Documentation/technical: describe bitmap lookup table extension
[git/debian.git] / diff-lib.c
blob7eb66a417aa08daa6f22753eba5c52f32361dfc8
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 #include "submodule.h"
14 #include "dir.h"
15 #include "fsmonitor.h"
16 #include "commit-reach.h"
19 * diff-files
23 * Has the work tree entity been removed?
25 * Return 1 if it was removed from the work tree, 0 if an entity to be
26 * compared with the cache entry ce still exists (the latter includes
27 * the case where a directory that is not a submodule repository
28 * exists for ce that is a submodule -- it is a submodule that is not
29 * checked out). Return negative for an error.
31 static int check_removed(const struct index_state *istate, const struct cache_entry *ce, struct stat *st)
33 assert(is_fsmonitor_refreshed(istate));
34 if (!(ce->ce_flags & CE_FSMONITOR_VALID) && lstat(ce->name, st) < 0) {
35 if (!is_missing_file_error(errno))
36 return -1;
37 return 1;
39 if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
40 return 1;
41 if (S_ISDIR(st->st_mode)) {
42 struct object_id sub;
45 * If ce is already a gitlink, we can have a plain
46 * directory (i.e. the submodule is not checked out),
47 * or a checked out submodule. Either case this is not
48 * a case where something was removed from the work tree,
49 * so we will return 0.
51 * Otherwise, if the directory is not a submodule
52 * repository, that means ce which was a blob turned into
53 * a directory --- the blob was removed!
55 if (!S_ISGITLINK(ce->ce_mode) &&
56 resolve_gitlink_ref(ce->name, "HEAD", &sub))
57 return 1;
59 return 0;
63 * Has a file changed or has a submodule new commits or a dirty work tree?
65 * Return 1 when changes are detected, 0 otherwise. If the DIRTY_SUBMODULES
66 * option is set, the caller does not only want to know if a submodule is
67 * modified at all but wants to know all the conditions that are met (new
68 * commits, untracked content and/or modified content).
70 static int match_stat_with_submodule(struct diff_options *diffopt,
71 const struct cache_entry *ce,
72 struct stat *st, unsigned ce_option,
73 unsigned *dirty_submodule)
75 int changed = ie_match_stat(diffopt->repo->index, ce, st, ce_option);
76 if (S_ISGITLINK(ce->ce_mode)) {
77 struct diff_flags orig_flags = diffopt->flags;
78 if (!diffopt->flags.override_submodule_config)
79 set_diffopt_flags_from_submodule_config(diffopt, ce->name);
80 if (diffopt->flags.ignore_submodules)
81 changed = 0;
82 else if (!diffopt->flags.ignore_dirty_submodules &&
83 (!changed || diffopt->flags.dirty_submodules))
84 *dirty_submodule = is_submodule_modified(ce->name,
85 diffopt->flags.ignore_untracked_in_submodules);
86 diffopt->flags = orig_flags;
88 return changed;
91 int run_diff_files(struct rev_info *revs, unsigned int option)
93 int entries, i;
94 int diff_unmerged_stage = revs->max_count;
95 unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
96 ? CE_MATCH_RACY_IS_DIRTY : 0);
97 uint64_t start = getnanotime();
98 struct index_state *istate = revs->diffopt.repo->index;
100 diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
102 refresh_fsmonitor(istate);
104 if (diff_unmerged_stage < 0)
105 diff_unmerged_stage = 2;
106 entries = istate->cache_nr;
107 for (i = 0; i < entries; i++) {
108 unsigned int oldmode, newmode;
109 struct cache_entry *ce = istate->cache[i];
110 int changed;
111 unsigned dirty_submodule = 0;
112 const struct object_id *old_oid, *new_oid;
114 if (diff_can_quit_early(&revs->diffopt))
115 break;
117 if (!ce_path_match(istate, ce, &revs->prune_data, NULL))
118 continue;
120 if (revs->diffopt.prefix &&
121 strncmp(ce->name, revs->diffopt.prefix, revs->diffopt.prefix_length))
122 continue;
124 if (ce_stage(ce)) {
125 struct combine_diff_path *dpath;
126 struct diff_filepair *pair;
127 unsigned int wt_mode = 0;
128 int num_compare_stages = 0;
129 size_t path_len;
130 struct stat st;
132 path_len = ce_namelen(ce);
134 dpath = xmalloc(combine_diff_path_size(5, path_len));
135 dpath->path = (char *) &(dpath->parent[5]);
137 dpath->next = NULL;
138 memcpy(dpath->path, ce->name, path_len);
139 dpath->path[path_len] = '\0';
140 oidclr(&dpath->oid);
141 memset(&(dpath->parent[0]), 0,
142 sizeof(struct combine_diff_parent)*5);
144 changed = check_removed(istate, ce, &st);
145 if (!changed)
146 wt_mode = ce_mode_from_stat(ce, st.st_mode);
147 else {
148 if (changed < 0) {
149 perror(ce->name);
150 continue;
152 wt_mode = 0;
154 dpath->mode = wt_mode;
156 while (i < entries) {
157 struct cache_entry *nce = istate->cache[i];
158 int stage;
160 if (strcmp(ce->name, nce->name))
161 break;
163 /* Stage #2 (ours) is the first parent,
164 * stage #3 (theirs) is the second.
