[3745] | 1 | /**************************************************************************
|
---|
| 2 | *
|
---|
| 3 | * mg_fast_comp_dict.c -- 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 | * $Id: mg_fast_comp_dict.c 16583 2008-07-29 10:20:36Z davidb $
|
---|
| 21 | *
|
---|
| 22 | **************************************************************************/
|
---|
| 23 |
|
---|
| 24 |
|
---|
| 25 | /*
|
---|
| 26 | $Log$
|
---|
| 27 | Revision 1.1 2003/02/20 21:18:24 mdewsnip
|
---|
| 28 | Addition of MG package for search and retrieval
|
---|
| 29 |
|
---|
| 30 | Revision 1.1 1999/08/10 21:18:06 sjboddie
|
---|
| 31 | renamed mg-1.3d directory mg
|
---|
| 32 |
|
---|
| 33 | Revision 1.2 1998/11/25 07:55:44 rjmcnab
|
---|
| 34 |
|
---|
| 35 | Modified mg to that you can specify the stemmer you want
|
---|
| 36 | to use via a command line option. You specify it to
|
---|
| 37 | mg_passes during the build process. The number of the
|
---|
| 38 | stemmer that you used is stored within the inverted
|
---|
| 39 | dictionary header and the stemmed dictionary header so
|
---|
| 40 | the correct stemmer is used in later stages of building
|
---|
| 41 | and querying.
|
---|
| 42 |
|
---|
| 43 | Revision 1.1 1998/11/17 09:34:57 rjmcnab
|
---|
| 44 | *** empty log message ***
|
---|
| 45 |
|
---|
| 46 | * Revision 1.3 1994/10/20 03:56:55 tes
|
---|
| 47 | * I have rewritten the boolean query optimiser and abstracted out the
|
---|
| 48 | * components of the boolean query.
|
---|
| 49 | *
|
---|
| 50 | * Revision 1.2 1994/09/20 04:41:47 tes
|
---|
| 51 | * For version 1.1
|
---|
| 52 | *
|
---|
| 53 | */
|
---|
| 54 |
|
---|
| 55 | static char *RCSID = "$Id: mg_fast_comp_dict.c 16583 2008-07-29 10:20:36Z davidb $";
|
---|
| 56 |
|
---|
| 57 | #include "sysfuncs.h"
|
---|
| 58 |
|
---|
| 59 | #include "filestats.h"
|
---|
| 60 | #include "huffman.h"
|
---|
| 61 | #include "timing.h"
|
---|
| 62 | #include "messages.h"
|
---|
| 63 | #include "memlib.h"
|
---|
| 64 | #include "netorder.h" /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 65 |
|
---|
| 66 | #include "local_strings.h"
|
---|
| 67 | #include "mg.h"
|
---|
| 68 | #include "text.h"
|
---|
| 69 | #include "invf.h"
|
---|
| 70 | #include "lists.h"
|
---|
| 71 | #include "backend.h"
|
---|
| 72 | #include "mg_files.h"
|
---|
| 73 | #include "locallib.h"
|
---|
| 74 | #include "words.h"
|
---|
| 75 |
|
---|
| 76 |
|
---|
| 77 |
|
---|
| 78 | #define ALIGN_SIZE(p,t) (((p) + (sizeof(t)-1)) & (~(sizeof(t)-1)))
|
---|
| 79 |
|
---|
| 80 | #define WORDNO(p, base) ((((char*)(p))-((char*)(base)))/sizeof(u_char*))
|
---|
| 81 | #define FIXUP(p) (fixup[WORDNO(p,buffer)/8] |= (1<<(WORDNO(p, buffer) & 7)))
|
---|
| 82 |
|
---|
| 83 | #define IS_FIXUP(p) ((fixup[WORDNO(p,buffer)/8] & (1<<(WORDNO(p, buffer) & 7))) != 0)
|
---|
| 84 |
|
---|
| 85 | #define FIXUP_VALS(vals) do { \
|
---|
| 86 | int i; \
|
---|
| 87 | for (i=0; i < MAX_HUFFCODE_LEN+1; i++) \
|
---|
| 88 | FIXUP(&vals[i]); \
|
---|
| 89 | } while(0)
|
---|
| 90 |
|
---|
| 91 |
|
---|
| 92 |
|
---|
| 93 | static u_long mem_for_comp_dict (char *filename);
|
---|
| 94 | static void load_comp_dict (char *filename);
|
---|
| 95 | static void save_fast_dict (char *filename);
|
---|
| 96 | static void unfixup_buffer (void);
|
---|
| 97 |
|
---|
| 98 |
|
---|
| 99 | static void *buffer;
|
---|
| 100 | static void *cur;
|
---|
| 101 | static u_char *fixup;
|
---|
| 102 | static u_long mem, fixup_mem;
|
---|
| 103 |
|
---|
| 104 | int main (int argc, char **argv)
|
---|
| 105 | {
|
---|
| 106 | ProgTime StartTime;
|
---|
| 107 | int ch;
|
---|
| 108 | char *filename = "";
|
---|
| 109 | opterr = 0;
|
---|
| 110 | msg_prefix = argv[0];
