git

1
#define USE_THE_REPOSITORY_VARIABLE
2

3
#include "git-compat-util.h"
4
#include "abspath.h"
5
#include "repository.h"
6
#include "config.h"
7
#include "submodule-config.h"
8
#include "submodule.h"
9
#include "dir.h"
10
#include "diff.h"
11
#include "commit.h"
12
#include "environment.h"
13
#include "gettext.h"
14
#include "hex.h"
15
#include "revision.h"
16
#include "run-command.h"
17
#include "diffcore.h"
18
#include "refs.h"
19
#include "string-list.h"
20
#include "oid-array.h"
21
#include "strvec.h"
22
#include "thread-utils.h"
23
#include "path.h"
24
#include "remote.h"
25
#include "worktree.h"
26
#include "parse-options.h"
27
#include "object-file.h"
28
#include "object-name.h"
29
#include "object-store-ll.h"
30
#include "commit-reach.h"
31
#include "read-cache-ll.h"
32
#include "setup.h"
33
#include "trace2.h"
34

35
static int config_update_recurse_submodules = RECURSE_SUBMODULES_OFF;
36
static int initialized_fetch_ref_tips;
37
static struct oid_array ref_tips_before_fetch;
38
static struct oid_array ref_tips_after_fetch;
39

40
/*
41
 * Check if the .gitmodules file is unmerged. Parsing of the .gitmodules file
42
 * will be disabled because we can't guess what might be configured in
43
 * .gitmodules unless the user resolves the conflict.
44
 */
45
int is_gitmodules_unmerged(struct index_state *istate)
46
{
47
	int pos = index_name_pos(istate, GITMODULES_FILE, strlen(GITMODULES_FILE));
48
	if (pos < 0) { /* .gitmodules not found or isn't merged */
49
		pos = -1 - pos;
50
		if (istate->cache_nr > pos) {  /* there is a .gitmodules */
51
			const struct cache_entry *ce = istate->cache[pos];
52
			if (ce_namelen(ce) == strlen(GITMODULES_FILE) &&
53
			    !strcmp(ce->name, GITMODULES_FILE))
54
				return 1;
55
		}
56
	}
57

58
	return 0;
59
}
60

61
/*
62
 * Check if the .gitmodules file is safe to write.
63
 *
64
 * Writing to the .gitmodules file requires that the file exists in the
65
 * working tree or, if it doesn't, that a brand new .gitmodules file is going
66
 * to be created (i.e. it's neither in the index nor in the current branch).
67
 *
68
 * It is not safe to write to .gitmodules if it's not in the working tree but
69
 * it is in the index or in the current branch, because writing new values
70
 * (and staging them) would blindly overwrite ALL the old content.
71
 */
72
int is_writing_gitmodules_ok(void)
73
{
74
	struct object_id oid;
75
	return file_exists(GITMODULES_FILE) ||
76
		(repo_get_oid(the_repository, GITMODULES_INDEX, &oid) < 0 && repo_get_oid(the_repository, GITMODULES_HEAD, &oid) < 0);
77
}
78

79
/*
80
 * Check if the .gitmodules file has unstaged modifications.  This must be
81
 * checked before allowing modifications to the .gitmodules file with the
82
 * intention to stage them later, because when continuing we would stage the
83
 * modifications the user didn't stage herself too. That might change in a
84
 * future version when we learn to stage the changes we do ourselves without
85
 * staging any previous modifications.
86
 */
87
int is_staging_gitmodules_ok(struct index_state *istate)
88
{
89
	int pos = index_name_pos(istate, GITMODULES_FILE, strlen(GITMODULES_FILE));
90

91
	if ((pos >= 0) && (pos < istate->cache_nr)) {
92
		struct stat st;
93
		if (lstat(GITMODULES_FILE, &st) == 0 &&
94
		    ie_modified(istate, istate->cache[pos], &st, 0) & DATA_CHANGED)
95
			return 0;
96
	}
97

98
	return 1;
99
}
100

101
static int for_each_remote_ref_submodule(const char *submodule,
102
					 each_ref_fn fn, void *cb_data)
103
{
104
	return refs_for_each_remote_ref(repo_get_submodule_ref_store(the_repository,
105
								     submodule),
106
					fn, cb_data);
107
}
108

109
/*
110
 * Try to update the "path" entry in the "submodule.<name>" section of the
111
 * .gitmodules file. Return 0 only if a .gitmodules file was found, a section
112
 * with the correct path=<oldpath> setting was found and we could update it.
113
 */
114
int update_path_in_gitmodules(const char *oldpath, const char *newpath)
115
{
116
	struct strbuf entry = STRBUF_INIT;
117
	const struct submodule *submodule;
118
	int ret;
119

120
	if (!file_exists(GITMODULES_FILE)) /* Do nothing without .gitmodules */
121
		return -1;
122

123
	if (is_gitmodules_unmerged(the_repository->index))
124
		die(_("Cannot change unmerged .gitmodules, resolve merge conflicts first"));
125

126
	submodule = submodule_from_path(the_repository, null_oid(), oldpath);
127
	if (!submodule || !submodule->name) {
128
		warning(_("Could not find section in .gitmodules where path=%s"), oldpath);
129
		return -1;
130
	}
131
	strbuf_addstr(&entry, "submodule.");
132
	strbuf_addstr(&entry, submodule->name);
133
	strbuf_addstr(&entry, ".path");
134
	ret = config_set_in_gitmodules_file_gently(entry.buf, newpath);
135
	strbuf_release(&entry);
136
	return ret;
137
}
138

139
/*
140
 * Try to remove the "submodule.<name>" section from .gitmodules where the given
141
 * path is configured. Return 0 only if a .gitmodules file was found, a section
142
 * with the correct path=<path> setting was found and we could remove it.
143
 */
144
int remove_path_from_gitmodules(const char *path)
145
{
146
	struct strbuf sect = STRBUF_INIT;
147
	const struct submodule *submodule;
148

149
	if (!file_exists(GITMODULES_FILE)) /* Do nothing without .gitmodules */
150
		return -1;
151

152
	if (is_gitmodules_unmerged(the_repository->index))
153
		die(_("Cannot change unmerged .gitmodules, resolve merge conflicts first"));
154

155
	submodule = submodule_from_path(the_repository, null_oid(), path);
156
	if (!submodule || !submodule->name) {
157
		warning(_("Could not find section in .gitmodules where path=%s"), path);
158
		return -1;
159
	}
160
	strbuf_addstr(&sect, "submodule.");
161
	strbuf_addstr(&sect, submodule->name);
162
	if (repo_config_rename_section_in_file(the_repository, GITMODULES_FILE, sect.buf, NULL) < 0) {
163
		/* Maybe the user already did that, don't error out here */
164
		warning(_("Could not remove .gitmodules entry for %s"), path);
165
		strbuf_release(&sect);
166
		return -1;
167
	}
168
	strbuf_release(&sect);
169
	return 0;
170
}
171

172
void stage_updated_gitmodules(struct index_state *istate)
173
{
174
	if (add_file_to_index(istate, GITMODULES_FILE, 0))
175
		die(_("staging updated .gitmodules failed"));
176
}
177

178
static struct string_list added_submodule_odb_paths = STRING_LIST_INIT_NODUP;
179

180
void add_submodule_odb_by_path(const char *path)
181
{
182
	string_list_insert(&added_submodule_odb_paths, xstrdup(path));
183
}
184

185
int register_all_submodule_odb_as_alternates(void)
186
{
187
	int i;
188
	int ret = added_submodule_odb_paths.nr;
189

190
	for (i = 0; i < added_submodule_odb_paths.nr; i++)
191
		add_to_alternates_memory(added_submodule_odb_paths.items[i].string);
192
	if (ret) {
193
		string_list_clear(&added_submodule_odb_paths, 0);
194
		trace2_data_intmax("submodule", the_repository,
195
				   "register_all_submodule_odb_as_alternates/registered", ret);
196
		if (git_env_bool("GIT_TEST_FATAL_REGISTER_SUBMODULE_ODB", 0))
197
			BUG("register_all_submodule_odb_as_alternates() called");
198
	}
199
	return ret;
200
}
201

202
void set_diffopt_flags_from_submodule_config(struct diff_options *diffopt,
203
					     const char *path)
204
{
205
	const struct submodule *submodule = submodule_from_path(the_repository,
206
								null_oid(),
207
								path);
208
	if (submodule) {
209
		const char *ignore;
210
		char *key;
211

212
		key = xstrfmt("submodule.%s.ignore", submodule->name);
213
		if (repo_config_get_string_tmp(the_repository, key, &ignore))
214
			ignore = submodule->ignore;
215
		free(key);
216

217
		if (ignore)
218
			handle_ignore_submodules_arg(diffopt, ignore);
219
		else if (is_gitmodules_unmerged(the_repository->index))
220
			diffopt->flags.ignore_submodules = 1;
221
	}
222
}
223

224
/* Cheap function that only determines if we're interested in submodules at all */
225
int git_default_submodule_config(const char *var, const char *value,
226
				 void *cb UNUSED)
227
{
228
	if (!strcmp(var, "submodule.recurse")) {
229
		int v = git_config_bool(var, value) ?
230
			RECURSE_SUBMODULES_ON : RECURSE_SUBMODULES_OFF;
231
		config_update_recurse_submodules = v;
232
	}
233
	return 0;
234
}
235

236
int option_parse_recurse_submodules_worktree_updater(const struct option *opt,
237
						     const char *arg, int unset)
238
{
239
	if (unset) {
240
		config_update_recurse_submodules = RECURSE_SUBMODULES_OFF;
241
		return 0;
242
	}
243
	if (arg)
244
		config_update_recurse_submodules =
245
			parse_update_recurse_submodules_arg(opt->long_name,
246
							    arg);
247
	else
248
		config_update_recurse_submodules = RECURSE_SUBMODULES_ON;
249

250
	return 0;
251
}
252

253
/*
254
 * Determine if a submodule has been initialized at a given 'path'
255
 */
256
/*
257
 * NEEDSWORK: Emit a warning if submodule.active exists, but is valueless,
258
 * ie, the config looks like: "[submodule] active\n".
259
 * Since that is an invalid pathspec, we should inform the user.
260
 */
261
int is_tree_submodule_active(struct repository *repo,
262
			     const struct object_id *treeish_name,
263
			     const char *path)
264
{
265
	int ret = 0;
266
	char *key = NULL;
267
	char *value = NULL;
268
	const struct string_list *sl;
269
	const struct submodule *module;
270

271
	module = submodule_from_path(repo, treeish_name, path);
272

273
	/* early return if there isn't a path->module mapping */
274
	if (!module)
275
		return 0;
276

277
	/* submodule.<name>.active is set */
278
	key = xstrfmt("submodule.%s.active", module->name);
279
	if (!repo_config_get_bool(repo, key, &ret)) {
280
		free(key);
281
		return ret;
282
	}
283
	free(key);
284

