strvec: declare the `strvec_push_nodup()` function globally
[git.git] / diff-lib.c
blob7a1eb637579c7df952394b45b269f5519c435225
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
5 #define USE_THE_REPOSITORY_VARIABLE
7 #include "git-compat-util.h"
8 #include "commit.h"
9 #include "diff.h"
10 #include "diffcore.h"
11 #include "gettext.h"
12 #include "hash.h"
13 #include "hex.h"
14 #include "object-name.h"
15 #include "read-cache.h"
16 #include "revision.h"
17 #include "cache-tree.h"
18 #include "unpack-trees.h"
19 #include "refs.h"
20 #include "repository.h"
21 #include "submodule.h"
22 #include "symlinks.h"
23 #include "trace.h"
24 #include "dir.h"
25 #include "fsmonitor.h"
26 #include "commit-reach.h"
29 * diff-files
33 * Has the work tree entity been removed?
35 * Return 1 if it was removed from the work tree, 0 if an entity to be
36 * compared with the cache entry ce still exists (the latter includes
37 * the case where a directory that is not a submodule repository
38 * exists for ce that is a submodule -- it is a submodule that is not
39 * checked out). Return negative for an error.
41 static int check_removed(const struct cache_entry *ce, struct stat *st)
43 int stat_err;
45 if (!(ce->ce_flags & CE_FSMONITOR_VALID))
46 stat_err = lstat(ce->name, st);
47 else
48 stat_err = fake_lstat(ce, st);
49 if (stat_err < 0) {
50 if (!is_missing_file_error(errno))
51 return -1;
52 return 1;
55 if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
56 return 1;
57 if (S_ISDIR(st->st_mode)) {
58 struct object_id sub;
61 * If ce is already a gitlink, we can have a plain
62 * directory (i.e. the submodule is not checked out),
63 * or a checked out submodule. Either case this is not
64 * a case where something was removed from the work tree,
65 * so we will return 0.
67 * Otherwise, if the directory is not a submodule
68 * repository, that means ce which was a blob turned into
69 * a directory --- the blob was removed!
71 if (!S_ISGITLINK(ce->ce_mode) &&
72 repo_resolve_gitlink_ref(the_repository, ce->name,
73 "HEAD", &sub))
74 return 1;
76 return 0;
80 * Has a file changed or has a submodule new commits or a dirty work tree?
82 * Return 1 when changes are detected, 0 otherwise. If the DIRTY_SUBMODULES
83 * option is set, the caller does not only want to know if a submodule is
84 * modified at all but wants to know all the conditions that are met (new
85 * commits, untracked content and/or modified content).
87 static int match_stat_with_submodule(struct diff_options *diffopt,
88 const struct cache_entry *ce,
89 struct stat *st, unsigned ce_option,
90 unsigned *dirty_submodule)
92 int changed = ie_match_stat(diffopt->repo->index, ce, st, ce_option);
93 if (S_ISGITLINK(ce->ce_mode)) {
94 struct diff_flags orig_flags = diffopt->flags;
95 if (!diffopt->flags.override_submodule_config)
96 set_diffopt_flags_from_submodule_config(diffopt, ce->name);
97 if (diffopt->flags.ignore_submodules)
98 changed = 0;
99 else if (!diffopt->flags.ignore_dirty_submodules &&
100 (!changed || diffopt->flags.dirty_submodules))
101 *dirty_submodule = is_submodule_modified(ce->name,
102 diffopt->flags.ignore_untracked_in_submodules);
103 diffopt->flags = orig_flags;
105 return changed;
108 void run_diff_files(struct rev_info *revs, unsigned int option)
110 int entries, i;
111 int diff_unmerged_stage = revs->max_count;
112 unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
113 ? CE_MATCH_RACY_IS_DIRTY : 0);
114 uint64_t start = getnanotime();
115 struct index_state *istate = revs->diffopt.repo->index;
117 diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
119 refresh_fsmonitor(istate);
121 if (diff_unmerged_stage < 0)
122 diff_unmerged_stage = 2;
123 entries = istate->cache_nr;
124 for (i = 0; i < entries; i++) {
125 unsigned int oldmode, newmode;
126 struct cache_entry *ce = istate->cache[i];
127 int changed;
128 unsigned dirty_submodule = 0;
129 const struct object_id *old_oid, *new_oid;
131 if (diff_can_quit_early(&revs->diffopt))
132 break;
135 * NEEDSWORK:
136 * Here we filter with pathspec but the result is further
137 * filtered out when --relative is in effect. To end-users,
138 * a pathspec element that matched only to paths outside the
139 * current directory is like not matching anything at all;
140 * the handling of ps_matched[] here may become problematic
141 * if/when we add the "--error-unmatch" option to "git diff".
