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 (!XDL_CALLOC_ARRAY(cf->rchash, cf->hsize)) {
  82
  83                 xdl_cha_free(&cf->ncha);
  84                 return -1;
  85         }
  86
  87         cf->alloc = size;
  88         if (!XDL_ALLOC_ARRAY(cf->rcrecs, cf->alloc)) {
  89
  90                 xdl_free(cf->rchash);
  91                 xdl_cha_free(&cf->ncha);
  92                 return -1;
  93         }
  94
  95         cf->count = 0;
  96
  97         return 0;
  98 }
  99
 100
 101 static void xdl_free_classifier(xdlclassifier_t *cf) {
 102
 103         xdl_free(cf->rcrecs);
 104         xdl_free(cf->rchash);
 105         xdl_cha_free(&cf->ncha);
 106 }
 107
 108
 109 static int xdl_classify_record(unsigned int pass, xdlclassifier_t *cf, xrecord_t **rhash,
 110                                unsigned int hbits, xrecord_t *rec) {
 111         long hi;
 112         char const *line;
 113         xdlclass_t *rcrec;
 114
 115         line = rec->ptr;
 116         hi = (long) XDL_HASHLONG(rec->ha, cf->hbits);
 117         for (rcrec = cf->rchash[hi]; rcrec; rcrec = rcrec->next)
 118                 if (rcrec->ha == rec->ha &&
 119                                 xdl_recmatch(rcrec->line, rcrec->size,
 120                                         rec->ptr, rec->size, cf->flags))
 121                         break;
 122
 123         if (!rcrec) {
 124                 if (!(rcrec = xdl_cha_alloc(&cf->ncha))) {
 125
 126                         return -1;
 127                 }
 128                 rcrec->idx = cf->count++;
 129                 if (XDL_ALLOC_GROW(cf->rcrecs, cf->count, cf->alloc))
 130                                 return -1;
 131                 cf->rcrecs[rcrec->idx] = rcrec;
 132                 rcrec->line = line;
 133                 rcrec->size = rec->size;
 134                 rcrec->ha = rec->ha;
 135                 rcrec->len1 = rcrec->len2 = 0;
 136                 rcrec->next = cf->rchash[hi];
 137                 cf->rchash[hi] = rcrec;
 138         }
 139
 140         (pass == 1) ? rcrec->len1++ : rcrec->len2++;
 141
 142         rec->ha = (unsigned long) rcrec->idx;
 143
 144         hi = (long) XDL_HASHLONG(rec->ha, hbits);
 145         rec->next = rhash[hi];
 146         rhash[hi] = rec;
 147
 148         return 0;
 149 }
 150
 151
 152 static int xdl_prepare_ctx(unsigned int pass, mmfile_t *mf, long narec, xpparam_t const *xpp,
 153                            xdlclassifier_t *cf, xdfile_t *xdf) {
 154         unsigned int hbits;
 155         long nrec, hsize, bsize;
 156         unsigned long hav;
 157         char const *blk, *cur, *top, *prev;
 158         xrecord_t *crec;
 159         xrecord_t **recs;
 160         xrecord_t **rhash;
 161         unsigned long *ha;
 162         char *rchg;
 163         long *rindex;
 164
 165         ha = NULL;
 166         rindex = NULL;
 167         rchg = NULL;
 168         rhash = NULL;
 169         recs = NULL;
 170
 171         if (xdl_cha_init(&xdf->rcha, sizeof(xrecord_t), narec / 4 + 1) < 0)
 172                 goto abort;
