refs/reftable: expose auto compaction via new flag
[alt-git.git] / diff-lib.c
blob5e8717c774eff4bb2364344c440a4efb4686e4f4
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "git-compat-util.h"
5 #include "commit.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "gettext.h"
9 #include "hash.h"
10 #include "hex.h"
11 #include "object-name.h"
12 #include "read-cache.h"
13 #include "revision.h"
14 #include "cache-tree.h"
15 #include "unpack-trees.h"
16 #include "refs.h"
17 #include "repository.h"
18 #include "submodule.h"
19 #include "symlinks.h"
20 #include "trace.h"
21 #include "dir.h"
22 #include "fsmonitor.h"
23 #include "commit-reach.h"
26 * diff-files
30 * Has the work tree entity been removed?
32 * Return 1 if it was removed from the work tree, 0 if an entity to be
33 * compared with the cache entry ce still exists (the latter includes
34 * the case where a directory that is not a submodule repository
35 * exists for ce that is a submodule -- it is a submodule that is not
36 * checked out). Return negative for an error.
38 static int check_removed(const struct cache_entry *ce, struct stat *st)
40 int stat_err;
42 if (!(ce->ce_flags & CE_FSMONITOR_VALID))
43 stat_err = lstat(ce->name, st);
44 else
45 stat_err = fake_lstat(ce, st);
46 if (stat_err < 0) {
47 if (!is_missing_file_error(errno))
48 return -1;
49 return 1;
52 if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
53 return 1;
54 if (S_ISDIR(st->st_mode)) {
55 struct object_id sub;
58 * If ce is already a gitlink, we can have a plain
59 * directory (i.e. the submodule is not checked out),
60 * or a checked out submodule. Either case this is not
61 * a case where something was removed from the work tree,
62 * so we will return 0.
64 * Otherwise, if the directory is not a submodule
65 * repository, that means ce which was a blob turned into
66 * a directory --- the blob was removed!
68 if (!S_ISGITLINK(ce->ce_mode) &&
69 resolve_gitlink_ref(ce->name, "HEAD", &sub))
70 return 1;
72 return 0;
76 * Has a file changed or has a submodule new commits or a dirty work tree?
78 * Return 1 when changes are detected, 0 otherwise. If the DIRTY_SUBMODULES
79 * option is set, the caller does not only want to know if a submodule is
80 * modified at all but wants to know all the conditions that are met (new
81 * commits, untracked content and/or modified content).
83 static int match_stat_with_submodule(struct diff_options *diffopt,
84 const struct cache_entry *ce,
85 struct stat *st, unsigned ce_option,
86 unsigned *dirty_submodule)
88 int changed = ie_match_stat(diffopt->repo->index, ce, st, ce_option);
89 if (S_ISGITLINK(ce->ce_mode)) {
90 struct diff_flags orig_flags = diffopt->flags;
91 if (!diffopt->flags.override_submodule_config)
92 set_diffopt_flags_from_submodule_config(diffopt, ce->name);
93 if (diffopt->flags.ignore_submodules)
94 changed = 0;
95 else if (!diffopt->flags.ignore_dirty_submodules &&
96 (!changed || diffopt->flags.dirty_submodules))
97 *dirty_submodule = is_submodule_modified(ce->name,
98 diffopt->flags.ignore_untracked_in_submodules);
99 diffopt->flags = orig_flags;
101 return changed;
104 void run_diff_files(struct rev_info *revs, unsigned int option)
106 int entries, i;
