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; cur < top; ) {
 172                         prev = cur;
 173                         hav = xdl_hash_record(&cur, top, xpp->flags);
 174                         if (nrec >= narec) {
 175                                 narec *= 2;
 176                                 if (!(rrecs = (xrecord_t **) xdl_realloc(recs, narec * sizeof(xrecord_t *))))
 177                                         goto abort;
 178                                 recs = rrecs;
 179                         }
 180                         if (!(crec = xdl_cha_alloc(&xdf->rcha)))
 181                                 goto abort;
 182                         crec->ptr = prev;
 183                         crec->size = (long) (cur - prev);
 184                         crec->ha = hav;
 185                         recs[nrec++] = crec;
 186
 187                         if (!(xpp->flags & XDF_HISTOGRAM_DIFF) &&
 188                                 xdl_classify_record(cf, rhash, hbits, crec) < 0)
 189                                 goto abort;
 190                 }
 191         }
 192
 193         if (!(rchg = (char *) xdl_malloc((nrec + 2) * sizeof(char))))
 194                 goto abort;
 195         memset(rchg, 0, (nrec + 2) * sizeof(char));
 196
 197         if (!(rindex = (long *) xdl_malloc((nrec + 1) * sizeof(long))))
 198                 goto abort;
 199         if (!(ha = (unsigned long *) xdl_malloc((nrec + 1) * sizeof(unsigned long))))
 200                 goto abort;
 201
 202         xdf->nrec = nrec;
 203         xdf->recs = recs;
 204         xdf->hbits = hbits;
 205         xdf->rhash = rhash;
 206         xdf->rchg = rchg + 1;
 207         xdf->rindex = rindex;
 208         xdf->nreff = 0;
 209         xdf->ha = ha;
 210         xdf->dstart = 0;
 211         xdf->dend = nrec - 1;
 212
 213         return 0;
 214
 215 abort:
 216         xdl_free(ha);
 217         xdl_free(rindex);
 218         xdl_free(rchg);
 219         xdl_free(rhash);
 220         xdl_free(recs);
 221         xdl_cha_free(&xdf->rcha);
 222         return -1;
 223 }
 224
 225
 226 static void xdl_free_ctx(xdfile_t *xdf) {
 227
 228         xdl_free(xdf->rhash);
 229         xdl_free(xdf->rindex);
 230         xdl_free(xdf->rchg - 1);
 231         xdl_free(xdf->ha);
 232         xdl_free(xdf->recs);
 233         xdl_cha_free(&xdf->rcha);
 234 }
 235
 236
 237 int xdl_prepare_env(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp,
 238                     xdfenv_t *xe) {
 239         long enl1, enl2, sample;
 240         xdlclassifier_t cf;
 241
 242         /*
 243          * For histogram diff, we can afford a smaller sample size and
 244          * thus a poorer estimate of the number of lines, as the hash
 245          * table (rhash) won't be filled up/grown. The number of lines
 246          * (nrecs) will be updated correctly anyway by
 247          * xdl_prepare_ctx().
