notes doc: split up run-on sentences
[git/gitster.git] / diff-lib.c
blob4169dd8cb1329a55c7f47674ed694293185f0b2a
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 "gettext.h"
10 #include "hex.h"
11 #include "revision.h"
12 #include "cache-tree.h"
13 #include "unpack-trees.h"
14 #include "refs.h"
15 #include "submodule.h"
16 #include "dir.h"
17 #include "fsmonitor.h"
18 #include "commit-reach.h"
21 * diff-files
25 * Has the work tree entity been removed?
27 * Return 1 if it was removed from the work tree, 0 if an entity to be
28 * compared with the cache entry ce still exists (the latter includes
29 * the case where a directory that is not a submodule repository
30 * exists for ce that is a submodule -- it is a submodule that is not
31 * checked out). Return negative for an error.
33 static int check_removed(const struct index_state *istate, const struct cache_entry *ce, struct stat *st)
35 assert(is_fsmonitor_refreshed(istate));
36 if (!(ce->ce_flags & CE_FSMONITOR_VALID) && lstat(ce->name, st) < 0) {
37 if (!is_missing_file_error(errno))
38 return -1;
39 return 1;
41 if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
42 return 1;
43 if (S_ISDIR(st->st_mode)) {
44 struct object_id sub;
47 * If ce is already a gitlink, we can have a plain
48 * directory (i.e. the submodule is not checked out),
49 * or a checked out submodule. Either case this is not
50 * a case where something was removed from the work tree,
51 * so we will return 0.
53 * Otherwise, if the directory is not a submodule
54 * repository, that means ce which was a blob turned into
55 * a directory --- the blob was removed!
57 if (!S_ISGITLINK(ce->ce_mode) &&
58 resolve_gitlink_ref(ce->name, "HEAD", &sub))
59 return 1;
61 return 0;
65 * Has a file changed or has a submodule new commits or a dirty work tree?
67 * Return 1 when changes are detected, 0 otherwise. If the DIRTY_SUBMODULES
68 * option is set, the caller does not only want to know if a submodule is
69 * modified at all but wants to know all the conditions that are met (new
70 * commits, untracked content and/or modified content).
72 static int match_stat_with_submodule(struct diff_options *diffopt,
73 const struct cache_entry *ce,
74 struct stat *st, unsigned ce_option,
75 unsigned *dirty_submodule)
77 int changed = ie_match_stat(diffopt->repo->index, ce, st, ce_option);
78 if (S_ISGITLINK(ce->ce_mode)) {
79 struct diff_flags orig_flags = diffopt->flags;
80 if (!diffopt->flags.override_submodule_config)
81 set_diffopt_flags_from_submodule_config(diffopt, ce->name);
82 if (diffopt->flags.ignore_submodules)
83 changed = 0;
84 else if (!diffopt->flags.ignore_dirty_submodules &&
85 (!changed || diffopt->flags.dirty_submodules))
86 *dirty_submodule = is_submodule_modified(ce->name,
87 diffopt->flags.ignore_untracked_in_submodules);
88 diffopt->flags = orig_flags;
90 return changed;
93 int run_diff_files(struct rev_info *revs, unsigned int option)
95 int entries, i;
96 int diff_unmerged_stage = revs->max_count;
97 unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
98 ? CE_MATCH_RACY_IS_DIRTY : 0);
99 uint64_t start = getnanotime();
100 struct index_state *istate = revs->diffopt.repo->index;
102 diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
104 refresh_fsmonitor(istate);
106 if (diff_unmerged_stage < 0)
107 diff_unmerged_stage = 2;
108 entries = istate->cache_nr;
109 for (i = 0; i < entries; i++) {
110 unsigned int oldmode, newmode;
111 struct cache_entry *ce = istate->cache[i];
112 int changed;
113 unsigned dirty_submodule = 0;
114 const struct object_id *old_oid, *new_oid;
116 if (diff_can_quit_early(&revs->diffopt))
117 break;
119 if (!ce_path_match(istate, ce, &revs->prune_data, NULL))
120 continue;
122 if (revs->diffopt.prefix &&
123 strncmp(ce->name, revs->diffopt.prefix, revs->diffopt.prefix_length))
124 continue;
126 if (ce_stage(ce)) {
127 struct combine_diff_path *dpath;
128 struct diff_filepair *pair;
129 unsigned int wt_mode = 0;
130 int num_compare_stages = 0;
131 size_t path_len;
132 struct stat st;
134 path_len = ce_namelen(ce);
136 dpath = xmalloc(combine_diff_path_size(5, path_len));
137 dpath->path = (char *) &(dpath->parent[5]);
139 dpath->next = NULL;
140 memcpy(dpath->path, ce->name, path_len);
141 dpath->path[path_len] = '\0';
142 oidclr(&dpath->oid);
143 memset(&(dpath->parent[0]), 0,
144 sizeof(struct combine_diff_parent)*5);
146 changed = check_removed(istate, ce, &st);
147 if (!changed)
148 wt_mode = ce_mode_from_stat(ce, st.st_mode);
149 else {
150 if (changed < 0) {
151 perror(ce->name);
152 continue;
154 wt_mode = 0;
156 dpath->mode = wt_mode;
158 while (i < entries) {
159 struct cache_entry *nce = istate->cache[i];
160 int stage;
162 if (strcmp(ce->name, nce->name))
163 break;
165 /* Stage #2 (ours) is the first parent,
166 * stage #3 (theirs) is the second.
