xdiff/xprepare.c

   1 /*
   2  *  LibXDiff by Davide Libenzi ( File Differential Library )
   3  *  Copyright (C) 2003  Davide Libenzi
   4  *
   5  *  This library is free software; you can redistribute it and/or
   6  *  modify it under the terms of the GNU Lesser General Public
   7  *  License as published by the Free Software Foundation; either
   8  *  version 2.1 of the License, or (at your option) any later version.
   9  *
  10  *  This library is distributed in the hope that it will be useful,
  11  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  *  Lesser General Public License for more details.
  14  *
  15  *  You should have received a copy of the GNU Lesser General Public
  16  *  License along with this library; if not, write to the Free Software
  17  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  18  *
  19  *  Davide Libenzi <davidel@xmailserver.org>
  20  *
  21  */
  22
  23 #include "xinclude.h"
  24
  25
  26 #define XDL_KPDIS_RUN 4
  27 #define XDL_MAX_EQLIMIT 1024
  28 #define XDL_SIMSCAN_WINDOW 100
  29
  30
  31 typedef struct s_xdlclass {
  32         struct s_xdlclass *next;
  33         unsigned long ha;
  34         char const *line;
  35         long size;
  36         long idx;
  37         long len1, len2;
  38 } xdlclass_t;
  39
  40 typedef struct s_xdlclassifier {
  41         unsigned int hbits;
  42         long hsize;
  43         xdlclass_t **rchash;
  44         chastore_t ncha;
  45         xdlclass_t **rcrecs;
  46         long alloc;
  47         long count;
  48         long flags;
  49 } xdlclassifier_t;
  50
  51
  52
  53
  54 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags);
  55 static void xdl_free_classifier(xdlclassifier_t *cf);
  56 static int xdl_classify_record(unsigned int pass, xdlclassifier_t *cf, xrecord_t **rhash,
  57                                unsigned int hbits, xrecord_t *rec);
  58 static int xdl_prepare_ctx(unsigned int pass, mmfile_t *mf, long narec, xpparam_t const *xpp,
  59                            xdlclassifier_t *cf, xdfile_t *xdf);
  60 static void xdl_free_ctx(xdfile_t *xdf);
  61 static int xdl_clean_mmatch(char const *dis, long i, long s, long e);
  62 static int xdl_cleanup_records(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2);
  63 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2);
  64 static int xdl_optimize_ctxs(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2);
  65
  66
  67
  68
  69 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags) {
  70         long i;
  71
  72         cf->flags = flags;
  73
  74         cf->hbits = xdl_hashbits((unsigned int) size);
  75         cf->hsize = 1 << cf->hbits;
  76
  77         if (xdl_cha_init(&cf->ncha, sizeof(xdlclass_t), size / 4 + 1) < 0) {
  78
  79                 return -1;
  80         }
  81         if (!(cf->rchash = (xdlclass_t **) xdl_malloc(cf->hsize * sizeof(xdlclass_t *)))) {
  82
  83                 xdl_cha_free(&cf->ncha);
  84                 return -1;
  85         }
  86         for (i = 0; i < cf->hsize; i++)
  87                 cf->rchash[i] = NULL;
  88
  89         cf->alloc = size;
  90         if (!(cf->rcrecs = (xdlclass_t **) xdl_malloc(cf->alloc * sizeof(xdlclass_t *)))) {
  91
  92                 xdl_free(cf->rchash);
  93                 xdl_cha_free(&cf->ncha);
  94                 return -1;
  95         }
  96
  97         cf->count = 0;
  98
  99         return 0;
 100 }
 101
 102
 103 static void xdl_free_classifier(xdlclassifier_t *cf) {
 104
 105         xdl_free(cf->rcrecs);
 106         xdl_free(cf->rchash);
 107         xdl_cha_free(&cf->ncha);
 108 }
 109
 110
 111 static int xdl_classify_record(unsigned int pass, xdlclassifier_t *cf, xrecord_t **rhash,
 112                                unsigned int hbits, xrecord_t *rec) {
 113         long hi;
 114         char const *line;
 115         xdlclass_t *rcrec;
 116         xdlclass_t **rcrecs;
 117
 118         line = rec->ptr;
 119         hi = (long) XDL_HASHLONG(rec->ha, cf->hbits);
 120         for (rcrec = cf->rchash[hi]; rcrec; rcrec = rcrec->next)
 121                 if (rcrec->ha == rec->ha &&
 122                                 xdl_recmatch(rcrec->line, rcrec->size,
 123                                         rec->ptr, rec->size, cf->flags))
 124                         break;
 125
 126         if (!rcrec) {
 127                 if (!(rcrec = xdl_cha_alloc(&cf->ncha))) {
 128
 129                         return -1;
 130                 }
 131                 rcrec->idx = cf->count++;
 132                 if (cf->count > cf->alloc) {
 133                         cf->alloc *= 2;
