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 /* Table of rename/copy destinations */
  10
  11 static struct diff_rename_dst {
  12         struct diff_filespec *two;
  13         struct diff_filepair *pair;
  14 } *rename_dst;
  15 static int rename_dst_nr, rename_dst_alloc;
  16
  17 static struct diff_rename_dst *locate_rename_dst(struct diff_filespec *two,
  18                                                  int insert_ok)
  19 {
  20         int first, last;
  21
  22         first = 0;
  23         last = rename_dst_nr;
  24         while (last > first) {
  25                 int next = (last + first) >> 1;
  26                 struct diff_rename_dst *dst = &(rename_dst[next]);
  27                 int cmp = strcmp(two->path, dst->two->path);
  28                 if (!cmp)
  29                         return dst;
  30                 if (cmp < 0) {
  31                         last = next;
  32                         continue;
  33                 }
  34                 first = next+1;
  35         }
  36         /* not found */
  37         if (!insert_ok)
  38                 return NULL;
  39         /* insert to make it at "first" */
  40         if (rename_dst_alloc <= rename_dst_nr) {
  41                 rename_dst_alloc = alloc_nr(rename_dst_alloc);
  42                 rename_dst = xrealloc(rename_dst,
  43                                       rename_dst_alloc * sizeof(*rename_dst));
  44         }
  45         rename_dst_nr++;
  46         if (first < rename_dst_nr)
  47                 memmove(rename_dst + first + 1, rename_dst + first,
  48                         (rename_dst_nr - first - 1) * sizeof(*rename_dst));
  49         rename_dst[first].two = two;
  50         rename_dst[first].pair = NULL;
  51         return &(rename_dst[first]);
  52 }
  53
  54 static struct diff_rename_src {
  55         struct diff_filespec *one;
  56         unsigned src_used : 1;
  57 } *rename_src;
  58 static int rename_src_nr, rename_src_alloc;
  59
  60 static struct diff_rename_src *locate_rename_src(struct diff_filespec *one,
  61                                                  int insert_ok)
  62 {
  63         int first, last;
  64
  65         first = 0;
  66         last = rename_src_nr;
  67         while (last > first) {
  68                 int next = (last + first) >> 1;
  69                 struct diff_rename_src *src = &(rename_src[next]);
  70                 int cmp = strcmp(one->path, src->one->path);
  71                 if (!cmp)
  72                         return src;
  73                 if (cmp < 0) {
  74                         last = next;
  75                         continue;
  76                 }
  77                 first = next+1;
  78         }
  79         /* not found */
  80         if (!insert_ok)
  81                 return NULL;
  82         /* insert to make it at "first" */
  83         if (rename_src_alloc <= rename_src_nr) {
  84                 rename_src_alloc = alloc_nr(rename_src_alloc);
  85                 rename_src = xrealloc(rename_src,
  86                                       rename_src_alloc * sizeof(*rename_src));
  87         }
  88         rename_src_nr++;
  89         if (first < rename_src_nr)
  90                 memmove(rename_src + first + 1, rename_src + first,
  91                         (rename_src_nr - first - 1) * sizeof(*rename_src));
  92         rename_src[first].one = one;
  93         rename_src[first].src_used = 0;
  94         return &(rename_src[first]);
  95 }
  96
  97 static int is_exact_match(struct diff_filespec *src, struct diff_filespec *dst)
  98 {
  99         if (src->sha1_valid && dst->sha1_valid &&
 100             !memcmp(src->sha1, dst->sha1, 20))
 101                 return 1;
 102         if (diff_populate_filespec(src) || diff_populate_filespec(dst))
 103                 /* this is an error but will be caught downstream */
 104                 return 0;
 105         if (src->size == dst->size &&
 106             !memcmp(src->data, dst->data, src->size))
 107                 return 1;
 108         return 0;
 109 }
 110
 111 struct diff_score {
 112         int src; /* index in rename_src */
 113         int dst; /* index in rename_dst */
 114         int score;
 115         int rank;
 116 };
 117
 118 static int estimate_similarity(struct diff_filespec *src,
 119                                struct diff_filespec *dst,
 120                                int minimum_score)
 121 {
 122         /* src points at a file that existed in the original tree (or
 123          * optionally a file in the destination tree) and dst points
 124          * at a newly created file.  They may be quite similar, in which
 125          * case we want to say src is renamed to dst or src is copied into
 126          * dst, and then some edit has been applied to dst.
 127          *
 128          * Compare them and return how similar they are, representing
 129          * the score as an integer between 0 and 10000, except
 130          * where they match exactly it is considered better than anything
 131          * else.
 132          */
 133         void *delta;
 134         unsigned long delta_size, base_size;
 135         int score;
 136
 137         /* We deal only with regular files.  Symlink renames are handled
 138          * only when they are exact matches --- in other words, no edits
 139          * after renaming.
