The sixteenth batch
[git/gitster.git] / diff-lib.c
blob244468dd1a2b30c9dfea91ecbb3f001ad204f5e0
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
5 #define USE_THE_REPOSITORY_VARIABLE
6 #define DISABLE_SIGN_COMPARE_WARNINGS
8 #include "git-compat-util.h"
9 #include "commit.h"
10 #include "diff.h"
11 #include "diffcore.h"
12 #include "gettext.h"
13 #include "hash.h"
14 #include "hex.h"
15 #include "object-name.h"
16 #include "read-cache.h"
17 #include "revision.h"
18 #include "cache-tree.h"
19 #include "unpack-trees.h"
20 #include "refs.h"
21 #include "repository.h"
22 #include "submodule.h"
23 #include "symlinks.h"
24 #include "trace.h"
25 #include "dir.h"
26 #include "fsmonitor.h"
27 #include "commit-reach.h"
30 * diff-files
34 * Has the work tree entity been removed?
36 * Return 1 if it was removed from the work tree, 0 if an entity to be
37 * compared with the cache entry ce still exists (the latter includes
38 * the case where a directory that is not a submodule repository
39 * exists for ce that is a submodule -- it is a submodule that is not
40 * checked out). Return negative for an error.
42 static int check_removed(const struct cache_entry *ce, struct stat *st)
44 int stat_err;
46 if (!(ce->ce_flags & CE_FSMONITOR_VALID))
47 stat_err = lstat(ce->name, st);
48 else
49 stat_err = fake_lstat(ce, st);
50 if (stat_err < 0) {
51 if (!is_missing_file_error(errno))
52 return -1;
53 return 1;
56 if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
57 return 1;
58 if (S_ISDIR(st->st_mode)) {
59 struct object_id sub;
62 * If ce is already a gitlink, we can have a plain
63 * directory (i.e. the submodule is not checked out),
64 * or a checked out submodule. Either case this is not
65 * a case where something was removed from the work tree,
66 * so we will return 0.
68 * Otherwise, if the directory is not a submodule
69 * repository, that means ce which was a blob turned into
70 * a directory --- the blob was removed!
72 if (!S_ISGITLINK(ce->ce_mode) &&
73 repo_resolve_gitlink_ref(the_repository, ce->name,
74 "HEAD", &sub))
75 return 1;
77 return 0;
81 * Has a file changed or has a submodule new commits or a dirty work tree?
83 * Return 1 when changes are detected, 0 otherwise. If the DIRTY_SUBMODULES
84 * option is set, the caller does not only want to know if a submodule is
85 * modified at all but wants to know all the conditions that are met (new
86 * commits, untracked content and/or modified content).
88 static int match_stat_with_submodule(struct diff_options *diffopt,
89 const struct cache_entry *ce,
90 struct stat *st, unsigned ce_option,
91 unsigned *dirty_submodule)
93 int changed = ie_match_stat(diffopt->repo->index, ce, st, ce_option);
94 if (S_ISGITLINK(ce->ce_mode)) {
95 struct diff_flags orig_flags = diffopt->flags;
96 if (!diffopt->flags.override_submodule_config)
97 set_diffopt_flags_from_submodule_config(diffopt, ce->name);
98 if (diffopt->flags.ignore_submodules)
99 changed = 0;
100 else if (!diffopt->flags.ignore_dirty_submodules &&
101 (!changed || diffopt->flags.dirty_submodules))
102 *dirty_submodule = is_submodule_modified(ce->name,
103 diffopt->flags.ignore_untracked_in_submodules);
104 diffopt->flags = orig_flags;
106 return changed;
109 void run_diff_files(struct rev_info *revs, unsigned int option)
111 int entries, i;
112 int diff_unmerged_stage = revs->max_count;
113 unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
114 ? CE_MATCH_RACY_IS_DIRTY : 0);
115 uint64_t start = getnanotime();
116 struct index_state *istate = revs->diffopt.repo->index;
118 diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
120 refresh_fsmonitor(istate);
122 if (diff_unmerged_stage < 0)
123 diff_unmerged_stage = 2;
124 entries = istate->cache_nr;
125 for (i = 0; i < entries; i++) {
126 unsigned int oldmode, newmode;
127 struct cache_entry *ce = istate->cache[i];
128 int changed;
129 unsigned dirty_submodule = 0;
130 const struct object_id *old_oid, *new_oid;
132 if (diff_can_quit_early(&revs->diffopt))
133 break;
136 * NEEDSWORK:
137 * Here we filter with pathspec but the result is further
138 * filtered out when --relative is in effect. To end-users,
139 * a pathspec element that matched only to paths outside the
140 * current directory is like not matching anything at all;
141 * the handling of ps_matched[] here may become problematic
142 * if/when we add the "--error-unmatch" option to "git diff".
