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         long len1, len2;
  40 } xdlclass_t;
  41
  42 typedef struct s_xdlclassifier {
  43         unsigned int hbits;
  44         long hsize;
  45         xdlclass_t **rchash;
  46         chastore_t ncha;
  47         xdlclass_t **rcrecs;
  48         long alloc;
  49         long count;
  50         long flags;
  51 } xdlclassifier_t;
  52
  53
  54
  55
  56 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags);
  57 static void xdl_free_classifier(xdlclassifier_t *cf);
  58 static int xdl_classify_record(unsigned int pass, xdlclassifier_t *cf, xrecord_t **rhash,
  59                                unsigned int hbits, xrecord_t *rec);
  60 static int xdl_prepare_ctx(unsigned int pass, mmfile_t *mf, long narec, xpparam_t const *xpp,
  61                            xdlclassifier_t *cf, xdfile_t *xdf);
  62 static void xdl_free_ctx(xdfile_t *xdf);
  63 static int xdl_clean_mmatch(char const *dis, long i, long s, long e);
  64 static int xdl_cleanup_records(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2);
  65 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2);
  66 static int xdl_optimize_ctxs(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2);
  67
  68
  69
  70
  71 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags) {
  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         memset(cf->rchash, 0, cf->hsize * sizeof(xdlclass_t *));
  87
  88         cf->alloc = size;
  89         if (!(cf->rcrecs = (xdlclass_t **) xdl_malloc(cf->alloc * sizeof(xdlclass_t *)))) {
  90
  91                 xdl_free(cf->rchash);
  92                 xdl_cha_free(&cf->ncha);
  93                 return -1;
  94         }
  95
  96         cf->count = 0;
  97
  98         return 0;
  99 }
 100
 101
 102 static void xdl_free_classifier(xdlclassifier_t *cf) {
 103
 104         xdl_free(cf->rcrecs);
 105         xdl_free(cf->rchash);
 106         xdl_cha_free(&cf->ncha);
 107 }
 108
 109
 110 static int xdl_classify_record(unsigned int pass, xdlclassifier_t *cf, xrecord_t **rhash,
 111                                unsigned int hbits, xrecord_t *rec) {
 112         long hi;
 113         char const *line;
 114         xdlclass_t *rcrec;
 115         xdlclass_t **rcrecs;
 116
 117         line = rec->ptr;
 118         hi = (long) XDL_HASHLONG(rec->ha, cf->hbits);
 119         for (rcrec = cf->rchash[hi]; rcrec; rcrec = rcrec->next)
 120                 if (rcrec->ha == rec->ha &&
 121                                 xdl_recmatch(rcrec->line, rcrec->size,
 122                                         rec->ptr, rec->size, cf->flags))
 123                         break;
 124
 125         if (!rcrec) {
 126                 if (!(rcrec = xdl_cha_alloc(&cf->ncha))) {
 127
 128                         return -1;
 129                 }
 130                 rcrec->idx = cf->count++;
 131                 if (cf->count > cf->alloc) {
 132                         cf->alloc *= 2;
 133                         if (!(rcrecs = (xdlclass_t **) xdl_realloc(cf->rcrecs, cf->alloc * sizeof(xdlclass_t *)))) {
 134
 135                                 return -1;
 136                         }
 137                         cf->rcrecs = rcrecs;
 138                 }
 139                 cf->rcrecs[rcrec->idx] = rcrec;
 140                 rcrec->line = line;
 141                 rcrec->size = rec->size;
 142                 rcrec->ha = rec->ha;
 143                 rcrec->len1 = rcrec->len2 = 0;
 144                 rcrec->next = cf->rchash[hi];
 145                 cf->rchash[hi] = rcrec;
 146         }
 147
 148         (pass == 1) ? rcrec->len1++ : rcrec->len2++;
 149
 150         rec->ha = (unsigned long) rcrec->idx;
 151
 152         hi = (long) XDL_HASHLONG(rec->ha, hbits);
 153         rec->next = rhash[hi];
 154         rhash[hi] = rec;
 155
 156         return 0;
 157 }
 158
 159
 160 static int xdl_prepare_ctx(unsigned int pass, mmfile_t *mf, long narec, xpparam_t const *xpp,
