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