285
	/* submodule.active is set */
286
	if (!repo_config_get_string_multi(repo, "submodule.active", &sl)) {
287
		struct pathspec ps;
288
		struct strvec args = STRVEC_INIT;
289
		const struct string_list_item *item;
290

291
		for_each_string_list_item(item, sl) {
292
			strvec_push(&args, item->string);
293
		}
294

295
		parse_pathspec(&ps, 0, 0, NULL, args.v);
296
		ret = match_pathspec(repo->index, &ps, path, strlen(path), 0, NULL, 1);
297

298
		strvec_clear(&args);
299
		clear_pathspec(&ps);
300
		return ret;
301
	}
302

303
	/* fallback to checking if the URL is set */
304
	key = xstrfmt("submodule.%s.url", module->name);
305
	ret = !repo_config_get_string(repo, key, &value);
306

307
	free(value);
308
	free(key);
309
	return ret;
310
}
311

312
int is_submodule_active(struct repository *repo, const char *path)
313
{
314
	return is_tree_submodule_active(repo, null_oid(), path);
315
}
316

317
int is_submodule_populated_gently(const char *path, int *return_error_code)
318
{
319
	int ret = 0;
320
	char *gitdir = xstrfmt("%s/.git", path);
321

322
	if (resolve_gitdir_gently(gitdir, return_error_code))
323
		ret = 1;
324

325
	free(gitdir);
326
	return ret;
327
}
328

329
/*
330
 * Dies if the provided 'prefix' corresponds to an unpopulated submodule
331
 */
332
void die_in_unpopulated_submodule(struct index_state *istate,
333
				  const char *prefix)
334
{
335
	int i, prefixlen;
336

337
	if (!prefix)
338
		return;
339

340
	prefixlen = strlen(prefix);
341

342
	for (i = 0; i < istate->cache_nr; i++) {
343
		struct cache_entry *ce = istate->cache[i];
344
		int ce_len = ce_namelen(ce);
345

346
		if (!S_ISGITLINK(ce->ce_mode))
347
			continue;
348
		if (prefixlen <= ce_len)
349
			continue;
350
		if (strncmp(ce->name, prefix, ce_len))
351
			continue;
352
		if (prefix[ce_len] != '/')
353
			continue;
354

355
		die(_("in unpopulated submodule '%s'"), ce->name);
356
	}
357
}
358

359
/*
360
 * Dies if any paths in the provided pathspec descends into a submodule
361
 */
362
void die_path_inside_submodule(struct index_state *istate,
363
			       const struct pathspec *ps)
364
{
365
	int i, j;
366

367
	for (i = 0; i < istate->cache_nr; i++) {
368
		struct cache_entry *ce = istate->cache[i];
369
		int ce_len = ce_namelen(ce);
370

371
		if (!S_ISGITLINK(ce->ce_mode))
372
			continue;
373

374
		for (j = 0; j < ps->nr ; j++) {
375
			const struct pathspec_item *item = &ps->items[j];
376

377
			if (item->len <= ce_len)
378
				continue;
379
			if (item->match[ce_len] != '/')
380
				continue;
381
			if (strncmp(ce->name, item->match, ce_len))
382
				continue;
383
			if (item->len == ce_len + 1)
384
				continue;
385

386
			die(_("Pathspec '%s' is in submodule '%.*s'"),
387
			    item->original, ce_len, ce->name);
388
		}
389
	}
390
}
391

392
enum submodule_update_type parse_submodule_update_type(const char *value)
393
{
394
	if (!strcmp(value, "none"))
395
		return SM_UPDATE_NONE;
396
	else if (!strcmp(value, "checkout"))
397
		return SM_UPDATE_CHECKOUT;
398
	else if (!strcmp(value, "rebase"))
399
		return SM_UPDATE_REBASE;
400
	else if (!strcmp(value, "merge"))
401
		return SM_UPDATE_MERGE;
402
	else if (*value == '!')
403
		return SM_UPDATE_COMMAND;
404
	else
405
		return SM_UPDATE_UNSPECIFIED;
406
}
407

408
int parse_submodule_update_strategy(const char *value,
409
		struct submodule_update_strategy *dst)
410
{
411
	enum submodule_update_type type;
412

413
	free((void*)dst->command);
414
	dst->command = NULL;
415

416
	type = parse_submodule_update_type(value);
417
	if (type == SM_UPDATE_UNSPECIFIED)
418
		return -1;
419

420
	dst->type = type;
421
	if (type == SM_UPDATE_COMMAND)
422
		dst->command = xstrdup(value + 1);
423

424
	return 0;
425
}
426

427
const char *submodule_update_type_to_string(enum submodule_update_type type)
428
{
429
	switch (type) {
430
	case SM_UPDATE_CHECKOUT:
431
		return "checkout";
432
	case SM_UPDATE_MERGE:
433
		return "merge";
434
	case SM_UPDATE_REBASE:
435
		return "rebase";
436
	case SM_UPDATE_NONE:
437
		return "none";
438
	case SM_UPDATE_UNSPECIFIED:
439
	case SM_UPDATE_COMMAND:
440
		BUG("init_submodule() should handle type %d", type);
441
	default:
442
		BUG("unexpected update strategy type: %d", type);
443
	}
444
}
445

446
void handle_ignore_submodules_arg(struct diff_options *diffopt,
447
				  const char *arg)
448
{
449
	diffopt->flags.ignore_submodule_set = 1;
450
	diffopt->flags.ignore_submodules = 0;
451
	diffopt->flags.ignore_untracked_in_submodules = 0;
452
	diffopt->flags.ignore_dirty_submodules = 0;
453

454
	if (!strcmp(arg, "all"))
455
		diffopt->flags.ignore_submodules = 1;
456
	else if (!strcmp(arg, "untracked"))
457
		diffopt->flags.ignore_untracked_in_submodules = 1;
458
	else if (!strcmp(arg, "dirty"))
459
		diffopt->flags.ignore_dirty_submodules = 1;
460
	else if (strcmp(arg, "none"))
461
		die(_("bad --ignore-submodules argument: %s"), arg);
462
	/*
463
	 * Please update _git_status() in git-completion.bash when you
464
	 * add new options
465
	 */
466
}
467

468
static int prepare_submodule_diff_summary(struct repository *r, struct rev_info *rev,
469
					  const char *path,
470
					  struct commit *left, struct commit *right,
471
					  struct commit_list *merge_bases)
472
{
473
	struct commit_list *list;
474

475
	repo_init_revisions(r, rev, NULL);
476
	setup_revisions(0, NULL, rev, NULL);
477
	rev->left_right = 1;
478
	rev->first_parent_only = 1;
479
	left->object.flags |= SYMMETRIC_LEFT;
480
	add_pending_object(rev, &left->object, path);
481
	add_pending_object(rev, &right->object, path);
482
	for (list = merge_bases; list; list = list->next) {
483
		list->item->object.flags |= UNINTERESTING;
484
		add_pending_object(rev, &list->item->object,
485
			oid_to_hex(&list->item->object.oid));
486
	}
487
	return prepare_revision_walk(rev);
488
}
489

490
static void print_submodule_diff_summary(struct repository *r, struct rev_info *rev, struct diff_options *o)
491
{
492
	static const char format[] = "  %m %s";
493
	struct strbuf sb = STRBUF_INIT;
494
	struct commit *commit;
495

496
	while ((commit = get_revision(rev))) {
497
		struct pretty_print_context ctx = {0};
498
		ctx.date_mode = rev->date_mode;
499
		ctx.output_encoding = get_log_output_encoding();
500
		strbuf_setlen(&sb, 0);
501
		repo_format_commit_message(r, commit, format, &sb,
502
				      &ctx);
503
		strbuf_addch(&sb, '\n');
504
		if (commit->object.flags & SYMMETRIC_LEFT)
505
			diff_emit_submodule_del(o, sb.buf);
506
		else
507
			diff_emit_submodule_add(o, sb.buf);
508
	}
509
	strbuf_release(&sb);
510
}
511

512
void prepare_submodule_repo_env(struct strvec *out)
513
{
514
	prepare_other_repo_env(out, DEFAULT_GIT_DIR_ENVIRONMENT);
515
}
516

517
static void prepare_submodule_repo_env_in_gitdir(struct strvec *out)
518
{
519
	prepare_other_repo_env(out, ".");
520
}
521

522
/*
523
 * Initialize a repository struct for a submodule based on the provided 'path'.
524
 *
525
 * Returns the repository struct on success,
526
 * NULL when the submodule is not present.
527
 */
528
static struct repository *open_submodule(const char *path)
529
{
530
	struct strbuf sb = STRBUF_INIT;
531
	struct repository *out = xmalloc(sizeof(*out));
532

533
	if (submodule_to_gitdir(&sb, path) || repo_init(out, sb.buf, NULL)) {
534
		strbuf_release(&sb);
535
		free(out);
536
		return NULL;
537
	}
538

539
	/* Mark it as a submodule */
540
	out->submodule_prefix = xstrdup(path);
541

542
	strbuf_release(&sb);
543
	return out;
544
}
545

546
/*
547
 * Helper function to display the submodule header line prior to the full
548
 * summary output.
549
 *
550
 * If it can locate the submodule git directory it will create a repository
551
 * handle for the submodule and lookup both the left and right commits and
552
 * put them into the left and right pointers.
553
 */
554
static void show_submodule_header(struct diff_options *o,
555
		const char *path,
556
		struct object_id *one, struct object_id *two,
557
		unsigned dirty_submodule,
558
		struct repository *sub,
559
		struct commit **left, struct commit **right,
560
		struct commit_list **merge_bases)
561
{
562
	const char *message = NULL;
563
	struct strbuf sb = STRBUF_INIT;
564
	int fast_forward = 0, fast_backward = 0;
565

566
	if (dirty_submodule & DIRTY_SUBMODULE_UNTRACKED)
567
		diff_emit_submodule_untracked(o, path);
568

569
	if (dirty_submodule & DIRTY_SUBMODULE_MODIFIED)
570
		diff_emit_submodule_modified(o, path);
571

572
	if (is_null_oid(one))
573
		message = "(new submodule)";
574
	else if (is_null_oid(two))
575
		message = "(submodule deleted)";
576

577
	if (!sub) {
578
		if (!message)
579
			message = "(commits not present)";
580
		goto output_header;
581
	}
582

583
	/*
584
	 * Attempt to lookup the commit references, and determine if this is
585
	 * a fast forward or fast backwards update.
586
	 */
587
	*left = lookup_commit_reference(sub, one);
588
	*right = lookup_commit_reference(sub, two);
589

590
	/*
591
	 * Warn about missing commits in the submodule project, but only if
592
	 * they aren't null.
593
	 */
594
	if ((!is_null_oid(one) && !*left) ||
595
	     (!is_null_oid(two) && !*right))
596
		message = "(commits not present)";
597

