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
  27 #define XDL_KPDIS_RUN 4
  28 #define XDL_MAX_EQLIMIT 1024
  29
  30
  31
  32 typedef struct s_xdlclass {
  33         struct s_xdlclass *next;
  34         unsigned long ha;
  35         char const *line;
  36         long size;
  37         long idx;
  38 } xdlclass_t;
  39
  40 typedef struct s_xdlclassifier {
  41         unsigned int hbits;
  42         long hsize;
  43         xdlclass_t **rchash;
  44         chastore_t ncha;
  45         long count;
  46         long flags;
  47 } xdlclassifier_t;
  48
  49
  50
  51
  52 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags);
  53 static void xdl_free_classifier(xdlclassifier_t *cf);
  54 static int xdl_classify_record(xdlclassifier_t *cf, xrecord_t **rhash, unsigned int hbits,
  55                                xrecord_t *rec);
  56 static int xdl_prepare_ctx(mmfile_t *mf, long narec, xpparam_t const *xpp,
  57                            xdlclassifier_t *cf, xdfile_t *xdf);
  58 static void xdl_free_ctx(xdfile_t *xdf);
  59 static int xdl_clean_mmatch(char const *dis, long i, long s, long e);
  60 static int xdl_cleanup_records(xdfile_t *xdf1, xdfile_t *xdf2);
  61 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2);
  62 static int xdl_optimize_ctxs(xdfile_t *xdf1, xdfile_t *xdf2);
  63
  64
  65
  66
  67 static int xdl_init_classifier(xdlclassifier_t *cf, long size, long flags) {
  68         long i;
  69
  70         cf->flags = flags;
  71
  72         cf->hbits = xdl_hashbits((unsigned int) size);
  73         cf->hsize = 1 << cf->hbits;
  74
  75         if (xdl_cha_init(&cf->ncha, sizeof(xdlclass_t), size / 4 + 1) < 0) {
  76
  77                 return -1;
  78         }
  79         if (!(cf->rchash = (xdlclass_t **) xdl_malloc(cf->hsize * sizeof(xdlclass_t *)))) {
  80
  81                 xdl_cha_free(&cf->ncha);
  82                 return -1;
  83         }
  84         for (i = 0; i < cf->hsize; i++)
  85                 cf->rchash[i] = NULL;
  86
  87         cf->count = 0;
  88
  89         return 0;
  90 }
  91
  92
  93 static void xdl_free_classifier(xdlclassifier_t *cf) {
  94
  95         xdl_free(cf->rchash);
  96         xdl_cha_free(&cf->ncha);
  97 }
  98
  99
 100 static int xdl_classify_record(xdlclassifier_t *cf, xrecord_t **rhash, unsigned int hbits,
 101                                xrecord_t *rec) {
 102         long hi;
 103         char const *line;
 104         xdlclass_t *rcrec;
 105
 106         line = rec->ptr;
 107         hi = (long) XDL_HASHLONG(rec->ha, cf->hbits);
 108         for (rcrec = cf->rchash[hi]; rcrec; rcrec = rcrec->next)
 109                 if (rcrec->ha == rec->ha &&
 110                                 xdl_recmatch(rcrec->line, rcrec->size,
 111                                         rec->ptr, rec->size, cf->flags))
 112                         break;
 113
 114         if (!rcrec) {
 115                 if (!(rcrec = xdl_cha_alloc(&cf->ncha))) {
 116
 117                         return -1;
 118                 }
 119                 rcrec->idx = cf->count++;
 120                 rcrec->line = line;
 121                 rcrec->size = rec->size;
 122                 rcrec->ha = rec->ha;
 123                 rcrec->next = cf->rchash[hi];
 124                 cf->rchash[hi] = rcrec;
 125         }
 126
 127         rec->ha = (unsigned long) rcrec->idx;
 128
 129         hi = (long) XDL_HASHLONG(rec->ha, hbits);
 130         rec->next = rhash[hi];
 131         rhash[hi] = rec;
 132
 133         return 0;
 134 }
 135
 136
 137 static int xdl_prepare_ctx(mmfile_t *mf, long narec, xpparam_t const *xpp,
 138                            xdlclassifier_t *cf, xdfile_t *xdf) {
 139         unsigned int hbits;
 140         long i, nrec, hsize, bsize;
 141         unsigned long hav;
 142         char const *blk, *cur, *top, *prev;
 143         xrecord_t *crec;
 144         xrecord_t **recs, **rrecs;
 145         xrecord_t **rhash;
 146         unsigned long *ha;
 147         char *rchg;
 148         long *rindex;
 149
 150         if (xdl_cha_init(&xdf->rcha, sizeof(xrecord_t), narec / 4 + 1) < 0) {
 151
 152                 return -1;
 153         }
 154         if (!(recs = (xrecord_t **) xdl_malloc(narec * sizeof(xrecord_t *)))) {
 155
 156                 xdl_cha_free(&xdf->rcha);
 157                 return -1;
 158         }
 159
 160         hbits = xdl_hashbits((unsigned int) narec);
 161         hsize = 1 << hbits;
 162         if (!(rhash = (xrecord_t **) xdl_malloc(hsize * sizeof(xrecord_t *)))) {
 163
 164                 xdl_free(recs);
 165                 xdl_cha_free(&xdf->rcha);
 166                 return -1;
 167         }
 168         for (i = 0; i < hsize; i++)
 169                 rhash[i] = NULL;
 170
 171         nrec = 0;
 172         if ((cur = blk = xdl_mmfile_first(mf, &bsize)) != NULL) {
 173                 for (top = blk + bsize;;) {
