN + 1][MAXWORDLEN + 1];
|
---|
| 458 |
|
---|
| 459 | lookback = cd->cdh.lookback;
|
---|
| 460 |
|
---|
[8692] | 461 | for (i = cfh->hd.mincodelen; i <= cfh->hd.maxcodelen; ++i)
|
---|
[3365] | 462 | {
|
---|
| 463 | ptrs_reqd += (cfh->hd.lencount[i] + ((1 << lookback) - 1)) >> lookback;
|
---|
| 464 | mem_reqd += cfh->huff_words_size[i];
|
---|
| 465 | }
|
---|
| 466 |
|
---|
| 467 | vals = (u_char **) getmem (ptrs_reqd * sizeof (*vals), sizeof (u_char *));
|
---|
| 468 |
|
---|
| 469 | values = (u_char ***) getmem ((MAX_HUFFCODE_LEN + 1) * sizeof (u_char **), sizeof (u_char **));
|
---|
| 470 |
|
---|
| 471 | next_word[0] = (u_char *) getmem (mem_reqd, sizeof (char));
|
---|
| 472 |
|
---|
| 473 | cd->MemForCompDict += ptrs_reqd * sizeof (*vals) +
|
---|
| 474 | (MAX_HUFFCODE_LEN + 1) * sizeof (u_char **) +
|
---|
| 475 | mem_reqd;
|
---|
| 476 |
|
---|
| 477 | values[0] = vals;
|
---|
| 478 | FIXUP (&values[0]);
|
---|
| 479 | values[0][0] = next_word[0];
|
---|
| 480 | FIXUP (&values[0][0]);
|
---|
[8692] | 481 | for (i = 1; i <= cfh->hd.maxcodelen; ++i)
|
---|
[3365] | 482 | {
|
---|
| 483 | int next_start = (values[i - 1] - vals) +
|
---|
| 484 | ((cfh->hd.lencount[i - 1] + ((1 << lookback) - 1)) >> lookback);
|
---|
| 485 | values[i] = &vals[next_start];
|
---|
| 486 | FIXUP (&values[i]);
|
---|
| 487 | next_word[i] = next_word[i - 1] + cfh->huff_words_size[i - 1];
|
---|
| 488 | values[i][0] = next_word[i];
|
---|
| 489 | FIXUP (&values[i][0]);
|
---|
| 490 | }
|
---|
| 491 |
|
---|
| 492 | memset (num_set, '\0', sizeof (num_set));
|
---|
| 493 |
|
---|
[8692] | 494 | for (i = 0; i < cfh->hd.num_codes; ++i)
|
---|
[3365] | 495 | {
|
---|
| 496 | register int val, copy;
|
---|
| 497 | register int len = cfh->hd.clens[i];
|
---|
| 498 | val = getc (dict);
|
---|
| 499 | copy = (val >> 4) & 0xf;
|
---|
| 500 | val &= 0xf;
|
---|
| 501 |
|
---|
| 502 | fread (word + copy + 1, sizeof (u_char), val, dict);
|
---|
| 503 | *word = val + copy;
|
---|
| 504 |
|
---|
| 505 | if ((num_set[len] & ((1 << lookback) - 1)) == 0)
|
---|
| 506 | {
|
---|
| 507 | values[len][num_set[len] >> lookback] = next_word[len];
|
---|
| 508 | FIXUP (&values[len][num_set[len] >> lookback]);
|
---|
| 509 | memcpy (next_word[len], word, *word + 1);
|
---|
| 510 | if (escape && i == cfh->hd.num_codes - 1)
|
---|
| 511 | {
|
---|
| 512 | *escape = next_word[len];
|
---|
| 513 | FIXUP (escape);
|
---|
| 514 | }
|
---|
| 515 | next_word[len] += *word + 1;
|
---|
| 516 | }
|
---|
| 517 | else
|
---|
| 518 | {
|
---|
| 519 | copy = prefixlen (last_word[len], word);
|
---|
| 520 | memcpy (next_word[len] + 1, word + copy + 1, *word - copy);
|
---|
| 521 | *next_word[len] = (copy << 4) + (*word - copy);
|
---|
| 522 | if (escape && i == cfh->hd.num_codes - 1)
|
---|
| 523 | {
|
---|
| 524 | *escape = next_word[len];
|
---|
| 525 | FIXUP (escape);
|
---|
| 526 | }
|
---|
| 527 | next_word[len] += (*word - copy) + 1;
|
---|
| 528 | }
|
---|
| 529 | memcpy (last_word[len], word, *word + 1);
|
---|
[8692] | 530 | ++num_set[len];
|
---|
[3365] | 531 | }
|
---|
| 532 | if (cfh->hd.clens)
|
---|
[8692] | 533 | delete []cfh->hd.clens;
|
---|
[3365] | 534 | cfh->hd.clens = NULL;