 134                         if (!(rcrecs = (xdlclass_t **) xdl_realloc(cf->rcrecs, cf->alloc * sizeof(xdlclass_t *)))) {
 135
 136                                 return -1;
 137                         }
 138                         cf->rcrecs = rcrecs;
 139                 }
 140                 cf->rcrecs[rcrec->idx] = rcrec;
 141                 rcrec->line = line;
 142                 rcrec->size = rec->size;
 143                 rcrec->ha = rec->ha;
 144                 rcrec->len1 = rcrec->len2 = 0;
 145                 rcrec->next = cf->rchash[hi];
 146                 cf->rchash[hi] = rcrec;
 147         }
 148
 149         (pass == 1) ? rcrec->len1++ : rcrec->len2++;
 150
 151         rec->ha = (unsigned long) rcrec->idx;
 152
 153         hi = (long) XDL_HASHLONG(rec->ha, hbits);
 154         rec->next = rhash[hi];
 155         rhash[hi] = rec;
 156
 157         return 0;
 158 }
 159
 160
 161 static int xdl_prepare_ctx(unsigned int pass, mmfile_t *mf, long narec, xpparam_t const *xpp,
 162                            xdlclassifier_t *cf, xdfile_t *xdf) {
 163         unsigned int hbits;
 164         long i, nrec, hsize, bsize;
 165         unsigned long hav;
 166         char const *blk, *cur, *top, *prev;
 167         xrecord_t *crec;
 168         xrecord_t **recs, **rrecs;
 169         xrecord_t **rhash;
 170         unsigned long *ha;
 171         char *rchg;
 172         long *rindex;
 173
 174         if (xdl_cha_init(&xdf->rcha, sizeof(xrecord_t), narec / 4 + 1) < 0) {
 175
 176                 return -1;
 177         }
 178         if (!(recs = (xrecord_t **) xdl_malloc(narec * sizeof(xrecord_t *)))) {
 179
 180                 xdl_cha_free(&xdf->rcha);
 181                 return -1;
 182         }
 183
 184         hbits = xdl_hashbits((unsigned int) narec);
 185         hsize = 1 << hbits;
 186         if (!(rhash = (xrecord_t **) xdl_malloc(hsize * sizeof(xrecord_t *)))) {
 187
 188                 xdl_free(recs);
 189                 xdl_cha_free(&xdf->rcha);
 190                 return -1;
 191         }
 192         for (i = 0; i < hsize; i++)
 193                 rhash[i] = NULL;
 194
 195         nrec = 0;
 196         if ((cur = blk = xdl_mmfile_first(mf, &bsize)) != NULL) {
 197                 for (top = blk + bsize;;) {
 198                         if (cur >= top) {
 199                                 if (!(cur = blk = xdl_mmfile_next(mf, &bsize)))
 200                                         break;
 201                                 top = blk + bsize;
 202                         }
 203                         prev = cur;
 204                         hav = xdl_hash_record(&cur, top, xpp->flags);
 205                         if (nrec >= narec) {
 206                                 narec *= 2;
 207                                 if (!(rrecs = (xrecord_t **) xdl_realloc(recs, narec * sizeof(xrecord_t *)))) {
 208
 209                                         xdl_free(rhash);
 210                                         xdl_free(recs);
 211                                         xdl_cha_free(&xdf->rcha);
 212                                         return -1;
 213                                 }
 214                                 recs = rrecs;
 215                         }
 216                         if (!(crec = xdl_cha_alloc(&xdf->rcha))) {
 217
 218                                 xdl_free(rhash);
 219                                 xdl_free(recs);
 220                                 xdl_cha_free(&xdf->rcha);
 221                                 return -1;
 222                         }
 223                         crec->ptr = prev;
 224                         crec->size = (long) (cur - prev);
 225                         crec->ha = hav;
 226                         recs[nrec++] = crec;
 227
 228                         if (xdl_classify_record(pass, cf, rhash, hbits, crec) < 0) {
 229
 230                                 xdl_free(rhash);
 231                                 xdl_free(recs);
 232                                 xdl_cha_free(&xdf->rcha);
 233                                 return -1;
 234                         }
 235                 }
 236         }
 237
 238         if (!(rchg = (char *) xdl_malloc((nrec + 2) * sizeof(char)))) {
 239
 240                 xdl_free(rhash);
 241                 xdl_free(recs);
 242                 xdl_cha_free(&xdf->rcha);
 243                 return -1;
 244         }
 245         memset(rchg, 0, (nrec + 2) * sizeof(char));
 246
 247         if (!(rindex = (long *) xdl_malloc((nrec + 1) * sizeof(long)))) {
 248
 249                 xdl_free(rchg);
 250                 xdl_free(rhash);
 251                 xdl_free(recs);
