sparse-checkout: consolidate cleanup when writing patterns
[git/gitster.git] / diff-lib.c
bloba680768ee7549a86e9387dc46d2ea94e41928c31
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 cache_entry *ce,
312 const struct object_id **oidp,
313 unsigned int *modep,
314 int cached, int match_missing,
315 unsigned *dirty_submodule, struct diff_options *diffopt)
317 const struct object_id *oid = &ce->oid;
318 unsigned int mode = ce->ce_mode;
320 if (!cached && !ce_uptodate(ce)) {
321 int changed;
322 struct stat st;
323 changed = check_removed(ce, &st);
324 if (changed < 0)
325 return -1;
326 else if (changed) {
327 if (match_missing) {
328 *oidp = oid;
329 *modep = mode;
330 return 0;
332 return -1;
334 changed = match_stat_with_submodule(diffopt, ce, &st,
335 0, dirty_submodule);
336 if (changed) {
337 mode = ce_mode_from_stat(ce, st.st_mode);
338 oid = null_oid();
342 *oidp = oid;
343 *modep = mode;
344 return 0;
347 static void show_new_file(struct rev_info *revs,
348 const struct cache_entry *new_file,
349 int cached, int match_missing)
351 const struct object_id *oid;
352 unsigned int mode;
353 unsigned dirty_submodule = 0;
355 if (new_file && S_ISSPARSEDIR(new_file->ce_mode)) {
356 diff_tree_oid(NULL, &new_file->oid, new_file->name, &revs->diffopt);
357 return;
361 * New file in the index: it might actually be different in
362 * the working tree.
364 if (get_stat_data(new_file, &oid, &mode, cached, match_missing,
365 &dirty_submodule, &revs->diffopt) < 0)
366 return;
368 diff_index_show_file(revs, "+", new_file, oid, !is_null_oid(oid), mode, dirty_submodule);
371 static int show_modified(struct rev_info *revs,
372 const struct cache_entry *old_entry,
373 const struct cache_entry *new_entry,
374 int report_missing,
375 int cached, int match_missing)
377 unsigned int mode, oldmode;
378 const struct object_id *oid;
379 unsigned dirty_submodule = 0;
381 assert(S_ISSPARSEDIR(old_entry->ce_mode) ==
382 S_ISSPARSEDIR(new_entry->ce_mode));
385 * If both are sparse directory entries, then expand the
386 * modifications to the file level. If only one was a sparse
387 * directory, then they appear as an add and delete instead of
388 * a modification.
390 if (S_ISSPARSEDIR(new_entry->ce_mode)) {
391 diff_tree_oid(&old_entry->oid, &new_entry->oid, new_entry->name, &revs->diffopt);
392 return 0;
395 if (get_stat_data(new_entry, &oid, &mode, cached, match_missing,
396 &dirty_submodule, &revs->diffopt) < 0) {
397 if (report_missing)
398 diff_index_show_file(revs, "-", old_entry,
399 &old_entry->oid, 1, old_entry->ce_mode,
401 return -1;
404 if (revs->combine_merges && !cached &&
405 (!oideq(oid, &old_entry->oid) || !oideq(&old_entry->oid, &new_entry->oid))) {
406 struct combine_diff_path *p;
407 int pathlen = ce_namelen(new_entry);
409 p = xmalloc(combine_diff_path_size(2, pathlen));
410 p->path = (char *) &p->parent[2];
411 p->next = NULL;
412 memcpy(p->path, new_entry->name, pathlen);
413 p->path[pathlen] = 0;
414 p->mode = mode;
415 oidclr(&p->oid, the_repository->hash_algo);
416 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
417 p->parent[0].status = DIFF_STATUS_MODIFIED;
418 p->parent[0].mode = new_entry->ce_mode;
419 oidcpy(&p->parent[0].oid, &new_entry->oid);
420 p->parent[1].status = DIFF_STATUS_MODIFIED;
421 p->parent[1].mode = old_entry->ce_mode;
422 oidcpy(&p->parent[1].oid, &old_entry->oid);
423 show_combined_diff(p, 2, revs);
424 free(p);
425 return 0;
428 oldmode = old_entry->ce_mode;
429 if (mode == oldmode && oideq(oid, &old_entry->oid) && !dirty_submodule &&
430 !revs->diffopt.flags.find_copies_harder)
431 return 0;
433 diff_change(&revs->diffopt, oldmode, mode,
434 &old_entry->oid, oid, 1, !is_null_oid(oid),
435 old_entry->name, 0, dirty_submodule);
436 return 0;
440 * This gets a mix of an existing index and a tree, one pathname entry
441 * at a time. The index entry may be a single stage-0 one, but it could
442 * also be multiple unmerged entries (in which case idx_pos/idx_nr will
443 * give you the position and number of entries in the index).