 248          */
 249         sample = xpp->flags & XDF_HISTOGRAM_DIFF ? XDL_GUESS_NLINES2 : XDL_GUESS_NLINES1;
 250
 251         enl1 = xdl_guess_lines(mf1, sample) + 1;
 252         enl2 = xdl_guess_lines(mf2, sample) + 1;
 253
 254         if (!(xpp->flags & XDF_HISTOGRAM_DIFF) &&
 255                 xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0) {
 256
 257                 return -1;
 258         }
 259
 260         if (xdl_prepare_ctx(mf1, enl1, xpp, &cf, &xe->xdf1) < 0) {
 261
 262                 xdl_free_classifier(&cf);
 263                 return -1;
 264         }
 265         if (xdl_prepare_ctx(mf2, enl2, xpp, &cf, &xe->xdf2) < 0) {
 266
 267                 xdl_free_ctx(&xe->xdf1);
 268                 xdl_free_classifier(&cf);
 269                 return -1;
 270         }
 271
 272         if (!(xpp->flags & XDF_HISTOGRAM_DIFF))
 273                 xdl_free_classifier(&cf);
 274
 275         if (!(xpp->flags & XDF_PATIENCE_DIFF) &&
 276                         !(xpp->flags & XDF_HISTOGRAM_DIFF) &&
 277                         xdl_optimize_ctxs(&xe->xdf1, &xe->xdf2) < 0) {
 278
 279                 xdl_free_ctx(&xe->xdf2);
 280                 xdl_free_ctx(&xe->xdf1);
 281                 return -1;
 282         }
 283
 284         return 0;
 285 }
 286
 287
 288 void xdl_free_env(xdfenv_t *xe) {
 289
 290         xdl_free_ctx(&xe->xdf2);
 291         xdl_free_ctx(&xe->xdf1);
 292 }
 293
 294
 295 static int xdl_clean_mmatch(char const *dis, long i, long s, long e) {
 296         long r, rdis0, rpdis0, rdis1, rpdis1;
 297
 298         /*
 299          * Limits the window the is examined during the similar-lines
 300          * scan. The loops below stops when dis[i - r] == 1 (line that
 301          * has no match), but there are corner cases where the loop
 302          * proceed all the way to the extremities by causing huge
 303          * performance penalties in case of big files.
 304          */
 305         if (i - s > XDL_SIMSCAN_WINDOW)
 306                 s = i - XDL_SIMSCAN_WINDOW;
 307         if (e - i > XDL_SIMSCAN_WINDOW)
 308                 e = i + XDL_SIMSCAN_WINDOW;
 309
 310         /*
 311          * Scans the lines before 'i' to find a run of lines that either
 312          * have no match (dis[j] == 0) or have multiple matches (dis[j] > 1).
 313          * Note that we always call this function with dis[i] > 1, so the
 314          * current line (i) is already a multimatch line.
 315          */
 316         for (r = 1, rdis0 = 0, rpdis0 = 1; (i - r) >= s; r++) {
 317                 if (!dis[i - r])
 318                         rdis0++;
 319                 else if (dis[i - r] == 2)
 320                         rpdis0++;
 321                 else
 322                         break;
 323         }
 324         /*
 325          * If the run before the line 'i' found only multimatch lines, we
 326          * return 0 and hence we don't make the current line (i) discarded.
 327          * We want to discard multimatch lines only when they appear in the
 328          * middle of runs with nomatch lines (dis[j] == 0).
 329          */
 330         if (rdis0 == 0)
 331                 return 0;
 332         for (r = 1, rdis1 = 0, rpdis1 = 1; (i + r) <= e; r++) {
 333                 if (!dis[i + r])
 334                         rdis1++;
 335                 else if (dis[i + r] == 2)
 336                         rpdis1++;
 337                 else
 338                         break;
 339         }
 340         /*
 341          * If the run after the line 'i' found only multimatch lines, we
 342          * return 0 and hence we don't make the current line (i) discarded.
 343          */
 344         if (rdis1 == 0)
 345                 return 0;
 346         rdis1 += rdis0;
 347         rpdis1 += rpdis0;
 348
 349         return rpdis1 * XDL_KPDIS_RUN < (rpdis1 + rdis1);
 350 }
 351
 352
 353 /*
 354  * Try to reduce the problem complexity, discard records that have no
 355  * matches on the other file. Also, lines that have multiple matches
 356  * might be potentially discarded if they happear in a run of discardable.