 173         if (!XDL_ALLOC_ARRAY(recs, narec))
 174                 goto abort;
 175
 176         hbits = xdl_hashbits((unsigned int) narec);
 177         hsize = 1 << hbits;
 178         if (!XDL_CALLOC_ARRAY(rhash, hsize))
 179                 goto abort;
 180
 181         nrec = 0;
 182         if ((cur = blk = xdl_mmfile_first(mf, &bsize))) {
 183                 for (top = blk + bsize; cur < top; ) {
 184                         prev = cur;
 185                         hav = xdl_hash_record(&cur, top, xpp->flags);
 186                         if (XDL_ALLOC_GROW(recs, nrec + 1, narec))
 187                                 goto abort;
 188                         if (!(crec = xdl_cha_alloc(&xdf->rcha)))
 189                                 goto abort;
 190                         crec->ptr = prev;
 191                         crec->size = (long) (cur - prev);
 192                         crec->ha = hav;
 193                         recs[nrec++] = crec;
 194                         if (xdl_classify_record(pass, cf, rhash, hbits, crec) < 0)
 195                                 goto abort;
 196                 }
 197         }
 198
 199         if (!XDL_CALLOC_ARRAY(rchg, nrec + 2))
 200                 goto abort;
 201
 202         if ((XDF_DIFF_ALG(xpp->flags) != XDF_PATIENCE_DIFF) &&
 203             (XDF_DIFF_ALG(xpp->flags) != XDF_HISTOGRAM_DIFF)) {
 204                 if (!XDL_ALLOC_ARRAY(rindex, nrec + 1))
 205                         goto abort;
 206                 if (!XDL_ALLOC_ARRAY(ha, nrec + 1))
 207                         goto abort;
 208         }
 209
 210         xdf->nrec = nrec;
 211         xdf->recs = recs;
 212         xdf->hbits = hbits;
 213         xdf->rhash = rhash;
 214         xdf->rchg = rchg + 1;
 215         xdf->rindex = rindex;
 216         xdf->nreff = 0;
 217         xdf->ha = ha;
 218         xdf->dstart = 0;
 219         xdf->dend = nrec - 1;
 220
 221         return 0;
 222
 223 abort:
 224         xdl_free(ha);
 225         xdl_free(rindex);
 226         xdl_free(rchg);
 227         xdl_free(rhash);
 228         xdl_free(recs);
 229         xdl_cha_free(&xdf->rcha);
 230         return -1;
 231 }
 232
 233
 234 static void xdl_free_ctx(xdfile_t *xdf) {
 235
 236         xdl_free(xdf->rhash);
 237         xdl_free(xdf->rindex);
 238         xdl_free(xdf->rchg - 1);
 239         xdl_free(xdf->ha);
 240         xdl_free(xdf->recs);
 241         xdl_cha_free(&xdf->rcha);
 242 }
 243
 244
 245 int xdl_prepare_env(mmfile_t *mf1, mmfile_t *mf2, xpparam_t const *xpp,
 246                     xdfenv_t *xe) {
 247         long enl1, enl2, sample;
 248         xdlclassifier_t cf;
 249
 250         memset(&cf, 0, sizeof(cf));
 251
 252         /*
 253          * For histogram diff, we can afford a smaller sample size and
 254          * thus a poorer estimate of the number of lines, as the hash
 255          * table (rhash) won't be filled up/grown. The number of lines
 256          * (nrecs) will be updated correctly anyway by
 257          * xdl_prepare_ctx().