107 int diff_unmerged_stage = revs->max_count;
108 unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
109 ? CE_MATCH_RACY_IS_DIRTY : 0);
110 uint64_t start = getnanotime();
111 struct index_state *istate = revs->diffopt.repo->index;
113 diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
115 refresh_fsmonitor(istate);
117 if (diff_unmerged_stage < 0)
118 diff_unmerged_stage = 2;
119 entries = istate->cache_nr;
120 for (i = 0; i < entries; i++) {
121 unsigned int oldmode, newmode;
122 struct cache_entry *ce = istate->cache[i];
123 int changed;
124 unsigned dirty_submodule = 0;
125 const struct object_id *old_oid, *new_oid;
127 if (diff_can_quit_early(&revs->diffopt))
128 break;
130 if (!ce_path_match(istate, ce, &revs->prune_data, NULL))
131 continue;
133 if (revs->diffopt.prefix &&
134 strncmp(ce->name, revs->diffopt.prefix, revs->diffopt.prefix_length))
135 continue;
137 if (ce_stage(ce)) {
138 struct combine_diff_path *dpath;
139 struct diff_filepair *pair;
140 unsigned int wt_mode = 0;
141 int num_compare_stages = 0;
142 size_t path_len;
143 struct stat st;
145 path_len = ce_namelen(ce);
147 dpath = xmalloc(combine_diff_path_size(5, path_len));
148 dpath->path = (char *) &(dpath->parent[5]);
150 dpath->next = NULL;
151 memcpy(dpath->path, ce->name, path_len);
152 dpath->path[path_len] = '\0';
153 oidclr(&dpath->oid);
154 memset(&(dpath->parent[0]), 0,
155 sizeof(struct combine_diff_parent)*5);
157 changed = check_removed(ce, &st);
158 if (!changed)
159 wt_mode = ce_mode_from_stat(ce, st.st_mode);
160 else {
161 if (changed < 0) {
162 perror(ce->name);
163 continue;
165 wt_mode = 0;
167 dpath->mode = wt_mode;
169 while (i < entries) {
170 struct cache_entry *nce = istate->cache[i];
171 int stage;
173 if (strcmp(ce->name, nce->name))
174 break;
176 /* Stage #2 (ours) is the first parent,
177 * stage #3 (theirs) is the second.
179 stage = ce_stage(nce);
180 if (2 <= stage) {
181 int mode = nce->ce_mode;
182 num_compare_stages++;
183 oidcpy(&dpath->parent[stage - 2].oid,
184 &nce->oid);
185 dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
186 dpath->parent[stage-2].status =
187 DIFF_STATUS_MODIFIED;
190 /* diff against the proper unmerged stage */
191 if (stage == diff_unmerged_stage)
192 ce = nce;
193 i++;
196 * Compensate for loop update
198 i--;
200 if (revs->combine_merges && num_compare_stages == 2) {
201 show_combined_diff(dpath, 2, revs);
202 free(dpath);
203 continue;
205 FREE_AND_NULL(dpath);
208 * Show the diff for the 'ce' if we found the one
209 * from the desired stage.
211 pair = diff_unmerge(&revs->diffopt, ce->name);
212 if (wt_mode)
213 pair->two->mode = wt_mode;
214 if (ce_stage(ce) != diff_unmerged_stage)
215 continue;
218 if (ce_uptodate(ce) || ce_skip_worktree(ce))
219 continue;
222 * When CE_VALID is set (via "update-index --assume-unchanged"
223 * or via adding paths while core.ignorestat is set to true),
224 * the user has promised that the working tree file for that
225 * path will not be modified. When CE_FSMONITOR_VALID is true,
226 * the fsmonitor knows that the path hasn't been modified since
227 * we refreshed the cached stat information. In either case,
228 * we do not have to stat to see if the path has been removed
229 * or modified.