144 if (!ce_path_match(istate, ce, &revs->prune_data, revs->ps_matched))
145 continue;
147 if (revs->diffopt.prefix &&
148 strncmp(ce->name, revs->diffopt.prefix, revs->diffopt.prefix_length))
149 continue;
151 if (ce_stage(ce)) {
152 struct combine_diff_path *dpath;
153 struct diff_filepair *pair;
154 unsigned int wt_mode = 0;
155 int num_compare_stages = 0;
156 struct stat st;
158 changed = check_removed(ce, &st);
159 if (!changed)
160 wt_mode = ce_mode_from_stat(ce, st.st_mode);
161 else {
162 if (changed < 0) {
163 perror(ce->name);
164 continue;
166 wt_mode = 0;
170 * Allocate space for two parents, which will come from
171 * index stages #2 and #3, if present. Below we'll fill
172 * these from (stage - 2).
174 dpath = combine_diff_path_new(ce->name, ce_namelen(ce),
175 wt_mode, null_oid(the_hash_algo), 2);
177 while (i < entries) {
178 struct cache_entry *nce = istate->cache[i];
179 int stage;
181 if (strcmp(ce->name, nce->name))
182 break;
184 /* Stage #2 (ours) is the first parent,
185 * stage #3 (theirs) is the second.
187 stage = ce_stage(nce);
188 if (2 <= stage) {
189 int mode = nce->ce_mode;
190 num_compare_stages++;
191 oidcpy(&dpath->parent[stage - 2].oid,
192 &nce->oid);
193 dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
194 dpath->parent[stage-2].status =
195 DIFF_STATUS_MODIFIED;
198 /* diff against the proper unmerged stage */
199 if (stage == diff_unmerged_stage)
200 ce = nce;
201 i++;
204 * Compensate for loop update
206 i--;
208 if (revs->combine_merges && num_compare_stages == 2) {
209 show_combined_diff(dpath, 2, revs);
210 free(dpath);
211 continue;
213 FREE_AND_NULL(dpath);
216 * Show the diff for the 'ce' if we found the one
217 * from the desired stage.
219 pair = diff_unmerge(&revs->diffopt, ce->name);
220 if (wt_mode)
221 pair->two->mode = wt_mode;
222 if (ce_stage(ce) != diff_unmerged_stage)
223 continue;
226 if (ce_uptodate(ce) || ce_skip_worktree(ce))
227 continue;
230 * When CE_VALID is set (via "update-index --assume-unchanged"
231 * or via adding paths while core.ignorestat is set to true),
232 * the user has promised that the working tree file for that
233 * path will not be modified. When CE_FSMONITOR_VALID is true,
234 * the fsmonitor knows that the path hasn't been modified since
235 * we refreshed the cached stat information. In either case,
236 * we do not have to stat to see if the path has been removed
237 * or modified.
