diffcore-rename.c

   1 /*
   2  * Copyright (C) 2005 Junio C Hamano
   3  */
   4 #include "cache.h"
   5 #include "diff.h"
   6 #include "diffcore.h"
   7 #include "delta.h"
   8
   9 struct diff_rename_pool {
  10         struct diff_filespec **s;
  11         int nr, alloc;
  12 };
  13
  14 static void diff_rename_pool_clear(struct diff_rename_pool *pool)
  15 {
  16         pool->s = NULL; pool->nr = pool->alloc = 0;
  17 }
  18
  19 static void diff_rename_pool_add(struct diff_rename_pool *pool,
  20                                  struct diff_filespec *s)
  21 {
  22         if (S_ISDIR(s->mode))
  23                 return;  /* rename/copy patch for tree does not make sense. */
  24
  25         if (pool->alloc <= pool->nr) {
  26                 pool->alloc = alloc_nr(pool->alloc);
  27                 pool->s = xrealloc(pool->s,
  28                                    sizeof(*(pool->s)) * pool->alloc);
  29         }
  30         pool->s[pool->nr] = s;
  31         pool->nr++;
  32 }
  33
  34 static int is_exact_match(struct diff_filespec *src, struct diff_filespec *dst)
  35 {
  36         if (src->sha1_valid && dst->sha1_valid &&
  37             !memcmp(src->sha1, dst->sha1, 20))
  38                 return 1;
  39         if (diff_populate_filespec(src) || diff_populate_filespec(dst))
  40                 /* this is an error but will be caught downstream */
  41                 return 0;
  42         if (src->size == dst->size &&
  43             !memcmp(src->data, dst->data, src->size))
  44                 return 1;
  45         return 0;
  46 }
  47
  48 struct diff_score {
  49         struct diff_filespec *src;
  50         struct diff_filespec *dst;
  51         int score;
  52         int rank;
  53 };
  54
  55 static int estimate_similarity(struct diff_filespec *src,
  56                                struct diff_filespec *dst,
  57                                int minimum_score)
  58 {
  59         /* src points at a file that existed in the original tree (or
  60          * optionally a file in the destination tree) and dst points
  61          * at a newly created file.  They may be quite similar, in which
  62          * case we want to say src is renamed to dst or src is copied into
  63          * dst, and then some edit has been applied to dst.
  64          *
  65          * Compare them and return how similar they are, representing
  66          * the score as an integer between 0 and 10000, except
  67          * where they match exactly it is considered better than anything
  68          * else.
  69          */
  70         void *delta;
  71         unsigned long delta_size, base_size;
  72         int score;
  73
  74         delta_size = ((src->size < dst->size) ?
  75                       (dst->size - src->size) : (src->size - dst->size));
  76         base_size = ((src->size < dst->size) ? src->size : dst->size);
  77
  78         /* We would not consider edits that change the file size so
  79          * drastically.  delta_size must be smaller than
  80          * (MAX_SCORE-minimum_score)/MAX_SCORE * min(src->size, dst->size).
  81          * Note that base_size == 0 case is handled here already
  82          * and the final score computation below would not have a
  83          * divide-by-zero issue.
  84          */
  85         if (base_size * (MAX_SCORE-minimum_score) < delta_size * MAX_SCORE)
  86                 return 0;
  87
  88         delta = diff_delta(src->data, src->size,
  89                            dst->data, dst->size,
  90                            &delta_size);
  91         /*
  92          * We currently punt here, but we may later end up parsing the
  93          * delta to really assess the extent of damage.  A big consecutive
  94          * remove would produce small delta_size that affects quite a
  95          * big portion of the file.
  96          */
  97         free(delta);
  98
  99         /*
 100          * Now we will give some score to it.  100% edit gets 0 points
 101          * and 0% edit gets MAX_SCORE points.
 102          */
 103         score = MAX_SCORE - (MAX_SCORE * delta_size / base_size);
 104         if (score < 0) return 0;
 105         if (MAX_SCORE < score) return MAX_SCORE;
 106         return score;
 107 }
 108
 109 static void record_rename_pair(struct diff_queue_struct *outq,
 110                                struct diff_filespec *src,
 111                                struct diff_filespec *dst,
 112                                int rank,
 113                                int score)
 114 {
 115         /* The rank is used to sort the final output, because there
 116          * are certain dependencies.
 117          *
 118          *  - rank #0 depends on deleted ones.
 119          *  - rank #1 depends on kept files before they are modified.
 120          *  - rank #2 depends on kept files after they are modified;
 121          *    currently not used.
 122          *
 123          * Therefore, the final output order should be:
 124          *
 125          *  1. rank #0 rename/copy diffs.