 357  */
 358 static int xdl_cleanup_records(xdfile_t *xdf1, xdfile_t *xdf2) {
 359         long i, nm, rhi, nreff, mlim;
 360         unsigned long hav;
 361         xrecord_t **recs;
 362         xrecord_t *rec;
 363         char *dis, *dis1, *dis2;
 364
 365         if (!(dis = (char *) xdl_malloc(xdf1->nrec + xdf2->nrec + 2))) {
 366
 367                 return -1;
 368         }
 369         memset(dis, 0, xdf1->nrec + xdf2->nrec + 2);
 370         dis1 = dis;
 371         dis2 = dis1 + xdf1->nrec + 1;
 372
 373         if ((mlim = xdl_bogosqrt(xdf1->nrec)) > XDL_MAX_EQLIMIT)
 374                 mlim = XDL_MAX_EQLIMIT;
 375         for (i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart]; i <= xdf1->dend; i++, recs++) {
 376                 hav = (*recs)->ha;
 377                 rhi = (long) XDL_HASHLONG(hav, xdf2->hbits);
 378                 for (nm = 0, rec = xdf2->rhash[rhi]; rec; rec = rec->next)
 379                         if (rec->ha == hav && ++nm == mlim)
 380                                 break;
 381                 dis1[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 382         }
 383
 384         if ((mlim = xdl_bogosqrt(xdf2->nrec)) > XDL_MAX_EQLIMIT)
 385                 mlim = XDL_MAX_EQLIMIT;
 386         for (i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart]; i <= xdf2->dend; i++, recs++) {
 387                 hav = (*recs)->ha;
 388                 rhi = (long) XDL_HASHLONG(hav, xdf1->hbits);
 389                 for (nm = 0, rec = xdf1->rhash[rhi]; rec; rec = rec->next)
 390                         if (rec->ha == hav && ++nm == mlim)
 391                                 break;
 392                 dis2[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 393         }
 394
 395         for (nreff = 0, i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart];
 396              i <= xdf1->dend; i++, recs++) {
 397                 if (dis1[i] == 1 ||
 398                     (dis1[i] == 2 && !xdl_clean_mmatch(dis1, i, xdf1->dstart, xdf1->dend))) {
 399                         xdf1->rindex[nreff] = i;
 400                         xdf1->ha[nreff] = (*recs)->ha;
 401                         nreff++;
 402                 } else
 403                         xdf1->rchg[i] = 1;
 404         }
 405         xdf1->nreff = nreff;
 406
 407         for (nreff = 0, i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart];
 408              i <= xdf2->dend; i++, recs++) {
 409                 if (dis2[i] == 1 ||
 410                     (dis2[i] == 2 && !xdl_clean_mmatch(dis2, i, xdf2->dstart, xdf2->dend))) {
 411                         xdf2->rindex[nreff] = i;
 412                         xdf2->ha[nreff] = (*recs)->ha;
 413                         nreff++;
 414                 } else
 415                         xdf2->rchg[i] = 1;
 416         }
 417         xdf2->nreff = nreff;
 418
 419         xdl_free(dis);
 420
 421         return 0;
 422 }
 423
 424
 425 /*
 426  * Early trim initial and terminal matching records.
 427  */
 428 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2) {
 429         long i, lim;
 430         xrecord_t **recs1, **recs2;
 431
 432         recs1 = xdf1->recs;
 433         recs2 = xdf2->recs;
 434         for (i = 0, lim = XDL_MIN(xdf1->nrec, xdf2->nrec); i < lim;
 435              i++, recs1++, recs2++)
 436                 if ((*recs1)->ha != (*recs2)->ha)
 437                         break;
 438
 439         xdf1->dstart = xdf2->dstart = i;
 440
 441         recs1 = xdf1->recs + xdf1->nrec - 1;
 442         recs2 = xdf2->recs + xdf2->nrec - 1;
 443         for (lim -= i, i = 0; i < lim; i++, recs1--, recs2--)
 444                 if ((*recs1)->ha != (*recs2)->ha)
 445                         break;
 446
 447         xdf1->dend = xdf1->nrec - i - 1;
 448         xdf2->dend = xdf2->nrec - i - 1;
 449
 450         return 0;
 451 }
 452
 453
 454 static int xdl_optimize_ctxs(xdfile_t *xdf1, xdfile_t *xdf2) {
 455
 456         if (xdl_trim_ends(xdf1, xdf2) < 0 ||
 457             xdl_cleanup_records(xdf1, xdf2) < 0) {
 458
 459                 return -1;
 460         }
 461
 462         return 0;
 463 }