|
---|
| 535 | return values;
|
---|
| 536 | }
|
---|
| 537 |
|
---|
| 538 |
|
---|
| 539 | static unsigned long **
|
---|
| 540 | Generate_Fast_Huffman_Vals (huff_data * data, u_long * mem)
|
---|
| 541 | {
|
---|
| 542 | int i;
|
---|
| 543 | unsigned long *fcode[MAX_HUFFCODE_LEN + 1];
|
---|
| 544 | unsigned long **values;
|
---|
| 545 | unsigned long *vals;
|
---|
| 546 |
|
---|
| 547 | if (!data)
|
---|
| 548 | return (NULL);
|
---|
| 549 | vals = (unsigned long *) getmem (data->num_codes * sizeof (*vals), sizeof (long *));
|
---|
| 550 | values = (unsigned long **) getmem ((MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *),
|
---|
| 551 | sizeof (long *));
|
---|
| 552 |
|
---|
| 553 | memset (values, '\0', (MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *));
|
---|
| 554 |
|
---|
| 555 | if (mem)
|
---|
| 556 | *mem += data->num_codes * sizeof (*vals) +
|
---|
| 557 | (MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *);
|
---|
| 558 |
|
---|
| 559 | fcode[0] = values[0] = &vals[0];
|
---|
| 560 | FIXUP (&values[0]);
|
---|
[8692] | 561 | for (i = 1; i <= data->maxcodelen; ++i)
|
---|
[3365] | 562 | {
|
---|
| 563 | fcode[i] = values[i] = &vals[(values[i - 1] - vals) + data->lencount[i - 1]];
|
---|
| 564 | FIXUP (&values[i]);
|
---|
| 565 | }
|
---|
| 566 |
|
---|
[8692] | 567 | for (i = 0; i < data->num_codes; ++i)
|
---|
[3365] | 568 | if (data->clens[i])
|
---|
| 569 | *fcode[(int) (data->clens[i])]++ = i;
|
---|
| 570 | return (values);
|
---|
| 571 | }
|
---|
| 572 |
|
---|
| 573 |
|
---|
| 574 | static void load_read_huffman(FILE *dict, compression_dict *cd, int which)
|
---|
| 575 | {
|
---|
| 576 | huff_data *hd;
|
---|
| 577 | u_long **vals;
|
---|
| 578 |
|
---|
| 579 | hd = (huff_data *) getmem (sizeof (huff_data), sizeof (char *));
|
---|
| 580 | cd->MemForCompDict += sizeof (huff_data);
|
---|
| 581 | Read_Huffman_Data (dict, hd, &cd->MemForCompDict, NULL);
|
---|
| 582 | vals = Generate_Fast_Huffman_Vals (hd, &cd->MemForCompDict);
|
---|
| 583 | cd->chars_huff[which] = hd;
|
---|
| 584 | FIXUP (&cd->chars_huff[which]);
|
---|
| 585 | cd->chars_vals[which] = vals;
|
---|
| 586 | FIXUP (&cd->chars_vals[which]);
|
---|
| 587 | if (hd->clens)
|
---|
[8692] | 588 | delete []hd->clens;
|
---|
[3365] | 589 | hd->clens = NULL;
|
---|
| 590 | }
|
---|
| 591 |
|
---|
| 592 | static void load_read_words(FILE *dict, compression_dict *cd, int which, int fixescape)
|
---|
| 593 | {
|
---|
| 594 | cd->cfh[which] = (comp_frags_header *) getmem (sizeof (*cd->cfh[which]), sizeof (u_char *));
|
---|
| 595 | cd->MemForCompDict += sizeof (*cd->cfh[which]);
|
---|
| 596 | Read_cfh (dict, cd->cfh[which], &cd->MemForCompDict, NULL);
|
---|
| 597 |
|
---|
| 598 | cd->values[which] = ReadInWords (dict, cd, cd->cfh[which], NULL);
|
---|
| 599 | FIXUP (&cd->cfh[which]);
|
---|
| 600 | FIXUP (&cd->values[which]);
|
---|
| 601 | if (fixescape)
|
---|
| 602 | { FIXUP(&cd->escape[which]);
|
---|
| 603 | }
|
---|
| 604 | else
|
---|
| 605 | {
|
---|
| 606 | cd->escape[which] = NULL;
|
---|
| 607 | }
|
---|
| 608 | }
|
---|
| 609 |
|
---|
| 610 | static void
|
---|
| 611 | load_comp_dict (char *filename)
|
---|
| 612 | {
