2 * Copyright (C) 2005 Junio C Hamano
8 #include "count-delta.h"
10 /* Table of rename/copy destinations */
12 static struct diff_rename_dst
{
13 struct diff_filespec
*two
;
14 struct diff_filepair
*pair
;
16 static int rename_dst_nr
, rename_dst_alloc
;
18 static struct diff_rename_dst
*locate_rename_dst(struct diff_filespec
*two
,
25 while (last
> first
) {
26 int next
= (last
+ first
) >> 1;
27 struct diff_rename_dst
*dst
= &(rename_dst
[next
]);
28 int cmp
= strcmp(two
->path
, dst
->two
->path
);
40 /* insert to make it at "first" */
41 if (rename_dst_alloc
<= rename_dst_nr
) {
42 rename_dst_alloc
= alloc_nr(rename_dst_alloc
);
43 rename_dst
= xrealloc(rename_dst
,
44 rename_dst_alloc
* sizeof(*rename_dst
));
47 if (first
< rename_dst_nr
)
48 memmove(rename_dst
+ first
+ 1, rename_dst
+ first
,
49 (rename_dst_nr
- first
- 1) * sizeof(*rename_dst
));
50 rename_dst
[first
].two
= two
;
51 rename_dst
[first
].pair
= NULL
;
52 return &(rename_dst
[first
]);
55 static struct diff_rename_src
{
56 struct diff_filespec
*one
;
57 unsigned src_used
: 1;
59 static int rename_src_nr
, rename_src_alloc
;
61 static struct diff_rename_src
*locate_rename_src(struct diff_filespec
*one
,
68 while (last
> first
) {
69 int next
= (last
+ first
) >> 1;
70 struct diff_rename_src
*src
= &(rename_src
[next
]);
71 int cmp
= strcmp(one
->path
, src
->one
->path
);
83 /* insert to make it at "first" */
84 if (rename_src_alloc
<= rename_src_nr
) {
85 rename_src_alloc
= alloc_nr(rename_src_alloc
);
86 rename_src
= xrealloc(rename_src
,
87 rename_src_alloc
* sizeof(*rename_src
));
90 if (first
< rename_src_nr
)
91 memmove(rename_src
+ first
+ 1, rename_src
+ first
,
92 (rename_src_nr
- first
- 1) * sizeof(*rename_src
));
93 rename_src
[first
].one
= one
;
94 rename_src
[first
].src_used
= 0;
95 return &(rename_src
[first
]);
98 static int is_exact_match(struct diff_filespec
*src
, struct diff_filespec
*dst
)
100 if (src
->sha1_valid
&& dst
->sha1_valid
&&
101 !memcmp(src
->sha1
, dst
->sha1
, 20))
103 if (diff_populate_filespec(src
) || diff_populate_filespec(dst
))
104 /* this is an error but will be caught downstream */
106 if (src
->size
== dst
->size
&&
107 !memcmp(src
->data
, dst
->data
, src
->size
))
113 int src
; /* index in rename_src */
114 int dst
; /* index in rename_dst */
118 static int estimate_similarity(struct diff_filespec
*src
,
119 struct diff_filespec
*dst
,
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.
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
134 unsigned long delta_size
, base_size
;
137 /* We deal only with regular files. Symlink renames are handled
138 * only when they are exact matches --- in other words, no edits
141 if (!S_ISREG(src
->mode
) || !S_ISREG(dst
->mode
))
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
);
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.
155 if (base_size
* (MAX_SCORE
-minimum_score
) < delta_size
* MAX_SCORE
)
158 delta
= diff_delta(src
->data
, src
->size
,
159 dst
->data
, dst
->size
,
162 /* A delta that has a lot of literal additions would have
163 * big delta_size no matter what else it does.
165 if (base_size
* (MAX_SCORE
-minimum_score
) < delta_size
* MAX_SCORE
)
168 /* Estimate the edit size by interpreting delta. */
169 delta_size
= count_delta(delta
, delta_size
);
171 if (delta_size
== UINT_MAX
)
175 * Now we will give some score to it. 100% edit gets 0 points
176 * and 0% edit gets MAX_SCORE points.
178 score
= MAX_SCORE
- (MAX_SCORE
* delta_size
/ base_size
);
179 if (score
< 0) return 0;
180 if (MAX_SCORE
< score
) return MAX_SCORE
;
184 static void record_rename_pair(struct diff_queue_struct
*renq
,
185 int dst_index
, int src_index
, int score
)
187 struct diff_filespec
*one
, *two
, *src
, *dst
;
188 struct diff_filepair
*dp
;
190 if (rename_dst
[dst_index
].pair
)
191 die("internal error: dst already matched.");
193 src
= rename_src
[src_index
].one
;
194 one
= alloc_filespec(src
->path
);
195 fill_filespec(one
, src
->sha1
, src
->mode
);
197 dst
= rename_dst
[dst_index
].two
;
198 two
= alloc_filespec(dst
->path
);
199 fill_filespec(two
, dst
->sha1
, dst
->mode
);
201 dp
= diff_queue(renq
, one
, two
);
204 rename_src
[src_index
].src_used
= 1;
205 rename_dst
[dst_index
].pair
= dp
;
209 * We sort the rename similarity matrix with the score, in descending
210 * order (more similar first).