166 stage = ce_stage(nce);
167 if (2 <= stage) {
168 int mode = nce->ce_mode;
169 num_compare_stages++;
170 oidcpy(&dpath->parent[stage - 2].oid,
171 &nce->oid);
172 dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
173 dpath->parent[stage-2].status =
174 DIFF_STATUS_MODIFIED;
177 /* diff against the proper unmerged stage */
178 if (stage == diff_unmerged_stage)
179 ce = nce;
180 i++;
183 * Compensate for loop update
185 i--;
187 if (revs->combine_merges && num_compare_stages == 2) {
188 show_combined_diff(dpath, 2, revs);
189 free(dpath);
190 continue;
192 FREE_AND_NULL(dpath);
195 * Show the diff for the 'ce' if we found the one
196 * from the desired stage.
198 pair = diff_unmerge(&revs->diffopt, ce->name);
199 if (wt_mode)
200 pair->two->mode = wt_mode;
201 if (ce_stage(ce) != diff_unmerged_stage)
202 continue;
205 if (ce_uptodate(ce) || ce_skip_worktree(ce))
206 continue;
209 * When CE_VALID is set (via "update-index --assume-unchanged"
210 * or via adding paths while core.ignorestat is set to true),
211 * the user has promised that the working tree file for that
212 * path will not be modified. When CE_FSMONITOR_VALID is true,
213 * the fsmonitor knows that the path hasn't been modified since
214 * we refreshed the cached stat information. In either case,
215 * we do not have to stat to see if the path has been removed
216 * or modified.
218 if (ce->ce_flags & (CE_VALID | CE_FSMONITOR_VALID)) {
219 changed = 0;
220 newmode = ce->ce_mode;
221 } else {
222 struct stat st;
224 changed = check_removed(istate, ce, &st);
225 if (changed) {
226 if (changed < 0) {
227 perror(ce->name);
228 continue;
230 diff_addremove(&revs->diffopt, '-', ce->ce_mode,
231 &ce->oid,
232 !is_null_oid(&ce->oid),
233 ce->name, 0);
234 continue;
235 } else if (revs->diffopt.ita_invisible_in_index &&
236 ce_intent_to_add(ce)) {
237 newmode = ce_mode_from_stat(ce, st.st_mode);
238 diff_addremove(&revs->diffopt, '+', newmode,
239 null_oid(), 0, ce->name, 0);
240 continue;
243 changed = match_stat_with_submodule(&revs->diffopt, ce, &st,
244 ce_option, &dirty_submodule);
245 newmode = ce_mode_from_stat(ce, st.st_mode);
248 if (!changed && !dirty_submodule) {
249 ce_mark_uptodate(ce);
250 mark_fsmonitor_valid(istate, ce);
251 if (!revs->diffopt.flags.find_copies_harder)
252 continue;
254 oldmode = ce->ce_mode;
255 old_oid = &ce->oid;
256 new_oid = changed ? null_oid() : &ce->oid;
257 diff_change(&revs->diffopt, oldmode, newmode,
258 old_oid, new_oid,
259 !is_null_oid(old_oid),
260 !is_null_oid(new_oid),
261 ce->name, 0, dirty_submodule);
264 diffcore_std(&revs->diffopt);
265 diff_flush(&revs->diffopt);
266 trace_performance_since(start, "diff-files");
267 return 0;
271 * diff-index
274 /* A file entry went away or appeared */
275 static void diff_index_show_file(struct rev_info *revs,
276 const char *prefix,
277 const struct cache_entry *ce,
278 const struct object_id *oid, int oid_valid,
279 unsigned int mode,
280 unsigned dirty_submodule)
282 diff_addremove(&revs->diffopt, prefix[0], mode,
283 oid, oid_valid, ce->name, dirty_submodule);
286 static int get_stat_data(const struct index_state *istate,
287 const struct cache_entry *ce,
288 const struct object_id **oidp,
289 unsigned int *modep,
290 int cached, int match_missing,
291 unsigned *dirty_submodule, struct diff_options *diffopt)
293 const struct object_id *oid = &ce->oid;
294 unsigned int mode = ce->ce_mode;
296 if (!cached && !ce_uptodate(ce)) {
297 int changed;
298 struct stat st;
299 changed = check_removed(istate, ce, &st);
300 if (changed < 0)
301 return -1;