 252                 xdl_cha_free(&xdf->rcha);
 253                 return -1;
 254         }
 255         if (!(ha = (unsigned long *) xdl_malloc((nrec + 1) * sizeof(unsigned long)))) {
 256
 257                 xdl_free(rindex);
 258                 xdl_free(rchg);
 259                 xdl_free(rhash);
 260                 xdl_free(recs);
 261                 xdl_cha_free(&xdf->rcha);
 262                 return -1;
 263         }
 264
 265         xdf->nrec = nrec;
 266         xdf->recs = recs;
 267         xdf->hbits = hbits;
 268         xdf->rhash = rhash;
 269         xdf->rchg = rchg + 1;
 270         xdf->rindex = rindex;
 271         xdf->nreff = 0;
 272         xdf->ha = ha;
 273         xdf->dstart = 0;
 274         xdf->dend = nrec - 1;
 275
 276         return 0;
 277 }
 278
 279
 280 static void xdl_free_ctx(xdfile_t *xdf) {
 281
 282         xdl_free(xdf->rhash);
 283         xdl_free(xdf->rindex);
 284         xdl_free(xdf->rchg - 1);
 285         xdl_free(xdf->ha);
 286         xdl_free(xdf->recs);
 287         xdl_cha_free(&xdf->rcha);
 288 }
 289
 290
 291 int xdl_prepare_env(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp,
 292                     xdfenv_t *xe) {
 293         long enl1, enl2;
 294         xdlclassifier_t cf;
 295
 296         enl1 = xdl_guess_lines(mf1) + 1;
 297         enl2 = xdl_guess_lines(mf2) + 1;
 298
 299         if (xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0) {
 300
 301                 return -1;
 302         }
 303
 304         if (xdl_prepare_ctx(1, mf1, enl1, xpp, &cf, &xe->xdf1) < 0) {
 305
 306                 xdl_free_classifier(&cf);
 307                 return -1;
 308         }
 309         if (xdl_prepare_ctx(2, mf2, enl2, xpp, &cf, &xe->xdf2) < 0) {
 310
 311                 xdl_free_ctx(&xe->xdf1);
 312                 xdl_free_classifier(&cf);
 313                 return -1;
 314         }
 315
 316         if (!(xpp->flags & XDF_PATIENCE_DIFF) &&
 317                         xdl_optimize_ctxs(&cf, &xe->xdf1, &xe->xdf2) < 0) {
 318
 319                 xdl_free_ctx(&xe->xdf2);
 320                 xdl_free_ctx(&xe->xdf1);
 321                 return -1;
 322         }
 323
 324         xdl_free_classifier(&cf);
 325
 326         return 0;
 327 }
 328
 329
 330 void xdl_free_env(xdfenv_t *xe) {
 331
 332         xdl_free_ctx(&xe->xdf2);
 333         xdl_free_ctx(&xe->xdf1);
 334 }
 335
 336
 337 static int xdl_clean_mmatch(char const *dis, long i, long s, long e) {
 338         long r, rdis0, rpdis0, rdis1, rpdis1;
 339
 340         /*
 341          * Limits the window the is examined during the similar-lines
 342          * scan. The loops below stops when dis[i - r] == 1 (line that
 343          * has no match), but there are corner cases where the loop
 344          * proceed all the way to the extremities by causing huge
 345          * performance penalties in case of big files.
 346          */
 347         if (i - s > XDL_SIMSCAN_WINDOW)
 348                 s = i - XDL_SIMSCAN_WINDOW;
 349         if (e - i > XDL_SIMSCAN_WINDOW)
 350                 e = i + XDL_SIMSCAN_WINDOW;
 351
 352         /*
 353          * Scans the lines before 'i' to find a run of lines that either
 354          * have no match (dis[j] == 0) or have multiple matches (dis[j] > 1).
 355          * Note that we always call this function with dis[i] > 1, so the
 356          * current line (i) is already a multimatch line.
 357          */
 358         for (r = 1, rdis0 = 0, rpdis0 = 1; (i - r) >= s; r++) {
 359                 if (!dis[i - r])
 360                         rdis0++;
 361                 else if (dis[i - r] == 2)
 362                         rpdis0++;
 363                 else
 364                         break;
 365         }
 366         /*
 367          * If the run before the line 'i' found only multimatch lines, we
 368          * return 0 and hence we don't make the current line (i) discarded.
 369          * We want to discard multimatch lines only when they appear in the
 370          * middle of runs with nomatch lines (dis[j] == 0).
 371          */
 372         if (rdis0 == 0)
 373                 return 0;
 374         for (r = 1, rdis1 = 0, rpdis1 = 1; (i + r) <= e; r++) {
 375                 if (!dis[i + r])
 376                         rdis1++;
 377                 else if (dis[i + r] == 2)
 378                         rpdis1++;
 379                 else
 380                         break;
 381         }
 382         /*
 383          * If the run after the line 'i' found only multimatch lines, we
 384          * return 0 and hence we don't make the current line (i) discarded.