231 if (ce->ce_flags & (CE_VALID | CE_FSMONITOR_VALID)) {
232 changed = 0;
233 newmode = ce->ce_mode;
234 } else {
235 struct stat st;
237 changed = check_removed(ce, &st);
238 if (changed) {
239 if (changed < 0) {
240 perror(ce->name);
241 continue;
243 diff_addremove(&revs->diffopt, '-', ce->ce_mode,
244 &ce->oid,
245 !is_null_oid(&ce->oid),
246 ce->name, 0);
247 continue;
248 } else if (revs->diffopt.ita_invisible_in_index &&
249 ce_intent_to_add(ce)) {
250 newmode = ce_mode_from_stat(ce, st.st_mode);
251 diff_addremove(&revs->diffopt, '+', newmode,
252 null_oid(), 0, ce->name, 0);
253 continue;
256 changed = match_stat_with_submodule(&revs->diffopt, ce, &st,
257 ce_option, &dirty_submodule);
258 newmode = ce_mode_from_stat(ce, st.st_mode);
261 if (!changed && !dirty_submodule) {
262 ce_mark_uptodate(ce);
263 mark_fsmonitor_valid(istate, ce);
264 if (!revs->diffopt.flags.find_copies_harder)
265 continue;
267 oldmode = ce->ce_mode;
268 old_oid = &ce->oid;
269 new_oid = changed ? null_oid() : &ce->oid;
270 diff_change(&revs->diffopt, oldmode, newmode,
271 old_oid, new_oid,
272 !is_null_oid(old_oid),
273 !is_null_oid(new_oid),
274 ce->name, 0, dirty_submodule);
277 diffcore_std(&revs->diffopt);
278 diff_flush(&revs->diffopt);
279 trace_performance_since(start, "diff-files");
283 * diff-index
286 /* A file entry went away or appeared */
287 static void diff_index_show_file(struct rev_info *revs,
288 const char *prefix,
289 const struct cache_entry *ce,
290 const struct object_id *oid, int oid_valid,
291 unsigned int mode,
292 unsigned dirty_submodule)
294 diff_addremove(&revs->diffopt, prefix[0], mode,
295 oid, oid_valid, ce->name, dirty_submodule);
298 static int get_stat_data(const struct index_state *istate,
299 const struct cache_entry *ce,
300 const struct object_id **oidp,
301 unsigned int *modep,
302 int cached, int match_missing,
303 unsigned *dirty_submodule, struct diff_options *diffopt)
305 const struct object_id *oid = &ce->oid;
306 unsigned int mode = ce->ce_mode;
308 if (!cached && !ce_uptodate(ce)) {
309 int changed;
310 struct stat st;
311 changed = check_removed(ce, &st);
312 if (changed < 0)
313 return -1;
314 else if (changed) {
315 if (match_missing) {
316 *oidp = oid;
317 *modep = mode;
318 return 0;
320 return -1;
322 changed = match_stat_with_submodule(diffopt, ce, &st,
323 0, dirty_submodule);
324 if (changed) {
325 mode = ce_mode_from_stat(ce, st.st_mode);
326 oid = null_oid();
330 *oidp = oid;
331 *modep = mode;
332 return 0;
335 static void show_new_file(struct rev_info *revs,
336 const struct cache_entry *new_file,
337 int cached, int match_missing)
339 const struct object_id *oid;
340 unsigned int mode;
341 unsigned dirty_submodule = 0;
342 struct index_state *istate = revs->diffopt.repo->index;
344 if (new_file && S_ISSPARSEDIR(new_file->ce_mode)) {
345 diff_tree_oid(NULL, &new_file->oid, new_file->name, &revs->diffopt);
346 return;
350 * New file in the index: it might actually be different in
351 * the working tree.
353 if (get_stat_data(istate, new_file, &oid, &mode, cached, match_missing,
354 &dirty_submodule, &revs->diffopt) < 0)
355 return;
357 diff_index_show_file(revs, "+", new_file, oid, !is_null_oid(oid), mode, dirty_submodule);
360 static int show_modified(struct rev_info *revs,
361 const struct cache_entry *old_entry,
362 const struct cache_entry *new_entry,
363 int report_missing,
364 int cached, int match_missing)
366 unsigned int mode, oldmode;
367 const struct object_id *oid;
368 unsigned dirty_submodule = 0;
369 struct index_state *istate = revs->diffopt.repo->index;
371 assert(S_ISSPARSEDIR(old_entry->ce_mode) ==
372 S_ISSPARSEDIR(new_entry->ce_mode));
375 * If both are sparse directory entries, then expand the
376 * modifications to the file level. If only one was a sparse
377 * directory, then they appear as an add and delete instead of
378 * a modification.
