[PATCH] Performance fix for pickaxe.
[git/dscho.git] / diffcore-rename.c
blobe8de49358667a51e9622d5423e4f8523a8f6907e
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"
9 struct diff_rename_pool {
10 struct diff_filespec **s;
11 int nr, alloc;
14 static void diff_rename_pool_clear(struct diff_rename_pool *pool)
16 pool->s = NULL; pool->nr = pool->alloc = 0;
19 static void diff_rename_pool_add(struct diff_rename_pool *pool,
20 struct diff_filespec *s)
22 if (S_ISDIR(s->mode))
23 return; /* no trees, please */
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);
30 pool->s[pool->nr] = s;
31 pool->nr++;
34 static int is_exact_match(struct diff_filespec *src, struct diff_filespec *dst)
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;
48 struct diff_score {
49 struct diff_filespec *src;
50 struct diff_filespec *dst;
51 int score;
52 int rank;
55 static int estimate_similarity(struct diff_filespec *src,
56 struct diff_filespec *dst,
57 int minimum_score)
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.
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.
70 void *delta;
71 unsigned long delta_size, base_size;
72 int score;
74 /* We deal only with regular files. Symlink renames are handled
75 * only when they are exact matches --- in other words, no edits
76 * after renaming.
78 if (!S_ISREG(src->mode) || !S_ISREG(dst->mode))
79 return 0;
81 delta_size = ((src->size < dst->size) ?
82 (dst->size - src->size) : (src->size - dst->size));
83 base_size = ((src->size < dst->size) ? src->size : dst->size);
85 /* We would not consider edits that change the file size so
86 * drastically. delta_size must be smaller than
87 * (MAX_SCORE-minimum_score)/MAX_SCORE * min(src->size, dst->size).
88 * Note that base_size == 0 case is handled here already
89 * and the final score computation below would not have a
90 * divide-by-zero issue.
92 if (base_size * (MAX_SCORE-minimum_score) < delta_size * MAX_SCORE)
93 return 0;
95 delta = diff_delta(src->data, src->size,
96 dst->data, dst->size,
97 &delta_size);
99 * We currently punt here, but we may later end up parsing the
100 * delta to really assess the extent of damage. A big consecutive
101 * remove would produce small delta_size that affects quite a
102 * big portion of the file.
104 free(delta);
107 * Now we will give some score to it. 100% edit gets 0 points
108 * and 0% edit gets MAX_SCORE points.
110 score = MAX_SCORE - (MAX_SCORE * delta_size / base_size);
111 if (score < 0) return 0;
112 if (MAX_SCORE < score) return MAX_SCORE;
113 return score;
116 static void record_rename_pair(struct diff_queue_struct *outq,
117 struct diff_filespec *src,
118 struct diff_filespec *dst,
119 int rank,
120 int score)
123 * These ranks are used to sort the final output, because there
124 * are certain dependencies:
126 * 1. rename/copy that depends on deleted ones.
127 * 2. deletions in the original.
128 * 3. rename/copy that depends on the pre-edit image of kept files.
129 * 4. additions, modifications and no-modifications in the original.
130 * 5. rename/copy that depends on the post-edit image of kept files
131 * (note that we currently do not detect such rename/copy).
133 * The downstream diffcore transformers are free to reorder
134 * the entries as long as they keep file pairs that has the
135 * same p->one->path in earlier rename_rank to appear before
136 * later ones. This ordering is used by the diff_flush()
137 * logic to tell renames from copies, and also used by the
138 * diffcore_prune() logic to omit unnecessary
139 * "no-modification" entries.
141 * To the final output routine, and in the diff-raw format
142 * output, a rename/copy that is based on a path that has a
143 * later entry that shares the same p->one->path and is not a
144 * deletion is a copy. Otherwise it is a rename.