 140          */
 141         if (!S_ISREG(src->mode) || !S_ISREG(dst->mode))
 142                 return 0;
 143
 144         delta_size = ((src->size < dst->size) ?
 145                       (dst->size - src->size) : (src->size - dst->size));
 146         base_size = ((src->size < dst->size) ? src->size : dst->size);
 147
 148         /* We would not consider edits that change the file size so
 149          * drastically.  delta_size must be smaller than
 150          * (MAX_SCORE-minimum_score)/MAX_SCORE * min(src->size, dst->size).
 151          * Note that base_size == 0 case is handled here already
 152          * and the final score computation below would not have a
 153          * divide-by-zero issue.
 154          */
 155         if (base_size * (MAX_SCORE-minimum_score) < delta_size * MAX_SCORE)
 156                 return 0;
 157
 158         delta = diff_delta(src->data, src->size,
 159                            dst->data, dst->size,
 160                            &delta_size);
 161         /*
 162          * We currently punt here, but we may later end up parsing the
 163          * delta to really assess the extent of damage.  A big consecutive
 164          * remove would produce small delta_size that affects quite a
 165          * big portion of the file.
 166          */
 167         free(delta);
 168
 169         /*
 170          * Now we will give some score to it.  100% edit gets 0 points
 171          * and 0% edit gets MAX_SCORE points.
 172          */
 173         score = MAX_SCORE - (MAX_SCORE * delta_size / base_size);
 174         if (score < 0) return 0;
 175         if (MAX_SCORE < score) return MAX_SCORE;
 176         return score;
 177 }
 178
 179 static void record_rename_pair(struct diff_queue_struct *renq,
 180                                int dst_index, int src_index, int score)
 181 {
 182         struct diff_filespec *one, *two, *src, *dst;
 183         struct diff_filepair *dp;
 184
 185         if (rename_dst[dst_index].pair)
 186                 die("internal error: dst already matched.");
 187
 188         src = rename_src[src_index].one;
 189         one = alloc_filespec(src->path);
 190         fill_filespec(one, src->sha1, src->mode);
 191
 192         dst = rename_dst[dst_index].two;
 193         two = alloc_filespec(dst->path);
 194         fill_filespec(two, dst->sha1, dst->mode);
 195
 196         dp = diff_queue(renq, one, two);
 197         dp->score = score;
 198
 199         rename_src[src_index].src_used = 1;
 200         rename_dst[dst_index].pair = dp;
 201 }
 202
 203 /*
 204  * We sort the rename similarity matrix with the score, in descending
 205  * order (more similar first).
 206  */
 207 static int score_compare(const void *a_, const void *b_)
 208 {
 209         const struct diff_score *a = a_, *b = b_;
 210         return b->score - a->score;
 211 }
 212
 213 int diff_scoreopt_parse(const char *opt)
 214 {
 215         int diglen, num, scale, i;
 216         if (opt[0] != '-' || (opt[1] != 'M' && opt[1] != 'C'))
 217                 return -1; /* that is not a -M nor -C option */
 218         diglen = strspn(opt+2, "0123456789");
 219         if (diglen == 0 || strlen(opt+2) != diglen)
 220                 return 0; /* use default */
 221         sscanf(opt+2, "%d", &num);
 222         for (i = 0, scale = 1; i < diglen; i++)
 223                 scale *= 10;
 224
 225         /* user says num divided by scale and we say internally that
 226          * is MAX_SCORE * num / scale.
 227          */
 228         return MAX_SCORE * num / scale;
 229 }
 230
 231 void diffcore_rename(int detect_rename, int minimum_score)
 232 {
 233         struct diff_queue_struct *q = &diff_queued_diff;
 234         struct diff_queue_struct renq, outq;
 235         struct diff_score *mx;
 236         int i, j;
 237         int num_create, num_src, dst_cnt;
 238
 239         if (!minimum_score)
 240                 minimum_score = DEFAULT_MINIMUM_SCORE;
 241         renq.queue = NULL;
 242         renq.nr = renq.alloc = 0;
 243
 244         for (i = 0; i < q->nr; i++) {
 245                 struct diff_filepair *p = q->queue[i];
 246                 if (!DIFF_FILE_VALID(p->one))
 247                         if (!DIFF_FILE_VALID(p->two))
 248                                 continue; /* unmerged */
 249                         else
 250                                 locate_rename_dst(p->two, 1);
 251                 else if (!DIFF_FILE_VALID(p->two))
 252                         locate_rename_src(p->one, 1);
 253                 else if (1 < detect_rename) /* find copy, too */
 254                         locate_rename_src(p->one, 1);
 255         }
 256         if (rename_dst_nr == 0)
 257                 goto cleanup; /* nothing to do */
 258
 259         /* We really want to cull the candidates list early
 260          * with cheap tests in order to avoid doing deltas.