598
	*merge_bases = NULL;
599
	if (repo_get_merge_bases(sub, *left, *right, merge_bases) < 0) {
600
		message = "(corrupt repository)";
601
		goto output_header;
602
	}
603

604
	if (*merge_bases) {
605
		if ((*merge_bases)->item == *left)
606
			fast_forward = 1;
607
		else if ((*merge_bases)->item == *right)
608
			fast_backward = 1;
609
	}
610

611
	if (oideq(one, two)) {
612
		strbuf_release(&sb);
613
		return;
614
	}
615

616
output_header:
617
	strbuf_addf(&sb, "Submodule %s ", path);
618
	strbuf_add_unique_abbrev(&sb, one, DEFAULT_ABBREV);
619
	strbuf_addstr(&sb, (fast_backward || fast_forward) ? ".." : "...");
620
	strbuf_add_unique_abbrev(&sb, two, DEFAULT_ABBREV);
621
	if (message)
622
		strbuf_addf(&sb, " %s\n", message);
623
	else
624
		strbuf_addf(&sb, "%s:\n", fast_backward ? " (rewind)" : "");
625
	diff_emit_submodule_header(o, sb.buf);
626

627
	strbuf_release(&sb);
628
}
629

630
void show_submodule_diff_summary(struct diff_options *o, const char *path,
631
		struct object_id *one, struct object_id *two,
632
		unsigned dirty_submodule)
633
{
634
	struct rev_info rev = REV_INFO_INIT;
635
	struct commit *left = NULL, *right = NULL;
636
	struct commit_list *merge_bases = NULL;
637
	struct repository *sub;
638

639
	sub = open_submodule(path);
640
	show_submodule_header(o, path, one, two, dirty_submodule,
641
			      sub, &left, &right, &merge_bases);
642

643
	/*
644
	 * If we don't have both a left and a right pointer, there is no
645
	 * reason to try and display a summary. The header line should contain
646
	 * all the information the user needs.
647
	 */
648
	if (!left || !right || !sub)
649
		goto out;
650

651
	/* Treat revision walker failure the same as missing commits */
652
	if (prepare_submodule_diff_summary(sub, &rev, path, left, right, merge_bases)) {
653
		diff_emit_submodule_error(o, "(revision walker failed)\n");
654
		goto out;
655
	}
656

657
	print_submodule_diff_summary(sub, &rev, o);
658

659
out:
660
	free_commit_list(merge_bases);
661
	release_revisions(&rev);
662
	clear_commit_marks(left, ~0);
663
	clear_commit_marks(right, ~0);
664
	if (sub) {
665
		repo_clear(sub);
666
		free(sub);
667
	}
668
}
669

670
void show_submodule_inline_diff(struct diff_options *o, const char *path,
671
		struct object_id *one, struct object_id *two,
672
		unsigned dirty_submodule)
673
{
674
	const struct object_id *old_oid = the_hash_algo->empty_tree, *new_oid = the_hash_algo->empty_tree;
675
	struct commit *left = NULL, *right = NULL;
676
	struct commit_list *merge_bases = NULL;
677
	struct child_process cp = CHILD_PROCESS_INIT;
678
	struct strbuf sb = STRBUF_INIT;
679
	struct repository *sub;
680

681
	sub = open_submodule(path);
682
	show_submodule_header(o, path, one, two, dirty_submodule,
683
			      sub, &left, &right, &merge_bases);
684

685
	/* We need a valid left and right commit to display a difference */
686
	if (!(left || is_null_oid(one)) ||
687
	    !(right || is_null_oid(two)))
688
		goto done;
689

690
	if (left)
691
		old_oid = one;
692
	if (right)
693
		new_oid = two;
694

695
	cp.git_cmd = 1;
696
	cp.dir = path;
697
	cp.out = -1;
698
	cp.no_stdin = 1;
699

700
	/* TODO: other options may need to be passed here. */
701
	strvec_pushl(&cp.args, "diff", "--submodule=diff", NULL);
702
	strvec_pushf(&cp.args, "--color=%s", want_color(o->use_color) ?
703
			 "always" : "never");
704

705
	if (o->flags.reverse_diff) {
706
		strvec_pushf(&cp.args, "--src-prefix=%s%s/",
707
			     o->b_prefix, path);
708
		strvec_pushf(&cp.args, "--dst-prefix=%s%s/",
709
			     o->a_prefix, path);
710
	} else {
711
		strvec_pushf(&cp.args, "--src-prefix=%s%s/",
712
			     o->a_prefix, path);
713
		strvec_pushf(&cp.args, "--dst-prefix=%s%s/",
714
			     o->b_prefix, path);
715
	}
716
	strvec_push(&cp.args, oid_to_hex(old_oid));
717
	/*
718
	 * If the submodule has modified content, we will diff against the
719
	 * work tree, under the assumption that the user has asked for the
720
	 * diff format and wishes to actually see all differences even if they
721
	 * haven't yet been committed to the submodule yet.
722
	 */
723
	if (!(dirty_submodule & DIRTY_SUBMODULE_MODIFIED))
724
		strvec_push(&cp.args, oid_to_hex(new_oid));
725

726
	prepare_submodule_repo_env(&cp.env);
727

728
	if (!is_directory(path)) {
729
		/* fall back to absorbed git dir, if any */
730
		if (!sub)
731
			goto done;
732
		cp.dir = sub->gitdir;
733
		strvec_push(&cp.env, GIT_DIR_ENVIRONMENT "=.");
734
		strvec_push(&cp.env, GIT_WORK_TREE_ENVIRONMENT "=.");
735
	}
736

737
	if (start_command(&cp)) {
738
		diff_emit_submodule_error(o, "(diff failed)\n");
739
		goto done;
740
	}
741

742
	while (strbuf_getwholeline_fd(&sb, cp.out, '\n') != EOF)
743
		diff_emit_submodule_pipethrough(o, sb.buf, sb.len);
744

745
	if (finish_command(&cp))
746
		diff_emit_submodule_error(o, "(diff failed)\n");
747

748
done:
749
	strbuf_release(&sb);
750
	free_commit_list(merge_bases);
751
	if (left)
752
		clear_commit_marks(left, ~0);
753
	if (right)
754
		clear_commit_marks(right, ~0);
755
	if (sub) {
756
		repo_clear(sub);
757
		free(sub);
758
	}
759
}
760

761
int should_update_submodules(void)
762
{
763
	return config_update_recurse_submodules == RECURSE_SUBMODULES_ON;
764
}
765

766
const struct submodule *submodule_from_ce(const struct cache_entry *ce)
767
{
768
	if (!S_ISGITLINK(ce->ce_mode))
769
		return NULL;
770

771
	if (!should_update_submodules())
772
		return NULL;
773

774
	return submodule_from_path(the_repository, null_oid(), ce->name);
775
}
776

777

778
struct collect_changed_submodules_cb_data {
779
	struct repository *repo;
780
	struct string_list *changed;
781
	const struct object_id *commit_oid;
782
};
783

784
/*
785
 * this would normally be two functions: default_name_from_path() and
786
 * path_from_default_name(). Since the default name is the same as
787
 * the submodule path we can get away with just one function which only
788
 * checks whether there is a submodule in the working directory at that
789
 * location.
790
 */
791
static const char *default_name_or_path(const char *path_or_name)
792
{
793
	int error_code;
794

795
	if (!is_submodule_populated_gently(path_or_name, &error_code))
796
		return NULL;
797

798
	return path_or_name;
799
}
800

801
/*
802
 * Holds relevant information for a changed submodule. Used as the .util
803
 * member of the changed submodule name string_list_item.
804
 *
805
 * (super_oid, path) allows the submodule config to be read from _some_
806
 * .gitmodules file. We store this information the first time we find a
807
 * superproject commit that points to the submodule, but this is
808
 * arbitrary - we can choose any (super_oid, path) that matches the
809
 * submodule's name.
810
 *
811
 * NEEDSWORK: Storing an arbitrary commit is undesirable because we can't
812
 * guarantee that we're reading the commit that the user would expect. A better
813
 * scheme would be to just fetch a submodule by its name. This requires two
814
 * steps:
815
 * - Create a function that behaves like repo_submodule_init(), but accepts a
816
 *   submodule name instead of treeish_name and path. This should be easy
817
 *   because repo_submodule_init() internally uses the submodule's name.
818
 *
819
 * - Replace most instances of 'struct submodule' (which is the .gitmodules
820
 *   config) with just the submodule name. This is OK because we expect
821
 *   submodule settings to be stored in .git/config (via "git submodule init"),
822
 *   not .gitmodules. This also lets us delete get_non_gitmodules_submodule(),
823
 *   which constructs a bogus 'struct submodule' for the sake of giving a
824
 *   placeholder name to a gitlink.
825
 */
826
struct changed_submodule_data {
827
	/*
828
	 * The first superproject commit in the rev walk that points to
829
	 * the submodule.
830
	 */
831
	const struct object_id *super_oid;
832
	/*
833
	 * Path to the submodule in the superproject commit referenced
834
	 * by 'super_oid'.
835
	 */
836
	char *path;
837
	/* The submodule commits that have changed in the rev walk. */
838
	struct oid_array new_commits;
839
};
840

841
static void changed_submodule_data_clear(struct changed_submodule_data *cs_data)
842
{
843
	oid_array_clear(&cs_data->new_commits);
844
	free(cs_data->path);
845
}
846

847
static void collect_changed_submodules_cb(struct diff_queue_struct *q,
848
					  struct diff_options *options UNUSED,
849
					  void *data)
850
{
851
	struct collect_changed_submodules_cb_data *me = data;
852
	struct string_list *changed = me->changed;
853
	const struct object_id *commit_oid = me->commit_oid;
854
	int i;
855

856
	for (i = 0; i < q->nr; i++) {
857
		struct diff_filepair *p = q->queue[i];
858
		const struct submodule *submodule;
859
		const char *name;
860
		struct string_list_item *item;
861
		struct changed_submodule_data *cs_data;
862

863
		if (!S_ISGITLINK(p->two->mode))
864
			continue;
865

866
		submodule = submodule_from_path(me->repo,
867
						commit_oid, p->two->path);
868
		if (submodule)
869
			name = submodule->name;
870
		else {
871
			name = default_name_or_path(p->two->path);
872
			/* make sure name does not collide with existing one */
873
			if (name)
874
				submodule = submodule_from_name(me->repo,
875
								commit_oid, name);
876
			if (submodule) {
877
				warning(_("Submodule in commit %s at path: "
878
					"'%s' collides with a submodule named "
879
					"the same. Skipping it."),
880
					oid_to_hex(commit_oid), p->two->path);
881
				name = NULL;
882
			}
883
		}
884

885
		if (!name)
886
			continue;
887

888
		item = string_list_insert(changed, name);
889
		if (item->util)
890
			cs_data = item->util;
891
		else {
892
			item->util = xcalloc(1, sizeof(struct changed_submodule_data));
893
			cs_data = item->util;
894
			cs_data->super_oid = commit_oid;
895
			cs_data->path = xstrdup(p->two->path);
896
		}
897
		oid_array_append(&cs_data->new_commits, &p->two->oid);
898
	}
899
}
900