302 else if (changed) {
303 if (match_missing) {
304 *oidp = oid;
305 *modep = mode;
306 return 0;
308 return -1;
310 changed = match_stat_with_submodule(diffopt, ce, &st,
311 0, dirty_submodule);
312 if (changed) {
313 mode = ce_mode_from_stat(ce, st.st_mode);
314 oid = null_oid();
318 *oidp = oid;
319 *modep = mode;
320 return 0;
323 static void show_new_file(struct rev_info *revs,
324 const struct cache_entry *new_file,
325 int cached, int match_missing)
327 const struct object_id *oid;
328 unsigned int mode;
329 unsigned dirty_submodule = 0;
330 struct index_state *istate = revs->diffopt.repo->index;
332 if (new_file && S_ISSPARSEDIR(new_file->ce_mode)) {
333 diff_tree_oid(NULL, &new_file->oid, new_file->name, &revs->diffopt);
334 return;
338 * New file in the index: it might actually be different in
339 * the working tree.
341 if (get_stat_data(istate, new_file, &oid, &mode, cached, match_missing,
342 &dirty_submodule, &revs->diffopt) < 0)
343 return;
345 diff_index_show_file(revs, "+", new_file, oid, !is_null_oid(oid), mode, dirty_submodule);
348 static int show_modified(struct rev_info *revs,
349 const struct cache_entry *old_entry,
350 const struct cache_entry *new_entry,
351 int report_missing,
352 int cached, int match_missing)
354 unsigned int mode, oldmode;
355 const struct object_id *oid;
356 unsigned dirty_submodule = 0;
357 struct index_state *istate = revs->diffopt.repo->index;
359 assert(S_ISSPARSEDIR(old_entry->ce_mode) ==
360 S_ISSPARSEDIR(new_entry->ce_mode));
363 * If both are sparse directory entries, then expand the
364 * modifications to the file level. If only one was a sparse
365 * directory, then they appear as an add and delete instead of
366 * a modification.
368 if (S_ISSPARSEDIR(new_entry->ce_mode)) {
369 diff_tree_oid(&old_entry->oid, &new_entry->oid, new_entry->name, &revs->diffopt);
370 return 0;
373 if (get_stat_data(istate, new_entry, &oid, &mode, cached, match_missing,
374 &dirty_submodule, &revs->diffopt) < 0) {
375 if (report_missing)
376 diff_index_show_file(revs, "-", old_entry,
377 &old_entry->oid, 1, old_entry->ce_mode,
379 return -1;
382 if (revs->combine_merges && !cached &&
383 (!oideq(oid, &old_entry->oid) || !oideq(&old_entry->oid, &new_entry->oid))) {
384 struct combine_diff_path *p;
385 int pathlen = ce_namelen(new_entry);
387 p = xmalloc(combine_diff_path_size(2, pathlen));
388 p->path = (char *) &p->parent[2];
389 p->next = NULL;
390 memcpy(p->path, new_entry->name, pathlen);
391 p->path[pathlen] = 0;
392 p->mode = mode;
393 oidclr(&p->oid);
394 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
395 p->parent[0].status = DIFF_STATUS_MODIFIED;
396 p->parent[0].mode = new_entry->ce_mode;
397 oidcpy(&p->parent[0].oid, &new_entry->oid);
398 p->parent[1].status = DIFF_STATUS_MODIFIED;
399 p->parent[1].mode = old_entry->ce_mode;
400 oidcpy(&p->parent[1].oid, &old_entry->oid);
401 show_combined_diff(p, 2, revs);
402 free(p);
403 return 0;
406 oldmode = old_entry->ce_mode;
407 if (mode == oldmode && oideq(oid, &old_entry->oid) && !dirty_submodule &&
408 !revs->diffopt.flags.find_copies_harder)
409 return 0;
411 diff_change(&revs->diffopt, oldmode, mode,
412 &old_entry->oid, oid, 1, !is_null_oid(oid),
413 old_entry->name, 0, dirty_submodule);
414 return 0;
418 * This gets a mix of an existing index and a tree, one pathname entry
419 * at a time. The index entry may be a single stage-0 one, but it could
420 * also be multiple unmerged entries (in which case idx_pos/idx_nr will
421 * give you the position and number of entries in the index).