143 if (!ce_path_match(istate, ce, &revs->prune_data, revs->ps_matched))
144 continue;
146 if (revs->diffopt.prefix &&
147 strncmp(ce->name, revs->diffopt.prefix, revs->diffopt.prefix_length))
148 continue;
150 if (ce_stage(ce)) {
151 struct combine_diff_path *dpath;
152 struct diff_filepair *pair;
153 unsigned int wt_mode = 0;
154 int num_compare_stages = 0;
155 size_t path_len;
156 struct stat st;
158 path_len = ce_namelen(ce);
160 dpath = xmalloc(combine_diff_path_size(5, path_len));
161 dpath->path = (char *) &(dpath->parent[5]);
163 dpath->next = NULL;
164 memcpy(dpath->path, ce->name, path_len);
165 dpath->path[path_len] = '\0';
166 oidclr(&dpath->oid, the_repository->hash_algo);
167 memset(&(dpath->parent[0]), 0,
168 sizeof(struct combine_diff_parent)*5);
170 changed = check_removed(ce, &st);
171 if (!changed)
172 wt_mode = ce_mode_from_stat(ce, st.st_mode);
173 else {
174 if (changed < 0) {
175 perror(ce->name);
176 continue;
178 wt_mode = 0;
180 dpath->mode = wt_mode;
182 while (i < entries) {
183 struct cache_entry *nce = istate->cache[i];
184 int stage;
186 if (strcmp(ce->name, nce->name))
187 break;
189 /* Stage #2 (ours) is the first parent,
190 * stage #3 (theirs) is the second.
192 stage = ce_stage(nce);
193 if (2 <= stage) {
194 int mode = nce->ce_mode;
195 num_compare_stages++;
196 oidcpy(&dpath->parent[stage - 2].oid,
197 &nce->oid);
198 dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
199 dpath->parent[stage-2].status =
200 DIFF_STATUS_MODIFIED;
203 /* diff against the proper unmerged stage */
204 if (stage == diff_unmerged_stage)
205 ce = nce;
206 i++;
209 * Compensate for loop update
211 i--;
213 if (revs->combine_merges && num_compare_stages == 2) {
214 show_combined_diff(dpath, 2, revs);
215 free(dpath);
216 continue;
218 FREE_AND_NULL(dpath);
221 * Show the diff for the 'ce' if we found the one
222 * from the desired stage.
224 pair = diff_unmerge(&revs->diffopt, ce->name);
225 if (wt_mode)
226 pair->two->mode = wt_mode;
227 if (ce_stage(ce) != diff_unmerged_stage)
228 continue;
231 if (ce_uptodate(ce) || ce_skip_worktree(ce))
232 continue;
235 * When CE_VALID is set (via "update-index --assume-unchanged"
236 * or via adding paths while core.ignorestat is set to true),
237 * the user has promised that the working tree file for that
238 * path will not be modified. When CE_FSMONITOR_VALID is true,
239 * the fsmonitor knows that the path hasn't been modified since
240 * we refreshed the cached stat information. In either case,
241 * we do not have to stat to see if the path has been removed
242 * or modified.