 258          */
 259         sample = (XDF_DIFF_ALG(xpp->flags) == XDF_HISTOGRAM_DIFF
 260                   ? XDL_GUESS_NLINES2 : XDL_GUESS_NLINES1);
 261
 262         enl1 = xdl_guess_lines(mf1, sample) + 1;
 263         enl2 = xdl_guess_lines(mf2, sample) + 1;
 264
 265         if (xdl_init_classifier(&cf, enl1 + enl2 + 1, xpp->flags) < 0)
 266                 return -1;
 267
 268         if (xdl_prepare_ctx(1, mf1, enl1, xpp, &cf, &xe->xdf1) < 0) {
 269
 270                 xdl_free_classifier(&cf);
 271                 return -1;
 272         }
 273         if (xdl_prepare_ctx(2, mf2, enl2, xpp, &cf, &xe->xdf2) < 0) {
 274
 275                 xdl_free_ctx(&xe->xdf1);
 276                 xdl_free_classifier(&cf);
 277                 return -1;
 278         }
 279
 280         if ((XDF_DIFF_ALG(xpp->flags) != XDF_PATIENCE_DIFF) &&
 281             (XDF_DIFF_ALG(xpp->flags) != XDF_HISTOGRAM_DIFF) &&
 282             xdl_optimize_ctxs(&cf, &xe->xdf1, &xe->xdf2) < 0) {
 283
 284                 xdl_free_ctx(&xe->xdf2);
 285                 xdl_free_ctx(&xe->xdf1);
 286                 xdl_free_classifier(&cf);
 287                 return -1;
 288         }
 289
 290         xdl_free_classifier(&cf);
 291
 292         return 0;
 293 }
 294
 295
 296 void xdl_free_env(xdfenv_t *xe) {
 297
 298         xdl_free_ctx(&xe->xdf2);
 299         xdl_free_ctx(&xe->xdf1);
 300 }
 301
 302
 303 static int xdl_clean_mmatch(char const *dis, long i, long s, long e) {
 304         long r, rdis0, rpdis0, rdis1, rpdis1;
 305
 306         /*
 307          * Limits the window the is examined during the similar-lines
 308          * scan. The loops below stops when dis[i - r] == 1 (line that
 309          * has no match), but there are corner cases where the loop
 310          * proceed all the way to the extremities by causing huge
 311          * performance penalties in case of big files.
 312          */
 313         if (i - s > XDL_SIMSCAN_WINDOW)
 314                 s = i - XDL_SIMSCAN_WINDOW;
 315         if (e - i > XDL_SIMSCAN_WINDOW)
 316                 e = i + XDL_SIMSCAN_WINDOW;
 317
 318         /*
 319          * Scans the lines before 'i' to find a run of lines that either
 320          * have no match (dis[j] == 0) or have multiple matches (dis[j] > 1).
 321          * Note that we always call this function with dis[i] > 1, so the
 322          * current line (i) is already a multimatch line.
 323          */
 324         for (r = 1, rdis0 = 0, rpdis0 = 1; (i - r) >= s; r++) {
 325                 if (!dis[i - r])
 326                         rdis0++;
 327                 else if (dis[i - r] == 2)
 328                         rpdis0++;
 329                 else
 330                         break;
 331         }
 332         /*
 333          * If the run before the line 'i' found only multimatch lines, we
 334          * return 0 and hence we don't make the current line (i) discarded.
 335          * We want to discard multimatch lines only when they appear in the
 336          * middle of runs with nomatch lines (dis[j] == 0).
 337          */
 338         if (rdis0 == 0)
 339                 return 0;
 340         for (r = 1, rdis1 = 0, rpdis1 = 1; (i + r) <= e; r++) {
 341                 if (!dis[i + r])
 342                         rdis1++;
 343                 else if (dis[i + r] == 2)
 344                         rpdis1++;
 345                 else
 346                         break;
 347         }
 348         /*
 349          * If the run after the line 'i' found only multimatch lines, we
 350          * return 0 and hence we don't make the current line (i) discarded.
 351          */
 352         if (rdis1 == 0)
 353                 return 0;
 354         rdis1 += rdis0;
 355         rpdis1 += rpdis0;
 356
 357         return rpdis1 * XDL_KPDIS_RUN < (rpdis1 + rdis1);
 358 }
 359
 360
 361 /*
 362  * Try to reduce the problem complexity, discard records that have no
 363  * matches on the other file. Also, lines that have multiple matches
 364  * might be potentially discarded if they happear in a run of discardable.