423 static void do_oneway_diff(struct unpack_trees_options *o,
424 const struct cache_entry *idx,
425 const struct cache_entry *tree)
427 struct rev_info *revs = o->unpack_data;
428 int match_missing, cached;
431 * i-t-a entries do not actually exist in the index (if we're
432 * looking at its content)
434 if (o->index_only &&
435 revs->diffopt.ita_invisible_in_index &&
436 idx && ce_intent_to_add(idx)) {
437 idx = NULL;
438 if (!tree)
439 return; /* nothing to diff.. */
442 /* if the entry is not checked out, don't examine work tree */
443 cached = o->index_only ||
444 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
446 match_missing = revs->match_missing;
448 if (cached && idx && ce_stage(idx)) {
449 struct diff_filepair *pair;
450 pair = diff_unmerge(&revs->diffopt, idx->name);
451 if (tree)
452 fill_filespec(pair->one, &tree->oid, 1,
453 tree->ce_mode);
454 return;
458 * Something added to the tree?
460 if (!tree) {
461 show_new_file(revs, idx, cached, match_missing);
462 return;
466 * Something removed from the tree?
468 if (!idx) {
469 diff_index_show_file(revs, "-", tree, &tree->oid, 1,
470 tree->ce_mode, 0);
471 return;
474 /* Show difference between old and new */
475 show_modified(revs, tree, idx, 1, cached, match_missing);
479 * The unpack_trees() interface is designed for merging, so
480 * the different source entries are designed primarily for
481 * the source trees, with the old index being really mainly
482 * used for being replaced by the result.
484 * For diffing, the index is more important, and we only have a
485 * single tree.
487 * We're supposed to advance o->pos to skip what we have already processed.
489 * This wrapper makes it all more readable, and takes care of all
490 * the fairly complex unpack_trees() semantic requirements, including
491 * the skipping, the path matching, the type conflict cases etc.
493 static int oneway_diff(const struct cache_entry * const *src,
494 struct unpack_trees_options *o)
496 const struct cache_entry *idx = src[0];
497 const struct cache_entry *tree = src[1];
498 struct rev_info *revs = o->unpack_data;
501 * Unpack-trees generates a DF/conflict entry if
502 * there was a directory in the index and a tree
503 * in the tree. From a diff standpoint, that's a
504 * delete of the tree and a create of the file.
506 if (tree == o->df_conflict_entry)
507 tree = NULL;
509 if (ce_path_match(revs->diffopt.repo->index,
510 idx ? idx : tree,
511 &revs->prune_data, NULL)) {
512 do_oneway_diff(o, idx, tree);
513 if (diff_can_quit_early(&revs->diffopt)) {
514 o->exiting_early = 1;
515 return -1;
519 return 0;
522 static int diff_cache(struct rev_info *revs,
523 const struct object_id *tree_oid,
524 const char *tree_name,
525 int cached)
527 struct tree *tree;
528 struct tree_desc t;
529 struct unpack_trees_options opts;
531 tree = parse_tree_indirect(tree_oid);
532 if (!tree)
533 return error("bad tree object %s",
534 tree_name ? tree_name : oid_to_hex(tree_oid));
535 memset(&opts, 0, sizeof(opts));
536 opts.head_idx = 1;
537 opts.index_only = cached;
538 opts.diff_index_cached = (cached &&
539 !revs->diffopt.flags.find_copies_harder);
540 opts.merge = 1;
541 opts.fn = oneway_diff;
542 opts.unpack_data = revs;
543 opts.src_index = revs->diffopt.repo->index;
544 opts.dst_index = NULL;
545 opts.pathspec = &revs->diffopt.pathspec;
546 opts.pathspec->recursive = 1;
548 init_tree_desc(&t, tree->buffer, tree->size);
549 return unpack_trees(1, &t, &opts);
552 void diff_get_merge_base(const struct rev_info *revs, struct object_id *mb)
554 int i;
555 struct commit *mb_child[2] = {0};
556 struct commit_list *merge_bases;
558 for (i = 0; i < revs->pending.nr; i++) {
559 struct object *obj = revs->pending.objects[i].item;
560 if (obj->flags)
561 die(_("--merge-base does not work with ranges"));
562 if (obj->type != OBJ_COMMIT)
563 die(_("--merge-base only works with commits"));
567 * This check must go after the for loop above because A...B
568 * ranges produce three pending commits, resulting in a
569 * misleading error message.