|
---|
| 111 | while ((ch = getopt (argc, argv, "f:d:h")) != -1)
|
---|
| 112 | switch (ch)
|
---|
| 113 | {
|
---|
| 114 | case 'f': /* input file */
|
---|
| 115 | filename = optarg;
|
---|
| 116 | break;
|
---|
| 117 | case 'd':
|
---|
| 118 | set_basepath (optarg);
|
---|
| 119 | break;
|
---|
| 120 | case 'h':
|
---|
| 121 | case '?':
|
---|
| 122 | fprintf (stderr, "usage: %s [-f input_file] [-d data directory] [-h]\n",
|
---|
| 123 | argv[0]);
|
---|
| 124 | exit (1);
|
---|
| 125 | }
|
---|
| 126 | GetTime (&StartTime);
|
---|
| 127 |
|
---|
| 128 | mem = mem_for_comp_dict (filename);
|
---|
| 129 |
|
---|
| 130 | fixup_mem = (ALIGN_SIZE (mem, u_char *) / sizeof (u_char *) + 7) / 8;
|
---|
| 131 |
|
---|
| 132 | cur = buffer = Xmalloc (mem);
|
---|
| 133 | bzero (buffer, mem);
|
---|
| 134 | fixup = Xmalloc (fixup_mem);
|
---|
| 135 | bzero (fixup, fixup_mem);
|
---|
| 136 |
|
---|
| 137 | Message ("Estimated memory for fast_dict %u", mem);
|
---|
| 138 | Message ("Estimated memory for fixups %u", fixup_mem);
|
---|
| 139 |
|
---|
| 140 | load_comp_dict (filename);
|
---|
| 141 |
|
---|
| 142 | Message ("Actual memory for fast_dict %u", (char *) cur - (char *) buffer);
|
---|
| 143 |
|
---|
| 144 | if ((u_long) cur > (u_long) buffer + mem)
|
---|
| 145 | FatalError (1, "The buffer was not big enough for the dictionary");
|
---|
| 146 |
|
---|
| 147 | {
|
---|
| 148 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 149 | compression_dict *cd = (compression_dict *) buffer;
|
---|
| 150 | int i, which;
|
---|
| 151 |
|
---|
| 152 | /* cfh */
|
---|
| 153 | for (which = 0; which <= 1; which++)
|
---|
| 154 | {
|
---|
| 155 | int j;
|
---|
| 156 |
|
---|
| 157 | HTONSI(cd->cfh[which]->hd.num_codes);
|
---|
| 158 | HTONSI(cd->cfh[which]->hd.mincodelen);
|
---|
| 159 | HTONSI(cd->cfh[which]->hd.maxcodelen);
|
---|
| 160 | for (j = 0; j < MAX_HUFFCODE_LEN + 1; j++)
|
---|
| 161 | {
|
---|
| 162 | HTONSI(cd->cfh[which]->hd.lencount[j]);
|
---|
| 163 | HTONUL(cd->cfh[which]->hd.min_code[j]);
|
---|
| 164 | }
|
---|
| 165 | HTONUL(cd->cfh[which]->uncompressed_size);
|
---|
| 166 | for (j = 0; j < MAX_HUFFCODE_LEN + 1; j++)
|
---|
| 167 | HTONUL(cd->cfh[which]->huff_words_size[j]);
|
---|
| 168 | }
|
---|
| 169 | HTONUL(cd->MemForCompDict);
|
---|
| 170 | /* ad */
|
---|
| 171 | if (cd->cdh.novel_method == MG_NOVEL_BINARY ||
|
---|
| 172 | cd->cdh.novel_method == MG_NOVEL_DELTA ||
|
---|
| 173 | cd->cdh.novel_method == MG_NOVEL_HYBRID ||
|
---|
| 174 | cd->cdh.novel_method == MG_NOVEL_HYBRID_MTF)
|
---|
| 175 | for (which = 0; which <= 1; which++)
|
---|
| 176 | {
|
---|
| 177 | int j;
|
---|
| 178 |
|
---|
| 179 | HTONUL(cd->ad->afh[which].num_frags);
|
---|
| 180 | HTONUL(cd->ad->afh[which].mem_for_frags);
|
---|
| 181 | for (j = 0; j < 33; j++)
|
---|
| 182 | {
|
---|
| 183 | HTONSI(cd->ad->blk_start[which][j]);
|
---|
| 184 | HTONSI(cd->ad->blk_end[which][j]);
|
---|
| 185 | }
|
---|
| 186 | }
|
---|
| 187 | /* cdh */
|
---|
| 188 | HTONUL(cd->cdh.dict_type);
|
---|
| 189 | HTONUL(cd->cdh.novel_method);
|
---|
| 190 | for (i = 0; i < TEXT_PARAMS; i++)
|
---|
| 191 | HTONUL(cd->cdh.params[which]);
|
---|
| 192 | HTONUL(cd->cdh.num_words[0]);
|
---|
| 193 | HTONUL(cd->cdh.num_words[1]);
|
---|
| 194 | HTONUL(cd->cdh.num_word_chars[0]);
|
---|
| 195 | HTONUL(cd->cdh.num_word_chars[1]);
|
---|
| 196 | HTONUL(cd->cdh.lookback);
|
---|
| 197 | HTONSI(cd->fast_loaded);
|
---|
| 198 | }
|
---|
| 199 |
|
---|
| 200 | unfixup_buffer ();
|
---|
| 201 |
|
---|
| 202 | save_fast_dict (filename);
|
---|
| 203 |
|
---|
| 204 | Message ("%s", ElapsedTime (&StartTime, NULL));
|
---|
| 205 | return 0;
|
---|
| 206 | }
|
---|
| 207 |