239 if (ce->ce_flags & (CE_VALID | CE_FSMONITOR_VALID)) {
240 changed = 0;
241 newmode = ce->ce_mode;
242 } else {
243 struct stat st;
245 changed = check_removed(ce, &st);
246 if (changed) {
247 if (changed < 0) {
248 perror(ce->name);
249 continue;
251 diff_addremove(&revs->diffopt, '-', ce->ce_mode,
252 &ce->oid,
253 !is_null_oid(&ce->oid),
254 ce->name, 0);
255 continue;
256 } else if (revs->diffopt.ita_invisible_in_index &&
257 ce_intent_to_add(ce)) {
258 newmode = ce_mode_from_stat(ce, st.st_mode);
259 diff_addremove(&revs->diffopt, '+', newmode,
260 null_oid(the_hash_algo), 0, ce->name, 0);
261 continue;
264 changed = match_stat_with_submodule(&revs->diffopt, ce, &st,
265 ce_option, &dirty_submodule);
266 newmode = ce_mode_from_stat(ce, st.st_mode);
269 if (!changed && !dirty_submodule) {
270 ce_mark_uptodate(ce);
271 mark_fsmonitor_valid(istate, ce);
272 if (!revs->diffopt.flags.find_copies_harder)
273 continue;
275 oldmode = ce->ce_mode;
276 old_oid = &ce->oid;
277 new_oid = changed ? null_oid(the_hash_algo) : &ce->oid;
278 diff_change(&revs->diffopt, oldmode, newmode,
279 old_oid, new_oid,
280 !is_null_oid(old_oid),
281 !is_null_oid(new_oid),
282 ce->name, 0, dirty_submodule);
285 diffcore_std(&revs->diffopt);
286 diff_flush(&revs->diffopt);
287 trace_performance_since(start, "diff-files");
291 * diff-index
294 /* A file entry went away or appeared */
295 static void diff_index_show_file(struct rev_info *revs,
296 const char *prefix,
297 const struct cache_entry *ce,
298 const struct object_id *oid, int oid_valid,
299 unsigned int mode,
300 unsigned dirty_submodule)
302 diff_addremove(&revs->diffopt, prefix[0], mode,
303 oid, oid_valid, ce->name, dirty_submodule);
306 static int get_stat_data(const struct cache_entry *ce,
307 const struct object_id **oidp,
308 unsigned int *modep,
309 int cached, int match_missing,
310 unsigned *dirty_submodule, struct diff_options *diffopt)
312 const struct object_id *oid = &ce->oid;
313 unsigned int mode = ce->ce_mode;
315 if (!cached && !ce_uptodate(ce)) {
316 int changed;
317 struct stat st;
318 changed = check_removed(ce, &st);
319 if (changed < 0)
320 return -1;
321 else if (changed) {
322 if (match_missing) {
323 *oidp = oid;
324 *modep = mode;
325 return 0;
327 return -1;
329 changed = match_stat_with_submodule(diffopt, ce, &st,
330 0, dirty_submodule);
331 if (changed) {
332 mode = ce_mode_from_stat(ce, st.st_mode);
333 oid = null_oid(the_hash_algo);
337 *oidp = oid;
338 *modep = mode;
339 return 0;
342 static void show_new_file(struct rev_info *revs,
343 const struct cache_entry *new_file,
344 int cached, int match_missing)
346 const struct object_id *oid;
347 unsigned int mode;
348 unsigned dirty_submodule = 0;
350 if (new_file && S_ISSPARSEDIR(new_file->ce_mode)) {
351 diff_tree_oid(NULL, &new_file->oid, new_file->name, &revs->diffopt);
352 return;
356 * New file in the index: it might actually be different in
357 * the working tree.
