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 #define XDL_GUESS_NLINES1 256
  30 #define XDL_GUESS_NLINES2 20
  31
  32
  33 typedef struct s_xdlclass {
  34         struct s_xdlclass *next;
  35         unsigned long ha;
  36         char const *line;
  37         long size;
  38         long idx;
  39 } xdlclass_t;
  40
  41 typedef struct s_xdlclassifier {
  42         unsigned int hbits;
  43         long hsize;
  44         xdlclass_t **rchash;
  45         chastore_t ncha;
  46         long count;
  47         long flags;
  48 } xdlclassifier_t;
  49
  50
  51
  52
  53 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags);
  54 static void xdl_free_classifier(xdlclassifier_t *cf);
  55 static int xdl_classify_record(xdlclassifier_t *cf, xrecord_t **rhash, unsigned int hbits,
  56                                xrecord_t *rec);
  57 static int xdl_prepare_ctx(mmfile_t *mf, long narec, xpparam_t const *xpp,
  58                            xdlclassifier_t *cf, xdfile_t *xdf);
  59 static void xdl_free_ctx(xdfile_t *xdf);
  60 static int xdl_clean_mmatch(char const *dis, long i, long s, long e);
  61 static int xdl_cleanup_records(xdfile_t *xdf1, xdfile_t *xdf2);
  62 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2);
  63 static int xdl_optimize_ctxs(xdfile_t *xdf1, xdfile_t *xdf2);
  64
  65
  66
  67
  68 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags) {
  69         cf->flags = flags;
  70
  71         cf->hbits = xdl_hashbits((unsigned int) size);
  72         cf->hsize = 1 << cf->hbits;
  73
  74         if (xdl_cha_init(&cf->ncha, sizeof(xdlclass_t), size / 4 + 1) < 0) {
  75
  76                 return -1;
  77         }
  78         if (!(cf->rchash = (xdlclass_t **) xdl_malloc(cf->hsize * sizeof(xdlclass_t *)))) {
  79
  80                 xdl_cha_free(&cf->ncha);
  81                 return -1;
  82         }
  83         memset(cf->rchash, 0, cf->hsize * sizeof(xdlclass_t *));
  84
  85         cf->count = 0;
  86
  87         return 0;
  88 }
  89
  90
  91 static void xdl_free_classifier(xdlclassifier_t *cf) {
  92
  93         xdl_free(cf->rchash);
  94         xdl_cha_free(&cf->ncha);
  95 }
  96
  97
  98 static int xdl_classify_record(xdlclassifier_t *cf, xrecord_t **rhash, unsigned int hbits,
  99                                xrecord_t *rec) {
 100         long hi;
 101         char const *line;
 102         xdlclass_t *rcrec;
 103
 104         line = rec->ptr;
 105         hi = (long) XDL_HASHLONG(rec->ha, cf->hbits);
 106         for (rcrec = cf->rchash[hi]; rcrec; rcrec = rcrec->next)
 107                 if (rcrec->ha == rec->ha &&
 108                                 xdl_recmatch(rcrec->line, rcrec->size,
 109                                         rec->ptr, rec->size, cf->flags))
 110                         break;
 111
 112         if (!rcrec) {
 113                 if (!(rcrec = xdl_cha_alloc(&cf->ncha))) {
 114
 115                         return -1;
 116                 }
 117                 rcrec->idx = cf->count++;
 118                 rcrec->line = line;
 119                 rcrec->size = rec->size;
 120                 rcrec->ha = rec->ha;
 121                 rcrec->next = cf->rchash[hi];
 122                 cf->rchash[hi] = rcrec;
 123         }
 124
 125         rec->ha = (unsigned long) rcrec->idx;
 126
 127         hi = (long) XDL_HASHLONG(rec->ha, hbits);
 128         rec->next = rhash[hi];
 129         rhash[hi] = rec;
 130
 131         return 0;
 132 }
 133
 134
 135 static int xdl_prepare_ctx(mmfile_t *mf, long narec, xpparam_t const *xpp,
 136                            xdlclassifier_t *cf, xdfile_t *xdf) {
 137         unsigned int hbits;
 138         long nrec, hsize, bsize;
 139         unsigned long hav;
 140         char const *blk, *cur, *top, *prev;
 141         xrecord_t *crec;
 142         xrecord_t **recs, **rrecs;
 143         xrecord_t **rhash;
 144         unsigned long *ha;
 145         char *rchg;
 146         long *rindex;
 147
 148         ha = NULL;
 149         rindex = NULL;