244 if (ce->ce_flags & (CE_VALID | CE_FSMONITOR_VALID)) {
245 changed = 0;
246 newmode = ce->ce_mode;
247 } else {
248 struct stat st;
250 changed = check_removed(ce, &st);
251 if (changed) {
252 if (changed < 0) {
253 perror(ce->name);
254 continue;
256 diff_addremove(&revs->diffopt, '-', ce->ce_mode,
257 &ce->oid,
258 !is_null_oid(&ce->oid),
259 ce->name, 0);
260 continue;
261 } else if (revs->diffopt.ita_invisible_in_index &&
262 ce_intent_to_add(ce)) {
263 newmode = ce_mode_from_stat(ce, st.st_mode);
264 diff_addremove(&revs->diffopt, '+', newmode,
265 null_oid(), 0, ce->name, 0);
266 continue;
269 changed = match_stat_with_submodule(&revs->diffopt, ce, &st,
270 ce_option, &dirty_submodule);
271 newmode = ce_mode_from_stat(ce, st.st_mode);
274 if (!changed && !dirty_submodule) {
275 ce_mark_uptodate(ce);
276 mark_fsmonitor_valid(istate, ce);
277 if (!revs->diffopt.flags.find_copies_harder)
278 continue;
280 oldmode = ce->ce_mode;
281 old_oid = &ce->oid;
282 new_oid = changed ? null_oid() : &ce->oid;
283 diff_change(&revs->diffopt, oldmode, newmode,
284 old_oid, new_oid,
285 !is_null_oid(old_oid),
286 !is_null_oid(new_oid),
287 ce->name, 0, dirty_submodule);
290 diffcore_std(&revs->diffopt);
291 diff_flush(&revs->diffopt);
292 trace_performance_since(start, "diff-files");
296 * diff-index
299 /* A file entry went away or appeared */
300 static void diff_index_show_file(struct rev_info *revs,
301 const char *prefix,
302 const struct cache_entry *ce,
303 const struct object_id *oid, int oid_valid,
304 unsigned int mode,
305 unsigned dirty_submodule)
307 diff_addremove(&revs->diffopt, prefix[0], mode,
308 oid, oid_valid, ce->name, dirty_submodule);
311 static int get_stat_data(const struct index_state *istate,
312 const struct cache_entry *ce,
313 const struct object_id **oidp,
314 unsigned int *modep,
315 int cached, int match_missing,
316 unsigned *dirty_submodule, struct diff_options *diffopt)
318 const struct object_id *oid = &ce->oid;
319 unsigned int mode = ce->ce_mode;
321 if (!cached && !ce_uptodate(ce)) {
322 int changed;
323 struct stat st;
324 changed = check_removed(ce, &st);
325 if (changed < 0)
326 return -1;
327 else if (changed) {
328 if (match_missing) {
329 *oidp = oid;
330 *modep = mode;
331 return 0;
333 return -1;
335 changed = match_stat_with_submodule(diffopt, ce, &st,
336 0, dirty_submodule);
337 if (changed) {
338 mode = ce_mode_from_stat(ce, st.st_mode);
339 oid = null_oid();
343 *oidp = oid;
344 *modep = mode;
345 return 0;
348 static void show_new_file(struct rev_info *revs,
349 const struct cache_entry *new_file,
350 int cached, int match_missing)
352 const struct object_id *oid;
353 unsigned int mode;
354 unsigned dirty_submodule = 0;
355 struct index_state *istate = revs->diffopt.repo->index;
357 if (new_file && S_ISSPARSEDIR(new_file->ce_mode)) {
358 diff_tree_oid(NULL, &new_file->oid, new_file->name, &revs->diffopt);
359 return;
363 * New file in the index: it might actually be different in
364 * the working tree.