445 static void do_oneway_diff(struct unpack_trees_options *o,
446 const struct cache_entry *idx,
447 const struct cache_entry *tree)
449 struct rev_info *revs = o->unpack_data;
450 int match_missing, cached;
453 * i-t-a entries do not actually exist in the index (if we're
454 * looking at its content)
456 if (o->index_only &&
457 revs->diffopt.ita_invisible_in_index &&
458 idx && ce_intent_to_add(idx)) {
459 idx = NULL;
460 if (!tree)
461 return; /* nothing to diff.. */
464 /* if the entry is not checked out, don't examine work tree */
465 cached = o->index_only ||
466 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
468 match_missing = revs->match_missing;
470 if (cached && idx && ce_stage(idx)) {
471 struct diff_filepair *pair;
472 pair = diff_unmerge(&revs->diffopt, idx->name);
473 if (tree)
474 fill_filespec(pair->one, &tree->oid, 1,
475 tree->ce_mode);
476 return;
480 * Something added to the tree?
482 if (!tree) {
483 show_new_file(revs, idx, cached, match_missing);
484 return;
488 * Something removed from the tree?
490 if (!idx) {
491 if (S_ISSPARSEDIR(tree->ce_mode)) {
492 diff_tree_oid(&tree->oid, NULL, tree->name, &revs->diffopt);
493 return;
496 diff_index_show_file(revs, "-", tree, &tree->oid, 1,
497 tree->ce_mode, 0);
498 return;
501 /* Show difference between old and new */
502 show_modified(revs, tree, idx, 1, cached, match_missing);
506 * The unpack_trees() interface is designed for merging, so
507 * the different source entries are designed primarily for
508 * the source trees, with the old index being really mainly
509 * used for being replaced by the result.
511 * For diffing, the index is more important, and we only have a
512 * single tree.
514 * We're supposed to advance o->pos to skip what we have already processed.
516 * This wrapper makes it all more readable, and takes care of all
517 * the fairly complex unpack_trees() semantic requirements, including
518 * the skipping, the path matching, the type conflict cases etc.
520 static int oneway_diff(const struct cache_entry * const *src,
521 struct unpack_trees_options *o)
523 const struct cache_entry *idx = src[0];
524 const struct cache_entry *tree = src[1];
525 struct rev_info *revs = o->unpack_data;
528 * Unpack-trees generates a DF/conflict entry if
529 * there was a directory in the index and a tree
530 * in the tree. From a diff standpoint, that's a
531 * delete of the tree and a create of the file.
533 if (tree == o->df_conflict_entry)
534 tree = NULL;
536 if (ce_path_match(revs->diffopt.repo->index,
537 idx ? idx : tree,
538 &revs->prune_data, NULL)) {
539 do_oneway_diff(o, idx, tree);
540 if (diff_can_quit_early(&revs->diffopt)) {
541 o->exiting_early = 1;
542 return -1;
546 return 0;
549 static int diff_cache(struct rev_info *revs,
550 const struct object_id *tree_oid,
551 const char *tree_name,
552 int cached)
554 struct tree *tree;
555 struct tree_desc t;
556 struct unpack_trees_options opts;
558 tree = parse_tree_indirect(tree_oid);
559 if (!tree)
560 return error("bad tree object %s",
561 tree_name ? tree_name : oid_to_hex(tree_oid));
562 memset(&opts, 0, sizeof(opts));
563 opts.head_idx = 1;
564 opts.index_only = cached;
565 opts.diff_index_cached = (cached &&
566 !revs->diffopt.flags.find_copies_harder);
567 opts.merge = 1;
568 opts.fn = oneway_diff;
569 opts.unpack_data = revs;
570 opts.src_index = revs->diffopt.repo->index;
571 opts.dst_index = NULL;
572 opts.pathspec = &revs->diffopt.pathspec;
573 opts.pathspec->recursive = 1;
575 init_tree_desc(&t, &tree->object.oid, tree->buffer, tree->size);
576 return unpack_trees(1, &t, &opts);
579 void diff_get_merge_base(const struct rev_info *revs, struct object_id *mb)
581 int i;
582 struct commit *mb_child[2] = {0};
583 struct commit_list *merge_bases = NULL;
585 for (i = 0; i < revs->pending.nr; i++) {
586 struct object *obj = revs->pending.objects[i].item;
587 if (obj->flags)
588 die(_("--merge-base does not work with ranges"));
592 * This check must go after the for loop above because A...B
593 * ranges produce three pending commits, resulting in a
594 * misleading error message.