|
---|
| 208 |
|
---|
| 209 |
|
---|
| 210 | static void
|
---|
| 211 | unfixup_buffer ()
|
---|
| 212 | {
|
---|
| 213 | u_long *p;
|
---|
| 214 | for (p = buffer; (u_long) p < (u_long) cur; p++)
|
---|
| 215 | {
|
---|
| 216 | if (IS_FIXUP (p))
|
---|
| 217 | *p = *p - (u_long) buffer;
|
---|
| 218 | }
|
---|
| 219 | }
|
---|
| 220 |
|
---|
| 221 |
|
---|
| 222 |
|
---|
| 223 |
|
---|
| 224 | static u_long
|
---|
| 225 | mem_for_aux_dict (compression_dict_header * cdh, char *filename)
|
---|
| 226 | {
|
---|
| 227 | int i;
|
---|
| 228 | u_long mem = sizeof (auxiliary_dict);
|
---|
| 229 | FILE *aux;
|
---|
| 230 |
|
---|
| 231 | aux = open_file (filename, TEXT_DICT_AUX_SUFFIX, "rb",
|
---|
| 232 | MAGIC_AUX_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 233 |
|
---|
| 234 | for (i = 0; i <= 1; i++)
|
---|
| 235 | {
|
---|
| 236 | aux_frags_header afh;
|
---|
| 237 | fread (&afh, sizeof (afh), 1, aux);
|
---|
| 238 | NTOHUL(afh.num_frags); /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 239 | NTOHUL(afh.mem_for_frags); /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 240 | mem += afh.num_frags * sizeof (u_char *);
|
---|
| 241 | mem = ALIGN_SIZE (mem + afh.mem_for_frags, u_char *);
|
---|
| 242 | fseek (aux, afh.mem_for_frags, SEEK_CUR);
|
---|
| 243 | }
|
---|
| 244 | fclose (aux);
|
---|
| 245 |
|
---|
| 246 | return mem;
|
---|
| 247 | }
|
---|
| 248 |
|
---|
| 249 |
|
---|
| 250 |
|
---|
| 251 | static u_long
|
---|
| 252 | mem_for_words (FILE * dict, compression_dict_header * cdh,
|
---|
| 253 | comp_frags_header * cfh)
|
---|
| 254 | {
|
---|
| 255 | u_long mem = 0;
|
---|
| 256 | u_long i, lookback;
|
---|
| 257 | int ptrs_reqd = 0;
|
---|
| 258 | int mem_reqd = 0;
|
---|
| 259 |
|
---|
| 260 | lookback = cdh->lookback;
|
---|
| 261 |
|
---|
| 262 | for (i = cfh->hd.mincodelen; i <= cfh->hd.maxcodelen; i++)
|
---|
| 263 | {
|
---|
| 264 | ptrs_reqd += (cfh->hd.lencount[i] + ((1 << lookback) - 1)) >> lookback;
|
---|
| 265 | mem_reqd += cfh->huff_words_size[i];
|
---|
| 266 | }
|
---|
| 267 |
|
---|
| 268 | mem += ptrs_reqd * sizeof (u_char *);
|
---|
| 269 | mem += (MAX_HUFFCODE_LEN + 1) * sizeof (u_char **);
|
---|
| 270 | mem += mem_reqd;
|
---|
| 271 |
|
---|
| 272 | for (i = 0; i < cfh->hd.num_codes; i++)
|
---|
| 273 | {
|
---|
| 274 | register int val;
|
---|
| 275 | for (val = getc (dict) & 0xf; val; val--)
|
---|
| 276 | getc (dict);
|
---|
| 277 | }
|
---|
| 278 | return ALIGN_SIZE (mem, u_char *);
|
---|
| 279 | }
|
---|
| 280 |
|
---|
| 281 |
|
---|
| 282 |
|
---|
| 283 |
|
---|
| 284 | static u_long
|
---|
| 285 | mem_for_comp_dict (char *filename)
|
---|
| 286 | {
|
---|
| 287 | u_long mem = sizeof (compression_dict);
|
---|
| 288 | compression_dict_header cdh;
|
---|
| 289 | comp_frags_header cfh;
|
---|
| 290 | FILE *dict;
|
---|
| 291 | int which;
|
---|
| 292 | int i; /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 293 |
|
---|
| 294 | dict = open_file (filename, TEXT_DICT_SUFFIX, "rb",
|
---|
| 295 | MAGIC_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 296 |
|
---|
| 297 | fread (&cdh, sizeof (cdh), 1, dict);
|
---|
| 298 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 299 | NTOHUL(cdh.dict_type);
|
---|
| 300 | NTOHUL(cdh.novel_method);
|
---|
| 301 | for (i = 0; i < TEXT_PARAMS; i++)
|
---|
| 302 | NTOHUL(cdh.params[i]);
|
---|
| 303 | NTOHUL(cdh.num_words[0]);
|
---|
| 304 | NTOHUL(cdh.num_words[1]);
|
---|
| 305 | NTOHUL(cdh.num_word_chars[0]);
|
---|
| 306 | NTOHUL(cdh.num_word_chars[1]);
|
---|
| 307 | NTOHUL(cdh.lookback);
|
---|
| 308 |
|
---|
| 309 | for (which = 0; which < 2; which++)
|
---|
| 310 | switch (cdh.dict_type)
|
---|
| 311 | {
|
---|
