Merge branch 'nd/sparse'
[git/git.git] / diff-lib.c
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "commit.h"
7 #include "diff.h"
8 #include "diffcore.h"
9 #include "revision.h"
10 #include "cache-tree.h"
11 #include "unpack-trees.h"
12 #include "refs.h"
13
14 /*
15 * diff-files
16 */
17
18 /*
19 * Has the work tree entity been removed?
20 *
21 * Return 1 if it was removed from the work tree, 0 if an entity to be
22 * compared with the cache entry ce still exists (the latter includes
23 * the case where a directory that is not a submodule repository
24 * exists for ce that is a submodule -- it is a submodule that is not
25 * checked out). Return negative for an error.
26 */
27 static int check_removed(const struct cache_entry *ce, struct stat *st)
28 {
29 if (lstat(ce->name, st) < 0) {
30 if (errno != ENOENT && errno != ENOTDIR)
31 return -1;
32 return 1;
33 }
34 if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
35 return 1;
36 if (S_ISDIR(st->st_mode)) {
37 unsigned char sub[20];
38
39 /*
40 * If ce is already a gitlink, we can have a plain
41 * directory (i.e. the submodule is not checked out),
42 * or a checked out submodule. Either case this is not
43 * a case where something was removed from the work tree,
44 * so we will return 0.
45 *
46 * Otherwise, if the directory is not a submodule
47 * repository, that means ce which was a blob turned into
48 * a directory --- the blob was removed!
49 */
50 if (!S_ISGITLINK(ce->ce_mode) &&
51 resolve_gitlink_ref(ce->name, "HEAD", sub))
52 return 1;
53 }
54 return 0;
55 }
56
57 int run_diff_files(struct rev_info *revs, unsigned int option)
58 {
59 int entries, i;
60 int diff_unmerged_stage = revs->max_count;
61 int silent_on_removed = option & DIFF_SILENT_ON_REMOVED;
62 unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
63 ? CE_MATCH_RACY_IS_DIRTY : 0);
64
65 diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
66
67 if (diff_unmerged_stage < 0)
68 diff_unmerged_stage = 2;
69 entries = active_nr;
70 for (i = 0; i < entries; i++) {
71 struct stat st;
72 unsigned int oldmode, newmode;
73 struct cache_entry *ce = active_cache[i];
74 int changed;
75
76 if (DIFF_OPT_TST(&revs->diffopt, QUICK) &&
77 DIFF_OPT_TST(&revs->diffopt, HAS_CHANGES))
78 break;
79
80 if (!ce_path_match(ce, revs->prune_data))
81 continue;
82
83 if (ce_stage(ce)) {
84 struct combine_diff_path *dpath;
85 int num_compare_stages = 0;
86 size_t path_len;
87
88 path_len = ce_namelen(ce);
89
90 dpath = xmalloc(combine_diff_path_size(5, path_len));
91 dpath->path = (char *) &(dpath->parent[5]);
92
93 dpath->next = NULL;
94 dpath->len = path_len;
95 memcpy(dpath->path, ce->name, path_len);
96 dpath->path[path_len] = '\0';
97 hashclr(dpath->sha1);
98 memset(&(dpath->parent[0]), 0,
99 sizeof(struct combine_diff_parent)*5);
100
101 changed = check_removed(ce, &st);
102 if (!changed)
103 dpath->mode = ce_mode_from_stat(ce, st.st_mode);
104 else {
105 if (changed < 0) {
106 perror(ce->name);
107 continue;
108 }
109 if (silent_on_removed)
110 continue;
111 }
112
113 while (i < entries) {
114 struct cache_entry *nce = active_cache[i];
115 int stage;
116
117 if (strcmp(ce->name, nce->name))
118 break;
119
120 /* Stage #2 (ours) is the first parent,
121 * stage #3 (theirs) is the second.