 161                            xdlclassifier_t *cf, xdfile_t *xdf) {
 162         unsigned int hbits;
 163         long nrec, hsize, bsize;
 164         unsigned long hav;
 165         char const *blk, *cur, *top, *prev;
 166         xrecord_t *crec;
 167         xrecord_t **recs, **rrecs;
 168         xrecord_t **rhash;
 169         unsigned long *ha;
 170         char *rchg;
 171         long *rindex;
 172
 173         ha = NULL;
 174         rindex = NULL;
 175         rchg = NULL;
 176         rhash = NULL;
 177         recs = NULL;
 178
 179         if (xdl_cha_init(&xdf->rcha, sizeof(xrecord_t), narec / 4 + 1) < 0)
 180                 goto abort;
 181         if (!(recs = (xrecord_t **) xdl_malloc(narec * sizeof(xrecord_t *))))
 182                 goto abort;
 183
 184         if (XDF_DIFF_ALG(xpp->flags) == XDF_HISTOGRAM_DIFF)
 185                 hbits = hsize = 0;
 186         else {
 187                 hbits = xdl_hashbits((unsigned int) narec);
 188                 hsize = 1 << hbits;
 189                 if (!(rhash = (xrecord_t **) xdl_malloc(hsize * sizeof(xrecord_t *))))
 190                         goto abort;
 191                 memset(rhash, 0, hsize * sizeof(xrecord_t *));
 192         }
 193
 194         nrec = 0;
 195         if ((cur = blk = xdl_mmfile_first(mf, &bsize)) != NULL) {
 196                 for (top = blk + bsize; cur < top; ) {
 197                         prev = cur;
 198                         hav = xdl_hash_record(&cur, top, xpp->flags);
 199                         if (nrec >= narec) {
 200                                 narec *= 2;
 201                                 if (!(rrecs = (xrecord_t **) xdl_realloc(recs, narec * sizeof(xrecord_t *))))
 202                                         goto abort;
 203                                 recs = rrecs;
 204                         }
 205                         if (!(crec = xdl_cha_alloc(&xdf->rcha)))
 206                                 goto abort;
 207                         crec->ptr = prev;
 208                         crec->size = (long) (cur - prev);
 209                         crec->ha = hav;
 210                         recs[nrec++] = crec;
 211
 212                         if ((XDF_DIFF_ALG(xpp->flags) != XDF_HISTOGRAM_DIFF) &&
 213                             xdl_classify_record(pass, cf, rhash, hbits, crec) < 0)
 214                                 goto abort;
 215                 }
 216         }
 217
 218         if (!(rchg = (char *) xdl_malloc((nrec + 2) * sizeof(char))))
 219                 goto abort;
 220         memset(rchg, 0, (nrec + 2) * sizeof(char));
 221
 222         if (!(rindex = (long *) xdl_malloc((nrec + 1) * sizeof(long))))
 223                 goto abort;
 224         if (!(ha = (unsigned long *) xdl_malloc((nrec + 1) * sizeof(unsigned long))))
 225                 goto abort;
 226
 227         xdf->nrec = nrec;
 228         xdf->recs = recs;
 229         xdf->hbits = hbits;
 230         xdf->rhash = rhash;
 231         xdf->rchg = rchg + 1;
 232         xdf->rindex = rindex;
 233         xdf->nreff = 0;
 234         xdf->ha = ha;
 235         xdf->dstart = 0;
 236         xdf->dend = nrec - 1;
 237
 238         return 0;
 239
 240 abort:
 241         xdl_free(ha);
 242         xdl_free(rindex);
 243         xdl_free(rchg);
 244         xdl_free(rhash);
 245         xdl_free(recs);
 246         xdl_cha_free(&xdf->rcha);
 247         return -1;
 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, sample;
 265         xdlclassifier_t cf;
 266
 267         memset(&cf, 0, sizeof(cf));
 268
 269         /*
 270          * For histogram diff, we can afford a smaller sample size and
 271          * thus a poorer estimate of the number of lines, as the hash
 272          * table (rhash) won't be filled up/grown. The number of lines
 273          * (nrecs) will be updated correctly anyway by
 274          * xdl_prepare_ctx().