| 312 | case MG_COMPLETE_DICTIONARY:
|
---|
| 313 | {
|
---|
| 314 | mem += sizeof (comp_frags_header);
|
---|
| 315 | Read_cfh (dict, &cfh, NULL, NULL);
|
---|
| 316 |
|
---|
| 317 | /* Don't need to count the space for the clens of the huffman data */
|
---|
| 318 |
|
---|
| 319 | mem += mem_for_words (dict, &cdh, &cfh);
|
---|
| 320 | if (cfh.hd.clens)
|
---|
| 321 | Xfree (cfh.hd.clens);
|
---|
| 322 | }
|
---|
| 323 | break;
|
---|
| 324 | case MG_PARTIAL_DICTIONARY:
|
---|
| 325 | {
|
---|
| 326 | huff_data hd;
|
---|
| 327 | if (cdh.num_words[which])
|
---|
| 328 | {
|
---|
| 329 | mem += sizeof (comp_frags_header);
|
---|
| 330 | Read_cfh (dict, &cfh, NULL, NULL);
|
---|
| 331 |
|
---|
| 332 | /* Don't need to count the space for the clens of the
|
---|
| 333 | huffman data */
|
---|
| 334 |
|
---|
| 335 | mem += mem_for_words (dict, &cdh, &cfh);
|
---|
| 336 | if (cfh.hd.clens)
|
---|
| 337 | Xfree (cfh.hd.clens);
|
---|
| 338 |
|
---|
| 339 | }
|
---|
| 340 |
|
---|
| 341 | mem += sizeof (hd);
|
---|
| 342 | Read_Huffman_Data (dict, &hd, NULL, NULL);
|
---|
| 343 | if (hd.clens)
|
---|
| 344 | Xfree (hd.clens);
|
---|
| 345 | mem += hd.num_codes * sizeof (unsigned long);
|
---|
| 346 | mem += (MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *);
|
---|
| 347 |
|
---|
| 348 | mem += sizeof (hd);
|
---|
| 349 | Read_Huffman_Data (dict, &hd, NULL, NULL);
|
---|
| 350 | if (hd.clens)
|
---|
| 351 | Xfree (hd.clens);
|
---|
| 352 | mem += hd.num_codes * sizeof (unsigned long);
|
---|
| 353 | mem += (MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *);
|
---|
| 354 | }
|
---|
| 355 | break;
|
---|
| 356 | case MG_SEED_DICTIONARY:
|
---|
| 357 | {
|
---|
| 358 | huff_data hd;
|
---|
| 359 | if (cdh.num_words[which])
|
---|
| 360 | {
|
---|
| 361 | mem += sizeof (comp_frags_header);
|
---|
| 362 | Read_cfh (dict, &cfh, NULL, NULL);
|
---|
| 363 |
|
---|
| 364 | /* Don't need to count the space for the clens of the
|
---|
| 365 | huffman data */
|
---|
| 366 |
|
---|
| 367 | mem += mem_for_words (dict, &cdh, &cfh);
|
---|
| 368 | if (cfh.hd.clens)
|
---|
| 369 | Xfree (cfh.hd.clens);
|
---|
| 370 |
|
---|
| 371 | }
|
---|
| 372 | switch (cdh.novel_method)
|
---|
| 373 | {
|
---|
| 374 | case MG_NOVEL_HUFFMAN_CHARS:
|
---|
| 375 | mem += sizeof (hd);
|
---|
| 376 | Read_Huffman_Data (dict, &hd, NULL, NULL);
|
---|
| 377 | if (hd.clens)
|
---|
| 378 | Xfree (hd.clens);
|
---|
| 379 | mem += hd.num_codes * sizeof (unsigned long);
|
---|
| 380 | mem += (MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *);
|
---|
| 381 |
|
---|
| 382 | mem += sizeof (hd);
|
---|
| 383 | Read_Huffman_Data (dict, &hd, NULL, NULL);
|
---|
| 384 | if (hd.clens)
|
---|
| 385 | Xfree (hd.clens);
|
---|
| 386 | mem += hd.num_codes * sizeof (unsigned long);
|
---|
| 387 | mem += (MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *);
|
---|
| 388 | break;
|
---|
| 389 | case MG_NOVEL_BINARY:
|
---|
| 390 | break;
|
---|
| 391 | case MG_NOVEL_DELTA:
|
---|
| 392 | break;
|
---|
| 393 | case MG_NOVEL_HYBRID:
|
---|
| 394 | break;
|
---|
| 395 | case MG_NOVEL_HYBRID_MTF:
|
---|
| 396 | break;
|
---|
| 397 | }
|
---|
| 398 | break;
|
---|
| 399 | }
|
---|
| 400 | }
|
---|
| 401 | fclose (dict);
|
---|
| 402 |
|
---|
| 403 | if (cdh.novel_method == MG_NOVEL_BINARY ||
|
---|
| 404 | cdh.novel_method == MG_NOVEL_DELTA ||
|
---|
| 405 | cdh.novel_method == MG_NOVEL_HYBRID ||
|
---|
| 406 | cdh.novel_method == MG_NOVEL_HYBRID_MTF)
|
---|
| 407 | mem += mem_for_aux_dict (&cdh, filename);
|
---|
| 408 |
|
---|
| 409 | return ALIGN_SIZE (mem, u_char *);
|
---|
| 410 | }
|
---|
| 411 |
|
---|
| 412 |
|
---|
| 413 |
|
---|
| 414 | void *
|
---|