366 if (get_stat_data(istate, new_file, &oid, &mode, cached, match_missing,
367 &dirty_submodule, &revs->diffopt) < 0)
368 return;
370 diff_index_show_file(revs, "+", new_file, oid, !is_null_oid(oid), mode, dirty_submodule);
373 static int show_modified(struct rev_info *revs,
374 const struct cache_entry *old_entry,
375 const struct cache_entry *new_entry,
376 int report_missing,
377 int cached, int match_missing)
379 unsigned int mode, oldmode;
380 const struct object_id *oid;
381 unsigned dirty_submodule = 0;
382 struct index_state *istate = revs->diffopt.repo->index;
384 assert(S_ISSPARSEDIR(old_entry->ce_mode) ==
385 S_ISSPARSEDIR(new_entry->ce_mode));
388 * If both are sparse directory entries, then expand the
389 * modifications to the file level. If only one was a sparse
390 * directory, then they appear as an add and delete instead of
391 * a modification.
393 if (S_ISSPARSEDIR(new_entry->ce_mode)) {
394 diff_tree_oid(&old_entry->oid, &new_entry->oid, new_entry->name, &revs->diffopt);
395 return 0;
398 if (get_stat_data(istate, new_entry, &oid, &mode, cached, match_missing,
399 &dirty_submodule, &revs->diffopt) < 0) {
400 if (report_missing)
401 diff_index_show_file(revs, "-", old_entry,
402 &old_entry->oid, 1, old_entry->ce_mode,
404 return -1;
407 if (revs->combine_merges && !cached &&
408 (!oideq(oid, &old_entry->oid) || !oideq(&old_entry->oid, &new_entry->oid))) {
409 struct combine_diff_path *p;
410 int pathlen = ce_namelen(new_entry);
412 p = xmalloc(combine_diff_path_size(2, pathlen));
413 p->path = (char *) &p->parent[2];
414 p->next = NULL;
415 memcpy(p->path, new_entry->name, pathlen);
416 p->path[pathlen] = 0;
417 p->mode = mode;
418 oidclr(&p->oid, the_repository->hash_algo);
419 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
420 p->parent[0].status = DIFF_STATUS_MODIFIED;
421 p->parent[0].mode = new_entry->ce_mode;
422 oidcpy(&p->parent[0].oid, &new_entry->oid);
423 p->parent[1].status = DIFF_STATUS_MODIFIED;
424 p->parent[1].mode = old_entry->ce_mode;
425 oidcpy(&p->parent[1].oid, &old_entry->oid);
426 show_combined_diff(p, 2, revs);
427 free(p);
428 return 0;
431 oldmode = old_entry->ce_mode;
432 if (mode == oldmode && oideq(oid, &old_entry->oid) && !dirty_submodule &&
433 !revs->diffopt.flags.find_copies_harder)
434 return 0;
436 diff_change(&revs->diffopt, oldmode, mode,
437 &old_entry->oid, oid, 1, !is_null_oid(oid),
438 old_entry->name, 0, dirty_submodule);
439 return 0;
443 * This gets a mix of an existing index and a tree, one pathname entry
444 * at a time. The index entry may be a single stage-0 one, but it could
445 * also be multiple unmerged entries (in which case idx_pos/idx_nr will
446 * give you the position and number of entries in the index).
448 static void do_oneway_diff(struct unpack_trees_options *o,
449 const struct cache_entry *idx,
450 const struct cache_entry *tree)
452 struct rev_info *revs = o->unpack_data;
453 int match_missing, cached;
456 * i-t-a entries do not actually exist in the index (if we're
457 * looking at its content)
459 if (o->index_only &&
460 revs->diffopt.ita_invisible_in_index &&
461 idx && ce_intent_to_add(idx)) {
462 idx = NULL;
463 if (!tree)
464 return; /* nothing to diff.. */
467 /* if the entry is not checked out, don't examine work tree */
468 cached = o->index_only ||
469 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
471 match_missing = revs->match_missing;
473 if (cached && idx && ce_stage(idx)) {
474 struct diff_filepair *pair;
475 pair = diff_unmerge(&revs->diffopt, idx->name);
476 if (tree)
477 fill_filespec(pair->one, &tree->oid, 1,
478 tree->ce_mode);
479 return;
483 * Something added to the tree?
485 if (!tree) {
486 show_new_file(revs, idx, cached, match_missing);
487 return;
491 * Something removed from the tree?