|
---|
| 613 | FILE *dict;
|
---|
| 614 | int which;
|
---|
| 615 | compression_dict *cd;
|
---|
| 616 |
|
---|
| 617 | cd = (compression_dict *) getmem (sizeof (compression_dict), sizeof (u_char *));
|
---|
| 618 | cd->fast_loaded = 1;
|
---|
| 619 |
|
---|
| 620 | dict = open_file (filename, TEXT_DICT_SUFFIX, "rb",
|
---|
| 621 | MAGIC_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 622 |
|
---|
| 623 | Read_cdh (dict, &cd->cdh, NULL, NULL);
|
---|
| 624 |
|
---|
[8692] | 625 | for (which = 0; which < 2; ++which)
|
---|
[3365] | 626 | switch (cd->cdh.dict_type)
|
---|
| 627 | {
|
---|
| 628 | case MG_COMPLETE_DICTIONARY:
|
---|
| 629 | {
|
---|
| 630 | load_read_words(dict, cd, which, 0);
|
---|
| 631 | }
|
---|
| 632 | break;
|
---|
| 633 | case MG_PARTIAL_DICTIONARY:
|
---|
| 634 | {
|
---|
| 635 | if (cd->cdh.num_words[which])
|
---|
| 636 | {
|
---|
| 637 | load_read_words(dict, cd, which, 1);
|
---|
| 638 | }
|
---|
| 639 |
|
---|
| 640 | load_read_huffman(dict, cd, which);
|
---|
| 641 | load_read_huffman(dict, cd, which);
|
---|
| 642 | }
|
---|
| 643 | break;
|
---|
| 644 | case MG_SEED_DICTIONARY:
|
---|
| 645 | {
|
---|
| 646 | if (cd->cdh.num_words[which])
|
---|
| 647 | {
|
---|
| 648 | load_read_words(dict, cd, which, 1);
|
---|
| 649 | }
|
---|
| 650 | switch (cd->cdh.novel_method)
|
---|
| 651 | {
|
---|
| 652 | case MG_NOVEL_HUFFMAN_CHARS:
|
---|
| 653 | load_read_huffman(dict, cd, which);
|
---|
| 654 | load_read_huffman(dict, cd, which);
|
---|
| 655 | break;
|
---|
| 656 | case MG_NOVEL_DELTA:
|
---|
| 657 | break;
|
---|
| 658 | case MG_NOVEL_HYBRID:
|
---|
| 659 | break;
|
---|
| 660 | }
|
---|
| 661 | break;
|
---|
| 662 | }
|
---|
| 663 | }
|
---|
| 664 | fclose (dict);
|
---|
| 665 |
|
---|
| 666 |
|
---|
| 667 | if (cd->cdh.novel_method == MG_NOVEL_DELTA ||
|
---|
| 668 | cd->cdh.novel_method == MG_NOVEL_HYBRID)
|
---|
| 669 | {
|
---|
| 670 | cd->ad = LoadAuxDict (&cd->cdh, filename);
|
---|
| 671 | FIXUP (&cd->ad);
|
---|
| 672 | }
|
---|
| 673 | }
|
---|
| 674 |
|
---|
| 675 | static void
|
---|
| 676 | save_fast_dict (char *filename)
|
---|
| 677 | {
|
---|
| 678 | FILE *fdict;
|
---|
| 679 |
|
---|
| 680 | fdict = create_file (filename, TEXT_DICT_FAST_SUFFIX, "wb",
|
---|
| 681 | MAGIC_FAST_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 682 |
|
---|
| 683 | {
|
---|
| 684 | u_long *p;
|
---|
[8692] | 685 | for (p = (u_long *) buffer; (u_long) p < (u_long) cur; ++p)
|
---|
[3365] | 686 | {
|
---|
| 687 | if (IS_FIXUP (p))
|
---|
| 688 | HTONUL(*p);
|
---|
| 689 | }
|
---|
| 690 | }
|
---|
| 691 |
|
---|
| 692 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 693 | HTONUL(mem);
|
---|
| 694 | HTONUL(fixup_mem);
|
---|
| 695 |
|
---|
| 696 | fwrite (&mem, sizeof (mem), 1, fdict);
|
---|
| 697 | fwrite (&fixup_mem, sizeof (fixup_mem), 1, fdict);
|
---|
| 698 |
|
---|
| 699 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 700 | NTOHUL(mem);
|
---|
| 701 | NTOHUL(fixup_mem);
|
---|
| 702 |
|
---|
| 703 | fwrite (buffer, sizeof (u_char), mem, fdict);
|
---|
| 704 | fwrite (fixup, sizeof (u_char), fixup_mem, fdict);
|
---|
| 705 |
|
---|
| 706 | fclose (fdict);
|
---|
| 707 | }
|
---|