380 if (S_ISSPARSEDIR(new_entry->ce_mode)) {
381 diff_tree_oid(&old_entry->oid, &new_entry->oid, new_entry->name, &revs->diffopt);
382 return 0;
385 if (get_stat_data(istate, new_entry, &oid, &mode, cached, match_missing,
386 &dirty_submodule, &revs->diffopt) < 0) {
387 if (report_missing)
388 diff_index_show_file(revs, "-", old_entry,
389 &old_entry->oid, 1, old_entry->ce_mode,
391 return -1;
394 if (revs->combine_merges && !cached &&
395 (!oideq(oid, &old_entry->oid) || !oideq(&old_entry->oid, &new_entry->oid))) {
396 struct combine_diff_path *p;
397 int pathlen = ce_namelen(new_entry);
399 p = xmalloc(combine_diff_path_size(2, pathlen));
400 p->path = (char *) &p->parent[2];
401 p->next = NULL;
402 memcpy(p->path, new_entry->name, pathlen);
403 p->path[pathlen] = 0;
404 p->mode = mode;
405 oidclr(&p->oid);
406 memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
407 p->parent[0].status = DIFF_STATUS_MODIFIED;
408 p->parent[0].mode = new_entry->ce_mode;
409 oidcpy(&p->parent[0].oid, &new_entry->oid);
410 p->parent[1].status = DIFF_STATUS_MODIFIED;
411 p->parent[1].mode = old_entry->ce_mode;
412 oidcpy(&p->parent[1].oid, &old_entry->oid);
413 show_combined_diff(p, 2, revs);
414 free(p);
415 return 0;
418 oldmode = old_entry->ce_mode;
419 if (mode == oldmode && oideq(oid, &old_entry->oid) && !dirty_submodule &&
420 !revs->diffopt.flags.find_copies_harder)
421 return 0;
423 diff_change(&revs->diffopt, oldmode, mode,
424 &old_entry->oid, oid, 1, !is_null_oid(oid),
425 old_entry->name, 0, dirty_submodule);
426 return 0;
430 * This gets a mix of an existing index and a tree, one pathname entry
431 * at a time. The index entry may be a single stage-0 one, but it could
432 * also be multiple unmerged entries (in which case idx_pos/idx_nr will
433 * give you the position and number of entries in the index).
435 static void do_oneway_diff(struct unpack_trees_options *o,
436 const struct cache_entry *idx,
437 const struct cache_entry *tree)
439 struct rev_info *revs = o->unpack_data;
440 int match_missing, cached;
443 * i-t-a entries do not actually exist in the index (if we're
444 * looking at its content)
446 if (o->index_only &&
447 revs->diffopt.ita_invisible_in_index &&
448 idx && ce_intent_to_add(idx)) {
449 idx = NULL;
450 if (!tree)
451 return; /* nothing to diff.. */
454 /* if the entry is not checked out, don't examine work tree */
455 cached = o->index_only ||
456 (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
458 match_missing = revs->match_missing;
460 if (cached && idx && ce_stage(idx)) {
461 struct diff_filepair *pair;
462 pair = diff_unmerge(&revs->diffopt, idx->name);
463 if (tree)
464 fill_filespec(pair->one, &tree->oid, 1,
465 tree->ce_mode);
466 return;
470 * Something added to the tree?
472 if (!tree) {
473 show_new_file(revs, idx, cached, match_missing);
474 return;
478 * Something removed from the tree?