168 stage = ce_stage(nce);
169 if (2 <= stage) {
170 int mode = nce->ce_mode;
171 num_compare_stages++;
172 oidcpy(&dpath->parent[stage - 2].oid,
173 &nce->oid);
174 dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
175 dpath->parent[stage-2].status =
176 DIFF_STATUS_MODIFIED;
179 /* diff against the proper unmerged stage */
180 if (stage == diff_unmerged_stage)
181 ce = nce;
182 i++;
185 * Compensate for loop update
187 i--;
189 if (revs->combine_merges && num_compare_stages == 2) {
190 show_combined_diff(dpath, 2, revs);
191 free(dpath);
192 continue;
194 FREE_AND_NULL(dpath);
197 * Show the diff for the 'ce' if we found the one
198 * from the desired stage.
200 pair = diff_unmerge(&revs->diffopt, ce->name);
201 if (wt_mode)
202 pair->two->mode = wt_mode;
203 if (ce_stage(ce) != diff_unmerged_stage)
204 continue;
207 if (ce_uptodate(ce) || ce_skip_worktree(ce))
208 continue;
211 * When CE_VALID is set (via "update-index --assume-unchanged"
212 * or via adding paths while core.ignorestat is set to true),
213 * the user has promised that the working tree file for that
214 * path will not be modified. When CE_FSMONITOR_VALID is true,
215 * the fsmonitor knows that the path hasn't been modified since
216 * we refreshed the cached stat information. In either case,
217 * we do not have to stat to see if the path has been removed
218 * or modified.
220 if (ce->ce_flags & (CE_VALID | CE_FSMONITOR_VALID)) {
221 changed = 0;
222 newmode = ce->ce_mode;
223 } else {
224 struct stat st;
226 changed = check_removed(istate, ce, &st);
227 if (changed) {
228 if (changed < 0) {
229 perror(ce->name);
230 continue;
232 diff_addremove(&revs->diffopt, '-', ce->ce_mode,
233 &ce->oid,
234 !is_null_oid(&ce->oid),
235 ce->name, 0);
236 continue;
237 } else if (revs->diffopt.ita_invisible_in_index &&
238 ce_intent_to_add(ce)) {
239 newmode = ce_mode_from_stat(ce, st.st_mode);
240 diff_addremove(&revs->diffopt, '+', newmode,
241 null_oid(), 0, ce->name, 0);
242 continue;
245 changed = match_stat_with_submodule(&revs->diffopt, ce, &st,
246 ce_option, &dirty_submodule);
247 newmode = ce_mode_from_stat(ce, st.st_mode);
250 if (!changed && !dirty_submodule) {
251 ce_mark_uptodate(ce);
252 mark_fsmonitor_valid(istate, ce);
253 if (!revs->diffopt.flags.find_copies_harder)
254 continue;
256 oldmode = ce->ce_mode;
257 old_oid = &ce->oid;
258 new_oid = changed ? null_oid() : &ce->oid;
259 diff_change(&revs->diffopt, oldmode, newmode,
260 old_oid, new_oid,
261 !is_null_oid(old_oid),
262 !is_null_oid(new_oid),
263 ce->name, 0, dirty_submodule);
266 diffcore_std(&revs->diffopt);
267 diff_flush(&revs->diffopt);
268 trace_performance_since(start, "diff-files");
269 return 0;
273 * diff-index
276 /* A file entry went away or appeared */
277 static void diff_index_show_file(struct rev_info *revs,
278 const char *prefix,
279 const struct cache_entry *ce,
280 const struct object_id *oid, int oid_valid,
281 unsigned int mode,
282 unsigned dirty_submodule)
284 diff_addremove(&revs->diffopt, prefix[0], mode,
285 oid, oid_valid, ce->name, dirty_submodule);
288 static int get_stat_data(const struct index_state *istate,
289 const struct cache_entry *ce,
290 const struct object_id **oidp,
291 unsigned int *modep,
292 int cached, int match_missing,
293 unsigned *dirty_submodule, struct diff_options *diffopt)
295 const struct object_id *oid = &ce->oid;
296 unsigned int mode = ce->ce_mode;
298 if (!cached && !ce_uptodate(ce)) {
299 int changed;
300 struct stat st;
301 changed = check_removed(istate, ce, &st);
302 if (changed < 0)
303 return -1;
