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, see
  17  *  <http://www.gnu.org/licenses/>.
  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         hbits = xdl_hashbits((unsigned int) narec);
 185         hsize = 1 << hbits;
 186         if (!(rhash = (xrecord_t **) xdl_malloc(hsize * sizeof(xrecord_t *))))
 187                 goto abort;
 188         memset(rhash, 0, hsize * sizeof(xrecord_t *));
 189
 190         nrec = 0;
 191         if ((cur = blk = xdl_mmfile_first(mf, &bsize))) {
 192                 for (top = blk + bsize; cur < top; ) {
 193                         prev = cur;
 194                         hav = xdl_hash_record(&cur, top, xpp->flags);
 195                         if (nrec >= narec) {
 196                                 narec *= 2;
 197                                 if (!(rrecs = (xrecord_t **) xdl_realloc(recs, narec * sizeof(xrecord_t *))))
 198                                         goto abort;
 199                                 recs = rrecs;
 200                         }
 201                         if (!(crec = xdl_cha_alloc(&xdf->rcha)))
 202                                 goto abort;
 203                         crec->ptr = prev;
 204                         crec->size = (long) (cur - prev);
 205                         crec->ha = hav;
 206                         recs[nrec++] = crec;
 207                         if (xdl_classify_record(pass, cf, rhash, hbits, crec) < 0)
 208                                 goto abort;
 209                 }
 210         }
 211
 212         if (!(rchg = (char *) xdl_malloc((nrec + 2) * sizeof(char))))
 213                 goto abort;
 214         memset(rchg, 0, (nrec + 2) * sizeof(char));
 215
 216         if ((XDF_DIFF_ALG(xpp->flags) != XDF_PATIENCE_DIFF) &&
 217             (XDF_DIFF_ALG(xpp->flags) != XDF_HISTOGRAM_DIFF)) {
 218                 if (!(rindex = xdl_malloc((nrec + 1) * sizeof(*rindex))))
 219                         goto abort;
 220                 if (!(ha = xdl_malloc((nrec + 1) * sizeof(*ha))))
 221                         goto abort;
 222         }
 223
 224         xdf->nrec = nrec;
 225         xdf->recs = recs;
 226         xdf->hbits = hbits;
 227         xdf->rhash = rhash;
 228         xdf->rchg = rchg + 1;
 229         xdf->rindex = rindex;
 230         xdf->nreff = 0;
 231         xdf->ha = ha;
 232         xdf->dstart = 0;
 233         xdf->dend = nrec - 1;
 234
 235         return 0;
 236
 237 abort:
 238         xdl_free(ha);
 239         xdl_free(rindex);
 240         xdl_free(rchg);
 241         xdl_free(rhash);
 242         xdl_free(recs);
 243         xdl_cha_free(&xdf->rcha);
 244         return -1;
 245 }
 246
 247
 248 static void xdl_free_ctx(xdfile_t *xdf) {
 249
 250         xdl_free(xdf->rhash);
 251         xdl_free(xdf->rindex);
 252         xdl_free(xdf->rchg - 1);
 253         xdl_free(xdf->ha);
 254         xdl_free(xdf->recs);
 255         xdl_cha_free(&xdf->rcha);
 256 }
 257
 258
 259 int xdl_prepare_env(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp,
 260                     xdfenv_t *xe) {
 261         long enl1, enl2, sample;
 262         xdlclassifier_t cf;
 263
 264         memset(&cf, 0, sizeof(cf));
 265
 266         /*
 267          * For histogram diff, we can afford a smaller sample size and
 268          * thus a poorer estimate of the number of lines, as the hash
 269          * table (rhash) won't be filled up/grown. The number of lines
 270          * (nrecs) will be updated correctly anyway by
 271          * xdl_prepare_ctx().