122 */
123 stage = ce_stage(nce);
124 if (2 <= stage) {
125 int mode = nce->ce_mode;
126 num_compare_stages++;
127 hashcpy(dpath->parent[stage-2].sha1, nce->sha1);
128 dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
129 dpath->parent[stage-2].status =
130 DIFF_STATUS_MODIFIED;
131 }
132
133 /* diff against the proper unmerged stage */
134 if (stage == diff_unmerged_stage)
135 ce = nce;
136 i++;
137 }
138 /*
139 * Compensate for loop update
140 */
141 i--;
142
143 if (revs->combine_merges && num_compare_stages == 2) {
144 show_combined_diff(dpath, 2,
145 revs->dense_combined_merges,
146 revs);
147 free(dpath);
148 continue;
149 }
150 free(dpath);
151 dpath = NULL;
152
153 /*
154 * Show the diff for the 'ce' if we found the one
155 * from the desired stage.
156 */
157 diff_unmerge(&revs->diffopt, ce->name, 0, null_sha1);
158 if (ce_stage(ce) != diff_unmerged_stage)
159 continue;
160 }
161
162 if (ce_uptodate(ce) || ce_skip_worktree(ce))
163 continue;
164
165 /* If CE_VALID is set, don't look at workdir for file removal */
166 changed = (ce->ce_flags & CE_VALID) ? 0 : check_removed(ce, &st);
167 if (changed) {
168 if (changed < 0) {
169 perror(ce->name);
170 continue;
171 }
172 if (silent_on_removed)
173 continue;
174 diff_addremove(&revs->diffopt, '-', ce->ce_mode,
175 ce->sha1, ce->name);
176 continue;
177 }
178 changed = ce_match_stat(ce, &st, ce_option);
179 if (!changed) {
180 ce_mark_uptodate(ce);
181 if (!DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
182 continue;
183 }
184 oldmode = ce->ce_mode;
185 newmode = ce_mode_from_stat(ce, st.st_mode);
186 diff_change(&revs->diffopt, oldmode, newmode,
187 ce->sha1, (changed ? null_sha1 : ce->sha1),
188 ce->name);
189
190 }
191 diffcore_std(&revs->diffopt);
192 diff_flush(&revs->diffopt);
193 return 0;
194 }
195
196 /*
197 * diff-index
198 */
199
200 /* A file entry went away or appeared */
201 static void diff_index_show_file(struct rev_info *revs,
202 const char *prefix,
203 struct cache_entry *ce,
204 const unsigned char *sha1, unsigned int mode)
205 {
206 diff_addremove(&revs->diffopt, prefix[0], mode,
207 sha1, ce->name);
208 }
209
210 static int get_stat_data(struct cache_entry *ce,
211 const unsigned char **sha1p,
212 unsigned int *modep,
213 int cached, int match_missing)
214 {
215 const unsigned char *sha1 = ce->sha1;
216 unsigned int mode = ce->ce_mode;
217
218 if (!cached && !ce_uptodate(ce)) {
219 int changed;
220 struct stat st;
221 changed = check_removed(ce, &st);
222 if (changed < 0)
223 return -1;
224 else if (changed) {
225 if (match_missing) {
226 *sha1p = sha1;
227 *modep = mode;
228 return 0;
229 }
230 return -1;
231 }
232 changed = ce_match_stat(ce, &st, 0);
233 if (changed) {
234 mode = ce_mode_from_stat(ce, st.st_mode);
235 sha1 = null_sha1;
236 }
237 }
238
239 *sha1p = sha1;
240 *modep = mode;
241 return 0;
242 }
243
244 static void show_new_file(struct rev_info *revs,
245 struct cache_entry *new,
246 int cached, int match_missing)
247 {
248 const unsigned char *sha1;
249 unsigned int mode;
250
251 /*
252 * New file in the index: it might actually be different in
253 * the working copy.