147 struct diff_filepair *dp = diff_queue(outq, src, dst);
148 dp->rename_rank = rank * 2 + 1;
149 dp->score = score;
150 dst->xfrm_flags |= RENAME_DST_MATCHED;
153 #if 0
154 static void debug_filespec(struct diff_filespec *s, int x, const char *one)
156 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
157 x, one,
158 s->path,
159 DIFF_FILE_VALID(s) ? "valid" : "invalid",
160 s->mode,
161 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
162 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
163 x, one,
164 s->size, s->xfrm_flags);
167 static void debug_filepair(const struct diff_filepair *p, int i)
169 debug_filespec(p->one, i, "one");
170 debug_filespec(p->two, i, "two");
171 fprintf(stderr, "pair rank %d, orig order %d, score %d\n",
172 p->rename_rank, p->orig_order, p->score);
175 static void debug_queue(const char *msg, struct diff_queue_struct *q)
177 int i;
178 if (msg)
179 fprintf(stderr, "%s\n", msg);
180 fprintf(stderr, "q->nr = %d\n", q->nr);
181 for (i = 0; i < q->nr; i++) {
182 struct diff_filepair *p = q->queue[i];
183 debug_filepair(p, i);
186 #else
187 #define debug_queue(a,b) do { ; /*nothing*/ } while(0)
188 #endif
191 * We sort the outstanding diff entries according to the rank (see
192 * comment at the beginning of record_rename_pair) and tiebreak with
193 * the order in the original input.
195 static int rank_compare(const void *a_, const void *b_)
197 const struct diff_filepair *a = *(const struct diff_filepair **)a_;
198 const struct diff_filepair *b = *(const struct diff_filepair **)b_;
199 int a_rank = a->rename_rank;
200 int b_rank = b->rename_rank;
202 if (a_rank != b_rank)
203 return a_rank - b_rank;
204 return a->orig_order - b->orig_order;
208 * We sort the rename similarity matrix with the score, in descending
209 * order (more similar first).
211 static int score_compare(const void *a_, const void *b_)
213 const struct diff_score *a = a_, *b = b_;
214 return b->score - a->score;
217 int diff_scoreopt_parse(const char *opt)
219 int diglen, num, scale, i;
220 if (opt[0] != '-' || (opt[1] != 'M' && opt[1] != 'C'))
221 return -1; /* that is not a -M nor -C option */
222 diglen = strspn(opt+2, "0123456789");
223 if (diglen == 0 || strlen(opt+2) != diglen)
224 return 0; /* use default */
225 sscanf(opt+2, "%d", &num);
226 for (i = 0, scale = 1; i < diglen; i++)
227 scale *= 10;
229 /* user says num divided by scale and we say internally that
230 * is MAX_SCORE * num / scale.
232 return MAX_SCORE * num / scale;
235 void diffcore_rename(int detect_rename, int minimum_score)
237 struct diff_queue_struct *q = &diff_queued_diff;
238 struct diff_queue_struct outq;
239 struct diff_rename_pool created, deleted, stay;
240 struct diff_rename_pool *(srcs[2]);
241 struct diff_score *mx;
242 int h, i, j;
243 int num_create, num_src, dst_cnt, src_cnt;
245 if (!minimum_score)
246 minimum_score = DEFAULT_MINIMUM_SCORE;
247 outq.queue = NULL;
248 outq.nr = outq.alloc = 0;
250 diff_rename_pool_clear(&created);
251 diff_rename_pool_clear(&deleted);
252 diff_rename_pool_clear(&stay);
254 srcs[0] = &deleted;
255 srcs[1] = &stay;
257 for (i = 0; i < q->nr; i++) {
258 struct diff_filepair *p = q->queue[i];
259 if (!DIFF_FILE_VALID(p->one))
260 if (!DIFF_FILE_VALID(p->two))
261 continue; /* unmerged */
262 else
263 diff_rename_pool_add(&created, p->two);
264 else if (!DIFF_FILE_VALID(p->two))
265 diff_rename_pool_add(&deleted, p->one);
266 else if (1 < detect_rename) /* find copy, too */
267 diff_rename_pool_add(&stay, p->one);
269 if (created.nr == 0)
270 goto cleanup; /* nothing to do */
272 /* We really want to cull the candidates list early
273 * with cheap tests in order to avoid doing deltas.
275 * With the current callers, we should not have already
276 * matched entries at this point, but it is nonetheless
277 * checked for sanity.
279 for (i = 0; i < created.nr; i++) {
280 if (created.s[i]->xfrm_flags & RENAME_DST_MATCHED)
281 continue; /* we have matched exactly already */
282 for (h = 0; h < sizeof(srcs)/sizeof(srcs[0]); h++) {
283 struct diff_rename_pool *p = srcs[h];
284 for (j = 0; j < p->nr; j++) {
285 if (!is_exact_match(p->s[j], created.s[i]))
286 continue;
287 record_rename_pair(&outq,
288 p->s[j], created.s[i], h,
289 MAX_SCORE);
290 break; /* we are done with this entry */
294 debug_queue("done detecting exact", &outq);
296 /* Have we run out the created file pool? If so we can avoid
297 * doing the delta matrix altogether.