901
/*
902
 * Collect the paths of submodules in 'changed' which have changed based on
903
 * the revisions as specified in 'argv'.  Each entry in 'changed' will also
904
 * have a corresponding 'struct oid_array' (in the 'util' field) which lists
905
 * what the submodule pointers were updated to during the change.
906
 */
907
static void collect_changed_submodules(struct repository *r,
908
				       struct string_list *changed,
909
				       struct strvec *argv)
910
{
911
	struct rev_info rev;
912
	const struct commit *commit;
913
	int save_warning;
914
	struct setup_revision_opt s_r_opt = {
915
		.assume_dashdash = 1,
916
	};
917

918
	save_warning = warn_on_object_refname_ambiguity;
919
	warn_on_object_refname_ambiguity = 0;
920
	repo_init_revisions(r, &rev, NULL);
921
	setup_revisions(argv->nr, argv->v, &rev, &s_r_opt);
922
	warn_on_object_refname_ambiguity = save_warning;
923
	if (prepare_revision_walk(&rev))
924
		die(_("revision walk setup failed"));
925

926
	while ((commit = get_revision(&rev))) {
927
		struct rev_info diff_rev;
928
		struct collect_changed_submodules_cb_data data;
929
		data.repo = r;
930
		data.changed = changed;
931
		data.commit_oid = &commit->object.oid;
932

933
		repo_init_revisions(r, &diff_rev, NULL);
934
		diff_rev.diffopt.output_format |= DIFF_FORMAT_CALLBACK;
935
		diff_rev.diffopt.format_callback = collect_changed_submodules_cb;
936
		diff_rev.diffopt.format_callback_data = &data;
937
		diff_rev.dense_combined_merges = 1;
938
		diff_tree_combined_merge(commit, &diff_rev);
939
		release_revisions(&diff_rev);
940
	}
941

942
	reset_revision_walk();
943
	release_revisions(&rev);
944
}
945

946
static void free_submodules_data(struct string_list *submodules)
947
{
948
	struct string_list_item *item;
949
	for_each_string_list_item(item, submodules)
950
		changed_submodule_data_clear(item->util);
951

952
	string_list_clear(submodules, 1);
953
}
954

955
static int has_remote(const char *refname UNUSED,
956
		      const char *referent UNUSED,
957
		      const struct object_id *oid UNUSED,
958
		      int flags UNUSED, void *cb_data UNUSED)
959
{
960
	return 1;
961
}
962

963
static int append_oid_to_argv(const struct object_id *oid, void *data)
964
{
965
	struct strvec *argv = data;
966
	strvec_push(argv, oid_to_hex(oid));
967
	return 0;
968
}
969

970
struct has_commit_data {
971
	struct repository *repo;
972
	int result;
973
	const char *path;
974
	const struct object_id *super_oid;
975
};
976

977
static int check_has_commit(const struct object_id *oid, void *data)
978
{
979
	struct has_commit_data *cb = data;
980
	struct repository subrepo;
981
	enum object_type type;
982

983
	if (repo_submodule_init(&subrepo, cb->repo, cb->path, cb->super_oid)) {
984
		cb->result = 0;
985
		/* subrepo failed to init, so don't clean it up. */
986
		return 0;
987
	}
988

989
	type = oid_object_info(&subrepo, oid, NULL);
990

991
	switch (type) {
992
	case OBJ_COMMIT:
993
		goto cleanup;
994
	case OBJ_BAD:
995
		/*
996
		 * Object is missing or invalid. If invalid, an error message
997
		 * has already been printed.
998
		 */
999
		cb->result = 0;
1000
		goto cleanup;
1001
	default:
1002
		die(_("submodule entry '%s' (%s) is a %s, not a commit"),
1003
		    cb->path, oid_to_hex(oid), type_name(type));
1004
	}
1005
cleanup:
1006
	repo_clear(&subrepo);
1007
	return 0;
1008
}
1009

1010
static int submodule_has_commits(struct repository *r,
1011
				 const char *path,
1012
				 const struct object_id *super_oid,
1013
				 struct oid_array *commits)
1014
{
1015
	struct has_commit_data has_commit = {
1016
		.repo = r,
1017
		.result = 1,
1018
		.path = path,
1019
		.super_oid = super_oid
1020
	};
1021

1022
	if (validate_submodule_path(path) < 0)
1023
		exit(128);
1024

1025
	oid_array_for_each_unique(commits, check_has_commit, &has_commit);
1026

1027
	if (has_commit.result) {
1028
		/*
1029
		 * Even if the submodule is checked out and the commit is
1030
		 * present, make sure it exists in the submodule's object store
1031
		 * and that it is reachable from a ref.
1032
		 */
1033
		struct child_process cp = CHILD_PROCESS_INIT;
1034
		struct strbuf out = STRBUF_INIT;
1035

1036
		strvec_pushl(&cp.args, "rev-list", "-n", "1", NULL);
1037
		oid_array_for_each_unique(commits, append_oid_to_argv, &cp.args);
1038
		strvec_pushl(&cp.args, "--not", "--all", NULL);
1039

1040
		prepare_submodule_repo_env(&cp.env);
1041
		cp.git_cmd = 1;
1042
		cp.no_stdin = 1;
1043
		cp.dir = path;
1044

1045
		if (capture_command(&cp, &out, GIT_MAX_HEXSZ + 1) || out.len)
1046
			has_commit.result = 0;
1047

1048
		strbuf_release(&out);
1049
	}
1050

1051
	return has_commit.result;
1052
}
1053

1054
static int submodule_needs_pushing(struct repository *r,
1055
				   const char *path,
1056
				   struct oid_array *commits)
1057
{
1058
	if (!submodule_has_commits(r, path, null_oid(), commits))
1059
		/*
1060
		 * NOTE: We do consider it safe to return "no" here. The
1061
		 * correct answer would be "We do not know" instead of
1062
		 * "No push needed", but it is quite hard to change
1063
		 * the submodule pointer without having the submodule
1064
		 * around. If a user did however change the submodules
1065
		 * without having the submodule around, this indicates
1066
		 * an expert who knows what they are doing or a
1067
		 * maintainer integrating work from other people. In
1068
		 * both cases it should be safe to skip this check.
1069
		 */
1070
		return 0;
1071

1072
	if (for_each_remote_ref_submodule(path, has_remote, NULL) > 0) {
1073
		struct child_process cp = CHILD_PROCESS_INIT;
1074
		struct strbuf buf = STRBUF_INIT;
1075
		int needs_pushing = 0;
1076

1077
		strvec_push(&cp.args, "rev-list");
1078
		oid_array_for_each_unique(commits, append_oid_to_argv, &cp.args);
1079
		strvec_pushl(&cp.args, "--not", "--remotes", "-n", "1" , NULL);
1080

1081
		prepare_submodule_repo_env(&cp.env);
1082
		cp.git_cmd = 1;
1083
		cp.no_stdin = 1;
1084
		cp.out = -1;
1085
		cp.dir = path;
1086
		if (start_command(&cp))
1087
			die(_("Could not run 'git rev-list <commits> --not --remotes -n 1' command in submodule %s"),
1088
					path);
1089
		if (strbuf_read(&buf, cp.out, the_hash_algo->hexsz + 1))
1090
			needs_pushing = 1;
1091
		finish_command(&cp);
1092
		close(cp.out);
1093
		strbuf_release(&buf);
1094
		return needs_pushing;
1095
	}
1096

1097
	return 0;
1098
}
1099

1100
int find_unpushed_submodules(struct repository *r,
1101
			     struct oid_array *commits,
1102
			     const char *remotes_name,
1103
			     struct string_list *needs_pushing)
1104
{
1105
	struct string_list submodules = STRING_LIST_INIT_DUP;
1106
	struct string_list_item *name;
1107
	struct strvec argv = STRVEC_INIT;
1108

1109
	/* argv.v[0] will be ignored by setup_revisions */
1110
	strvec_push(&argv, "find_unpushed_submodules");
1111
	oid_array_for_each_unique(commits, append_oid_to_argv, &argv);
1112
	strvec_push(&argv, "--not");
1113
	strvec_pushf(&argv, "--remotes=%s", remotes_name);
1114

1115
	collect_changed_submodules(r, &submodules, &argv);
1116

1117
	for_each_string_list_item(name, &submodules) {
1118
		struct changed_submodule_data *cs_data = name->util;
1119
		const struct submodule *submodule;
1120
		const char *path = NULL;
1121

1122
		submodule = submodule_from_name(r, null_oid(), name->string);
1123
		if (submodule)
1124
			path = submodule->path;
1125
		else
1126
			path = default_name_or_path(name->string);
1127

1128
		if (!path)
1129
			continue;
1130

1131
		if (submodule_needs_pushing(r, path, &cs_data->new_commits))
1132
			string_list_insert(needs_pushing, path);
1133
	}
1134

1135
	free_submodules_data(&submodules);
1136
	strvec_clear(&argv);
1137

1138
	return needs_pushing->nr;
1139
}
1140

1141
static int push_submodule(const char *path,
1142
			  const struct remote *remote,
1143
			  const struct refspec *rs,
1144
			  const struct string_list *push_options,
1145
			  int dry_run)
1146
{
1147
	if (validate_submodule_path(path) < 0)
1148
		exit(128);
1149

1150
	if (for_each_remote_ref_submodule(path, has_remote, NULL) > 0) {
1151
		struct child_process cp = CHILD_PROCESS_INIT;
1152
		strvec_push(&cp.args, "push");
1153
		/*
1154
		 * When recursing into a submodule, treat any "only" configurations as "on-
1155
		 * demand", since "only" would not work (we need all submodules to be pushed
1156
		 * in order to be able to push the superproject).
1157
		 */
1158
		strvec_push(&cp.args, "--recurse-submodules=only-is-on-demand");
1159
		if (dry_run)
1160
			strvec_push(&cp.args, "--dry-run");
1161

1162
		if (push_options && push_options->nr) {
1163
			const struct string_list_item *item;
1164
			for_each_string_list_item(item, push_options)
1165
				strvec_pushf(&cp.args, "--push-option=%s",
1166
					     item->string);
1167
		}
1168

1169
		if (remote->origin != REMOTE_UNCONFIGURED) {
1170
			int i;
1171
			strvec_push(&cp.args, remote->name);
1172
			for (i = 0; i < rs->raw_nr; i++)
1173
				strvec_push(&cp.args, rs->raw[i]);
1174
		}
1175

1176
		prepare_submodule_repo_env(&cp.env);
1177
		cp.git_cmd = 1;
1178
		cp.no_stdin = 1;
1179
		cp.dir = path;
1180
		if (run_command(&cp))
1181
			return 0;
1182
		close(cp.out);
1183
	}
1184