359 if (get_stat_data(new_file, &oid, &mode, cached, match_missing,
360 &dirty_submodule, &revs->diffopt) < 0)
361 return;
363 diff_index_show_file(revs, "+", new_file, oid, !is_null_oid(oid), mode, dirty_submodule);
366 static int show_modified(struct rev_info *revs,
367 const struct cache_entry *old_entry,
368 const struct cache_entry *new_entry,
369 int report_missing,
370 int cached, int match_missing)
372 unsigned int mode, oldmode;
373 const struct object_id *oid;
374 unsigned dirty_submodule = 0;
376 assert(S_ISSPARSEDIR(old_entry->ce_mode) ==
377 S_ISSPARSEDIR(new_entry->ce_mode));
380 * If both are sparse directory entries, then expand the
381 * modifications to the file level. If only one was a sparse
382 * directory, then they appear as an add and delete instead of
383 * a modification.
385 if (S_ISSPARSEDIR(new_entry->ce_mode)) {
386 diff_tree_oid(&old_entry->oid, &new_entry->oid, new_entry->name, &revs->diffopt);
387 return 0;
390 if (get_stat_data(new_entry, &oid, &mode, cached, match_missing,
391 &dirty_submodule, &revs->diffopt) < 0) {
392 if (report_missing)
393 diff_index_show_file(revs, "-", old_entry,
394 &old_entry->oid, 1, old_entry->ce_mode,
396 return -1;
399 if (revs->combine_merges && !cached &&
400 (!oideq(oid, &old_entry->oid) || !oideq(&old_entry->oid, &new_entry->oid))) {
401 struct combine_diff_path *p;
403 p = combine_diff_path_new(new_entry->name,
404 ce_namelen(new_entry),
405 mode, null_oid(the_hash_algo), 2);
406 p->parent[0].status = DIFF_STATUS_MODIFIED;
407 p->parent[0].mode = new_entry->ce_mode;
408 oidcpy(&p->parent[0].oid, &new_entry->oid);
409 p->parent[1].status = DIFF_STATUS_MODIFIED;
410 p->parent[1].mode = old_entry->ce_mode;
411 oidcpy(&p->parent[1].oid, &old_entry->oid);
412 show_combined_diff(p, 2, revs);
413 free(p);
414 return 0;
417 oldmode = old_entry->ce_mode;
418 if (mode == oldmode && oideq(oid, &old_entry->oid) && !dirty_submodule &&
419 !revs->diffopt.flags.find_copies_harder)
420 return 0;
422 diff_change(&revs->diffopt, oldmode, mode,
423 &old_entry->oid, oid, 1, !is_null_oid(oid),
424 old_entry->name, 0, dirty_submodule);
425 return 0;
429 * This gets a mix of an existing index and a tree, one pathname entry
430 * at a time. The index entry may be a single stage-0 one, but it could
431 * also be multiple unmerged entries (in which case idx_pos/idx_nr will
432 * give you the position and number of entries in the index).
434 static void do_oneway_diff(struct unpack_trees_options *o,
435 const struct cache_entry *idx,
436 const struct cache_entry *tree)
438 struct rev_info *revs = o->unpack_data;
439 int match_missing, cached;
442 * i-t-a entries do not actually exist in the index (if we're
443 * looking at its content)
445 if (o->index_only &&
446 revs->diffopt.ita_invisible_in_index &&
447 idx && ce_intent_to_add(idx)) {
448 idx = NULL;
449 if (!tree)
450 return; /* nothing to diff.. */
453 /* if the entry is not checked out, don't examine work tree */
454 cached = o->index_only ||
455 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
457 match_missing = revs->match_missing;
459 if (cached && idx && ce_stage(idx)) {
460 struct diff_filepair *pair;
461 pair = diff_unmerge(&revs->diffopt, idx->name);
462 if (tree)
463 fill_filespec(pair->one, &tree->oid, 1,
464 tree->ce_mode);
465 return;
469 * Something added to the tree?
471 if (!tree) {
472 show_new_file(revs, idx, cached, match_missing);
473 return;
477 * Something removed from the tree?