299 if (outq.nr == created.nr)
300 goto flush_rest;
302 num_create = (created.nr - outq.nr);
303 num_src = deleted.nr + stay.nr;
304 mx = xmalloc(sizeof(*mx) * num_create * num_src);
305 for (dst_cnt = i = 0; i < created.nr; i++) {
306 int base = dst_cnt * num_src;
307 if (created.s[i]->xfrm_flags & RENAME_DST_MATCHED)
308 continue; /* dealt with exact match already. */
309 for (src_cnt = h = 0; h < sizeof(srcs)/sizeof(srcs[0]); h++) {
310 struct diff_rename_pool *p = srcs[h];
311 for (j = 0; j < p->nr; j++, src_cnt++) {
312 struct diff_score *m = &mx[base + src_cnt];
313 m->src = p->s[j];
314 m->dst = created.s[i];
315 m->score = estimate_similarity(m->src, m->dst,
316 minimum_score);
317 m->rank = h;
320 dst_cnt++;
322 /* cost matrix sorted by most to least similar pair */
323 qsort(mx, num_create * num_src, sizeof(*mx), score_compare);
324 for (i = 0; i < num_create * num_src; i++) {
325 if (mx[i].dst->xfrm_flags & RENAME_DST_MATCHED)
326 continue; /* alreayd done, either exact or fuzzy. */
327 if (mx[i].score < minimum_score)
328 break; /* there is not any more diffs applicable. */
329 record_rename_pair(&outq,
330 mx[i].src, mx[i].dst, mx[i].rank,
331 mx[i].score);
333 free(mx);
334 debug_queue("done detecting fuzzy", &outq);
336 flush_rest:
337 /* At this point, we have found some renames and copies and they
338 * are kept in outq. The original list is still in *q.
340 * Scan the original list and move them into the outq; we will sort
341 * outq and swap it into the queue supplied to pass that to
342 * downstream, so we assign the sort keys in this loop.
344 * See comments at the top of record_rename_pair for numbers used
345 * to assign rename_rank.
347 for (i = 0; i < q->nr; i++) {
348 struct diff_filepair *dp, *p = q->queue[i];
349 if (!DIFF_FILE_VALID(p->one)) {
350 /* creation or unmerged entries */
351 dp = diff_queue(&outq, p->one, p->two);
352 dp->rename_rank = 4;
354 else if (!DIFF_FILE_VALID(p->two)) {
355 /* deletion */
356 dp = diff_queue(&outq, p->one, p->two);
357 dp->rename_rank = 2;
359 else {
360 /* modification, or stay as is */
361 dp = diff_queue(&outq, p->one, p->two);
362 dp->rename_rank = 4;
364 free(p);
366 debug_queue("done copying original", &outq);
368 /* Sort outq */
369 qsort(outq.queue, outq.nr, sizeof(outq.queue[0]), rank_compare);
371 debug_queue("done sorting", &outq);
373 free(q->queue);
374 q->nr = q->alloc = 0;
375 q->queue = NULL;
377 /* Copy it out to q, removing duplicates. */
378 for (i = 0; i < outq.nr; i++) {
379 struct diff_filepair *p = outq.queue[i];
380 if (!DIFF_FILE_VALID(p->one)) {
381 /* created or unmerged */
382 if (p->two->xfrm_flags & RENAME_DST_MATCHED)
383 ; /* rename/copy created it already */
384 else
385 diff_queue(q, p->one, p->two);
387 else if (!DIFF_FILE_VALID(p->two)) {
388 /* deleted */
389 if (p->one->xfrm_flags & RENAME_SRC_GONE)
390 ; /* rename/copy deleted it already */
391 else
392 diff_queue(q, p->one, p->two);
394 else if (strcmp(p->one->path, p->two->path)) {
395 /* rename or copy */
396 struct diff_filepair *dp =
397 diff_queue(q, p->one, p->two);
398 dp->score = p->score;
400 /* if we have a later entry that is a rename/copy
401 * that depends on p->one, then we copy here.
402 * otherwise we rename it.
404 if (!diff_needs_to_stay(&outq, i+1, p->one))
405 /* this is the last one, so mark it as gone.
407 p->one->xfrm_flags |= RENAME_SRC_GONE;
409 else
410 /* otherwise it is a modified (or "stay") entry */
411 diff_queue(q, p->one, p->two);
412 free(p);
415 free(outq.queue);
416 debug_queue("done collapsing", q);
418 cleanup:
419 free(created.s);
420 free(deleted.s);
421 free(stay.s);
422 return;