 150         rchg = NULL;
 151         rhash = NULL;
 152         recs = NULL;
 153
 154         if (xdl_cha_init(&xdf->rcha, sizeof(xrecord_t), narec / 4 + 1) < 0)
 155                 goto abort;
 156         if (!(recs = (xrecord_t **) xdl_malloc(narec * sizeof(xrecord_t *))))
 157                 goto abort;
 158
 159         if (xpp->flags & XDF_HISTOGRAM_DIFF)
 160                 hbits = hsize = 0;
 161         else {
 162                 hbits = xdl_hashbits((unsigned int) narec);
 163                 hsize = 1 << hbits;
 164                 if (!(rhash = (xrecord_t **) xdl_malloc(hsize * sizeof(xrecord_t *))))
 165                         goto abort;
 166                 memset(rhash, 0, hsize * sizeof(xrecord_t *));
 167         }
 168
 169         nrec = 0;
 170         if ((cur = blk = xdl_mmfile_first(mf, &bsize)) != NULL) {
 171                 for (top = blk + bsize;;) {
 172                         if (cur >= top) {
 173                                 if (!(cur = blk = xdl_mmfile_next(mf, &bsize)))
 174                                         break;
 175                                 top = blk + bsize;
 176                         }
 177                         prev = cur;
 178                         hav = xdl_hash_record(&cur, top, xpp->flags);
 179                         if (nrec >= narec) {
 180                                 narec *= 2;
 181                                 if (!(rrecs = (xrecord_t **) xdl_realloc(recs, narec * sizeof(xrecord_t *))))
 182                                         goto abort;
 183                                 recs = rrecs;
 184                         }
 185                         if (!(crec = xdl_cha_alloc(&xdf->rcha)))
 186                                 goto abort;
 187                         crec->ptr = prev;
 188                         crec->size = (long) (cur - prev);
 189                         crec->ha = hav;
 190                         recs[nrec++] = crec;
 191
 192                         if (!(xpp->flags & XDF_HISTOGRAM_DIFF) &&
 193                                 xdl_classify_record(cf, rhash, hbits, crec) < 0)
 194                                 goto abort;
 195                 }
 196         }
 197
 198         if (!(rchg = (char *) xdl_malloc((nrec + 2) * sizeof(char))))
 199                 goto abort;
 200         memset(rchg, 0, (nrec + 2) * sizeof(char));
 201
 202         if (!(rindex = (long *) xdl_malloc((nrec + 1) * sizeof(long))))
 203                 goto abort;
 204         if (!(ha = (unsigned long *) xdl_malloc((nrec + 1) * sizeof(unsigned long))))
 205                 goto abort;
 206
 207         xdf->nrec = nrec;
 208         xdf->recs = recs;
 209         xdf->hbits = hbits;
 210         xdf->rhash = rhash;
 211         xdf->rchg = rchg + 1;
 212         xdf->rindex = rindex;
 213         xdf->nreff = 0;
 214         xdf->ha = ha;
 215         xdf->dstart = 0;
 216         xdf->dend = nrec - 1;
 217
 218         return 0;
 219
 220 abort:
 221         xdl_free(ha);
 222         xdl_free(rindex);
 223         xdl_free(rchg);
 224         xdl_free(rhash);
 225         xdl_free(recs);
 226         xdl_cha_free(&xdf->rcha);
 227         return -1;
 228 }
 229
 230
 231 static void xdl_free_ctx(xdfile_t *xdf) {
 232
 233         xdl_free(xdf->rhash);
 234         xdl_free(xdf->rindex);
 235         xdl_free(xdf->rchg - 1);
 236         xdl_free(xdf->ha);
 237         xdl_free(xdf->recs);
 238         xdl_cha_free(&xdf->rcha);
 239 }
 240
 241
 242 int xdl_prepare_env(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp,
 243                     xdfenv_t *xe) {
 244         long enl1, enl2, sample;
 245         xdlclassifier_t cf;
 246
 247         /*
 248          * For histogram diff, we can afford a smaller sample size and
 249          * thus a poorer estimate of the number of lines, as the hash
 250          * table (rhash) won't be filled up/grown. The number of lines
 251          * (nrecs) will be updated correctly anyway by
 252          * xdl_prepare_ctx().