1185
	return 1;
1186
}
1187

1188
/*
1189
 * Perform a check in the submodule to see if the remote and refspec work.
1190
 * Die if the submodule can't be pushed.
1191
 */
1192
static void submodule_push_check(const char *path, const char *head,
1193
				 const struct remote *remote,
1194
				 const struct refspec *rs)
1195
{
1196
	struct child_process cp = CHILD_PROCESS_INIT;
1197
	int i;
1198

1199
	if (validate_submodule_path(path) < 0)
1200
		exit(128);
1201

1202
	strvec_push(&cp.args, "submodule--helper");
1203
	strvec_push(&cp.args, "push-check");
1204
	strvec_push(&cp.args, head);
1205
	strvec_push(&cp.args, remote->name);
1206

1207
	for (i = 0; i < rs->raw_nr; i++)
1208
		strvec_push(&cp.args, rs->raw[i]);
1209

1210
	prepare_submodule_repo_env(&cp.env);
1211
	cp.git_cmd = 1;
1212
	cp.no_stdin = 1;
1213
	cp.no_stdout = 1;
1214
	cp.dir = path;
1215

1216
	/*
1217
	 * Simply indicate if 'submodule--helper push-check' failed.
1218
	 * More detailed error information will be provided by the
1219
	 * child process.
1220
	 */
1221
	if (run_command(&cp))
1222
		die(_("process for submodule '%s' failed"), path);
1223
}
1224

1225
int push_unpushed_submodules(struct repository *r,
1226
			     struct oid_array *commits,
1227
			     const struct remote *remote,
1228
			     const struct refspec *rs,
1229
			     const struct string_list *push_options,
1230
			     int dry_run)
1231
{
1232
	int i, ret = 1;
1233
	struct string_list needs_pushing = STRING_LIST_INIT_DUP;
1234

1235
	if (!find_unpushed_submodules(r, commits,
1236
				      remote->name, &needs_pushing))
1237
		return 1;
1238

1239
	/*
1240
	 * Verify that the remote and refspec can be propagated to all
1241
	 * submodules.  This check can be skipped if the remote and refspec
1242
	 * won't be propagated due to the remote being unconfigured (e.g. a URL
1243
	 * instead of a remote name).
1244
	 */
1245
	if (remote->origin != REMOTE_UNCONFIGURED) {
1246
		char *head;
1247
		struct object_id head_oid;
1248

1249
		head = refs_resolve_refdup(get_main_ref_store(the_repository),
1250
					   "HEAD", 0, &head_oid, NULL);
1251
		if (!head)
1252
			die(_("Failed to resolve HEAD as a valid ref."));
1253

1254
		for (i = 0; i < needs_pushing.nr; i++)
1255
			submodule_push_check(needs_pushing.items[i].string,
1256
					     head, remote, rs);
1257
		free(head);
1258
	}
1259

1260
	/* Actually push the submodules */
1261
	for (i = 0; i < needs_pushing.nr; i++) {
1262
		const char *path = needs_pushing.items[i].string;
1263
		fprintf(stderr, _("Pushing submodule '%s'\n"), path);
1264
		if (!push_submodule(path, remote, rs,
1265
				    push_options, dry_run)) {
1266
			fprintf(stderr, _("Unable to push submodule '%s'\n"), path);
1267
			ret = 0;
1268
		}
1269
	}
1270

1271
	string_list_clear(&needs_pushing, 0);
1272

1273
	return ret;
1274
}
1275

1276
static int append_oid_to_array(const char *ref UNUSED,
1277
			       const char *referent UNUSED,
1278
			       const struct object_id *oid,
1279
			       int flags UNUSED, void *data)
1280
{
1281
	struct oid_array *array = data;
1282
	oid_array_append(array, oid);
1283
	return 0;
1284
}
1285

1286
void check_for_new_submodule_commits(struct object_id *oid)
1287
{
1288
	if (!initialized_fetch_ref_tips) {
1289
		refs_for_each_ref(get_main_ref_store(the_repository),
1290
				  append_oid_to_array, &ref_tips_before_fetch);
1291
		initialized_fetch_ref_tips = 1;
1292
	}
1293

1294
	oid_array_append(&ref_tips_after_fetch, oid);
1295
}
1296

1297
/*
1298
 * Returns 1 if there is at least one submodule gitdir in
1299
 * $GIT_DIR/modules and 0 otherwise. This follows
1300
 * submodule_name_to_gitdir(), which looks for submodules in
1301
 * $GIT_DIR/modules, not $GIT_COMMON_DIR.
1302
 *
1303
 * A submodule can be moved to $GIT_DIR/modules manually by running "git
1304
 * submodule absorbgitdirs", or it may be initialized there by "git
1305
 * submodule update".
1306
 */
1307
static int repo_has_absorbed_submodules(struct repository *r)
1308
{
1309
	int ret;
1310
	struct strbuf buf = STRBUF_INIT;
1311

1312
	strbuf_repo_git_path(&buf, r, "modules/");
1313
	ret = file_exists(buf.buf) && !is_empty_dir(buf.buf);
1314
	strbuf_release(&buf);
1315
	return ret;
1316
}
1317

1318
static void calculate_changed_submodule_paths(struct repository *r,
1319
		struct string_list *changed_submodule_names)
1320
{
1321
	struct strvec argv = STRVEC_INIT;
1322
	struct string_list_item *name;
1323

1324
	/* No need to check if no submodules would be fetched */
1325
	if (!submodule_from_path(r, NULL, NULL) &&
1326
	    !repo_has_absorbed_submodules(r))
1327
		return;
1328

1329
	strvec_push(&argv, "--"); /* argv[0] program name */
1330
	oid_array_for_each_unique(&ref_tips_after_fetch,
1331
				   append_oid_to_argv, &argv);
1332
	strvec_push(&argv, "--not");
1333
	oid_array_for_each_unique(&ref_tips_before_fetch,
1334
				   append_oid_to_argv, &argv);
1335

1336
	/*
1337
	 * Collect all submodules (whether checked out or not) for which new
1338
	 * commits have been recorded upstream in "changed_submodule_names".
1339
	 */
1340
	collect_changed_submodules(r, changed_submodule_names, &argv);
1341

1342
	for_each_string_list_item(name, changed_submodule_names) {
1343
		struct changed_submodule_data *cs_data = name->util;
1344
		const struct submodule *submodule;
1345
		const char *path = NULL;
1346

1347
		submodule = submodule_from_name(r, null_oid(), name->string);
1348
		if (submodule)
1349
			path = submodule->path;
1350
		else
1351
			path = default_name_or_path(name->string);
1352

1353
		if (!path)
1354
			continue;
1355

1356
		if (submodule_has_commits(r, path, null_oid(), &cs_data->new_commits)) {
1357
			changed_submodule_data_clear(cs_data);
1358
			*name->string = '\0';
1359
		}
1360
	}
1361

1362
	string_list_remove_empty_items(changed_submodule_names, 1);
1363

1364
	strvec_clear(&argv);
1365
	oid_array_clear(&ref_tips_before_fetch);
1366
	oid_array_clear(&ref_tips_after_fetch);
1367
	initialized_fetch_ref_tips = 0;
1368
}
1369

1370
int submodule_touches_in_range(struct repository *r,
1371
			       struct object_id *excl_oid,
1372
			       struct object_id *incl_oid)
1373
{
1374
	struct string_list subs = STRING_LIST_INIT_DUP;
1375
	struct strvec args = STRVEC_INIT;
1376
	int ret;
1377

1378
	/* No need to check if there are no submodules configured */
1379
	if (!submodule_from_path(r, NULL, NULL))
1380
		return 0;
1381

1382
	strvec_push(&args, "--"); /* args[0] program name */
1383
	strvec_push(&args, oid_to_hex(incl_oid));
1384
	if (!is_null_oid(excl_oid)) {
1385
		strvec_push(&args, "--not");
1386
		strvec_push(&args, oid_to_hex(excl_oid));
1387
	}
1388

1389
	collect_changed_submodules(r, &subs, &args);
1390
	ret = subs.nr;
1391

1392
	strvec_clear(&args);
1393

1394
	free_submodules_data(&subs);
1395
	return ret;
1396
}
1397

1398
struct submodule_parallel_fetch {
1399
	/*
1400
	 * The index of the last index entry processed by
1401
	 * get_fetch_task_from_index().
1402
	 */
1403
	int index_count;
1404
	/*
1405
	 * The index of the last string_list entry processed by
1406
	 * get_fetch_task_from_changed().
1407
	 */
1408
	int changed_count;
1409
	struct strvec args;
1410
	struct repository *r;
1411
	const char *prefix;
1412
	int command_line_option;
1413
	int default_option;
1414
	int quiet;
1415
	int result;
1416

1417
	/*
1418
	 * Names of submodules that have new commits. Generated by
1419
	 * walking the newly fetched superproject commits.
1420
	 */
1421
	struct string_list changed_submodule_names;
1422
	/*
1423
	 * Names of submodules that have already been processed. Lets us
1424
	 * avoid fetching the same submodule more than once.
1425
	 */
1426
	struct string_list seen_submodule_names;
1427

1428
	/* Pending fetches by OIDs */
1429
	struct fetch_task **oid_fetch_tasks;
1430
	int oid_fetch_tasks_nr, oid_fetch_tasks_alloc;
1431

1432
	struct strbuf submodules_with_errors;
1433
};
1434
#define SPF_INIT { \
1435
	.args = STRVEC_INIT, \
1436
	.changed_submodule_names = STRING_LIST_INIT_DUP, \
1437
	.seen_submodule_names = STRING_LIST_INIT_DUP, \
1438
	.submodules_with_errors = STRBUF_INIT, \
1439
}
1440

1441
static int get_fetch_recurse_config(const struct submodule *submodule,
1442
				    struct submodule_parallel_fetch *spf)
1443
{
1444
	if (spf->command_line_option != RECURSE_SUBMODULES_DEFAULT)
1445
		return spf->command_line_option;
1446

1447
	if (submodule) {
1448
		char *key;
1449
		const char *value;
1450

1451
		int fetch_recurse = submodule->fetch_recurse;
1452
		key = xstrfmt("submodule.%s.fetchRecurseSubmodules", submodule->name);
1453
		if (!repo_config_get_string_tmp(spf->r, key, &value)) {
1454
			fetch_recurse = parse_fetch_recurse_submodules_arg(key, value);
1455
		}
1456
		free(key);
1457

1458
		if (fetch_recurse != RECURSE_SUBMODULES_NONE)
1459
			/* local config overrules everything except commandline */
1460
			return fetch_recurse;
1461
	}
1462

1463
	return spf->default_option;
1464
}
1465

1466
/*
1467
 * Fetch in progress (if callback data) or
1468
 * pending (if in oid_fetch_tasks in struct submodule_parallel_fetch)
1469
 */
1470
struct fetch_task {
1471
	struct repository *repo;
1472
	const struct submodule *sub;
1473
	unsigned free_sub : 1; /* Do we need to free the submodule? */
1474
	const char *default_argv; /* The default fetch mode. */
1475
	struct strvec git_args; /* Args for the child git process. */
1476