| 415 | getmem (u_long size, int align)
|
---|
| 416 | {
|
---|
| 417 | void *res;
|
---|
| 418 | cur = (void *) (((u_long) cur + (align - 1)) & (~(align - 1)));
|
---|
| 419 | res = cur;
|
---|
| 420 | cur = (char *) cur + size;
|
---|
| 421 | if ((long) cur > (long) buffer + mem)
|
---|
| 422 | FatalError (1, "The buffer was not big enough for the dictionary");
|
---|
| 423 | return res;
|
---|
| 424 | }
|
---|
| 425 |
|
---|
| 426 |
|
---|
| 427 |
|
---|
| 428 |
|
---|
| 429 |
|
---|
| 430 |
|
---|
| 431 |
|
---|
| 432 |
|
---|
| 433 |
|
---|
| 434 |
|
---|
| 435 |
|
---|
| 436 | static auxiliary_dict *
|
---|
| 437 | LoadAuxDict (compression_dict_header * cdh,
|
---|
| 438 | char *filename)
|
---|
| 439 | {
|
---|
| 440 | auxiliary_dict *ad;
|
---|
| 441 | int i;
|
---|
| 442 | FILE *aux;
|
---|
| 443 |
|
---|
| 444 | aux = open_file (filename, TEXT_DICT_AUX_SUFFIX, "rb",
|
---|
| 445 | MAGIC_AUX_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 446 |
|
---|
| 447 | ad = getmem (sizeof (auxiliary_dict), sizeof (u_char *));
|
---|
| 448 |
|
---|
| 449 | for (i = 0; i <= 1; i++)
|
---|
| 450 | {
|
---|
| 451 | int j;
|
---|
| 452 | u_char *pos;
|
---|
| 453 |
|
---|
| 454 | fread (&ad->afh[i], sizeof (aux_frags_header), 1, aux);
|
---|
| 455 |
|
---|
| 456 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 457 | NTOHUL(ad->afh[i].num_frags);
|
---|
| 458 | NTOHUL(ad->afh[i].mem_for_frags);
|
---|
| 459 |
|
---|
| 460 | ad->word_data[i] = getmem (ad->afh[i].mem_for_frags, 1);
|
---|
| 461 | FIXUP (&ad->word_data[i]);
|
---|
| 462 |
|
---|
| 463 | ad->words[i] = getmem (ad->afh[i].num_frags * sizeof (u_char *),
|
---|
| 464 | sizeof (u_char *));
|
---|
| 465 | FIXUP (&ad->words[i]);
|
---|
| 466 |
|
---|
| 467 | fread (ad->word_data[i], ad->afh[i].mem_for_frags, sizeof (u_char), aux);
|
---|
| 468 |
|
---|
| 469 | pos = ad->word_data[i];
|
---|
| 470 | for (j = 0; j < ad->afh[i].num_frags; j++)
|
---|
| 471 | {
|
---|
| 472 | ad->words[i][j] = pos;
|
---|
| 473 | FIXUP (&ad->words[i][j]);
|
---|
| 474 | pos += *pos + 1;
|
---|
| 475 | }
|
---|
| 476 | if (cdh->novel_method == MG_NOVEL_HYBRID ||
|
---|
| 477 | cdh->novel_method == MG_NOVEL_HYBRID_MTF)
|
---|
| 478 | {
|
---|
| 479 | int num;
|
---|
| 480 | num = 1;
|
---|
| 481 | ad->blk_start[i][0] = 0;
|
---|
| 482 | ad->blk_end[i][0] = cdh->num_words[i] - 1;
|
---|
| 483 | while (num < 33)
|
---|
| 484 | {
|
---|
| 485 | ad->blk_start[i][num] = ad->blk_end[i][num - 1] + 1;
|
---|
| 486 | ad->blk_end[i][num] = ad->blk_start[i][num] +
|
---|
| 487 | (ad->blk_end[i][num - 1] - ad->blk_start[i][num - 1]) * 2;
|
---|
| 488 | num++;
|
---|
| 489 | }
|
---|
| 490 | }
|
---|
| 491 | }
|
---|
| 492 | fclose (aux);
|
---|
| 493 | return (ad);
|
---|
| 494 | }
|
---|
| 495 |
|
---|
| 496 |
|
---|
| 497 |
|
---|
| 498 |
|
---|
| 499 |
|
---|
| 500 | static u_char ***
|
---|
| 501 | ReadInWords (FILE * dict, compression_dict * cd,
|
---|
| 502 | comp_frags_header * cfh, u_char ** escape)
|
---|
| 503 | {
|
---|
| 504 | int i, lookback;
|
---|
| 505 | int ptrs_reqd = 0;
|
---|
| 506 | int mem_reqd = 0;
|
---|
| 507 | int num_set[MAX_HUFFCODE_LEN + 1];
|
---|
| 508 | u_char *next_word[MAX_HUFFCODE_LEN + 1];
|
---|
| 509 | u_char **vals;
|
---|
| 510 | u_char ***values;
|
---|
| 511 | u_char word[MAXWORDLEN + 1];
|
---|
| 512 | u_char last_word[MAX_HUFFCODE_LEN + 1][MAXWORDLEN + 1];
|
---|
| 513 |
|
---|
| 514 | lookback = cd->cdh.lookback;
|
---|
| 515 |
|
---|
| 516 | for (i = cfh->hd.mincodelen; i <= cfh->hd.maxcodelen; i++)
|
---|
| 517 | {
|
---|
| 518 | ptrs_reqd += (cfh->hd.lencount[i] + ((1 << lookback) - 1)) >> lookback;
|
---|
| 519 | mem_reqd += cfh->huff_words_size[i];
|