596 if (revs->pending.nr < 1 || revs->pending.nr > 2)
597 BUG("unexpected revs->pending.nr: %d", revs->pending.nr);
599 for (i = 0; i < revs->pending.nr; i++)
600 mb_child[i] = lookup_commit_reference(the_repository, &revs->pending.objects[i].item->oid);
601 if (revs->pending.nr == 1) {
602 struct object_id oid;
604 if (repo_get_oid(the_repository, "HEAD", &oid))
605 die(_("unable to get HEAD"));
607 mb_child[1] = lookup_commit_reference(the_repository, &oid);
610 if (repo_get_merge_bases(the_repository, mb_child[0], mb_child[1], &merge_bases) < 0)
611 exit(128);
612 if (!merge_bases)
613 die(_("no merge base found"));
614 if (merge_bases->next)
615 die(_("multiple merge bases found"));
617 oidcpy(mb, &merge_bases->item->object.oid);
619 free_commit_list(merge_bases);
622 void run_diff_index(struct rev_info *revs, unsigned int option)
624 struct object_array_entry *ent;
625 int cached = !!(option & DIFF_INDEX_CACHED);
626 int merge_base = !!(option & DIFF_INDEX_MERGE_BASE);
627 struct object_id oid;
628 const char *name;
629 char merge_base_hex[GIT_MAX_HEXSZ + 1];
630 struct index_state *istate = revs->diffopt.repo->index;
632 if (revs->pending.nr != 1)
633 BUG("run_diff_index must be passed exactly one tree");
635 trace_performance_enter();
636 ent = revs->pending.objects;
638 refresh_fsmonitor(istate);
640 if (merge_base) {
641 diff_get_merge_base(revs, &oid);
642 name = oid_to_hex_r(merge_base_hex, &oid);
643 } else {
644 oidcpy(&oid, &ent->item->oid);
645 name = ent->name;
648 if (diff_cache(revs, &oid, name, cached))
649 exit(128);
651 diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
652 diffcore_fix_diff_index();
653 diffcore_std(&revs->diffopt);
654 diff_flush(&revs->diffopt);
655 trace_performance_leave("diff-index");
658 int do_diff_cache(const struct object_id *tree_oid, struct diff_options *opt)
660 struct rev_info revs;
662 repo_init_revisions(opt->repo, &revs, NULL);
663 copy_pathspec(&revs.prune_data, &opt->pathspec);
664 revs.diffopt = *opt;
665 revs.diffopt.no_free = 1;
667 if (diff_cache(&revs, tree_oid, NULL, 1))
668 exit(128);
670 release_revisions(&revs);
671 return 0;
674 int index_differs_from(struct repository *r,
675 const char *def, const struct diff_flags *flags,
676 int ita_invisible_in_index)
678 struct rev_info rev;
679 struct setup_revision_opt opt;
680 unsigned has_changes;
682 repo_init_revisions(r, &rev, NULL);
683 memset(&opt, 0, sizeof(opt));
684 opt.def = def;
685 setup_revisions(0, NULL, &rev, &opt);
686 rev.diffopt.flags.quick = 1;
687 rev.diffopt.flags.exit_with_status = 1;
688 if (flags) {
689 diff_flags_or(&rev.diffopt.flags, flags);
691 * Now that flags are merged, honor override_submodule_config
692 * and ignore_submodules from passed flags.
694 if (flags->override_submodule_config)
695 rev.diffopt.flags.ignore_submodules = flags->ignore_submodules;
697 rev.diffopt.ita_invisible_in_index = ita_invisible_in_index;
698 run_diff_index(&rev, DIFF_INDEX_CACHED);
699 has_changes = rev.diffopt.flags.has_changes;
700 release_revisions(&rev);
701 return (has_changes != 0);
704 static struct strbuf *idiff_prefix_cb(struct diff_options *opt UNUSED, void *data)
706 return data;
709 void show_interdiff(const struct object_id *oid1, const struct object_id *oid2,
710 int indent, struct diff_options *diffopt)
712 struct diff_options opts;
713 struct strbuf prefix = STRBUF_INIT;
715 memcpy(&opts, diffopt, sizeof(opts));
716 opts.output_format = DIFF_FORMAT_PATCH;
717 opts.output_prefix = idiff_prefix_cb;
718 strbuf_addchars(&prefix, ' ', indent);
719 opts.output_prefix_data = &prefix;
720 diff_setup_done(&opts);
722 diff_tree_oid(oid1, oid2, "", &opts);
723 diffcore_std(&opts);
724 diff_flush(&opts);
726 strbuf_release(&prefix);