1477
	struct oid_array *commits; /* Ensure these commits are fetched */
1478
};
1479

1480
/**
1481
 * When a submodule is not defined in .gitmodules, we cannot access it
1482
 * via the regular submodule-config. Create a fake submodule, which we can
1483
 * work on.
1484
 */
1485
static const struct submodule *get_non_gitmodules_submodule(const char *path)
1486
{
1487
	struct submodule *ret = NULL;
1488
	const char *name = default_name_or_path(path);
1489

1490
	if (!name)
1491
		return NULL;
1492

1493
	ret = xmalloc(sizeof(*ret));
1494
	memset(ret, 0, sizeof(*ret));
1495
	ret->path = name;
1496
	ret->name = name;
1497

1498
	return (const struct submodule *) ret;
1499
}
1500

1501
static void fetch_task_free(struct fetch_task *p)
1502
{
1503
	if (p->free_sub)
1504
		free((void*)p->sub);
1505
	p->free_sub = 0;
1506
	p->sub = NULL;
1507

1508
	if (p->repo)
1509
		repo_clear(p->repo);
1510
	FREE_AND_NULL(p->repo);
1511

1512
	strvec_clear(&p->git_args);
1513
	free(p);
1514
}
1515

1516
static struct repository *get_submodule_repo_for(struct repository *r,
1517
						 const char *path,
1518
						 const struct object_id *treeish_name)
1519
{
1520
	struct repository *ret = xmalloc(sizeof(*ret));
1521

1522
	if (repo_submodule_init(ret, r, path, treeish_name)) {
1523
		free(ret);
1524
		return NULL;
1525
	}
1526

1527
	return ret;
1528
}
1529

1530
static struct fetch_task *fetch_task_create(struct submodule_parallel_fetch *spf,
1531
					    const char *path,
1532
					    const struct object_id *treeish_name)
1533
{
1534
	struct fetch_task *task = xmalloc(sizeof(*task));
1535
	memset(task, 0, sizeof(*task));
1536

1537
	if (validate_submodule_path(path) < 0)
1538
		exit(128);
1539

1540
	task->sub = submodule_from_path(spf->r, treeish_name, path);
1541

1542
	if (!task->sub) {
1543
		/*
1544
		 * No entry in .gitmodules? Technically not a submodule,
1545
		 * but historically we supported repositories that happen to be
1546
		 * in-place where a gitlink is. Keep supporting them.
1547
		 */
1548
		task->sub = get_non_gitmodules_submodule(path);
1549
		if (!task->sub)
1550
			goto cleanup;
1551

1552
		task->free_sub = 1;
1553
	}
1554

1555
	if (string_list_lookup(&spf->seen_submodule_names, task->sub->name))
1556
		goto cleanup;
1557

1558
	switch (get_fetch_recurse_config(task->sub, spf))
1559
	{
1560
	default:
1561
	case RECURSE_SUBMODULES_DEFAULT:
1562
	case RECURSE_SUBMODULES_ON_DEMAND:
1563
		if (!task->sub ||
1564
			!string_list_lookup(
1565
				&spf->changed_submodule_names,
1566
				task->sub->name))
1567
			goto cleanup;
1568
		task->default_argv = "on-demand";
1569
		break;
1570
	case RECURSE_SUBMODULES_ON:
1571
		task->default_argv = "yes";
1572
		break;
1573
	case RECURSE_SUBMODULES_OFF:
1574
		goto cleanup;
1575
	}
1576

1577
	task->repo = get_submodule_repo_for(spf->r, path, treeish_name);
1578

1579
	return task;
1580

1581
 cleanup:
1582
	fetch_task_free(task);
1583
	return NULL;
1584
}
1585

1586
static struct fetch_task *
1587
get_fetch_task_from_index(struct submodule_parallel_fetch *spf,
1588
			  struct strbuf *err)
1589
{
1590
	for (; spf->index_count < spf->r->index->cache_nr; spf->index_count++) {
1591
		const struct cache_entry *ce =
1592
			spf->r->index->cache[spf->index_count];
1593
		struct fetch_task *task;
1594

1595
		if (!S_ISGITLINK(ce->ce_mode))
1596
			continue;
1597

1598
		task = fetch_task_create(spf, ce->name, null_oid());
1599
		if (!task)
1600
			continue;
1601

1602
		if (task->repo) {
1603
			if (!spf->quiet)
1604
				strbuf_addf(err, _("Fetching submodule %s%s\n"),
1605
					    spf->prefix, ce->name);
1606

1607
			spf->index_count++;
1608
			return task;
1609
		} else {
1610
			struct strbuf empty_submodule_path = STRBUF_INIT;
1611

1612
			fetch_task_free(task);
1613

1614
			/*
1615
			 * An empty directory is normal,
1616
			 * the submodule is not initialized
1617
			 */
1618
			strbuf_addf(&empty_submodule_path, "%s/%s/",
1619
							spf->r->worktree,
1620
							ce->name);
1621
			if (S_ISGITLINK(ce->ce_mode) &&
1622
			    !is_empty_dir(empty_submodule_path.buf)) {
1623
				spf->result = 1;
1624
				strbuf_addf(err,
1625
					    _("Could not access submodule '%s'\n"),
1626
					    ce->name);
1627
			}
1628
			strbuf_release(&empty_submodule_path);
1629
		}
1630
	}
1631
	return NULL;
1632
}
1633

1634
static struct fetch_task *
1635
get_fetch_task_from_changed(struct submodule_parallel_fetch *spf,
1636
			    struct strbuf *err)
1637
{
1638
	for (; spf->changed_count < spf->changed_submodule_names.nr;
1639
	     spf->changed_count++) {
1640
		struct string_list_item item =
1641
			spf->changed_submodule_names.items[spf->changed_count];
1642
		struct changed_submodule_data *cs_data = item.util;
1643
		struct fetch_task *task;
1644

1645
		if (!is_tree_submodule_active(spf->r, cs_data->super_oid,cs_data->path))
1646
			continue;
1647

1648
		task = fetch_task_create(spf, cs_data->path,
1649
					 cs_data->super_oid);
1650
		if (!task)
1651
			continue;
1652

1653
		if (!task->repo) {
1654
			strbuf_addf(err, _("Could not access submodule '%s' at commit %s\n"),
1655
				    cs_data->path,
1656
				    repo_find_unique_abbrev(the_repository, cs_data->super_oid, DEFAULT_ABBREV));
1657

1658
			fetch_task_free(task);
1659
			continue;
1660
		}
1661

1662
		if (!spf->quiet)
1663
			strbuf_addf(err,
1664
				    _("Fetching submodule %s%s at commit %s\n"),
1665
				    spf->prefix, task->sub->path,
1666
				    repo_find_unique_abbrev(the_repository, cs_data->super_oid,
1667
							    DEFAULT_ABBREV));
1668

1669
		spf->changed_count++;
1670
		/*
1671
		 * NEEDSWORK: Submodules set/unset a value for
1672
		 * core.worktree when they are populated/unpopulated by
1673
		 * "git checkout" (and similar commands, see
1674
		 * submodule_move_head() and
1675
		 * connect_work_tree_and_git_dir()), but if the
1676
		 * submodule is unpopulated in another way (e.g. "git
1677
		 * rm", "rm -r"), core.worktree will still be set even
1678
		 * though the directory doesn't exist, and the child
1679
		 * process will crash while trying to chdir into the
1680
		 * nonexistent directory.
1681
		 *
1682
		 * In this case, we know that the submodule has no
1683
		 * working tree, so we can work around this by
1684
		 * setting "--work-tree=." (--bare does not work because
1685
		 * worktree settings take precedence over bare-ness).
1686
		 * However, this is not necessarily true in other cases,
1687
		 * so a generalized solution is still necessary.
1688
		 *
1689
		 * Possible solutions:
1690
		 * - teach "git [add|rm]" to unset core.worktree and
1691
		 *   discourage users from removing submodules without
1692
		 *   using a Git command.
1693
		 * - teach submodule child processes to ignore stale
1694
		 *   core.worktree values.
1695
		 */
1696
		strvec_push(&task->git_args, "--work-tree=.");
1697
		return task;
1698
	}
1699
	return NULL;
1700
}
1701

1702
static int get_next_submodule(struct child_process *cp, struct strbuf *err,
1703
			      void *data, void **task_cb)
1704
{
1705
	struct submodule_parallel_fetch *spf = data;
1706
	struct fetch_task *task =
1707
		get_fetch_task_from_index(spf, err);
1708
	if (!task)
1709
		task = get_fetch_task_from_changed(spf, err);
1710

1711
	if (task) {
1712
		child_process_init(cp);
1713
		cp->dir = task->repo->gitdir;
1714
		prepare_submodule_repo_env_in_gitdir(&cp->env);
1715
		cp->git_cmd = 1;
1716
		strvec_init(&cp->args);
1717
		if (task->git_args.nr)
1718
			strvec_pushv(&cp->args, task->git_args.v);
1719
		strvec_pushv(&cp->args, spf->args.v);
1720
		strvec_push(&cp->args, task->default_argv);
1721
		strvec_pushf(&cp->args, "--submodule-prefix=%s%s/",
1722
			     spf->prefix, task->sub->path);
1723

1724
		*task_cb = task;
1725

1726
		string_list_insert(&spf->seen_submodule_names, task->sub->name);
1727
		return 1;
1728
	}
1729

1730
	if (spf->oid_fetch_tasks_nr) {
1731
		struct fetch_task *task =
1732
			spf->oid_fetch_tasks[spf->oid_fetch_tasks_nr - 1];
1733
		spf->oid_fetch_tasks_nr--;
1734

1735
		child_process_init(cp);
1736
		prepare_submodule_repo_env_in_gitdir(&cp->env);
1737
		cp->git_cmd = 1;
1738
		cp->dir = task->repo->gitdir;
1739

1740
		strvec_init(&cp->args);
1741
		strvec_pushv(&cp->args, spf->args.v);
1742
		strvec_push(&cp->args, "on-demand");
1743
		strvec_pushf(&cp->args, "--submodule-prefix=%s%s/",
1744
			     spf->prefix, task->sub->path);
1745

1746
		/* NEEDSWORK: have get_default_remote from submodule--helper */
1747
		strvec_push(&cp->args, "origin");
1748
		oid_array_for_each_unique(task->commits,
1749
					  append_oid_to_argv, &cp->args);
1750

1751
		*task_cb = task;
1752
		return 1;
1753
	}
1754

1755
	return 0;
1756
}
1757

1758
static int fetch_start_failure(struct strbuf *err UNUSED,
1759
			       void *cb, void *task_cb)
1760
{
1761
	struct submodule_parallel_fetch *spf = cb;
1762
	struct fetch_task *task = task_cb;
1763