 272          */
 273         sample = (XDF_DIFF_ALG(xpp->flags) == XDF_HISTOGRAM_DIFF
 274                   ? XDL_GUESS_NLINES2 : XDL_GUESS_NLINES1);
 275
 276         enl1 = xdl_guess_lines(mf1, sample) + 1;
 277         enl2 = xdl_guess_lines(mf2, sample) + 1;
 278
 279         if (xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0)
 280                 return -1;
 281
 282         if (xdl_prepare_ctx(1, mf1, enl1, xpp, &cf, &xe->xdf1) < 0) {
 283
 284                 xdl_free_classifier(&cf);
 285                 return -1;
 286         }
 287         if (xdl_prepare_ctx(2, mf2, enl2, xpp, &cf, &xe->xdf2) < 0) {
 288
 289                 xdl_free_ctx(&xe->xdf1);
 290                 xdl_free_classifier(&cf);
 291                 return -1;
 292         }
 293
 294         if ((XDF_DIFF_ALG(xpp->flags) != XDF_PATIENCE_DIFF) &&
 295             (XDF_DIFF_ALG(xpp->flags) != XDF_HISTOGRAM_DIFF) &&
 296             xdl_optimize_ctxs(&cf, &xe->xdf1, &xe->xdf2) < 0) {
 297
 298                 xdl_free_ctx(&xe->xdf2);
 299                 xdl_free_ctx(&xe->xdf1);
 300                 xdl_free_classifier(&cf);
 301                 return -1;
 302         }
 303
 304         xdl_free_classifier(&cf);
 305
 306         return 0;
 307 }
 308
 309
 310 void xdl_free_env(xdfenv_t *xe) {
 311
 312         xdl_free_ctx(&xe->xdf2);
 313         xdl_free_ctx(&xe->xdf1);
 314 }
 315
 316
 317 static int xdl_clean_mmatch(char const *dis, long i, long s, long e) {
 318         long r, rdis0, rpdis0, rdis1, rpdis1;
 319
 320         /*
 321          * Limits the window the is examined during the similar-lines
 322          * scan. The loops below stops when dis[i - r] == 1 (line that
 323          * has no match), but there are corner cases where the loop
 324          * proceed all the way to the extremities by causing huge
 325          * performance penalties in case of big files.
 326          */
 327         if (i - s > XDL_SIMSCAN_WINDOW)
 328                 s = i - XDL_SIMSCAN_WINDOW;
 329         if (e - i > XDL_SIMSCAN_WINDOW)
 330                 e = i + XDL_SIMSCAN_WINDOW;
 331
 332         /*
 333          * Scans the lines before 'i' to find a run of lines that either
 334          * have no match (dis[j] == 0) or have multiple matches (dis[j] > 1).
 335          * Note that we always call this function with dis[i] > 1, so the
 336          * current line (i) is already a multimatch line.
 337          */
 338         for (r = 1, rdis0 = 0, rpdis0 = 1; (i - r) >= s; r++) {
 339                 if (!dis[i - r])
 340                         rdis0++;
 341                 else if (dis[i - r] == 2)
 342                         rpdis0++;
 343                 else
 344                         break;
 345         }
 346         /*
 347          * If the run before the line 'i' found only multimatch lines, we
 348          * return 0 and hence we don't make the current line (i) discarded.
 349          * We want to discard multimatch lines only when they appear in the
 350          * middle of runs with nomatch lines (dis[j] == 0).
 351          */
 352         if (rdis0 == 0)
 353                 return 0;
 354         for (r = 1, rdis1 = 0, rpdis1 = 1; (i + r) <= e; r++) {
 355                 if (!dis[i + r])
 356                         rdis1++;
 357                 else if (dis[i + r] == 2)
 358                         rpdis1++;
 359                 else
 360                         break;
 361         }
 362         /*
 363          * If the run after the line 'i' found only multimatch lines, we
 364          * return 0 and hence we don't make the current line (i) discarded.
 365          */
 366         if (rdis1 == 0)
 367                 return 0;
 368         rdis1 += rdis0;
 369         rpdis1 += rpdis0;
 370
 371         return rpdis1 * XDL_KPDIS_RUN < (rpdis1 + rdis1);
 372 }
 373
 374
 375 /*
 376  * Try to reduce the problem complexity, discard records that have no
 377  * matches on the other file. Also, lines that have multiple matches
 378  * might be potentially discarded if they happear in a run of discardable.