 253          */
 254         sample = xpp->flags & XDF_HISTOGRAM_DIFF ? XDL_GUESS_NLINES2 : XDL_GUESS_NLINES1;
 255
 256         enl1 = xdl_guess_lines(mf1, sample) + 1;
 257         enl2 = xdl_guess_lines(mf2, sample) + 1;
 258
 259         if (!(xpp->flags & XDF_HISTOGRAM_DIFF) &&
 260                 xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0) {
 261
 262                 return -1;
 263         }
 264
 265         if (xdl_prepare_ctx(mf1, enl1, xpp, &cf, &xe->xdf1) < 0) {
 266
 267                 xdl_free_classifier(&cf);
 268                 return -1;
 269         }
 270         if (xdl_prepare_ctx(mf2, enl2, xpp, &cf, &xe->xdf2) < 0) {
 271
 272                 xdl_free_ctx(&xe->xdf1);
 273                 xdl_free_classifier(&cf);
 274                 return -1;
 275         }
 276
 277         if (!(xpp->flags & XDF_HISTOGRAM_DIFF))
 278                 xdl_free_classifier(&cf);
 279
 280         if (!(xpp->flags & XDF_PATIENCE_DIFF) &&
 281                         !(xpp->flags & XDF_HISTOGRAM_DIFF) &&
 282                         xdl_optimize_ctxs(&xe->xdf1, &xe->xdf2) < 0) {
 283
 284                 xdl_free_ctx(&xe->xdf2);
 285                 xdl_free_ctx(&xe->xdf1);
 286                 return -1;
 287         }
 288
 289         return 0;
 290 }
 291
 292
 293 void xdl_free_env(xdfenv_t *xe) {
 294
 295         xdl_free_ctx(&xe->xdf2);
 296         xdl_free_ctx(&xe->xdf1);
 297 }
 298
 299
 300 static int xdl_clean_mmatch(char const *dis, long i, long s, long e) {
 301         long r, rdis0, rpdis0, rdis1, rpdis1;
 302
 303         /*
 304          * Limits the window the is examined during the similar-lines
 305          * scan. The loops below stops when dis[i - r] == 1 (line that
 306          * has no match), but there are corner cases where the loop
 307          * proceed all the way to the extremities by causing huge
 308          * performance penalties in case of big files.
 309          */
 310         if (i - s > XDL_SIMSCAN_WINDOW)
 311                 s = i - XDL_SIMSCAN_WINDOW;
 312         if (e - i > XDL_SIMSCAN_WINDOW)
 313                 e = i + XDL_SIMSCAN_WINDOW;
 314
 315         /*
 316          * Scans the lines before 'i' to find a run of lines that either
 317          * have no match (dis[j] == 0) or have multiple matches (dis[j] > 1).
 318          * Note that we always call this function with dis[i] > 1, so the
 319          * current line (i) is already a multimatch line.
 320          */
 321         for (r = 1, rdis0 = 0, rpdis0 = 1; (i - r) >= s; r++) {
 322                 if (!dis[i - r])
 323                         rdis0++;
 324                 else if (dis[i - r] == 2)
 325                         rpdis0++;
 326                 else
 327                         break;
 328         }
 329         /*
 330          * If the run before the line 'i' found only multimatch lines, we
 331          * return 0 and hence we don't make the current line (i) discarded.
 332          * We want to discard multimatch lines only when they appear in the
 333          * middle of runs with nomatch lines (dis[j] == 0).
 334          */
 335         if (rdis0 == 0)
 336                 return 0;
 337         for (r = 1, rdis1 = 0, rpdis1 = 1; (i + r) <= e; r++) {
 338                 if (!dis[i + r])
 339                         rdis1++;
 340                 else if (dis[i + r] == 2)
 341                         rpdis1++;
 342                 else
 343                         break;
 344         }
 345         /*
 346          * If the run after the line 'i' found only multimatch lines, we
 347          * return 0 and hence we don't make the current line (i) discarded.
 348          */
 349         if (rdis1 == 0)
 350                 return 0;
 351         rdis1 += rdis0;
 352         rpdis1 += rpdis0;
 353
 354         return rpdis1 * XDL_KPDIS_RUN < (rpdis1 + rdis1);
 355 }
 356
 357
 358 /*
 359  * Try to reduce the problem complexity, discard records that have no
 360  * matches on the other file. Also, lines that have multiple matches
 361  * might be potentially discarded if they happear in a run of discardable.