1764
	spf->result = 1;
1765

1766
	fetch_task_free(task);
1767
	return 0;
1768
}
1769

1770
static int commit_missing_in_sub(const struct object_id *oid, void *data)
1771
{
1772
	struct repository *subrepo = data;
1773

1774
	enum object_type type = oid_object_info(subrepo, oid, NULL);
1775

1776
	return type != OBJ_COMMIT;
1777
}
1778

1779
static int fetch_finish(int retvalue, struct strbuf *err UNUSED,
1780
			void *cb, void *task_cb)
1781
{
1782
	struct submodule_parallel_fetch *spf = cb;
1783
	struct fetch_task *task = task_cb;
1784

1785
	struct string_list_item *it;
1786
	struct changed_submodule_data *cs_data;
1787

1788
	if (!task || !task->sub)
1789
		BUG("callback cookie bogus");
1790

1791
	if (retvalue) {
1792
		/*
1793
		 * NEEDSWORK: This indicates that the overall fetch
1794
		 * failed, even though there may be a subsequent fetch
1795
		 * by commit hash that might work. It may be a good
1796
		 * idea to not indicate failure in this case, and only
1797
		 * indicate failure if the subsequent fetch fails.
1798
		 */
1799
		spf->result = 1;
1800

1801
		strbuf_addf(&spf->submodules_with_errors, "\t%s\n",
1802
			    task->sub->name);
1803
	}
1804

1805
	/* Is this the second time we process this submodule? */
1806
	if (task->commits)
1807
		goto out;
1808

1809
	it = string_list_lookup(&spf->changed_submodule_names, task->sub->name);
1810
	if (!it)
1811
		/* Could be an unchanged submodule, not contained in the list */
1812
		goto out;
1813

1814
	cs_data = it->util;
1815
	oid_array_filter(&cs_data->new_commits,
1816
			 commit_missing_in_sub,
1817
			 task->repo);
1818

1819
	/* Are there commits we want, but do not exist? */
1820
	if (cs_data->new_commits.nr) {
1821
		task->commits = &cs_data->new_commits;
1822
		ALLOC_GROW(spf->oid_fetch_tasks,
1823
			   spf->oid_fetch_tasks_nr + 1,
1824
			   spf->oid_fetch_tasks_alloc);
1825
		spf->oid_fetch_tasks[spf->oid_fetch_tasks_nr] = task;
1826
		spf->oid_fetch_tasks_nr++;
1827
		return 0;
1828
	}
1829

1830
out:
1831
	fetch_task_free(task);
1832
	return 0;
1833
}
1834

1835
int fetch_submodules(struct repository *r,
1836
		     const struct strvec *options,
1837
		     const char *prefix, int command_line_option,
1838
		     int default_option,
1839
		     int quiet, int max_parallel_jobs)
1840
{
1841
	int i;
1842
	struct submodule_parallel_fetch spf = SPF_INIT;
1843
	const struct run_process_parallel_opts opts = {
1844
		.tr2_category = "submodule",
1845
		.tr2_label = "parallel/fetch",
1846

1847
		.processes = max_parallel_jobs,
1848

1849
		.get_next_task = get_next_submodule,
1850
		.start_failure = fetch_start_failure,
1851
		.task_finished = fetch_finish,
1852
		.data = &spf,
1853
	};
1854

1855
	spf.r = r;
1856
	spf.command_line_option = command_line_option;
1857
	spf.default_option = default_option;
1858
	spf.quiet = quiet;
1859
	spf.prefix = prefix;
1860

1861
	if (!r->worktree)
1862
		goto out;
1863

1864
	if (repo_read_index(r) < 0)
1865
		die(_("index file corrupt"));
1866

1867
	strvec_push(&spf.args, "fetch");
1868
	for (i = 0; i < options->nr; i++)
1869
		strvec_push(&spf.args, options->v[i]);
1870
	strvec_push(&spf.args, "--recurse-submodules-default");
1871
	/* default value, "--submodule-prefix" and its value are added later */
1872

1873
	calculate_changed_submodule_paths(r, &spf.changed_submodule_names);
1874
	string_list_sort(&spf.changed_submodule_names);
1875
	run_processes_parallel(&opts);
1876

1877
	if (spf.submodules_with_errors.len > 0)
1878
		fprintf(stderr, _("Errors during submodule fetch:\n%s"),
1879
			spf.submodules_with_errors.buf);
1880

1881

1882
	strvec_clear(&spf.args);
1883
out:
1884
	free_submodules_data(&spf.changed_submodule_names);
1885
	string_list_clear(&spf.seen_submodule_names, 0);
1886
	return spf.result;
1887
}
1888

1889
unsigned is_submodule_modified(const char *path, int ignore_untracked)
1890
{
1891
	struct child_process cp = CHILD_PROCESS_INIT;
1892
	struct strbuf buf = STRBUF_INIT;
1893
	FILE *fp;
1894
	unsigned dirty_submodule = 0;
1895
	const char *git_dir;
1896
	int ignore_cp_exit_code = 0;
1897

1898
	if (validate_submodule_path(path) < 0)
1899
		exit(128);
1900

1901
	strbuf_addf(&buf, "%s/.git", path);
1902
	git_dir = read_gitfile(buf.buf);
1903
	if (!git_dir)
1904
		git_dir = buf.buf;
1905
	if (!is_git_directory(git_dir)) {
1906
		if (is_directory(git_dir))
1907
			die(_("'%s' not recognized as a git repository"), git_dir);
1908
		strbuf_release(&buf);
1909
		/* The submodule is not checked out, so it is not modified */
1910
		return 0;
1911
	}
1912
	strbuf_reset(&buf);
1913

1914
	strvec_pushl(&cp.args, "status", "--porcelain=2", NULL);
1915
	if (ignore_untracked)
1916
		strvec_push(&cp.args, "-uno");
1917

1918
	prepare_submodule_repo_env(&cp.env);
1919
	cp.git_cmd = 1;
1920
	cp.no_stdin = 1;
1921
	cp.out = -1;
1922
	cp.dir = path;
1923
	if (start_command(&cp))
1924
		die(_("Could not run 'git status --porcelain=2' in submodule %s"), path);
1925

1926
	fp = xfdopen(cp.out, "r");
1927
	while (strbuf_getwholeline(&buf, fp, '\n') != EOF) {
1928
		/* regular untracked files */
1929
		if (buf.buf[0] == '?')
1930
			dirty_submodule |= DIRTY_SUBMODULE_UNTRACKED;
1931

1932
		if (buf.buf[0] == 'u' ||
1933
		    buf.buf[0] == '1' ||
1934
		    buf.buf[0] == '2') {
1935
			/* T = line type, XY = status, SSSS = submodule state */
1936
			if (buf.len < strlen("T XY SSSS"))
1937
				BUG("invalid status --porcelain=2 line %s",
1938
				    buf.buf);
1939

1940
			if (buf.buf[5] == 'S' && buf.buf[8] == 'U')
1941
				/* nested untracked file */
1942
				dirty_submodule |= DIRTY_SUBMODULE_UNTRACKED;
1943

1944
			if (buf.buf[0] == 'u' ||
1945
			    buf.buf[0] == '2' ||
1946
			    memcmp(buf.buf + 5, "S..U", 4))
1947
				/* other change */
1948
				dirty_submodule |= DIRTY_SUBMODULE_MODIFIED;
1949
		}
1950

1951
		if ((dirty_submodule & DIRTY_SUBMODULE_MODIFIED) &&
1952
		    ((dirty_submodule & DIRTY_SUBMODULE_UNTRACKED) ||
1953
		     ignore_untracked)) {
1954
			/*
1955
			 * We're not interested in any further information from
1956
			 * the child any more, neither output nor its exit code.
1957
			 */
1958
			ignore_cp_exit_code = 1;
1959
			break;
1960
		}
1961
	}
1962
	fclose(fp);
1963

1964
	if (finish_command(&cp) && !ignore_cp_exit_code)
1965
		die(_("'git status --porcelain=2' failed in submodule %s"), path);
1966

1967
	strbuf_release(&buf);
1968
	return dirty_submodule;
1969
}
1970

1971
int submodule_uses_gitfile(const char *path)
1972
{
1973
	struct child_process cp = CHILD_PROCESS_INIT;
1974
	struct strbuf buf = STRBUF_INIT;
1975
	const char *git_dir;
1976

1977
	if (validate_submodule_path(path) < 0)
1978
		exit(128);
1979

1980
	strbuf_addf(&buf, "%s/.git", path);
1981
	git_dir = read_gitfile(buf.buf);
1982
	if (!git_dir) {
1983
		strbuf_release(&buf);
1984
		return 0;
1985
	}
1986
	strbuf_release(&buf);
1987

1988
	/* Now test that all nested submodules use a gitfile too */
1989
	strvec_pushl(&cp.args,
1990
		     "submodule", "foreach", "--quiet",	"--recursive",
1991
		     "test -f .git", NULL);
1992

1993
	prepare_submodule_repo_env(&cp.env);
1994
	cp.git_cmd = 1;
1995
	cp.no_stdin = 1;
1996
	cp.no_stderr = 1;
1997
	cp.no_stdout = 1;
1998
	cp.dir = path;
1999
	if (run_command(&cp))
2000
		return 0;
2001

2002
	return 1;
2003
}
2004

2005
/*
2006
 * Check if it is a bad idea to remove a submodule, i.e. if we'd lose data
2007
 * when doing so.
2008
 *
2009
 * Return 1 if we'd lose data, return 0 if the removal is fine,
2010
 * and negative values for errors.
2011
 */
2012
int bad_to_remove_submodule(const char *path, unsigned flags)
2013
{
2014
	ssize_t len;
2015
	struct child_process cp = CHILD_PROCESS_INIT;
2016
	struct strbuf buf = STRBUF_INIT;
2017
	int ret = 0;
2018

2019
	if (validate_submodule_path(path) < 0)
2020
		exit(128);
2021

2022
	if (!file_exists(path) || is_empty_dir(path))
2023
		return 0;
2024

2025
	if (!submodule_uses_gitfile(path))
2026
		return 1;
2027

2028
	strvec_pushl(&cp.args, "status", "--porcelain",
2029
		     "--ignore-submodules=none", NULL);
2030

2031
	if (flags & SUBMODULE_REMOVAL_IGNORE_UNTRACKED)
2032
		strvec_push(&cp.args, "-uno");
2033
	else
2034
		strvec_push(&cp.args, "-uall");
2035

2036
	if (!(flags & SUBMODULE_REMOVAL_IGNORE_IGNORED_UNTRACKED))
2037
		strvec_push(&cp.args, "--ignored");
2038