304 else if (changed) {
305 if (match_missing) {
306 *oidp = oid;
307 *modep = mode;
308 return 0;
310 return -1;
312 changed = match_stat_with_submodule(diffopt, ce, &st,
313 0, dirty_submodule);
314 if (changed) {
315 mode = ce_mode_from_stat(ce, st.st_mode);
316 oid = null_oid();
320 *oidp = oid;
321 *modep = mode;
322 return 0;
325 static void show_new_file(struct rev_info *revs,
326 const struct cache_entry *new_file,
327 int cached, int match_missing)
329 const struct object_id *oid;
330 unsigned int mode;
331 unsigned dirty_submodule = 0;
332 struct index_state *istate = revs->diffopt.repo->index;
334 if (new_file && S_ISSPARSEDIR(new_file->ce_mode)) {
335 diff_tree_oid(NULL, &new_file->oid, new_file->name, &revs->diffopt);
336 return;
340 * New file in the index: it might actually be different in
341 * the working tree.
343 if (get_stat_data(istate, new_file, &oid, &mode, cached, match_missing,
344 &dirty_submodule, &revs->diffopt) < 0)
345 return;
347 diff_index_show_file(revs, "+", new_file, oid, !is_null_oid(oid), mode, dirty_submodule);
350 static int show_modified(struct rev_info *revs,
351 const struct cache_entry *old_entry,
352 const struct cache_entry *new_entry,
353 int report_missing,
354 int cached, int match_missing)
356 unsigned int mode, oldmode;
357 const struct object_id *oid;
358 unsigned dirty_submodule = 0;
359 struct index_state *istate = revs->diffopt.repo->index;
361 assert(S_ISSPARSEDIR(old_entry->ce_mode) ==
362 S_ISSPARSEDIR(new_entry->ce_mode));
365 * If both are sparse directory entries, then expand the
366 * modifications to the file level. If only one was a sparse
367 * directory, then they appear as an add and delete instead of
368 * a modification.
370 if (S_ISSPARSEDIR(new_entry->ce_mode)) {
371 diff_tree_oid(&old_entry->oid, &new_entry->oid, new_entry->name, &revs->diffopt);
372 return 0;
375 if (get_stat_data(istate, new_entry, &oid, &mode, cached, match_missing,
376 &dirty_submodule, &revs->diffopt) < 0) {
377 if (report_missing)
378 diff_index_show_file(revs, "-", old_entry,
379 &old_entry->oid, 1, old_entry->ce_mode,
381 return -1;
384 if (revs->combine_merges && !cached &&
385 (!oideq(oid, &old_entry->oid) || !oideq(&old_entry->oid, &new_entry->oid))) {
386 struct combine_diff_path *p;
387 int pathlen = ce_namelen(new_entry);
389 p = xmalloc(combine_diff_path_size(2, pathlen));
390 p->path = (char *) &p->parent[2];
391 p->next = NULL;
392 memcpy(p->path, new_entry->name, pathlen);
393 p->path[pathlen] = 0;
394 p->mode = mode;
395 oidclr(&p->oid);
396 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
397 p->parent[0].status = DIFF_STATUS_MODIFIED;
398 p->parent[0].mode = new_entry->ce_mode;
399 oidcpy(&p->parent[0].oid, &new_entry->oid);
400 p->parent[1].status = DIFF_STATUS_MODIFIED;
401 p->parent[1].mode = old_entry->ce_mode;
402 oidcpy(&p->parent[1].oid, &old_entry->oid);
403 show_combined_diff(p, 2, revs);
404 free(p);
405 return 0;
408 oldmode = old_entry->ce_mode;
409 if (mode == oldmode && oideq(oid, &old_entry->oid) && !dirty_submodule &&
410 !revs->diffopt.flags.find_copies_harder)
411 return 0;
413 diff_change(&revs->diffopt, oldmode, mode,
414 &old_entry->oid, oid, 1, !is_null_oid(oid),
415 old_entry->name, 0, dirty_submodule);
416 return 0;
420 * This gets a mix of an existing index and a tree, one pathname entry
421 * at a time. The index entry may be a single stage-0 one, but it could
422 * also be multiple unmerged entries (in which case idx_pos/idx_nr will
423 * give you the position and number of entries in the index).