---|
| 520 | }
|
---|
| 521 |
|
---|
| 522 | vals = getmem (ptrs_reqd * sizeof (*vals), sizeof (u_char *));
|
---|
| 523 |
|
---|
| 524 | values = getmem ((MAX_HUFFCODE_LEN + 1) * sizeof (u_char **), sizeof (u_char **));
|
---|
| 525 |
|
---|
| 526 | next_word[0] = getmem (mem_reqd, sizeof (char));
|
---|
| 527 |
|
---|
| 528 | cd->MemForCompDict += ptrs_reqd * sizeof (*vals) +
|
---|
| 529 | (MAX_HUFFCODE_LEN + 1) * sizeof (u_char **) +
|
---|
| 530 | mem_reqd;
|
---|
| 531 |
|
---|
| 532 | values[0] = vals;
|
---|
| 533 | FIXUP (&values[0]);
|
---|
| 534 | values[0][0] = next_word[0];
|
---|
| 535 | FIXUP (&values[0][0]);
|
---|
| 536 | for (i = 1; i <= cfh->hd.maxcodelen; i++)
|
---|
| 537 | {
|
---|
| 538 | int next_start = (values[i - 1] - vals) +
|
---|
| 539 | ((cfh->hd.lencount[i - 1] + ((1 << lookback) - 1)) >> lookback);
|
---|
| 540 | values[i] = &vals[next_start];
|
---|
| 541 | FIXUP (&values[i]);
|
---|
| 542 | next_word[i] = next_word[i - 1] + cfh->huff_words_size[i - 1];
|
---|
| 543 | values[i][0] = next_word[i];
|
---|
| 544 | FIXUP (&values[i][0]);
|
---|
| 545 | }
|
---|
| 546 |
|
---|
| 547 | bzero ((char *) num_set, sizeof (num_set));
|
---|
| 548 |
|
---|
| 549 | for (i = 0; i < cfh->hd.num_codes; i++)
|
---|
| 550 | {
|
---|
| 551 | register int val, copy;
|
---|
| 552 | register int len = cfh->hd.clens[i];
|
---|
| 553 | val = getc (dict);
|
---|
| 554 | copy = (val >> 4) & 0xf;
|
---|
| 555 | val &= 0xf;
|
---|
| 556 |
|
---|
| 557 | fread (word + copy + 1, sizeof (u_char), val, dict);
|
---|
| 558 | *word = val + copy;
|
---|
| 559 |
|
---|
| 560 | if ((num_set[len] & ((1 << lookback) - 1)) == 0)
|
---|
| 561 | {
|
---|
| 562 | values[len][num_set[len] >> lookback] = next_word[len];
|
---|
| 563 | FIXUP (&values[len][num_set[len] >> lookback]);
|
---|
| 564 | memcpy (next_word[len], word, *word + 1);
|
---|
| 565 | if (escape && i == cfh->hd.num_codes - 1)
|
---|
| 566 | {
|
---|
| 567 | *escape = next_word[len];
|
---|
| 568 | FIXUP (escape);
|
---|
| 569 | }
|
---|
| 570 | next_word[len] += *word + 1;
|
---|
| 571 | }
|
---|
| 572 | else
|
---|
| 573 | {
|
---|
| 574 | copy = prefixlen (last_word[len], word);
|
---|
| 575 | memcpy (next_word[len] + 1, word + copy + 1, *word - copy);
|
---|
| 576 | *next_word[len] = (copy << 4) + (*word - copy);
|
---|
| 577 | if (escape && i == cfh->hd.num_codes - 1)
|
---|
| 578 | {
|
---|
| 579 | *escape = next_word[len];
|
---|
| 580 | FIXUP (escape);
|
---|
| 581 | }
|
---|
| 582 | next_word[len] += (*word - copy) + 1;
|
---|
| 583 | }
|
---|
| 584 | memcpy (last_word[len], word, *word + 1);
|
---|
| 585 | num_set[len]++;
|
---|
| 586 | }
|
---|
| 587 | if (cfh->hd.clens)
|
---|
| 588 | Xfree (cfh->hd.clens);
|
---|
| 589 | cfh->hd.clens = NULL;
|
---|
| 590 | return values;
|
---|
| 591 | }
|
---|
| 592 |
|
---|
| 593 |
|
---|
| 594 | static unsigned long **
|
---|
| 595 | Generate_Fast_Huffman_Vals (huff_data * data, u_long * mem)
|
---|
| 596 | {
|
---|
| 597 | int i;
|
---|
| 598 | unsigned long *fcode[MAX_HUFFCODE_LEN + 1];
|
---|
| 599 | unsigned long **values;
|
---|
| 600 | unsigned long *vals;
|
---|
| 601 |
|
---|
| 602 | if (!data)
|
---|
| 603 | return (NULL);
|
---|
| 604 | vals = getmem (data->num_codes * sizeof (*vals), sizeof (long *));
|
---|
| 605 | values = getmem ((MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *),
|
---|
| 606 | sizeof (long *));
|
---|
| 607 |
|
---|
| 608 | bzero ((char *) values, (MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *));
|
---|
| 609 |
|
---|
| 610 | if (mem)
|
---|
| 611 | *mem += data->num_codes * sizeof (*vals) +
|
---|
| 612 | (MAX_HUFFCODE_LEN + 1) * sizeof (unsigned long *);
|
---|
| 613 |