254 */
255 if (get_stat_data(new, &sha1, &mode, cached, match_missing) < 0)
256 return;
257
258 diff_index_show_file(revs, "+", new, sha1, mode);
259 }
260
261 static int show_modified(struct rev_info *revs,
262 struct cache_entry *old,
263 struct cache_entry *new,
264 int report_missing,
265 int cached, int match_missing)
266 {
267 unsigned int mode, oldmode;
268 const unsigned char *sha1;
269
270 if (get_stat_data(new, &sha1, &mode, cached, match_missing) < 0) {
271 if (report_missing)
272 diff_index_show_file(revs, "-", old,
273 old->sha1, old->ce_mode);
274 return -1;
275 }
276
277 if (revs->combine_merges && !cached &&
278 (hashcmp(sha1, old->sha1) || hashcmp(old->sha1, new->sha1))) {
279 struct combine_diff_path *p;
280 int pathlen = ce_namelen(new);
281
282 p = xmalloc(combine_diff_path_size(2, pathlen));
283 p->path = (char *) &p->parent[2];
284 p->next = NULL;
285 p->len = pathlen;
286 memcpy(p->path, new->name, pathlen);
287 p->path[pathlen] = 0;
288 p->mode = mode;
289 hashclr(p->sha1);
290 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
291 p->parent[0].status = DIFF_STATUS_MODIFIED;
292 p->parent[0].mode = new->ce_mode;
293 hashcpy(p->parent[0].sha1, new->sha1);
294 p->parent[1].status = DIFF_STATUS_MODIFIED;
295 p->parent[1].mode = old->ce_mode;
296 hashcpy(p->parent[1].sha1, old->sha1);
297 show_combined_diff(p, 2, revs->dense_combined_merges, revs);
298 free(p);
299 return 0;
300 }
301
302 oldmode = old->ce_mode;
303 if (mode == oldmode && !hashcmp(sha1, old->sha1) &&
304 !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
305 return 0;
306
307 diff_change(&revs->diffopt, oldmode, mode,
308 old->sha1, sha1, old->name);
309 return 0;
310 }
311
312 /*
313 * This gets a mix of an existing index and a tree, one pathname entry
314 * at a time. The index entry may be a single stage-0 one, but it could
315 * also be multiple unmerged entries (in which case idx_pos/idx_nr will
316 * give you the position and number of entries in the index).
317 */
318 static void do_oneway_diff(struct unpack_trees_options *o,
319 struct cache_entry *idx,
320 struct cache_entry *tree)
321 {
322 struct rev_info *revs = o->unpack_data;
323 int match_missing, cached;
324
325 /* if the entry is not checked out, don't examine work tree */
326 cached = o->index_only ||
327 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
328 /*
329 * Backward compatibility wart - "diff-index -m" does
330 * not mean "do not ignore merges", but "match_missing".
331 *
332 * But with the revision flag parsing, that's found in
333 * "!revs->ignore_merges".
334 */
335 match_missing = !revs->ignore_merges;
336
337 if (cached && idx && ce_stage(idx)) {
338 diff_unmerge(&revs->diffopt, idx->name, idx->ce_mode,
339 idx->sha1);
340 return;
341 }
342
343 /*
344 * Something added to the tree?
345 */
346 if (!tree) {
347 show_new_file(revs, idx, cached, match_missing);
348 return;
349 }
350
351 /*
352 * Something removed from the tree?
353 */
354 if (!idx) {
355 diff_index_show_file(revs, "-", tree, tree->sha1, tree->ce_mode);
356 return;
357 }
358
359 /* Show difference between old and new */
360 show_modified(revs, tree, idx, 1, cached, match_missing);
361 }
362
363 static inline void skip_same_name(struct cache_entry *ce, struct unpack_trees_options *o)
364 {
365 int len = ce_namelen(ce);
366 const struct index_state *index = o->src_index;
367
368 while (o->pos < index->cache_nr) {
369 struct cache_entry *next = index->cache[o->pos];
370 if (len != ce_namelen(next))
371 break;
372 if (memcmp(ce->name, next->name, len))
373 break;
374 o->pos++;
375 }
376 }
377
378 /*
379 * The unpack_trees() interface is designed for merging, so
380 * the different source entries are designed primarily for
381 * the source trees, with the old index being really mainly
382 * used for being replaced by the result.
383 *
384 * For diffing, the index is more important, and we only have a
385 * single tree.
386 *
387 * We're supposed to advance o->pos to skip what we have already processed.
388 *
389 * This wrapper makes it all more readable, and takes care of all
390 * the fairly complex unpack_trees() semantic requirements, including
391 * the skipping, the path matching, the type conflict cases etc.