479 if (!idx) {
480 if (S_ISSPARSEDIR(tree->ce_mode)) {
481 diff_tree_oid(&tree->oid, NULL, tree->name, &revs->diffopt);
482 return;
485 diff_index_show_file(revs, "-", tree, &tree->oid, 1,
486 tree->ce_mode, 0);
487 return;
490 /* Show difference between old and new */
491 show_modified(revs, tree, idx, 1, cached, match_missing);
495 * The unpack_trees() interface is designed for merging, so
496 * the different source entries are designed primarily for
497 * the source trees, with the old index being really mainly
498 * used for being replaced by the result.
500 * For diffing, the index is more important, and we only have a
501 * single tree.
503 * We're supposed to advance o->pos to skip what we have already processed.
505 * This wrapper makes it all more readable, and takes care of all
506 * the fairly complex unpack_trees() semantic requirements, including
507 * the skipping, the path matching, the type conflict cases etc.
509 static int oneway_diff(const struct cache_entry * const *src,
510 struct unpack_trees_options *o)
512 const struct cache_entry *idx = src[0];
513 const struct cache_entry *tree = src[1];
514 struct rev_info *revs = o->unpack_data;
517 * Unpack-trees generates a DF/conflict entry if
518 * there was a directory in the index and a tree
519 * in the tree. From a diff standpoint, that's a
520 * delete of the tree and a create of the file.
522 if (tree == o->df_conflict_entry)
523 tree = NULL;
525 if (ce_path_match(revs->diffopt.repo->index,
526 idx ? idx : tree,
527 &revs->prune_data, NULL)) {
528 do_oneway_diff(o, idx, tree);
529 if (diff_can_quit_early(&revs->diffopt)) {
530 o->exiting_early = 1;
531 return -1;
535 return 0;
538 static int diff_cache(struct rev_info *revs,
539 const struct object_id *tree_oid,
540 const char *tree_name,
541 int cached)
543 struct tree *tree;
544 struct tree_desc t;
545 struct unpack_trees_options opts;
547 tree = parse_tree_indirect(tree_oid);
548 if (!tree)
549 return error("bad tree object %s",
550 tree_name ? tree_name : oid_to_hex(tree_oid));
551 memset(&opts, 0, sizeof(opts));
552 opts.head_idx = 1;
553 opts.index_only = cached;
554 opts.diff_index_cached = (cached &&
555 !revs->diffopt.flags.find_copies_harder);
556 opts.merge = 1;
557 opts.fn = oneway_diff;
558 opts.unpack_data = revs;
559 opts.src_index = revs->diffopt.repo->index;
560 opts.dst_index = NULL;
561 opts.pathspec = &revs->diffopt.pathspec;
562 opts.pathspec->recursive = 1;
564 init_tree_desc(&t, &tree->object.oid, tree->buffer, tree->size);
565 return unpack_trees(1, &t, &opts);
568 void diff_get_merge_base(const struct rev_info *revs, struct object_id *mb)
570 int i;
571 struct commit *mb_child[2] = {0};
572 struct commit_list *merge_bases = NULL;
574 for (i = 0; i < revs->pending.nr; i++) {
575 struct object *obj = revs->pending.objects[i].item;
576 if (obj->flags)
577 die(_("--merge-base does not work with ranges"));
581 * This check must go after the for loop above because A...B
582 * ranges produce three pending commits, resulting in a
583 * misleading error message.