 379  */
 380 static int xdl_cleanup_records(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2) {
 381         long i, nm, nreff, mlim;
 382         xrecord_t **recs;
 383         xdlclass_t *rcrec;
 384         char *dis, *dis1, *dis2;
 385
 386         if (!(dis = (char *) xdl_malloc(xdf1->nrec + xdf2->nrec + 2))) {
 387
 388                 return -1;
 389         }
 390         memset(dis, 0, xdf1->nrec + xdf2->nrec + 2);
 391         dis1 = dis;
 392         dis2 = dis1 + xdf1->nrec + 1;
 393
 394         if ((mlim = xdl_bogosqrt(xdf1->nrec)) > XDL_MAX_EQLIMIT)
 395                 mlim = XDL_MAX_EQLIMIT;
 396         for (i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart]; i <= xdf1->dend; i++, recs++) {
 397                 rcrec = cf->rcrecs[(*recs)->ha];
 398                 nm = rcrec ? rcrec->len2 : 0;
 399                 dis1[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 400         }
 401
 402         if ((mlim = xdl_bogosqrt(xdf2->nrec)) > XDL_MAX_EQLIMIT)
 403                 mlim = XDL_MAX_EQLIMIT;
 404         for (i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart]; i <= xdf2->dend; i++, recs++) {
 405                 rcrec = cf->rcrecs[(*recs)->ha];
 406                 nm = rcrec ? rcrec->len1 : 0;
 407                 dis2[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 408         }
 409
 410         for (nreff = 0, i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart];
 411              i <= xdf1->dend; i++, recs++) {
 412                 if (dis1[i] == 1 ||
 413                     (dis1[i] == 2 && !xdl_clean_mmatch(dis1, i, xdf1->dstart, xdf1->dend))) {
 414                         xdf1->rindex[nreff] = i;
 415                         xdf1->ha[nreff] = (*recs)->ha;
 416                         nreff++;
 417                 } else
 418                         xdf1->rchg[i] = 1;
 419         }
 420         xdf1->nreff = nreff;
 421
 422         for (nreff = 0, i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart];
 423              i <= xdf2->dend; i++, recs++) {
 424                 if (dis2[i] == 1 ||
 425                     (dis2[i] == 2 && !xdl_clean_mmatch(dis2, i, xdf2->dstart, xdf2->dend))) {
 426                         xdf2->rindex[nreff] = i;
 427                         xdf2->ha[nreff] = (*recs)->ha;
 428                         nreff++;
 429                 } else
 430                         xdf2->rchg[i] = 1;
 431         }
 432         xdf2->nreff = nreff;
 433
 434         xdl_free(dis);
 435
 436         return 0;
 437 }
 438
 439
 440 /*
 441  * Early trim initial and terminal matching records.
 442  */
 443 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2) {
 444         long i, lim;
 445         xrecord_t **recs1, **recs2;
 446
 447         recs1 = xdf1->recs;
 448         recs2 = xdf2->recs;
 449         for (i = 0, lim = XDL_MIN(xdf1->nrec, xdf2->nrec); i < lim;
 450              i++, recs1++, recs2++)
 451                 if ((*recs1)->ha != (*recs2)->ha)
 452                         break;
 453
 454         xdf1->dstart = xdf2->dstart = i;
 455
 456         recs1 = xdf1->recs + xdf1->nrec - 1;
 457         recs2 = xdf2->recs + xdf2->nrec - 1;
 458         for (lim -= i, i = 0; i < lim; i++, recs1--, recs2--)
 459                 if ((*recs1)->ha != (*recs2)->ha)
 460                         break;
 461
 462         xdf1->dend = xdf1->nrec - i - 1;
 463         xdf2->dend = xdf2->nrec - i - 1;
 464
 465         return 0;
 466 }
 467
 468
 469 static int xdl_optimize_ctxs(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2) {
 470
 471         if (xdl_trim_ends(xdf1, xdf2) < 0 ||
 472             xdl_cleanup_records(cf, xdf1, xdf2) < 0) {
 473
 474                 return -1;
 475         }
 476
 477         return 0;
 478 }