 126          *  2. deletions in the original.
 127          *  3. rank #1 rename/copy diffs.
 128          *  4. additions and modifications in the original.
 129          *  5. rank #2 rename/copy diffs; currently not used.
 130          *
 131          * To achieve this sort order, we give xform_work the number
 132          * above.
 133          */
 134         struct diff_filepair *dp = diff_queue(outq, src, dst);
 135         dp->xfrm_work = (rank * 2 + 1) | (score<<RENAME_SCORE_SHIFT);
 136         dst->xfrm_flags |= RENAME_DST_MATCHED;
 137 }
 138
 139 #if 0
 140 static void debug_filespec(struct diff_filespec *s, int x, const char *one)
 141 {
 142         fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
 143                 x, one,
 144                 s->path,
 145                 DIFF_FILE_VALID(s) ? "valid" : "invalid",
 146                 s->mode,
 147                 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
 148         fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
 149                 x, one,
 150                 s->size, s->xfrm_flags);
 151 }
 152
 153 static void debug_filepair(const struct diff_filepair *p, int i)
 154 {
 155         debug_filespec(p->one, i, "one");
 156         debug_filespec(p->two, i, "two");
 157         fprintf(stderr, "pair flags %d, orig order %d, score %d\n",
 158                 (p->xfrm_work & ((1<<RENAME_SCORE_SHIFT) - 1)),
 159                 p->orig_order,
 160                 (p->xfrm_work >> RENAME_SCORE_SHIFT));
 161 }
 162
 163 static void debug_queue(const char *msg, struct diff_queue_struct *q)
 164 {
 165         int i;
 166         if (msg)
 167                 fprintf(stderr, "%s\n", msg);
 168         fprintf(stderr, "q->nr = %d\n", q->nr);
 169         for (i = 0; i < q->nr; i++) {
 170                 struct diff_filepair *p = q->queue[i];
 171                 debug_filepair(p, i);
 172         }
 173 }
 174 #else
 175 #define debug_queue(a,b) do { ; /*nothing*/ } while(0)
 176 #endif
 177
 178 /*
 179  * We sort the outstanding diff entries according to the rank (see
 180  * comment at the beginning of record_rename_pair) and tiebreak with
 181  * the order in the original input.
 182  */
 183 static int rank_compare(const void *a_, const void *b_)
 184 {
 185         const struct diff_filepair *a = *(const struct diff_filepair **)a_;
 186         const struct diff_filepair *b = *(const struct diff_filepair **)b_;
 187         int a_rank = a->xfrm_work & ((1<<RENAME_SCORE_SHIFT) - 1);
 188         int b_rank = b->xfrm_work & ((1<<RENAME_SCORE_SHIFT) - 1);
 189
 190         if (a_rank != b_rank)
 191                 return a_rank - b_rank;
 192         return a->orig_order - b->orig_order;
 193 }
 194
 195 /*
 196  * We sort the rename similarity matrix with the score, in descending
 197  * order (more similar first).
 198  */
 199 static int score_compare(const void *a_, const void *b_)
 200 {
 201         const struct diff_score *a = a_, *b = b_;
 202         return b->score - a->score;
 203 }
 204
 205 static int needs_to_stay(struct diff_queue_struct *q, int i,
 206                          struct diff_filespec *it)
 207 {
 208         /* If it will be used in later entry (either stay or used
 209          * as the source of rename/copy), we need to copy, not rename.
 210          */
 211         while (i < q->nr) {
 212                 struct diff_filepair *p = q->queue[i++];
 213                 if (!DIFF_FILE_VALID(p->two))
 214                         continue; /* removed is fine */
 215                 if (strcmp(p->one->path, it->path))
 216                         continue; /* not relevant */
 217
 218                 /* p has its src set to *it and it is not a delete;
 219                  * it will be used for in-place change or rename/copy,
 220                  * so we cannot rename it out.
 221                  */
 222                 return 1;
 223         }
 224         return 0;
 225 }
 226
 227 int diff_scoreopt_parse(const char *opt)
 228 {
 229         int diglen, num, scale, i;
 230         if (opt[0] != '-' || (opt[1] != 'M' && opt[1] != 'C'))
 231                 return -1; /* that is not a -M nor -C option */
 232         diglen = strspn(opt+2, "0123456789");
 233         if (diglen == 0 || strlen(opt+2) != diglen)
 234                 return 0; /* use default */
 235         sscanf(opt+2, "%d", &num);
 236         for (i = 0, scale = 1; i < diglen; i++)
 237                 scale *= 10;
 238
 239         /* user says num divided by scale and we say internally that
 240          * is MAX_SCORE * num / scale.