 261          */
 262         for (i = 0; i < rename_dst_nr; i++) {
 263                 struct diff_filespec *two = rename_dst[i].two;
 264                 for (j = 0; j < rename_src_nr; j++) {
 265                         struct diff_filespec *one = rename_src[j].one;
 266                         if (!is_exact_match(one, two))
 267                                 continue;
 268                         record_rename_pair(&renq, i, j, MAX_SCORE);
 269                         break; /* we are done with this entry */
 270                 }
 271         }
 272         diff_debug_queue("done detecting exact", &renq);
 273
 274         /* Have we run out the created file pool?  If so we can avoid
 275          * doing the delta matrix altogether.
 276          */
 277         if (renq.nr == rename_dst_nr)
 278                 goto flush_rest;
 279
 280         num_create = (rename_dst_nr - renq.nr);
 281         num_src = rename_src_nr;
 282         mx = xmalloc(sizeof(*mx) * num_create * num_src);
 283         for (dst_cnt = i = 0; i < rename_dst_nr; i++) {
 284                 int base = dst_cnt * num_src;
 285                 struct diff_filespec *two = rename_dst[i].two;
 286                 if (rename_dst[i].pair)
 287                         continue; /* dealt with exact match already. */
 288                 for (j = 0; j < rename_src_nr; j++) {
 289                         struct diff_filespec *one = rename_src[j].one;
 290                         struct diff_score *m = &mx[base+j];
 291                         m->src = j;
 292                         m->dst = i;
 293                         m->score = estimate_similarity(one, two,
 294                                                        minimum_score);
 295                 }
 296                 dst_cnt++;
 297         }
 298         /* cost matrix sorted by most to least similar pair */
 299         qsort(mx, num_create * num_src, sizeof(*mx), score_compare);
 300         for (i = 0; i < num_create * num_src; i++) {
 301                 struct diff_rename_dst *dst = &rename_dst[mx[i].dst];
 302                 if (dst->pair)
 303                         continue; /* already done, either exact or fuzzy. */
 304                 if (mx[i].score < minimum_score)
 305                         break; /* there is not any more diffs applicable. */
 306                 record_rename_pair(&renq, mx[i].dst, mx[i].src, mx[i].score);
 307         }
 308         free(mx);
 309         diff_debug_queue("done detecting fuzzy", &renq);
 310
 311  flush_rest:
 312         /* At this point, we have found some renames and copies and they
 313          * are kept in renq.  The original list is still in *q.
 314          *
 315          * Scan the original list and move them into the outq; we will sort
 316          * outq and swap it into the queue supplied to pass that to
 317          * downstream, so we assign the sort keys in this loop.
 318          *
 319          * See comments at the top of record_rename_pair for numbers used
 320          * to assign rename_rank.
 321          */
 322         outq.queue = NULL;
 323         outq.nr = outq.alloc = 0;
 324         for (i = 0; i < q->nr; i++) {
 325                 struct diff_filepair *p = q->queue[i];
 326                 struct diff_rename_src *src = locate_rename_src(p->one, 0);
 327                 struct diff_rename_dst *dst = locate_rename_dst(p->two, 0);
 328                 struct diff_filepair *pair_to_free = NULL;
 329
 330                 if (dst) {
 331                         /* creation */
 332                         if (dst->pair) {
 333                                 /* renq has rename/copy already to produce
 334                                  * this file, so we do not emit the creation
 335                                  * record in the output.
 336                                  */
 337                                 diff_q(&outq, dst->pair);
 338                                 pair_to_free = p;
 339                         }
 340                         else
 341                                 /* no matching rename/copy source, so record
 342                                  * this as a creation.
 343                                  */
 344                                 diff_q(&outq, p);
 345                 }
 346                 else if (!diff_unmodified_pair(p))
 347                         /* all the other cases need to be recorded as is */
 348                         diff_q(&outq, p);
 349                 else {
 350                         /* unmodified pair needs to be recorded only if
 351                          * it is used as the source of rename/copy
 352                          */
 353                         if (src && src->src_used)
 354                                 diff_q(&outq, p);
 355                         else
 356                                 pair_to_free = p;
 357                 }
 358                 if (pair_to_free) {
 359                         diff_free_filespec_data(pair_to_free->one);
 360                         diff_free_filespec_data(pair_to_free->two);
 361                         free(pair_to_free);
 362                 }
 363         }
 364         diff_debug_queue("done copying original", &outq);
 365
 366         free(renq.queue);
 367         free(q->queue);
 368         *q = outq;
 369         diff_debug_queue("done collapsing", q);
 370
 371  cleanup:
 372         free(rename_dst);
 373         rename_dst = NULL;
 374         rename_dst_nr = rename_dst_alloc = 0;
 375         free(rename_src);
 376         rename_src = NULL;
 377         rename_src_nr = rename_src_alloc = 0;
 378         return;
 379 }