 365  */
 366 static int xdl_cleanup_records(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2) {
 367         long i, nm, nreff, mlim;
 368         xrecord_t **recs;
 369         xdlclass_t *rcrec;
 370         char *dis, *dis1, *dis2;
 371
 372         if (!XDL_CALLOC_ARRAY(dis, xdf1->nrec + xdf2->nrec + 2))
 373                 return -1;
 374         dis1 = dis;
 375         dis2 = dis1 + xdf1->nrec + 1;
 376
 377         if ((mlim = xdl_bogosqrt(xdf1->nrec)) > XDL_MAX_EQLIMIT)
 378                 mlim = XDL_MAX_EQLIMIT;
 379         for (i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart]; i <= xdf1->dend; i++, recs++) {
 380                 rcrec = cf->rcrecs[(*recs)->ha];
 381                 nm = rcrec ? rcrec->len2 : 0;
 382                 dis1[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 383         }
 384
 385         if ((mlim = xdl_bogosqrt(xdf2->nrec)) > XDL_MAX_EQLIMIT)
 386                 mlim = XDL_MAX_EQLIMIT;
 387         for (i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart]; i <= xdf2->dend; i++, recs++) {
 388                 rcrec = cf->rcrecs[(*recs)->ha];
 389                 nm = rcrec ? rcrec->len1 : 0;
 390                 dis2[i] = (nm == 0) ? 0: (nm >= mlim) ? 2: 1;
 391         }
 392
 393         for (nreff = 0, i = xdf1->dstart, recs = &xdf1->recs[xdf1->dstart];
 394              i <= xdf1->dend; i++, recs++) {
 395                 if (dis1[i] == 1 ||
 396                     (dis1[i] == 2 && !xdl_clean_mmatch(dis1, i, xdf1->dstart, xdf1->dend))) {
 397                         xdf1->rindex[nreff] = i;
 398                         xdf1->ha[nreff] = (*recs)->ha;
 399                         nreff++;
 400                 } else
 401                         xdf1->rchg[i] = 1;
 402         }
 403         xdf1->nreff = nreff;
 404
 405         for (nreff = 0, i = xdf2->dstart, recs = &xdf2->recs[xdf2->dstart];
 406              i <= xdf2->dend; i++, recs++) {
 407                 if (dis2[i] == 1 ||
 408                     (dis2[i] == 2 && !xdl_clean_mmatch(dis2, i, xdf2->dstart, xdf2->dend))) {
 409                         xdf2->rindex[nreff] = i;
 410                         xdf2->ha[nreff] = (*recs)->ha;
 411                         nreff++;
 412                 } else
 413                         xdf2->rchg[i] = 1;
 414         }
 415         xdf2->nreff = nreff;
 416
 417         xdl_free(dis);
 418
 419         return 0;
 420 }
 421
 422
 423 /*
 424  * Early trim initial and terminal matching records.
 425  */
 426 static int xdl_trim_ends(xdfile_t *xdf1, xdfile_t *xdf2) {
 427         long i, lim;
 428         xrecord_t **recs1, **recs2;
 429
 430         recs1 = xdf1->recs;
 431         recs2 = xdf2->recs;
 432         for (i = 0, lim = XDL_MIN(xdf1->nrec, xdf2->nrec); i < lim;
 433              i++, recs1++, recs2++)
 434                 if ((*recs1)->ha != (*recs2)->ha)
 435                         break;
 436
 437         xdf1->dstart = xdf2->dstart = i;
 438
 439         recs1 = xdf1->recs + xdf1->nrec - 1;
 440         recs2 = xdf2->recs + xdf2->nrec - 1;
 441         for (lim -= i, i = 0; i < lim; i++, recs1--, recs2--)
 442                 if ((*recs1)->ha != (*recs2)->ha)
 443                         break;
 444
 445         xdf1->dend = xdf1->nrec - i - 1;
 446         xdf2->dend = xdf2->nrec - i - 1;
 447
 448         return 0;
 449 }
 450
 451
 452 static int xdl_optimize_ctxs(xdlclassifier_t *cf, xdfile_t *xdf1, xdfile_t *xdf2) {
 453
 454         if (xdl_trim_ends(xdf1, xdf2) < 0 ||
 455             xdl_cleanup_records(cf, xdf1, xdf2) < 0) {
 456
 457                 return -1;
 458         }
 459
 460         return 0;
 461 }