425 static void do_oneway_diff(struct unpack_trees_options *o,
426 const struct cache_entry *idx,
427 const struct cache_entry *tree)
429 struct rev_info *revs = o->unpack_data;
430 int match_missing, cached;
433 * i-t-a entries do not actually exist in the index (if we're
434 * looking at its content)
436 if (o->index_only &&
437 revs->diffopt.ita_invisible_in_index &&
438 idx && ce_intent_to_add(idx)) {
439 idx = NULL;
440 if (!tree)
441 return; /* nothing to diff.. */
444 /* if the entry is not checked out, don't examine work tree */
445 cached = o->index_only ||
446 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
448 match_missing = revs->match_missing;
450 if (cached && idx && ce_stage(idx)) {
451 struct diff_filepair *pair;
452 pair = diff_unmerge(&revs->diffopt, idx->name);
453 if (tree)
454 fill_filespec(pair->one, &tree->oid, 1,
455 tree->ce_mode);
456 return;
460 * Something added to the tree?
462 if (!tree) {
463 show_new_file(revs, idx, cached, match_missing);
464 return;
468 * Something removed from the tree?
470 if (!idx) {
471 if (S_ISSPARSEDIR(tree->ce_mode)) {
472 diff_tree_oid(&tree->oid, NULL, tree->name, &revs->diffopt);
473 return;
476 diff_index_show_file(revs, "-", tree, &tree->oid, 1,
477 tree->ce_mode, 0);
478 return;
481 /* Show difference between old and new */
482 show_modified(revs, tree, idx, 1, cached, match_missing);
486 * The unpack_trees() interface is designed for merging, so
487 * the different source entries are designed primarily for
488 * the source trees, with the old index being really mainly
489 * used for being replaced by the result.
491 * For diffing, the index is more important, and we only have a
492 * single tree.
494 * We're supposed to advance o->pos to skip what we have already processed.
496 * This wrapper makes it all more readable, and takes care of all
497 * the fairly complex unpack_trees() semantic requirements, including
498 * the skipping, the path matching, the type conflict cases etc.
500 static int oneway_diff(const struct cache_entry * const *src,
501 struct unpack_trees_options *o)
503 const struct cache_entry *idx = src[0];
504 const struct cache_entry *tree = src[1];
505 struct rev_info *revs = o->unpack_data;
508 * Unpack-trees generates a DF/conflict entry if
509 * there was a directory in the index and a tree
510 * in the tree. From a diff standpoint, that's a
511 * delete of the tree and a create of the file.
513 if (tree == o->df_conflict_entry)
514 tree = NULL;
516 if (ce_path_match(revs->diffopt.repo->index,
517 idx ? idx : tree,
518 &revs->prune_data, NULL)) {
519 do_oneway_diff(o, idx, tree);
520 if (diff_can_quit_early(&revs->diffopt)) {
521 o->exiting_early = 1;
522 return -1;
526 return 0;
529 static int diff_cache(struct rev_info *revs,
530 const struct object_id *tree_oid,
531 const char *tree_name,
532 int cached)
534 struct tree *tree;
535 struct tree_desc t;
536 struct unpack_trees_options opts;
538 tree = parse_tree_indirect(tree_oid);
539 if (!tree)
540 return error("bad tree object %s",
541 tree_name ? tree_name : oid_to_hex(tree_oid));
542 memset(&opts, 0, sizeof(opts));
543 opts.head_idx = 1;
544 opts.index_only = cached;
545 opts.diff_index_cached = (cached &&
546 !revs->diffopt.flags.find_copies_harder);
547 opts.merge = 1;
548 opts.fn = oneway_diff;
549 opts.unpack_data = revs;
550 opts.src_index = revs->diffopt.repo->index;
551 opts.dst_index = NULL;
552 opts.pathspec = &revs->diffopt.pathspec;
553 opts.pathspec->recursive = 1;
555 init_tree_desc(&t, tree->buffer, tree->size);
556 return unpack_trees(1, &t, &opts);
559 void diff_get_merge_base(const struct rev_info *revs, struct object_id *mb)
561 int i;
562 struct commit *mb_child[2] = {0};
563 struct commit_list *merge_bases;
565 for (i = 0; i < revs->pending.nr; i++) {
566 struct object *obj = revs->pending.objects[i].item;
567 if (obj->flags)
568 die(_("--merge-base does not work with ranges"));
569 if (obj->type != OBJ_COMMIT)
570 die(_("--merge-base only works with commits"));
574 * This check must go after the for loop above because A...B
575 * ranges produce three pending commits, resulting in a
576 * misleading error message.