212 static int score_compare(const void *a_
, const void *b_
)
214 const struct diff_score
*a
= a_
, *b
= b_
;
215 return b
->score
- a
->score
;
218 int diff_scoreopt_parse(const char *opt
)
220 int diglen
, num
, scale
, i
;
221 if (opt
[0] != '-' || (opt
[1] != 'M' && opt
[1] != 'C'))
222 return -1; /* that is not a -M nor -C option */
223 diglen
= strspn(opt
+2, "0123456789");
224 if (diglen
== 0 || strlen(opt
+2) != diglen
)
225 return 0; /* use default */
226 sscanf(opt
+2, "%d", &num
);
227 for (i
= 0, scale
= 1; i
< diglen
; i
++)
230 /* user says num divided by scale and we say internally that
231 * is MAX_SCORE * num / scale.
233 return MAX_SCORE
* num
/ scale
;
236 void diffcore_rename(int detect_rename
, int minimum_score
)
238 struct diff_queue_struct
*q
= &diff_queued_diff
;
239 struct diff_queue_struct renq
, outq
;
240 struct diff_score
*mx
;
242 int num_create
, num_src
, dst_cnt
;
245 minimum_score
= DEFAULT_MINIMUM_SCORE
;
247 renq
.nr
= renq
.alloc
= 0;
249 for (i
= 0; i
< q
->nr
; i
++) {
250 struct diff_filepair
*p
= q
->queue
[i
];
251 if (!DIFF_FILE_VALID(p
->one
))
252 if (!DIFF_FILE_VALID(p
->two
))
253 continue; /* unmerged */
255 locate_rename_dst(p
->two
, 1);
256 else if (!DIFF_FILE_VALID(p
->two
))
257 locate_rename_src(p
->one
, 1);
258 else if (1 < detect_rename
) /* find copy, too */
259 locate_rename_src(p
->one
, 1);
261 if (rename_dst_nr
== 0)
262 goto cleanup
; /* nothing to do */
264 /* We really want to cull the candidates list early
265 * with cheap tests in order to avoid doing deltas.
267 for (i
= 0; i
< rename_dst_nr
; i
++) {
268 struct diff_filespec
*two
= rename_dst
[i
].two
;
269 for (j
= 0; j
< rename_src_nr
; j
++) {
270 struct diff_filespec
*one
= rename_src
[j
].one
;
271 if (!is_exact_match(one
, two
))
273 record_rename_pair(&renq
, i
, j
, MAX_SCORE
);
274 break; /* we are done with this entry */
277 diff_debug_queue("done detecting exact", &renq
);
279 /* Have we run out the created file pool? If so we can avoid
280 * doing the delta matrix altogether.
282 if (renq
.nr
== rename_dst_nr
)
285 num_create
= (rename_dst_nr
- renq
.nr
);
286 num_src
= rename_src_nr
;
287 mx
= xmalloc(sizeof(*mx
) * num_create
* num_src
);
288 for (dst_cnt
= i
= 0; i
< rename_dst_nr
; i
++) {
289 int base
= dst_cnt
* num_src
;
290 struct diff_filespec
*two
= rename_dst
[i
].two
;
291 if (rename_dst
[i
].pair
)
292 continue; /* dealt with exact match already. */
293 for (j
= 0; j
< rename_src_nr
; j
++) {
294 struct diff_filespec
*one
= rename_src
[j
].one
;
295 struct diff_score
*m
= &mx
[base
+j
];
298 m
->score
= estimate_similarity(one
, two
,
303 /* cost matrix sorted by most to least similar pair */
304 qsort(mx
, num_create
* num_src
, sizeof(*mx
), score_compare
);
305 for (i
= 0; i
< num_create
* num_src
; i
++) {
306 struct diff_rename_dst
*dst
= &rename_dst
[mx
[i
].dst
];
308 continue; /* already done, either exact or fuzzy. */
309 if (mx
[i
].score
< minimum_score
)
310 break; /* there is not any more diffs applicable. */
311 record_rename_pair(&renq
, mx
[i
].dst
, mx
[i
].src
, mx
[i
].score
);
314 diff_debug_queue("done detecting fuzzy", &renq
);
317 /* At this point, we have found some renames and copies and they
318 * are kept in renq. The original list is still in *q.
320 * Scan the original list and move them into the outq; we will sort
321 * outq and swap it into the queue supplied to pass that to
322 * downstream, so we assign the sort keys in this loop.
324 * See comments at the top of record_rename_pair for numbers used
325 * to assign rename_rank.
328 outq
.nr
= outq
.alloc
= 0;
329 for (i
= 0; i
< q
->nr
; i
++) {
330 struct diff_filepair
*p
= q
->queue
[i
];
331 struct diff_rename_src
*src
= locate_rename_src(p
->one
, 0);
332 struct diff_rename_dst
*dst
= locate_rename_dst(p
->two
, 0);
333 struct diff_filepair
*pair_to_free
= NULL
;
338 /* renq has rename/copy already to produce
339 * this file, so we do not emit the creation
340 * record in the output.
342 diff_q(&outq
, dst
->pair
);
346 /* no matching rename/copy source, so record
347 * this as a creation.
351 else if (!diff_unmodified_pair(p
))
352 /* all the other cases need to be recorded as is */
355 /* unmodified pair needs to be recorded only if
356 * it is used as the source of rename/copy
358 if (src
&& src
->src_used
)
364 diff_free_filepair(pair_to_free
);
366 diff_debug_queue("done copying original", &outq
);
371 diff_debug_queue("done collapsing", q
);
376 rename_dst_nr
= rename_dst_alloc
= 0;
379 rename_src_nr
= rename_src_alloc
= 0;