 241          */
 242         return MAX_SCORE * num / scale;
 243 }
 244
 245 void diffcore_rename(int detect_rename, int minimum_score)
 246 {
 247         struct diff_queue_struct *q = &diff_queued_diff;
 248         struct diff_queue_struct outq;
 249         struct diff_rename_pool created, deleted, stay;
 250         struct diff_rename_pool *(srcs[2]);
 251         struct diff_score *mx;
 252         int h, i, j;
 253         int num_create, num_src, dst_cnt, src_cnt;
 254
 255         if (!minimum_score)
 256                 minimum_score = DEFAULT_MINIMUM_SCORE;
 257         outq.queue = NULL;
 258         outq.nr = outq.alloc = 0;
 259
 260         diff_rename_pool_clear(&created);
 261         diff_rename_pool_clear(&deleted);
 262         diff_rename_pool_clear(&stay);
 263
 264         srcs[0] = &deleted;
 265         srcs[1] = &stay;
 266
 267         for (i = 0; i < q->nr; i++) {
 268                 struct diff_filepair *p = q->queue[i];
 269                 if (!DIFF_FILE_VALID(p->one))
 270                         if (!DIFF_FILE_VALID(p->two))
 271                                 continue; /* ignore nonsense */
 272                         else
 273                                 diff_rename_pool_add(&created, p->two);
 274                 else if (!DIFF_FILE_VALID(p->two))
 275                         diff_rename_pool_add(&deleted, p->one);
 276                 else if (1 < detect_rename) /* find copy, too */
 277                         diff_rename_pool_add(&stay, p->one);
 278         }
 279         if (created.nr == 0)
 280                 goto cleanup; /* nothing to do */
 281
 282         /* We really want to cull the candidates list early
 283          * with cheap tests in order to avoid doing deltas.
 284          *
 285          * With the current callers, we should not have already
 286          * matched entries at this point, but it is nonetheless
 287          * checked for sanity.
 288          */
 289         for (i = 0; i < created.nr; i++) {
 290                 if (created.s[i]->xfrm_flags & RENAME_DST_MATCHED)
 291                         continue; /* we have matched exactly already */
 292                 for (h = 0; h < sizeof(srcs)/sizeof(srcs[0]); h++) {
 293                         struct diff_rename_pool *p = srcs[h];
 294                         for (j = 0; j < p->nr; j++) {
 295                                 if (!is_exact_match(p->s[j], created.s[i]))
 296                                         continue;
 297                                 record_rename_pair(&outq,
 298                                                    p->s[j], created.s[i], h,
 299                                                    MAX_SCORE);
 300                                 break; /* we are done with this entry */
 301                         }
 302                 }
 303         }
 304         debug_queue("done detecting exact", &outq);
 305
 306         /* Have we run out the created file pool?  If so we can avoid
 307          * doing the delta matrix altogether.
 308          */
 309         if (outq.nr == created.nr)
 310                 goto flush_rest;
 311
 312         num_create = (created.nr - outq.nr);
 313         num_src = deleted.nr + stay.nr;
 314         mx = xmalloc(sizeof(*mx) * num_create * num_src);
 315         for (dst_cnt = i = 0; i < created.nr; i++) {
 316                 int base = dst_cnt * num_src;
 317                 if (created.s[i]->xfrm_flags & RENAME_DST_MATCHED)
 318                         continue; /* dealt with exact match already. */
 319                 for (src_cnt = h = 0; h < sizeof(srcs)/sizeof(srcs[0]); h++) {
 320                         struct diff_rename_pool *p = srcs[h];
 321                         for (j = 0; j < p->nr; j++, src_cnt++) {
 322                                 struct diff_score *m = &mx[base + src_cnt];
 323                                 m->src = p->s[j];
 324                                 m->dst = created.s[i];
 325                                 m->score = estimate_similarity(m->src, m->dst,
 326                                                                minimum_score);
 327                                 m->rank = h;
 328                         }
 329                 }
 330                 dst_cnt++;
 331         }
 332         /* cost matrix sorted by most to least similar pair */
 333         qsort(mx, num_create * num_src, sizeof(*mx), score_compare);
 334         for (i = 0; i < num_create * num_src; i++) {
 335                 if (mx[i].dst->xfrm_flags & RENAME_DST_MATCHED)
 336                         continue; /* alreayd done, either exact or fuzzy. */
 337                 if (mx[i].score < minimum_score)
 338                         break; /* there is not any more diffs applicable. */
 339                 record_rename_pair(&outq,
 340                                   mx[i].src, mx[i].dst, mx[i].rank,
 341                                   mx[i].score);
 342         }
 343         free(mx);
 344         debug_queue("done detecting fuzzy", &outq);
 345
 346  flush_rest:
 347         /* At this point, we have found some renames and copies and they
 348          * are kept in outq.  The original list is still in *q.