493 if (!idx) {
494 if (S_ISSPARSEDIR(tree->ce_mode)) {
495 diff_tree_oid(&tree->oid, NULL, tree->name, &revs->diffopt);
496 return;
499 diff_index_show_file(revs, "-", tree, &tree->oid, 1,
500 tree->ce_mode, 0);
501 return;
504 /* Show difference between old and new */
505 show_modified(revs, tree, idx, 1, cached, match_missing);
509 * The unpack_trees() interface is designed for merging, so
510 * the different source entries are designed primarily for
511 * the source trees, with the old index being really mainly
512 * used for being replaced by the result.
514 * For diffing, the index is more important, and we only have a
515 * single tree.
517 * We're supposed to advance o->pos to skip what we have already processed.
519 * This wrapper makes it all more readable, and takes care of all
520 * the fairly complex unpack_trees() semantic requirements, including
521 * the skipping, the path matching, the type conflict cases etc.
523 static int oneway_diff(const struct cache_entry * const *src,
524 struct unpack_trees_options *o)
526 const struct cache_entry *idx = src[0];
527 const struct cache_entry *tree = src[1];
528 struct rev_info *revs = o->unpack_data;
531 * Unpack-trees generates a DF/conflict entry if
532 * there was a directory in the index and a tree
533 * in the tree. From a diff standpoint, that's a
534 * delete of the tree and a create of the file.
536 if (tree == o->df_conflict_entry)
537 tree = NULL;
539 if (ce_path_match(revs->diffopt.repo->index,
540 idx ? idx : tree,
541 &revs->prune_data, NULL)) {
542 do_oneway_diff(o, idx, tree);
543 if (diff_can_quit_early(&revs->diffopt)) {
544 o->exiting_early = 1;
545 return -1;
549 return 0;
552 static int diff_cache(struct rev_info *revs,
553 const struct object_id *tree_oid,
554 const char *tree_name,
555 int cached)
557 struct tree *tree;
558 struct tree_desc t;
559 struct unpack_trees_options opts;
561 tree = parse_tree_indirect(tree_oid);
562 if (!tree)
563 return error("bad tree object %s",
564 tree_name ? tree_name : oid_to_hex(tree_oid));
565 memset(&opts, 0, sizeof(opts));
566 opts.head_idx = 1;
567 opts.index_only = cached;
568 opts.diff_index_cached = (cached &&
569 !revs->diffopt.flags.find_copies_harder);
570 opts.merge = 1;
571 opts.fn = oneway_diff;
572 opts.unpack_data = revs;
573 opts.src_index = revs->diffopt.repo->index;
574 opts.dst_index = NULL;
575 opts.pathspec = &revs->diffopt.pathspec;
576 opts.pathspec->recursive = 1;
578 init_tree_desc(&t, &tree->object.oid, tree->buffer, tree->size);
579 return unpack_trees(1, &t, &opts);
582 void diff_get_merge_base(const struct rev_info *revs, struct object_id *mb)
584 int i;
585 struct commit *mb_child[2] = {0};
586 struct commit_list *merge_bases = NULL;
588 for (i = 0; i < revs->pending.nr; i++) {
589 struct object *obj = revs->pending.objects[i].item;
590 if (obj->flags)
591 die(_("--merge-base does not work with ranges"));
595 * This check must go after the for loop above because A...B
596 * ranges produce three pending commits, resulting in a
597 * misleading error message.