392 */
393 static int oneway_diff(struct cache_entry **src, struct unpack_trees_options *o)
394 {
395 struct cache_entry *idx = src[0];
396 struct cache_entry *tree = src[1];
397 struct rev_info *revs = o->unpack_data;
398
399 if (idx && ce_stage(idx))
400 skip_same_name(idx, o);
401
402 /*
403 * Unpack-trees generates a DF/conflict entry if
404 * there was a directory in the index and a tree
405 * in the tree. From a diff standpoint, that's a
406 * delete of the tree and a create of the file.
407 */
408 if (tree == o->df_conflict_entry)
409 tree = NULL;
410
411 if (ce_path_match(idx ? idx : tree, revs->prune_data))
412 do_oneway_diff(o, idx, tree);
413
414 return 0;
415 }
416
417 int run_diff_index(struct rev_info *revs, int cached)
418 {
419 struct object *ent;
420 struct tree *tree;
421 const char *tree_name;
422 struct unpack_trees_options opts;
423 struct tree_desc t;
424
425 ent = revs->pending.objects[0].item;
426 tree_name = revs->pending.objects[0].name;
427 tree = parse_tree_indirect(ent->sha1);
428 if (!tree)
429 return error("bad tree object %s", tree_name);
430
431 memset(&opts, 0, sizeof(opts));
432 opts.head_idx = 1;
433 opts.index_only = cached;
434 opts.diff_index_cached = (cached &&
435 !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER));
436 opts.merge = 1;
437 opts.fn = oneway_diff;
438 opts.unpack_data = revs;
439 opts.src_index = &the_index;
440 opts.dst_index = NULL;
441
442 init_tree_desc(&t, tree->buffer, tree->size);
443 if (unpack_trees(1, &t, &opts))
444 exit(128);
445
446 diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
447 diffcore_std(&revs->diffopt);
448 diff_flush(&revs->diffopt);
449 return 0;
450 }
451
452 int do_diff_cache(const unsigned char *tree_sha1, struct diff_options *opt)
453 {
454 struct tree *tree;
455 struct rev_info revs;
456 int i;
457 struct cache_entry **dst;
458 struct cache_entry *last = NULL;
459 struct unpack_trees_options opts;
460 struct tree_desc t;
461
462 /*
463 * This is used by git-blame to run diff-cache internally;
464 * it potentially needs to repeatedly run this, so we will
465 * start by removing the higher order entries the last round
466 * left behind.
467 */
468 dst = active_cache;
469 for (i = 0; i < active_nr; i++) {
470 struct cache_entry *ce = active_cache[i];
471 if (ce_stage(ce)) {
472 if (last && !strcmp(ce->name, last->name))
473 continue;
474 cache_tree_invalidate_path(active_cache_tree,
475 ce->name);
476 last = ce;
477 ce->ce_flags |= CE_REMOVE;
478 }
479 *dst++ = ce;
480 }
481 active_nr = dst - active_cache;
482
483 init_revisions(&revs, NULL);
484 revs.prune_data = opt->paths;
485 tree = parse_tree_indirect(tree_sha1);
486 if (!tree)
487 die("bad tree object %s", sha1_to_hex(tree_sha1));
488
489 memset(&opts, 0, sizeof(opts));
490 opts.head_idx = 1;
491 opts.index_only = 1;
492 opts.diff_index_cached = !DIFF_OPT_TST(opt, FIND_COPIES_HARDER);
493 opts.merge = 1;
494 opts.fn = oneway_diff;
495 opts.unpack_data = &revs;
496 opts.src_index = &the_index;
497 opts.dst_index = &the_index;
498
499 init_tree_desc(&t, tree->buffer, tree->size);
500 if (unpack_trees(1, &t, &opts))
501 exit(128);
502 return 0;
503 }
504
505 int index_differs_from(const char *def, int diff_flags)
506 {
507 struct rev_info rev;
508
509 init_revisions(&rev, NULL);
510 setup_revisions(0, NULL, &rev, def);
511 DIFF_OPT_SET(&rev.diffopt, QUICK);
512 DIFF_OPT_SET(&rev.diffopt, EXIT_WITH_STATUS);
513 rev.diffopt.flags |= diff_flags;
514 run_diff_index(&rev, 1);
515 if (rev.pending.alloc)
516 free(rev.pending.objects);
517 return (DIFF_OPT_TST(&rev.diffopt, HAS_CHANGES) != 0);
518 }