480 if (!idx) {
481 if (S_ISSPARSEDIR(tree->ce_mode)) {
482 diff_tree_oid(&tree->oid, NULL, tree->name, &revs->diffopt);
483 return;
486 diff_index_show_file(revs, "-", tree, &tree->oid, 1,
487 tree->ce_mode, 0);
488 return;
491 /* Show difference between old and new */
492 show_modified(revs, tree, idx, 1, cached, match_missing);
496 * The unpack_trees() interface is designed for merging, so
497 * the different source entries are designed primarily for
498 * the source trees, with the old index being really mainly
499 * used for being replaced by the result.
501 * For diffing, the index is more important, and we only have a
502 * single tree.
504 * We're supposed to advance o->pos to skip what we have already processed.
506 * This wrapper makes it all more readable, and takes care of all
507 * the fairly complex unpack_trees() semantic requirements, including
508 * the skipping, the path matching, the type conflict cases etc.
510 static int oneway_diff(const struct cache_entry * const *src,
511 struct unpack_trees_options *o)
513 const struct cache_entry *idx = src[0];
514 const struct cache_entry *tree = src[1];
515 struct rev_info *revs = o->unpack_data;
518 * Unpack-trees generates a DF/conflict entry if
519 * there was a directory in the index and a tree
520 * in the tree. From a diff standpoint, that's a
521 * delete of the tree and a create of the file.
523 if (tree == o->df_conflict_entry)
524 tree = NULL;
526 if (ce_path_match(revs->diffopt.repo->index,
527 idx ? idx : tree,
528 &revs->prune_data, NULL)) {
529 do_oneway_diff(o, idx, tree);
530 if (diff_can_quit_early(&revs->diffopt)) {
531 o->exiting_early = 1;
532 return -1;
536 return 0;
539 static int diff_cache(struct rev_info *revs,
540 const struct object_id *tree_oid,
541 const char *tree_name,
542 int cached)
544 struct tree *tree;
545 struct tree_desc t;
546 struct unpack_trees_options opts;
548 tree = parse_tree_indirect(tree_oid);
549 if (!tree)
550 return error("bad tree object %s",
551 tree_name ? tree_name : oid_to_hex(tree_oid));
552 memset(&opts, 0, sizeof(opts));
553 opts.head_idx = 1;
554 opts.index_only = cached;
555 opts.diff_index_cached = (cached &&
556 !revs->diffopt.flags.find_copies_harder);
557 opts.merge = 1;
558 opts.fn = oneway_diff;
559 opts.unpack_data = revs;
560 opts.src_index = revs->diffopt.repo->index;
561 opts.dst_index = NULL;
562 opts.pathspec = &revs->diffopt.pathspec;
563 opts.pathspec->recursive = 1;
565 init_tree_desc(&t, tree->buffer, tree->size);
566 return unpack_trees(1, &t, &opts);
569 void diff_get_merge_base(const struct rev_info *revs, struct object_id *mb)
571 int i;
572 struct commit *mb_child[2] = {0};
573 struct commit_list *merge_bases = NULL;
575 for (i = 0; i < revs->pending.nr; i++) {
576 struct object *obj = revs->pending.objects[i].item;
577 if (obj->flags)
578 die(_("--merge-base does not work with ranges"));
582 * This check must go after the for loop above because A...B
583 * ranges produce three pending commits, resulting in a
584 * misleading error message.