578 if (revs->pending.nr < 1 || revs->pending.nr > 2)
579 BUG("unexpected revs->pending.nr: %d", revs->pending.nr);
581 for (i = 0; i < revs->pending.nr; i++)
582 mb_child[i] = lookup_commit_reference(the_repository, &revs->pending.objects[i].item->oid);
583 if (revs->pending.nr == 1) {
584 struct object_id oid;
586 if (repo_get_oid(the_repository, "HEAD", &oid))
587 die(_("unable to get HEAD"));
589 mb_child[1] = lookup_commit_reference(the_repository, &oid);
592 merge_bases = repo_get_merge_bases(the_repository, mb_child[0], mb_child[1]);
593 if (!merge_bases)
594 die(_("no merge base found"));
595 if (merge_bases->next)
596 die(_("multiple merge bases found"));
598 oidcpy(mb, &merge_bases->item->object.oid);
600 free_commit_list(merge_bases);
603 int run_diff_index(struct rev_info *revs, unsigned int option)
605 struct object_array_entry *ent;
606 int cached = !!(option & DIFF_INDEX_CACHED);
607 int merge_base = !!(option & DIFF_INDEX_MERGE_BASE);
608 struct object_id oid;
609 const char *name;
610 char merge_base_hex[GIT_MAX_HEXSZ + 1];
611 struct index_state *istate = revs->diffopt.repo->index;
613 if (revs->pending.nr != 1)
614 BUG("run_diff_index must be passed exactly one tree");
616 trace_performance_enter();
617 ent = revs->pending.objects;
619 refresh_fsmonitor(istate);
621 if (merge_base) {
622 diff_get_merge_base(revs, &oid);
623 name = oid_to_hex_r(merge_base_hex, &oid);
624 } else {
625 oidcpy(&oid, &ent->item->oid);
626 name = ent->name;
629 if (diff_cache(revs, &oid, name, cached))
630 exit(128);
632 diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
633 diffcore_fix_diff_index();
634 diffcore_std(&revs->diffopt);
635 diff_flush(&revs->diffopt);
636 trace_performance_leave("diff-index");
637 return 0;
640 int do_diff_cache(const struct object_id *tree_oid, struct diff_options *opt)
642 struct rev_info revs;
644 repo_init_revisions(opt->repo, &revs, NULL);
645 copy_pathspec(&revs.prune_data, &opt->pathspec);
646 diff_setup_done(&revs.diffopt);
647 revs.diffopt = *opt;
649 if (diff_cache(&revs, tree_oid, NULL, 1))
650 exit(128);
651 release_revisions(&revs);
652 return 0;
655 int index_differs_from(struct repository *r,
656 const char *def, const struct diff_flags *flags,
657 int ita_invisible_in_index)
659 struct rev_info rev;
660 struct setup_revision_opt opt;
661 unsigned has_changes;
663 repo_init_revisions(r, &rev, NULL);
664 memset(&opt, 0, sizeof(opt));
665 opt.def = def;
666 setup_revisions(0, NULL, &rev, &opt);
667 rev.diffopt.flags.quick = 1;
668 rev.diffopt.flags.exit_with_status = 1;
669 if (flags)
670 diff_flags_or(&rev.diffopt.flags, flags);
671 rev.diffopt.ita_invisible_in_index = ita_invisible_in_index;
672 run_diff_index(&rev, 1);
673 has_changes = rev.diffopt.flags.has_changes;
674 release_revisions(&rev);
675 return (has_changes != 0);
678 static struct strbuf *idiff_prefix_cb(struct diff_options *opt UNUSED, void *data)
680 return data;
683 void show_interdiff(const struct object_id *oid1, const struct object_id *oid2,
684 int indent, struct diff_options *diffopt)
686 struct diff_options opts;
687 struct strbuf prefix = STRBUF_INIT;
689 memcpy(&opts, diffopt, sizeof(opts));
690 opts.output_format = DIFF_FORMAT_PATCH;
691 opts.output_prefix = idiff_prefix_cb;
692 strbuf_addchars(&prefix, ' ', indent);
693 opts.output_prefix_data = &prefix;
694 diff_setup_done(&opts);
696 diff_tree_oid(oid1, oid2, "", &opts);
697 diffcore_std(&opts);
698 diff_flush(&opts);
700 strbuf_release(&prefix);