|
---|
| 614 | fcode[0] = values[0] = &vals[0];
|
---|
| 615 | FIXUP (&values[0]);
|
---|
| 616 | for (i = 1; i <= data->maxcodelen; i++)
|
---|
| 617 | {
|
---|
| 618 | fcode[i] = values[i] = &vals[(values[i - 1] - vals) + data->lencount[i - 1]];
|
---|
| 619 | FIXUP (&values[i]);
|
---|
| 620 | }
|
---|
| 621 |
|
---|
| 622 | for (i = 0; i < data->num_codes; i++)
|
---|
| 623 | if (data->clens[i])
|
---|
| 624 | *fcode[(int) (data->clens[i])]++ = i;
|
---|
| 625 | return (values);
|
---|
| 626 | }
|
---|
| 627 |
|
---|
| 628 |
|
---|
| 629 |
|
---|
| 630 | static void
|
---|
| 631 | load_comp_dict (char *filename)
|
---|
| 632 | {
|
---|
| 633 | FILE *dict;
|
---|
| 634 | int which;
|
---|
| 635 | compression_dict *cd;
|
---|
| 636 |
|
---|
| 637 | cd = getmem (sizeof (compression_dict), sizeof (u_char *));
|
---|
| 638 | cd->fast_loaded = 1;
|
---|
| 639 |
|
---|
| 640 | dict = open_file (filename, TEXT_DICT_SUFFIX, "rb",
|
---|
| 641 | MAGIC_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 642 |
|
---|
| 643 | Read_cdh (dict, &cd->cdh, NULL, NULL);
|
---|
| 644 |
|
---|
| 645 | for (which = 0; which < 2; which++)
|
---|
| 646 | switch (cd->cdh.dict_type)
|
---|
| 647 | {
|
---|
| 648 | case MG_COMPLETE_DICTIONARY:
|
---|
| 649 | {
|
---|
| 650 | cd->cfh[which] = getmem (sizeof (*cd->cfh[which]), sizeof (u_char *));
|
---|
| 651 | cd->MemForCompDict += sizeof (*cd->cfh[which]);
|
---|
| 652 | Read_cfh (dict, cd->cfh[which], &cd->MemForCompDict, NULL);
|
---|
| 653 |
|
---|
| 654 | cd->values[which] = ReadInWords (dict, cd, cd->cfh[which], NULL);
|
---|
| 655 | FIXUP (&cd->cfh[which]);
|
---|
| 656 | FIXUP (&cd->values[which]);
|
---|
| 657 | cd->escape[which] = NULL;
|
---|
| 658 | }
|
---|
| 659 | break;
|
---|
| 660 | case MG_PARTIAL_DICTIONARY:
|
---|
| 661 | {
|
---|
| 662 | huff_data *hd;
|
---|
| 663 | u_long **vals;
|
---|
| 664 | if (cd->cdh.num_words[which])
|
---|
| 665 | {
|
---|
| 666 | cd->cfh[which] = getmem (sizeof (*cd->cfh[which]), sizeof (u_char *));
|
---|
| 667 | cd->MemForCompDict += sizeof (*cd->cfh[which]);
|
---|
| 668 | Read_cfh (dict, cd->cfh[which], &cd->MemForCompDict, NULL);
|
---|
| 669 | cd->values[which] = ReadInWords (dict, cd, cd->cfh[which],
|
---|
| 670 | &cd->escape[which]);
|
---|
| 671 | FIXUP (&cd->cfh[which]);
|
---|
| 672 | FIXUP (&cd->values[which]);
|
---|
| 673 | FIXUP (&cd->escape[which]);
|
---|
| 674 | }
|
---|
| 675 |
|
---|
| 676 | hd = getmem (sizeof (huff_data), sizeof (char *));
|
---|
| 677 | cd->MemForCompDict += sizeof (huff_data);
|
---|
| 678 | Read_Huffman_Data (dict, hd, &cd->MemForCompDict, NULL);
|
---|
| 679 | vals = Generate_Fast_Huffman_Vals (hd, &cd->MemForCompDict);
|
---|
| 680 | cd->chars_huff[which] = hd;
|
---|
| 681 | FIXUP (&cd->chars_huff[which]);
|
---|
| 682 | cd->chars_vals[which] = vals;
|
---|
| 683 | FIXUP (&cd->chars_vals[which]);
|
---|
| 684 | if (hd->clens)
|
---|
| 685 | Xfree (hd->clens);
|
---|
| 686 | hd->clens = NULL;
|
---|
| 687 |
|
---|
| 688 | hd = getmem (sizeof (huff_data), sizeof (char *));
|
---|
| 689 | cd->MemForCompDict += sizeof (huff_data);
|
---|
| 690 | Read_Huffman_Data (dict, hd, &cd->MemForCompDict, NULL);
|
---|
| 691 | vals = Generate_Fast_Huffman_Vals (hd, &cd->MemForCompDict);
|
---|
| 692 | cd->lens_huff[which] = hd;
|
---|
| 693 | FIXUP (&cd->lens_huff[which]);
|
---|
| 694 | cd->lens_vals[which] = vals;
|
---|
| 695 | FIXUP (&cd->lens_vals[which]);
|
---|
| 696 | if (hd->clens)
|
---|
| 697 | Xfree (hd->clens);
|
---|
| 698 | hd->clens = NULL;
|
---|
| 699 | }
|
---|
| 700 | break;
|
---|
| 701 | case MG_SEED_DICTIONARY:
|
---|
| 702 | {
|
---|
| 703 | huff_data *hd;
|
---|
| 704 | u_long **vals;
|