 385          */
 386         if (rdis1 == 0)
 387                 return 0;
 388         rdis1 += rdis0;
 389         rpdis1 += rpdis0;
 390
 391         return rpdis1 * XDL_KPDIS_RUN < (rpdis1 + rdis1);
 392 }
 393
 394
 395 /*
 396  * Try to reduce the problem complexity, discard records that have no
 397  * matches on the other file. Also, lines that have multiple matches
 398  * might be potentially discarded if they happear in a run of discardable.
 399  */
 400 static int xdl_cleanup_records(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2) {
 401         long i, nm, nreff, mlim;
 402         xrecord_t **recs;
 403         xdlclass_t *rcrec;
 404         char *dis, *dis1, *dis2;
 405
 406         if (!(dis = (char *) xdl_malloc(xdf1->nrec + xdf2->nrec + 2))) {
 407
 408                 return -1;
 409         }
 410         memset(dis, 0, xdf1->nrec + xdf2->nrec + 2);
 411         dis1 = dis;
 412         dis2 = dis1 + xdf1->nrec + 1;
 413
 414         if ((mlim = xdl_bogosqrt(xdf1->nrec)) > XDL_MAX_EQLIMIT)
 415                 mlim = XDL_MAX_EQLIMIT;
 416         for (i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart]; i <= xdf1->dend; i++, recs++) {
 417                 rcrec = cf->rcrecs[(*recs)->ha];
 418                 nm = rcrec ? rcrec->len2 : 0;
 419                 dis1[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 420         }
 421
 422         if ((mlim = xdl_bogosqrt(xdf2->nrec)) > XDL_MAX_EQLIMIT)
 423                 mlim = XDL_MAX_EQLIMIT;
 424         for (i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart]; i <= xdf2->dend; i++, recs++) {
 425                 rcrec = cf->rcrecs[(*recs)->ha];
 426                 nm = rcrec ? rcrec->len1 : 0;
 427                 dis2[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 428         }
 429
 430         for (nreff = 0, i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart];
 431              i <= xdf1->dend; i++, recs++) {
 432                 if (dis1[i] == 1 ||
 433                     (dis1[i] == 2 && !xdl_clean_mmatch(dis1, i, xdf1->dstart, xdf1->dend))) {
 434                         xdf1->rindex[nreff] = i;
 435                         xdf1->ha[nreff] = (*recs)->ha;
 436                         nreff++;
 437                 } else
 438                         xdf1->rchg[i] = 1;
 439         }
 440         xdf1->nreff = nreff;
 441
 442         for (nreff = 0, i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart];
 443              i <= xdf2->dend; i++, recs++) {
 444                 if (dis2[i] == 1 ||
 445                     (dis2[i] == 2 && !xdl_clean_mmatch(dis2, i, xdf2->dstart, xdf2->dend))) {
 446                         xdf2->rindex[nreff] = i;
 447                         xdf2->ha[nreff] = (*recs)->ha;
 448                         nreff++;
 449                 } else
 450                         xdf2->rchg[i] = 1;
 451         }
 452         xdf2->nreff = nreff;
 453
 454         xdl_free(dis);
 455
 456         return 0;
 457 }
 458
 459
 460 /*
 461  * Early trim initial and terminal matching records.
 462  */
 463 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2) {
 464         long i, lim;
 465         xrecord_t **recs1, **recs2;
 466
 467         recs1 = xdf1->recs;
 468         recs2 = xdf2->recs;
 469         for (i = 0, lim = XDL_MIN(xdf1->nrec, xdf2->nrec); i < lim;
 470              i++, recs1++, recs2++)
 471                 if ((*recs1)->ha != (*recs2)->ha)
 472                         break;
 473
 474         xdf1->dstart = xdf2->dstart = i;
 475
 476         recs1 = xdf1->recs + xdf1->nrec - 1;
 477         recs2 = xdf2->recs + xdf2->nrec - 1;
 478         for (lim -= i, i = 0; i < lim; i++, recs1--, recs2--)
 479                 if ((*recs1)->ha != (*recs2)->ha)
 480                         break;
 481
 482         xdf1->dend = xdf1->nrec - i - 1;
 483         xdf2->dend = xdf2->nrec - i - 1;
 484
 485         return 0;
 486 }
 487
 488
 489 static int xdl_optimize_ctxs(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2) {
 490
 491         if (xdl_trim_ends(xdf1, xdf2) < 0 ||
 492             xdl_cleanup_records(cf, xdf1, xdf2) < 0) {
 493
 494                 return -1;
 495         }
 496
 497         return 0;
 498 }