 362  */
 363 static int xdl_cleanup_records(xdfile_t *xdf1, xdfile_t *xdf2) {
 364         long i, nm, rhi, nreff, mlim;
 365         unsigned long hav;
 366         xrecord_t **recs;
 367         xrecord_t *rec;
 368         char *dis, *dis1, *dis2;
 369
 370         if (!(dis = (char *) xdl_malloc(xdf1->nrec + xdf2->nrec + 2))) {
 371
 372                 return -1;
 373         }
 374         memset(dis, 0, xdf1->nrec + xdf2->nrec + 2);
 375         dis1 = dis;
 376         dis2 = dis1 + xdf1->nrec + 1;
 377
 378         if ((mlim = xdl_bogosqrt(xdf1->nrec)) > XDL_MAX_EQLIMIT)
 379                 mlim = XDL_MAX_EQLIMIT;
 380         for (i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart]; i <= xdf1->dend; i++, recs++) {
 381                 hav = (*recs)->ha;
 382                 rhi = (long) XDL_HASHLONG(hav, xdf2->hbits);
 383                 for (nm = 0, rec = xdf2->rhash[rhi]; rec; rec = rec->next)
 384                         if (rec->ha == hav && ++nm == mlim)
 385                                 break;
 386                 dis1[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 387         }
 388
 389         if ((mlim = xdl_bogosqrt(xdf2->nrec)) > XDL_MAX_EQLIMIT)
 390                 mlim = XDL_MAX_EQLIMIT;
 391         for (i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart]; i <= xdf2->dend; i++, recs++) {
 392                 hav = (*recs)->ha;
 393                 rhi = (long) XDL_HASHLONG(hav, xdf1->hbits);
 394                 for (nm = 0, rec = xdf1->rhash[rhi]; rec; rec = rec->next)
 395                         if (rec->ha == hav && ++nm == mlim)
 396                                 break;
 397                 dis2[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 398         }
 399
 400         for (nreff = 0, i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart];
 401              i <= xdf1->dend; i++, recs++) {
 402                 if (dis1[i] == 1 ||
 403                     (dis1[i] == 2 && !xdl_clean_mmatch(dis1, i, xdf1->dstart, xdf1->dend))) {
 404                         xdf1->rindex[nreff] = i;
 405                         xdf1->ha[nreff] = (*recs)->ha;
 406                         nreff++;
 407                 } else
 408                         xdf1->rchg[i] = 1;
 409         }
 410         xdf1->nreff = nreff;
 411
 412         for (nreff = 0, i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart];
 413              i <= xdf2->dend; i++, recs++) {
 414                 if (dis2[i] == 1 ||
 415                     (dis2[i] == 2 && !xdl_clean_mmatch(dis2, i, xdf2->dstart, xdf2->dend))) {
 416                         xdf2->rindex[nreff] = i;
 417                         xdf2->ha[nreff] = (*recs)->ha;
 418                         nreff++;
 419                 } else
 420                         xdf2->rchg[i] = 1;
 421         }
 422         xdf2->nreff = nreff;
 423
 424         xdl_free(dis);
 425
 426         return 0;
 427 }
 428
 429
 430 /*
 431  * Early trim initial and terminal matching records.
 432  */
 433 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2) {
 434         long i, lim;
 435         xrecord_t **recs1, **recs2;
 436
 437         recs1 = xdf1->recs;
 438         recs2 = xdf2->recs;
 439         for (i = 0, lim = XDL_MIN(xdf1->nrec, xdf2->nrec); i < lim;
 440              i++, recs1++, recs2++)
 441                 if ((*recs1)->ha != (*recs2)->ha)
 442                         break;
 443
 444         xdf1->dstart = xdf2->dstart = i;
 445
 446         recs1 = xdf1->recs + xdf1->nrec - 1;
 447         recs2 = xdf2->recs + xdf2->nrec - 1;
 448         for (lim -= i, i = 0; i < lim; i++, recs1--, recs2--)
 449                 if ((*recs1)->ha != (*recs2)->ha)
 450                         break;
 451
 452         xdf1->dend = xdf1->nrec - i - 1;
 453         xdf2->dend = xdf2->nrec - i - 1;
 454
 455         return 0;
 456 }
 457
 458
 459 static int xdl_optimize_ctxs(xdfile_t *xdf1, xdfile_t *xdf2) {
 460
 461         if (xdl_trim_ends(xdf1, xdf2) < 0 ||
 462             xdl_cleanup_records(xdf1, xdf2) < 0) {
 463
 464                 return -1;
 465         }
 466
 467         return 0;
 468 }