---|
| 705 | if (cd->cdh.num_words[which])
|
---|
| 706 | {
|
---|
| 707 | cd->cfh[which] = getmem (sizeof (*cd->cfh[which]), sizeof (u_char *));
|
---|
| 708 | cd->MemForCompDict += sizeof (*cd->cfh[which]);
|
---|
| 709 | Read_cfh (dict, cd->cfh[which], &cd->MemForCompDict, NULL);
|
---|
| 710 | cd->values[which] = ReadInWords (dict, cd, cd->cfh[which],
|
---|
| 711 | &cd->escape[which]);
|
---|
| 712 | FIXUP (&cd->cfh[which]);
|
---|
| 713 | FIXUP (&cd->values[which]);
|
---|
| 714 | FIXUP (&cd->escape[which]);
|
---|
| 715 | }
|
---|
| 716 | switch (cd->cdh.novel_method)
|
---|
| 717 | {
|
---|
| 718 | case MG_NOVEL_HUFFMAN_CHARS:
|
---|
| 719 | hd = getmem (sizeof (huff_data), sizeof (char *));
|
---|
| 720 | cd->MemForCompDict += sizeof (huff_data);
|
---|
| 721 | Read_Huffman_Data (dict, hd, &cd->MemForCompDict, NULL);
|
---|
| 722 | vals = Generate_Fast_Huffman_Vals (hd, &cd->MemForCompDict);
|
---|
| 723 | cd->chars_huff[which] = hd;
|
---|
| 724 | FIXUP (&cd->chars_huff[which]);
|
---|
| 725 | cd->chars_vals[which] = vals;
|
---|
| 726 | FIXUP (&cd->chars_vals[which]);
|
---|
| 727 | if (hd->clens)
|
---|
| 728 | Xfree (hd->clens);
|
---|
| 729 | hd->clens = NULL;
|
---|
| 730 |
|
---|
| 731 | hd = getmem (sizeof (huff_data), sizeof (char *));
|
---|
| 732 | cd->MemForCompDict += sizeof (huff_data);
|
---|
| 733 | Read_Huffman_Data (dict, hd, &cd->MemForCompDict, NULL);
|
---|
| 734 | vals = Generate_Fast_Huffman_Vals (hd, &cd->MemForCompDict);
|
---|
| 735 | cd->lens_huff[which] = hd;
|
---|
| 736 | FIXUP (&cd->lens_huff[which]);
|
---|
| 737 | cd->lens_vals[which] = vals;
|
---|
| 738 | FIXUP (&cd->lens_vals[which]);
|
---|
| 739 | if (hd->clens)
|
---|
| 740 | Xfree (hd->clens);
|
---|
| 741 | hd->clens = NULL;
|
---|
| 742 | break;
|
---|
| 743 | case MG_NOVEL_BINARY:
|
---|
| 744 | break;
|
---|
| 745 | case MG_NOVEL_DELTA:
|
---|
| 746 | break;
|
---|
| 747 | case MG_NOVEL_HYBRID:
|
---|
| 748 | break;
|
---|
| 749 | case MG_NOVEL_HYBRID_MTF:
|
---|
| 750 | break;
|
---|
| 751 | }
|
---|
| 752 | break;
|
---|
| 753 | }
|
---|
| 754 | }
|
---|
| 755 | fclose (dict);
|
---|
| 756 |
|
---|
| 757 |
|
---|
| 758 | if (cd->cdh.novel_method == MG_NOVEL_BINARY ||
|
---|
| 759 | cd->cdh.novel_method == MG_NOVEL_DELTA ||
|
---|
| 760 | cd->cdh.novel_method == MG_NOVEL_HYBRID ||
|
---|
| 761 | cd->cdh.novel_method == MG_NOVEL_HYBRID_MTF)
|
---|
| 762 | {
|
---|
| 763 | cd->ad = LoadAuxDict (&cd->cdh, filename);
|
---|
| 764 | FIXUP (&cd->ad);
|
---|
| 765 | }
|
---|
| 766 | }
|
---|
| 767 |
|
---|
| 768 |
|
---|
| 769 | static void
|
---|
| 770 | save_fast_dict (char *filename)
|
---|
| 771 | {
|
---|
| 772 | FILE *fdict;
|
---|
| 773 |
|
---|
| 774 | fdict = create_file (filename, TEXT_DICT_FAST_SUFFIX, "wb",
|
---|
| 775 | MAGIC_FAST_DICT, MG_ABORT); /* [RPAP - Feb 97: WIN32 Port] */
|
---|
| 776 |
|
---|
| 777 | {
|
---|
| 778 | u_long *p;
|
---|
| 779 | for (p = buffer; (u_long) p < (u_long) cur; p++)
|
---|
| 780 | {
|
---|
| 781 | if (IS_FIXUP (p))
|
---|
| 782 | HTONUL(*p);
|
---|
| 783 | }
|
---|
| 784 | }
|
---|
| 785 |
|
---|
| 786 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 787 | HTONUL(mem);
|
---|
| 788 | HTONUL(fixup_mem);
|
---|
| 789 |
|
---|
| 790 | fwrite (&mem, sizeof (mem), 1, fdict);
|
---|
| 791 | fwrite (&fixup_mem, sizeof (fixup_mem), 1, fdict);
|
---|
| 792 |
|
---|
| 793 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
| 794 | NTOHUL(mem);
|
---|
| 795 | NTOHUL(fixup_mem);
|
---|
| 796 |
|
---|
| 797 | fwrite (buffer, sizeof (u_char), mem, fdict);
|
---|
| 798 | fwrite (fixup, sizeof (u_char), fixup_mem, fdict);
|
---|
| 799 |
|
---|
| 800 | fclose (fdict);
|
---|
| 801 | }
|
---|