 275          */
 276         sample = (XDF_DIFF_ALG(xpp->flags) == XDF_HISTOGRAM_DIFF
 277                   ? XDL_GUESS_NLINES2 : XDL_GUESS_NLINES1);
 278
 279         enl1 = xdl_guess_lines(mf1, sample) + 1;
 280         enl2 = xdl_guess_lines(mf2, sample) + 1;
 281
 282         if (XDF_DIFF_ALG(xpp->flags) != XDF_HISTOGRAM_DIFF &&
 283             xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0)
 284                 return -1;
 285
 286         if (xdl_prepare_ctx(1, mf1, enl1, xpp, &cf, &xe->xdf1) < 0) {
 287
 288                 xdl_free_classifier(&cf);
 289                 return -1;
 290         }
 291         if (xdl_prepare_ctx(2, mf2, enl2, xpp, &cf, &xe->xdf2) < 0) {
 292
 293                 xdl_free_ctx(&xe->xdf1);
 294                 xdl_free_classifier(&cf);
 295                 return -1;
 296         }
 297
 298         if ((XDF_DIFF_ALG(xpp->flags) != XDF_PATIENCE_DIFF) &&
 299             (XDF_DIFF_ALG(xpp->flags) != XDF_HISTOGRAM_DIFF) &&
 300             xdl_optimize_ctxs(&cf, &xe->xdf1, &xe->xdf2) < 0) {
 301
 302                 xdl_free_ctx(&xe->xdf2);
 303                 xdl_free_ctx(&xe->xdf1);
 304                 xdl_free_classifier(&cf);
 305                 return -1;
 306         }
 307
 308         if (XDF_DIFF_ALG(xpp->flags) != XDF_HISTOGRAM_DIFF)
 309                 xdl_free_classifier(&cf);
 310
 311         return 0;
 312 }
 313
 314
 315 void xdl_free_env(xdfenv_t *xe) {
 316
 317         xdl_free_ctx(&xe->xdf2);
 318         xdl_free_ctx(&xe->xdf1);
 319 }
 320
 321
 322 static int xdl_clean_mmatch(char const *dis, long i, long s, long e) {
 323         long r, rdis0, rpdis0, rdis1, rpdis1;
 324
 325         /*
 326          * Limits the window the is examined during the similar-lines
 327          * scan. The loops below stops when dis[i - r] == 1 (line that
 328          * has no match), but there are corner cases where the loop
 329          * proceed all the way to the extremities by causing huge
 330          * performance penalties in case of big files.
 331          */
 332         if (i - s > XDL_SIMSCAN_WINDOW)
 333                 s = i - XDL_SIMSCAN_WINDOW;
 334         if (e - i > XDL_SIMSCAN_WINDOW)
 335                 e = i + XDL_SIMSCAN_WINDOW;
 336
 337         /*
 338          * Scans the lines before 'i' to find a run of lines that either
 339          * have no match (dis[j] == 0) or have multiple matches (dis[j] > 1).
 340          * Note that we always call this function with dis[i] > 1, so the
 341          * current line (i) is already a multimatch line.
 342          */
 343         for (r = 1, rdis0 = 0, rpdis0 = 1; (i - r) >= s; r++) {
 344                 if (!dis[i - r])
 345                         rdis0++;
 346                 else if (dis[i - r] == 2)
 347                         rpdis0++;
 348                 else
 349                         break;
 350         }
 351         /*
 352          * If the run before the line 'i' found only multimatch lines, we
 353          * return 0 and hence we don't make the current line (i) discarded.
 354          * We want to discard multimatch lines only when they appear in the
 355          * middle of runs with nomatch lines (dis[j] == 0).
 356          */
 357         if (rdis0 == 0)
 358                 return 0;
 359         for (r = 1, rdis1 = 0, rpdis1 = 1; (i + r) <= e; r++) {
 360                 if (!dis[i + r])
 361                         rdis1++;
 362                 else if (dis[i + r] == 2)
 363                         rpdis1++;
 364                 else
 365                         break;
 366         }
 367         /*
 368          * If the run after the line 'i' found only multimatch lines, we
 369          * return 0 and hence we don't make the current line (i) discarded.
 370          */
 371         if (rdis1 == 0)
 372                 return 0;
 373         rdis1 += rdis0;
 374         rpdis1 += rpdis0;
 375
 376         return rpdis1 * XDL_KPDIS_RUN < (rpdis1 + rdis1);
 377 }
 378
 379
 380 /*
 381  * Try to reduce the problem complexity, discard records that have no
 382  * matches on the other file. Also, lines that have multiple matches
 383  * might be potentially discarded if they happear in a run of discardable.