585 if (revs->pending.nr < 1 || revs->pending.nr > 2)
586 BUG("unexpected revs->pending.nr: %d", revs->pending.nr);
588 for (i = 0; i < revs->pending.nr; i++)
589 mb_child[i] = lookup_commit_reference(the_repository, &revs->pending.objects[i].item->oid);
590 if (revs->pending.nr == 1) {
591 struct object_id oid;
593 if (repo_get_oid(the_repository, "HEAD", &oid))
594 die(_("unable to get HEAD"));
596 mb_child[1] = lookup_commit_reference(the_repository, &oid);
599 if (repo_get_merge_bases(the_repository, mb_child[0], mb_child[1], &merge_bases) < 0)
600 exit(128);
601 if (!merge_bases)
602 die(_("no merge base found"));
603 if (merge_bases->next)
604 die(_("multiple merge bases found"));
606 oidcpy(mb, &merge_bases->item->object.oid);
608 free_commit_list(merge_bases);
611 void run_diff_index(struct rev_info *revs, unsigned int option)
613 struct object_array_entry *ent;
614 int cached = !!(option & DIFF_INDEX_CACHED);
615 int merge_base = !!(option & DIFF_INDEX_MERGE_BASE);
616 struct object_id oid;
617 const char *name;
618 char merge_base_hex[GIT_MAX_HEXSZ + 1];
619 struct index_state *istate = revs->diffopt.repo->index;
621 if (revs->pending.nr != 1)
622 BUG("run_diff_index must be passed exactly one tree");
624 trace_performance_enter();
625 ent = revs->pending.objects;
627 refresh_fsmonitor(istate);
629 if (merge_base) {
630 diff_get_merge_base(revs, &oid);
631 name = oid_to_hex_r(merge_base_hex, &oid);
632 } else {
633 oidcpy(&oid, &ent->item->oid);
634 name = ent->name;
637 if (diff_cache(revs, &oid, name, cached))
638 exit(128);
640 diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
641 diffcore_fix_diff_index();
642 diffcore_std(&revs->diffopt);
643 diff_flush(&revs->diffopt);
644 trace_performance_leave("diff-index");
647 int do_diff_cache(const struct object_id *tree_oid, struct diff_options *opt)
649 struct rev_info revs;
651 repo_init_revisions(opt->repo, &revs, NULL);
652 copy_pathspec(&revs.prune_data, &opt->pathspec);
653 diff_free(&revs.diffopt);
654 revs.diffopt = *opt;
655 revs.diffopt.no_free = 1;
657 if (diff_cache(&revs, tree_oid, NULL, 1))
658 exit(128);
660 release_revisions(&revs);
661 return 0;
664 int index_differs_from(struct repository *r,
665 const char *def, const struct diff_flags *flags,
666 int ita_invisible_in_index)
668 struct rev_info rev;
669 struct setup_revision_opt opt;
670 unsigned has_changes;
672 repo_init_revisions(r, &rev, NULL);
673 memset(&opt, 0, sizeof(opt));
674 opt.def = def;
675 setup_revisions(0, NULL, &rev, &opt);
676 rev.diffopt.flags.quick = 1;
677 rev.diffopt.flags.exit_with_status = 1;
678 if (flags) {
679 diff_flags_or(&rev.diffopt.flags, flags);
681 * Now that flags are merged, honor override_submodule_config
682 * and ignore_submodules from passed flags.
684 if (flags->override_submodule_config)
685 rev.diffopt.flags.ignore_submodules = flags->ignore_submodules;
687 rev.diffopt.ita_invisible_in_index = ita_invisible_in_index;
688 run_diff_index(&rev, DIFF_INDEX_CACHED);
689 has_changes = rev.diffopt.flags.has_changes;
690 release_revisions(&rev);
691 return (has_changes != 0);
694 static const char *idiff_prefix_cb(struct diff_options *opt UNUSED, void *data)
696 return data;
699 void show_interdiff(const struct object_id *oid1, const struct object_id *oid2,
700 int indent, struct diff_options *diffopt)
702 struct diff_options opts;
703 struct strbuf prefix = STRBUF_INIT;
705 memcpy(&opts, diffopt, sizeof(opts));
706 opts.output_format = DIFF_FORMAT_PATCH;
707 opts.output_prefix = idiff_prefix_cb;
708 strbuf_addchars(&prefix, ' ', indent);
709 opts.output_prefix_data = prefix.buf;
710 diff_setup_done(&opts);
712 diff_tree_oid(oid1, oid2, "", &opts);
713 diffcore_std(&opts);
714 diff_flush(&opts);
716 strbuf_release(&prefix);