 349          *
 350          * Scan the original list and move them into the outq; we will sort
 351          * outq and swap it into the queue supplied to pass that to
 352          * downstream, so we assign the sort keys in this loop.
 353          *
 354          * See comments at the top of record_rename_pair for numbers used
 355          * to assign xfrm_work.
 356          *
 357          * Note that we have not annotated the diff_filepair with any comment
 358          * so there is nothing other than p to free.
 359          */
 360         for (i = 0; i < q->nr; i++) {
 361                 struct diff_filepair *dp, *p = q->queue[i];
 362                 if (!DIFF_FILE_VALID(p->one)) {
 363                         if (DIFF_FILE_VALID(p->two)) {
 364                                 /* creation */
 365                                 dp = diff_queue(&outq, p->one, p->two);
 366                                 dp->xfrm_work = 4;
 367                         }
 368                         /* otherwise it is a nonsense; just ignore it */
 369                 }
 370                 else if (!DIFF_FILE_VALID(p->two)) {
 371                         /* deletion */
 372                         dp = diff_queue(&outq, p->one, p->two);
 373                         dp->xfrm_work = 2;
 374                 }
 375                 else {
 376                         /* modification, or stay as is */
 377                         dp = diff_queue(&outq, p->one, p->two);
 378                         dp->xfrm_work = 4;
 379                 }
 380                 free(p);
 381         }
 382         debug_queue("done copying original", &outq);
 383
 384         /* Sort outq */
 385         qsort(outq.queue, outq.nr, sizeof(outq.queue[0]), rank_compare);
 386
 387         debug_queue("done sorting", &outq);
 388
 389         free(q->queue);
 390         q->nr = q->alloc = 0;
 391         q->queue = NULL;
 392
 393         /* Copy it out to q, removing duplicates. */
 394         for (i = 0; i < outq.nr; i++) {
 395                 struct diff_filepair *p = outq.queue[i];
 396                 if (!DIFF_FILE_VALID(p->one)) {
 397                         /* created */
 398                         if (p->two->xfrm_flags & RENAME_DST_MATCHED)
 399                                 ; /* rename/copy created it already */
 400                         else
 401                                 diff_queue(q, p->one, p->two);
 402                 }
 403                 else if (!DIFF_FILE_VALID(p->two)) {
 404                         /* deleted */
 405                         if (p->one->xfrm_flags & RENAME_SRC_GONE)
 406                                 ; /* rename/copy deleted it already */
 407                         else
 408                                 diff_queue(q, p->one, p->two);
 409                 }
 410                 else if (strcmp(p->one->path, p->two->path)) {
 411                         /* rename or copy */
 412                         struct diff_filepair *dp =
 413                                 diff_queue(q, p->one, p->two);
 414                         int msglen = (strlen(p->one->path) +
 415                                       strlen(p->two->path) + 100);
 416                         int score = (p->xfrm_work >> RENAME_SCORE_SHIFT);
 417                         dp->xfrm_msg = xmalloc(msglen);
 418
 419                         /* if we have a later entry that is a rename/copy
 420                          * that depends on p->one, then we copy here.
 421                          * otherwise we rename it.
 422                          */
 423                         if (needs_to_stay(&outq, i+1, p->one)) {
 424                                 /* copy it */
 425                                 sprintf(dp->xfrm_msg,
 426                                         "similarity index %d%%\n"
 427                                         "copy from %s\n"
 428                                         "copy to %s\n",
 429                                         (int)(0.5 + score * 100 / MAX_SCORE),
 430                                         p->one->path, p->two->path);
 431                         }
 432                         else {
 433                                 /* rename it, and mark it as gone. */
 434                                 p->one->xfrm_flags |= RENAME_SRC_GONE;
 435                                 sprintf(dp->xfrm_msg,
 436                                         "similarity index %d%%\n"
 437                                         "rename old %s\n"
 438                                         "rename new %s\n",
 439                                         (int)(0.5 + score * 100 / MAX_SCORE),
 440                                         p->one->path, p->two->path);
 441                         }
 442                 }
 443                 else
 444                         /* otherwise it is a modified (or stayed) entry */
 445                         diff_queue(q, p->one, p->two);
 446                 free(p);
 447         }
 448
 449         free(outq.queue);
 450         debug_queue("done collapsing", q);
 451
 452  cleanup:
 453         free(created.s);
 454         free(deleted.s);
 455         free(stay.s);
 456         return;
 457 }