571 if (revs->pending.nr < 1 || revs->pending.nr > 2)
572 BUG("unexpected revs->pending.nr: %d", revs->pending.nr);
574 for (i = 0; i < revs->pending.nr; i++)
575 mb_child[i] = lookup_commit_reference(the_repository, &revs->pending.objects[i].item->oid);
576 if (revs->pending.nr == 1) {
577 struct object_id oid;
579 if (get_oid("HEAD", &oid))
580 die(_("unable to get HEAD"));
582 mb_child[1] = lookup_commit_reference(the_repository, &oid);
585 merge_bases = repo_get_merge_bases(the_repository, mb_child[0], mb_child[1]);
586 if (!merge_bases)
587 die(_("no merge base found"));
588 if (merge_bases->next)
589 die(_("multiple merge bases found"));
591 oidcpy(mb, &merge_bases->item->object.oid);
593 free_commit_list(merge_bases);
596 int run_diff_index(struct rev_info *revs, unsigned int option)
598 struct object_array_entry *ent;
599 int cached = !!(option & DIFF_INDEX_CACHED);
600 int merge_base = !!(option & DIFF_INDEX_MERGE_BASE);
601 struct object_id oid;
602 const char *name;
603 char merge_base_hex[GIT_MAX_HEXSZ + 1];
604 struct index_state *istate = revs->diffopt.repo->index;
606 if (revs->pending.nr != 1)
607 BUG("run_diff_index must be passed exactly one tree");
609 trace_performance_enter();
610 ent = revs->pending.objects;
612 refresh_fsmonitor(istate);
614 if (merge_base) {
615 diff_get_merge_base(revs, &oid);
616 name = oid_to_hex_r(merge_base_hex, &oid);
617 } else {
618 oidcpy(&oid, &ent->item->oid);
619 name = ent->name;
622 if (diff_cache(revs, &oid, name, cached))
623 exit(128);
625 diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
626 diffcore_fix_diff_index();
627 diffcore_std(&revs->diffopt);
628 diff_flush(&revs->diffopt);
629 trace_performance_leave("diff-index");
630 return 0;
633 int do_diff_cache(const struct object_id *tree_oid, struct diff_options *opt)
635 struct rev_info revs;
637 repo_init_revisions(opt->repo, &revs, NULL);
638 copy_pathspec(&revs.prune_data, &opt->pathspec);
639 diff_setup_done(&revs.diffopt);
640 revs.diffopt = *opt;
642 if (diff_cache(&revs, tree_oid, NULL, 1))
643 exit(128);
644 release_revisions(&revs);
645 return 0;
648 int index_differs_from(struct repository *r,
649 const char *def, const struct diff_flags *flags,
650 int ita_invisible_in_index)
652 struct rev_info rev;
653 struct setup_revision_opt opt;
654 unsigned has_changes;
656 repo_init_revisions(r, &rev, NULL);
657 memset(&opt, 0, sizeof(opt));
658 opt.def = def;
659 setup_revisions(0, NULL, &rev, &opt);
660 rev.diffopt.flags.quick = 1;
661 rev.diffopt.flags.exit_with_status = 1;
662 if (flags)
663 diff_flags_or(&rev.diffopt.flags, flags);
664 rev.diffopt.ita_invisible_in_index = ita_invisible_in_index;
665 run_diff_index(&rev, 1);
666 has_changes = rev.diffopt.flags.has_changes;
667 release_revisions(&rev);
668 return (has_changes != 0);
671 static struct strbuf *idiff_prefix_cb(struct diff_options *opt, void *data)
673 return data;
676 void show_interdiff(const struct object_id *oid1, const struct object_id *oid2,
677 int indent, struct diff_options *diffopt)
679 struct diff_options opts;
680 struct strbuf prefix = STRBUF_INIT;
682 memcpy(&opts, diffopt, sizeof(opts));
683 opts.output_format = DIFF_FORMAT_PATCH;
684 opts.output_prefix = idiff_prefix_cb;
685 strbuf_addchars(&prefix, ' ', indent);
686 opts.output_prefix_data = &prefix;
687 diff_setup_done(&opts);
689 diff_tree_oid(oid1, oid2, "", &opts);
690 diffcore_std(&opts);
691 diff_flush(&opts);
693 strbuf_release(&prefix);