599 if (revs->pending.nr < 1 || revs->pending.nr > 2)
600 BUG("unexpected revs->pending.nr: %d", revs->pending.nr);
602 for (i = 0; i < revs->pending.nr; i++)
603 mb_child[i] = lookup_commit_reference(the_repository, &revs->pending.objects[i].item->oid);
604 if (revs->pending.nr == 1) {
605 struct object_id oid;
607 if (repo_get_oid(the_repository, "HEAD", &oid))
608 die(_("unable to get HEAD"));
610 mb_child[1] = lookup_commit_reference(the_repository, &oid);
613 if (repo_get_merge_bases(the_repository, mb_child[0], mb_child[1], &merge_bases) < 0)
614 exit(128);
615 if (!merge_bases)
616 die(_("no merge base found"));
617 if (merge_bases->next)
618 die(_("multiple merge bases found"));
620 oidcpy(mb, &merge_bases->item->object.oid);
622 free_commit_list(merge_bases);
625 void run_diff_index(struct rev_info *revs, unsigned int option)
627 struct object_array_entry *ent;
628 int cached = !!(option & DIFF_INDEX_CACHED);
629 int merge_base = !!(option & DIFF_INDEX_MERGE_BASE);
630 struct object_id oid;
631 const char *name;
632 char merge_base_hex[GIT_MAX_HEXSZ + 1];
633 struct index_state *istate = revs->diffopt.repo->index;
635 if (revs->pending.nr != 1)
636 BUG("run_diff_index must be passed exactly one tree");
638 trace_performance_enter();
639 ent = revs->pending.objects;
641 refresh_fsmonitor(istate);
643 if (merge_base) {
644 diff_get_merge_base(revs, &oid);
645 name = oid_to_hex_r(merge_base_hex, &oid);
646 } else {
647 oidcpy(&oid, &ent->item->oid);
648 name = ent->name;
651 if (diff_cache(revs, &oid, name, cached))
652 exit(128);
654 diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
655 diffcore_fix_diff_index();
656 diffcore_std(&revs->diffopt);
657 diff_flush(&revs->diffopt);
658 trace_performance_leave("diff-index");
661 int do_diff_cache(const struct object_id *tree_oid, struct diff_options *opt)
663 struct rev_info revs;
665 repo_init_revisions(opt->repo, &revs, NULL);
666 copy_pathspec(&revs.prune_data, &opt->pathspec);
667 revs.diffopt = *opt;
668 revs.diffopt.no_free = 1;
670 if (diff_cache(&revs, tree_oid, NULL, 1))
671 exit(128);
673 release_revisions(&revs);
674 return 0;
677 int index_differs_from(struct repository *r,
678 const char *def, const struct diff_flags *flags,
679 int ita_invisible_in_index)
681 struct rev_info rev;
682 struct setup_revision_opt opt;
683 unsigned has_changes;
685 repo_init_revisions(r, &rev, NULL);
686 memset(&opt, 0, sizeof(opt));
687 opt.def = def;
688 setup_revisions(0, NULL, &rev, &opt);
689 rev.diffopt.flags.quick = 1;
690 rev.diffopt.flags.exit_with_status = 1;
691 if (flags) {
692 diff_flags_or(&rev.diffopt.flags, flags);
694 * Now that flags are merged, honor override_submodule_config
695 * and ignore_submodules from passed flags.
697 if (flags->override_submodule_config)
698 rev.diffopt.flags.ignore_submodules = flags->ignore_submodules;
700 rev.diffopt.ita_invisible_in_index = ita_invisible_in_index;
701 run_diff_index(&rev, DIFF_INDEX_CACHED);
702 has_changes = rev.diffopt.flags.has_changes;
703 release_revisions(&rev);
704 return (has_changes != 0);
707 static struct strbuf *idiff_prefix_cb(struct diff_options *opt UNUSED, void *data)
709 return data;
712 void show_interdiff(const struct object_id *oid1, const struct object_id *oid2,
713 int indent, struct diff_options *diffopt)
715 struct diff_options opts;
716 struct strbuf prefix = STRBUF_INIT;
718 memcpy(&opts, diffopt, sizeof(opts));
719 opts.output_format = DIFF_FORMAT_PATCH;
720 opts.output_prefix = idiff_prefix_cb;
721 strbuf_addchars(&prefix, ' ', indent);
722 opts.output_prefix_data = &prefix;
723 diff_setup_done(&opts);
725 diff_tree_oid(oid1, oid2, "", &opts);
726 diffcore_std(&opts);
727 diff_flush(&opts);
729 strbuf_release(&prefix);