 174                         if (cur >= top) {
 175                                 if (!(cur = blk = xdl_mmfile_next(mf, &bsize)))
 176                                         break;
 177                                 top = blk + bsize;
 178                         }
 179                         prev = cur;
 180                         hav = xdl_hash_record(&cur, top, xpp->flags);
 181                         if (nrec >= narec) {
 182                                 narec *= 2;
 183                                 if (!(rrecs = (xrecord_t **) xdl_realloc(recs, narec * sizeof(xrecord_t *)))) {
 184
 185                                         xdl_free(rhash);
 186                                         xdl_free(recs);
 187                                         xdl_cha_free(&xdf->rcha);
 188                                         return -1;
 189                                 }
 190                                 recs = rrecs;
 191                         }
 192                         if (!(crec = xdl_cha_alloc(&xdf->rcha))) {
 193
 194                                 xdl_free(rhash);
 195                                 xdl_free(recs);
 196                                 xdl_cha_free(&xdf->rcha);
 197                                 return -1;
 198                         }
 199                         crec->ptr = prev;
 200                         crec->size = (long) (cur - prev);
 201                         crec->ha = hav;
 202                         recs[nrec++] = crec;
 203
 204                         if (xdl_classify_record(cf, rhash, hbits, crec) < 0) {
 205
 206                                 xdl_free(rhash);
 207                                 xdl_free(recs);
 208                                 xdl_cha_free(&xdf->rcha);
 209                                 return -1;
 210                         }
 211                 }
 212         }
 213
 214         if (!(rchg = (char *) xdl_malloc((nrec + 2) * sizeof(char)))) {
 215
 216                 xdl_free(rhash);
 217                 xdl_free(recs);
 218                 xdl_cha_free(&xdf->rcha);
 219                 return -1;
 220         }
 221         memset(rchg, 0, (nrec + 2) * sizeof(char));
 222
 223         if (!(rindex = (long *) xdl_malloc((nrec + 1) * sizeof(long)))) {
 224
 225                 xdl_free(rchg);
 226                 xdl_free(rhash);
 227                 xdl_free(recs);
 228                 xdl_cha_free(&xdf->rcha);
 229                 return -1;
 230         }
 231         if (!(ha = (unsigned long *) xdl_malloc((nrec + 1) * sizeof(unsigned long)))) {
 232
 233                 xdl_free(rindex);
 234                 xdl_free(rchg);
 235                 xdl_free(rhash);
 236                 xdl_free(recs);
 237                 xdl_cha_free(&xdf->rcha);
 238                 return -1;
 239         }
 240
 241         xdf->nrec = nrec;
 242         xdf->recs = recs;
 243         xdf->hbits = hbits;
 244         xdf->rhash = rhash;
 245         xdf->rchg = rchg + 1;
 246         xdf->rindex = rindex;
 247         xdf->nreff = 0;
 248         xdf->ha = ha;
 249         xdf->dstart = 0;
 250         xdf->dend = nrec - 1;
 251
 252         return 0;
 253 }
 254
 255
 256 static void xdl_free_ctx(xdfile_t *xdf) {
 257
 258         xdl_free(xdf->rhash);
 259         xdl_free(xdf->rindex);
 260         xdl_free(xdf->rchg - 1);
 261         xdl_free(xdf->ha);
 262         xdl_free(xdf->recs);
 263         xdl_cha_free(&xdf->rcha);
 264 }
 265
 266
 267 int xdl_prepare_env(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp,
 268                     xdfenv_t *xe) {
 269         long enl1, enl2;
 270         xdlclassifier_t cf;
 271
 272         enl1 = xdl_guess_lines(mf1) + 1;
 273         enl2 = xdl_guess_lines(mf2) + 1;
 274
 275         if (xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0) {
 276
 277                 return -1;
 278         }
 279
 280         if (xdl_prepare_ctx(mf1, enl1, xpp, &cf, &xe->xdf1) < 0) {
 281
 282                 xdl_free_classifier(&cf);
 283                 return -1;
 284         }
 285         if (xdl_prepare_ctx(mf2, enl2, xpp, &cf, &xe->xdf2) < 0) {
 286
 287                 xdl_free_ctx(&xe->xdf1);
 288                 xdl_free_classifier(&cf);
 289                 return -1;
 290         }
 291
 292         xdl_free_classifier(&cf);
 293
 294         if (xdl_optimize_ctxs(&xe->xdf1, &xe->xdf2) < 0) {
 295
 296                 xdl_free_ctx(&xe->xdf2);
 297                 xdl_free_ctx(&xe->xdf1);
 298                 return -1;
 299         }
 300
 301         return 0;
 302 }
 303
 304
 305 void xdl_free_env(xdfenv_t *xe) {
 306
 307         xdl_free_ctx(&xe->xdf2);
 308         xdl_free_ctx(&xe->xdf1);
 309 }
 310
 311
 312 static int xdl_clean_mmatch(char const *dis, long i, long s, long e) {
 313         long r, rdis0, rpdis0, rdis1, rpdis1;
 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 }
 469