 384  */
 385 static int xdl_cleanup_records(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2) {
 386         long i, nm, nreff, mlim;
 387         xrecord_t **recs;
 388         xdlclass_t *rcrec;
 389         char *dis, *dis1, *dis2;
 390
 391         if (!(dis = (char *) xdl_malloc(xdf1->nrec + xdf2->nrec + 2))) {
 392
 393                 return -1;
 394         }
 395         memset(dis, 0, xdf1->nrec + xdf2->nrec + 2);
 396         dis1 = dis;
 397         dis2 = dis1 + xdf1->nrec + 1;
 398
 399         if ((mlim = xdl_bogosqrt(xdf1->nrec)) > XDL_MAX_EQLIMIT)
 400                 mlim = XDL_MAX_EQLIMIT;
 401         for (i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart]; i <= xdf1->dend; i++, recs++) {
 402                 rcrec = cf->rcrecs[(*recs)->ha];
 403                 nm = rcrec ? rcrec->len2 : 0;
 404                 dis1[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 405         }
 406
 407         if ((mlim = xdl_bogosqrt(xdf2->nrec)) > XDL_MAX_EQLIMIT)
 408                 mlim = XDL_MAX_EQLIMIT;
 409         for (i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart]; i <= xdf2->dend; i++, recs++) {
 410                 rcrec = cf->rcrecs[(*recs)->ha];
 411                 nm = rcrec ? rcrec->len1 : 0;
 412                 dis2[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 413         }
 414
 415         for (nreff = 0, i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart];
 416              i <= xdf1->dend; i++, recs++) {
 417                 if (dis1[i] == 1 ||
 418                     (dis1[i] == 2 && !xdl_clean_mmatch(dis1, i, xdf1->dstart, xdf1->dend))) {
 419                         xdf1->rindex[nreff] = i;
 420                         xdf1->ha[nreff] = (*recs)->ha;
 421                         nreff++;
 422                 } else
 423                         xdf1->rchg[i] = 1;
 424         }
 425         xdf1->nreff = nreff;
 426
 427         for (nreff = 0, i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart];
 428              i <= xdf2->dend; i++, recs++) {
 429                 if (dis2[i] == 1 ||
 430                     (dis2[i] == 2 && !xdl_clean_mmatch(dis2, i, xdf2->dstart, xdf2->dend))) {
 431                         xdf2->rindex[nreff] = i;
 432                         xdf2->ha[nreff] = (*recs)->ha;
 433                         nreff++;
 434                 } else
 435                         xdf2->rchg[i] = 1;
 436         }
 437         xdf2->nreff = nreff;
 438
 439         xdl_free(dis);
 440
 441         return 0;
 442 }
 443
 444
 445 /*
 446  * Early trim initial and terminal matching records.
 447  */
 448 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2) {
 449         long i, lim;
 450         xrecord_t **recs1, **recs2;
 451
 452         recs1 = xdf1->recs;
 453         recs2 = xdf2->recs;
 454         for (i = 0, lim = XDL_MIN(xdf1->nrec, xdf2->nrec); i < lim;
 455              i++, recs1++, recs2++)
 456                 if ((*recs1)->ha != (*recs2)->ha)
 457                         break;
 458
 459         xdf1->dstart = xdf2->dstart = i;
 460
 461         recs1 = xdf1->recs + xdf1->nrec - 1;
 462         recs2 = xdf2->recs + xdf2->nrec - 1;
 463         for (lim -= i, i = 0; i < lim; i++, recs1--, recs2--)
 464                 if ((*recs1)->ha != (*recs2)->ha)
 465                         break;
 466
 467         xdf1->dend = xdf1->nrec - i - 1;
 468         xdf2->dend = xdf2->nrec - i - 1;
 469
 470         return 0;
 471 }
 472
 473
 474 static int xdl_optimize_ctxs(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2) {
 475
 476         if (xdl_trim_ends(xdf1, xdf2) < 0 ||
 477             xdl_cleanup_records(cf, xdf1, xdf2) < 0) {
 478
 479                 return -1;
 480         }
 481
 482         return 0;
 483 }