586 if (revs->pending.nr < 1 || revs->pending.nr > 2)
587 BUG("unexpected revs->pending.nr: %d", revs->pending.nr);
589 for (i = 0; i < revs->pending.nr; i++)
590 mb_child[i] = lookup_commit_reference(the_repository, &revs->pending.objects[i].item->oid);
591 if (revs->pending.nr == 1) {
592 struct object_id oid;
594 if (repo_get_oid(the_repository, "HEAD", &oid))
595 die(_("unable to get HEAD"));
597 mb_child[1] = lookup_commit_reference(the_repository, &oid);
600 if (repo_get_merge_bases(the_repository, mb_child[0], mb_child[1], &merge_bases) < 0)
601 exit(128);
602 if (!merge_bases)
603 die(_("no merge base found"));
604 if (merge_bases->next)
605 die(_("multiple merge bases found"));
607 oidcpy(mb, &merge_bases->item->object.oid);
609 free_commit_list(merge_bases);
612 void run_diff_index(struct rev_info *revs, unsigned int option)
614 struct object_array_entry *ent;
615 int cached = !!(option & DIFF_INDEX_CACHED);
616 int merge_base = !!(option & DIFF_INDEX_MERGE_BASE);
617 struct object_id oid;
618 const char *name;
619 char merge_base_hex[GIT_MAX_HEXSZ + 1];
620 struct index_state *istate = revs->diffopt.repo->index;
622 if (revs->pending.nr != 1)
623 BUG("run_diff_index must be passed exactly one tree");
625 trace_performance_enter();
626 ent = revs->pending.objects;
628 refresh_fsmonitor(istate);
630 if (merge_base) {
631 diff_get_merge_base(revs, &oid);
632 name = oid_to_hex_r(merge_base_hex, &oid);
633 } else {
634 oidcpy(&oid, &ent->item->oid);
635 name = ent->name;
638 if (diff_cache(revs, &oid, name, cached))
639 exit(128);
641 diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
642 diffcore_fix_diff_index();
643 diffcore_std(&revs->diffopt);
644 diff_flush(&revs->diffopt);
645 trace_performance_leave("diff-index");
648 int do_diff_cache(const struct object_id *tree_oid, struct diff_options *opt)
650 struct rev_info revs;
652 repo_init_revisions(opt->repo, &revs, NULL);
653 copy_pathspec(&revs.prune_data, &opt->pathspec);
654 diff_setup_done(&revs.diffopt);
655 revs.diffopt = *opt;
657 if (diff_cache(&revs, tree_oid, NULL, 1))
658 exit(128);
659 release_revisions(&revs);
660 return 0;
663 int index_differs_from(struct repository *r,
664 const char *def, const struct diff_flags *flags,
665 int ita_invisible_in_index)
667 struct rev_info rev;
668 struct setup_revision_opt opt;
669 unsigned has_changes;
671 repo_init_revisions(r, &rev, NULL);
672 memset(&opt, 0, sizeof(opt));
673 opt.def = def;
674 setup_revisions(0, NULL, &rev, &opt);
675 rev.diffopt.flags.quick = 1;
676 rev.diffopt.flags.exit_with_status = 1;
677 if (flags) {
678 diff_flags_or(&rev.diffopt.flags, flags);
680 * Now that flags are merged, honor override_submodule_config
681 * and ignore_submodules from passed flags.
683 if (flags->override_submodule_config)
684 rev.diffopt.flags.ignore_submodules = flags->ignore_submodules;
686 rev.diffopt.ita_invisible_in_index = ita_invisible_in_index;
687 run_diff_index(&rev, DIFF_INDEX_CACHED);
688 has_changes = rev.diffopt.flags.has_changes;
689 release_revisions(&rev);
690 return (has_changes != 0);
693 static struct strbuf *idiff_prefix_cb(struct diff_options *opt UNUSED, void *data)
695 return data;
698 void show_interdiff(const struct object_id *oid1, const struct object_id *oid2,
699 int indent, struct diff_options *diffopt)
701 struct diff_options opts;
702 struct strbuf prefix = STRBUF_INIT;
704 memcpy(&opts, diffopt, sizeof(opts));
705 opts.output_format = DIFF_FORMAT_PATCH;
706 opts.output_prefix = idiff_prefix_cb;
707 strbuf_addchars(&prefix, ' ', indent);
708 opts.output_prefix_data = &prefix;
709 diff_setup_done(&opts);
711 diff_tree_oid(oid1, oid2, "", &opts);
712 diffcore_std(&opts);
713 diff_flush(&opts);
715 strbuf_release(&prefix);