2039
	prepare_submodule_repo_env(&cp.env);
2040
	cp.git_cmd = 1;
2041
	cp.no_stdin = 1;
2042
	cp.out = -1;
2043
	cp.dir = path;
2044
	if (start_command(&cp)) {
2045
		if (flags & SUBMODULE_REMOVAL_DIE_ON_ERROR)
2046
			die(_("could not start 'git status' in submodule '%s'"),
2047
				path);
2048
		ret = -1;
2049
		goto out;
2050
	}
2051

2052
	len = strbuf_read(&buf, cp.out, 1024);
2053
	if (len > 2)
2054
		ret = 1;
2055
	close(cp.out);
2056

2057
	if (finish_command(&cp)) {
2058
		if (flags & SUBMODULE_REMOVAL_DIE_ON_ERROR)
2059
			die(_("could not run 'git status' in submodule '%s'"),
2060
				path);
2061
		ret = -1;
2062
	}
2063
out:
2064
	strbuf_release(&buf);
2065
	return ret;
2066
}
2067

2068
void submodule_unset_core_worktree(const struct submodule *sub)
2069
{
2070
	struct strbuf config_path = STRBUF_INIT;
2071

2072
	if (validate_submodule_path(sub->path) < 0)
2073
		exit(128);
2074

2075
	submodule_name_to_gitdir(&config_path, the_repository, sub->name);
2076
	strbuf_addstr(&config_path, "/config");
2077

2078
	if (git_config_set_in_file_gently(config_path.buf, "core.worktree", NULL, NULL))
2079
		warning(_("Could not unset core.worktree setting in submodule '%s'"),
2080
			  sub->path);
2081

2082
	strbuf_release(&config_path);
2083
}
2084

2085
static int submodule_has_dirty_index(const struct submodule *sub)
2086
{
2087
	struct child_process cp = CHILD_PROCESS_INIT;
2088

2089
	if (validate_submodule_path(sub->path) < 0)
2090
		exit(128);
2091

2092
	prepare_submodule_repo_env(&cp.env);
2093

2094
	cp.git_cmd = 1;
2095
	strvec_pushl(&cp.args, "diff-index", "--quiet",
2096
		     "--cached", "HEAD", NULL);
2097
	cp.no_stdin = 1;
2098
	cp.no_stdout = 1;
2099
	cp.dir = sub->path;
2100
	if (start_command(&cp))
2101
		die(_("could not recurse into submodule '%s'"), sub->path);
2102

2103
	return finish_command(&cp);
2104
}
2105

2106
static void submodule_reset_index(const char *path, const char *super_prefix)
2107
{
2108
	struct child_process cp = CHILD_PROCESS_INIT;
2109

2110
	if (validate_submodule_path(path) < 0)
2111
		exit(128);
2112

2113
	prepare_submodule_repo_env(&cp.env);
2114

2115
	cp.git_cmd = 1;
2116
	cp.no_stdin = 1;
2117
	cp.dir = path;
2118

2119
	/* TODO: determine if this might overwright untracked files */
2120
	strvec_pushl(&cp.args, "read-tree", "-u", "--reset", NULL);
2121
	strvec_pushf(&cp.args, "--super-prefix=%s%s/",
2122
		     (super_prefix ? super_prefix : ""), path);
2123

2124
	strvec_push(&cp.args, empty_tree_oid_hex(the_repository->hash_algo));
2125

2126
	if (run_command(&cp))
2127
		die(_("could not reset submodule index"));
2128
}
2129

2130
/**
2131
 * Moves a submodule at a given path from a given head to another new head.
2132
 * For edge cases (a submodule coming into existence or removing a submodule)
2133
 * pass NULL for old or new respectively.
2134
 */
2135
int submodule_move_head(const char *path, const char *super_prefix,
2136
			const char *old_head, const char *new_head,
2137
			unsigned flags)
2138
{
2139
	int ret = 0;
2140
	struct child_process cp = CHILD_PROCESS_INIT;
2141
	const struct submodule *sub;
2142
	int *error_code_ptr, error_code;
2143

2144
	if (!is_submodule_active(the_repository, path))
2145
		return 0;
2146

2147
	if (flags & SUBMODULE_MOVE_HEAD_FORCE)
2148
		/*
2149
		 * Pass non NULL pointer to is_submodule_populated_gently
2150
		 * to prevent die()-ing. We'll use connect_work_tree_and_git_dir
2151
		 * to fixup the submodule in the force case later.
2152
		 */
2153
		error_code_ptr = &error_code;
2154
	else
2155
		error_code_ptr = NULL;
2156

2157
	if (old_head && !is_submodule_populated_gently(path, error_code_ptr))
2158
		return 0;
2159

2160
	sub = submodule_from_path(the_repository, null_oid(), path);
2161

2162
	if (!sub)
2163
		BUG("could not get submodule information for '%s'", path);
2164

2165
	if (old_head && !(flags & SUBMODULE_MOVE_HEAD_FORCE)) {
2166
		/* Check if the submodule has a dirty index. */
2167
		if (submodule_has_dirty_index(sub))
2168
			return error(_("submodule '%s' has dirty index"), path);
2169
	}
2170

2171
	if (!(flags & SUBMODULE_MOVE_HEAD_DRY_RUN)) {
2172
		if (old_head) {
2173
			if (!submodule_uses_gitfile(path))
2174
				absorb_git_dir_into_superproject(path,
2175
								 super_prefix);
2176
			else {
2177
				char *dotgit = xstrfmt("%s/.git", path);
2178
				char *git_dir = xstrdup(read_gitfile(dotgit));
2179

2180
				free(dotgit);
2181
				if (validate_submodule_git_dir(git_dir,
2182
							       sub->name) < 0)
2183
					die(_("refusing to create/use '%s' in "
2184
					      "another submodule's git dir"),
2185
					    git_dir);
2186
				free(git_dir);
2187
			}
2188
		} else {
2189
			struct strbuf gitdir = STRBUF_INIT;
2190
			submodule_name_to_gitdir(&gitdir, the_repository,
2191
						 sub->name);
2192
			if (validate_submodule_git_dir(gitdir.buf,
2193
						       sub->name) < 0)
2194
				die(_("refusing to create/use '%s' in another "
2195
				      "submodule's git dir"),
2196
				    gitdir.buf);
2197
			connect_work_tree_and_git_dir(path, gitdir.buf, 0);
2198
			strbuf_release(&gitdir);
2199

2200
			/* make sure the index is clean as well */
2201
			submodule_reset_index(path, super_prefix);
2202
		}
2203

2204
		if (old_head && (flags & SUBMODULE_MOVE_HEAD_FORCE)) {
2205
			struct strbuf gitdir = STRBUF_INIT;
2206
			submodule_name_to_gitdir(&gitdir, the_repository,
2207
						 sub->name);
2208
			connect_work_tree_and_git_dir(path, gitdir.buf, 1);
2209
			strbuf_release(&gitdir);
2210
		}
2211
	}
2212

2213
	prepare_submodule_repo_env(&cp.env);
2214

2215
	cp.git_cmd = 1;
2216
	cp.no_stdin = 1;
2217
	cp.dir = path;
2218

2219
	strvec_pushl(&cp.args, "read-tree", "--recurse-submodules", NULL);
2220
	strvec_pushf(&cp.args, "--super-prefix=%s%s/",
2221
		     (super_prefix ? super_prefix : ""), path);
2222

2223
	if (flags & SUBMODULE_MOVE_HEAD_DRY_RUN)
2224
		strvec_push(&cp.args, "-n");
2225
	else
2226
		strvec_push(&cp.args, "-u");
2227

2228
	if (flags & SUBMODULE_MOVE_HEAD_FORCE)
2229
		strvec_push(&cp.args, "--reset");
2230
	else
2231
		strvec_push(&cp.args, "-m");
2232

2233
	if (!(flags & SUBMODULE_MOVE_HEAD_FORCE))
2234
		strvec_push(&cp.args, old_head ? old_head : empty_tree_oid_hex(the_repository->hash_algo));
2235

2236
	strvec_push(&cp.args, new_head ? new_head : empty_tree_oid_hex(the_repository->hash_algo));
2237

2238
	if (run_command(&cp)) {
2239
		ret = error(_("Submodule '%s' could not be updated."), path);
2240
		goto out;
2241
	}
2242

2243
	if (!(flags & SUBMODULE_MOVE_HEAD_DRY_RUN)) {
2244
		if (new_head) {
2245
			child_process_init(&cp);
2246
			/* also set the HEAD accordingly */
2247
			cp.git_cmd = 1;
2248
			cp.no_stdin = 1;
2249
			cp.dir = path;
2250

2251
			prepare_submodule_repo_env(&cp.env);
2252
			strvec_pushl(&cp.args, "update-ref", "HEAD",
2253
				     "--no-deref", new_head, NULL);
2254

2255
			if (run_command(&cp)) {
2256
				ret = -1;
2257
				goto out;
2258
			}
2259
		} else {
2260
			struct strbuf sb = STRBUF_INIT;
2261

2262
			strbuf_addf(&sb, "%s/.git", path);
2263
			unlink_or_warn(sb.buf);
2264
			strbuf_release(&sb);
2265

2266
			if (is_empty_dir(path))
2267
				rmdir_or_warn(path);
2268

2269
			submodule_unset_core_worktree(sub);
2270
		}
2271
	}
2272
out:
2273
	return ret;
2274
}
2275

2276
int validate_submodule_git_dir(char *git_dir, const char *submodule_name)
2277
{
2278
	size_t len = strlen(git_dir), suffix_len = strlen(submodule_name);
2279
	char *p;
2280
	int ret = 0;
2281

2282
	if (len <= suffix_len || (p = git_dir + len - suffix_len)[-1] != '/' ||
2283
	    strcmp(p, submodule_name))
2284
		BUG("submodule name '%s' not a suffix of git dir '%s'",
2285
		    submodule_name, git_dir);
2286

2287
	/*
2288
	 * We prevent the contents of sibling submodules' git directories to
2289
	 * clash.
2290
	 *
2291
	 * Example: having a submodule named `hippo` and another one named
2292
	 * `hippo/hooks` would result in the git directories
2293
	 * `.git/modules/hippo/` and `.git/modules/hippo/hooks/`, respectively,
2294
	 * but the latter directory is already designated to contain the hooks
2295
	 * of the former.
2296
	 */
2297
	for (; *p; p++) {
2298
		if (is_dir_sep(*p)) {
2299
			char c = *p;
2300

2301
			*p = '\0';
2302
			if (is_git_directory(git_dir))
2303
				ret = -1;
2304
			*p = c;
2305

2306
			if (ret < 0)
2307
				return error(_("submodule git dir '%s' is "
2308
					       "inside git dir '%.*s'"),
2309
					     git_dir,
2310
					     (int)(p - git_dir), git_dir);
2311
		}
2312
	}
2313

2314
	return 0;
2315
}
2316

2317
int validate_submodule_path(const char *path)
2318
{
2319
	char *p = xstrdup(path);
2320
	struct stat st;
2321
	int i, ret = 0;
2322
	char sep;
2323

2324
	for (i = 0; !ret && p[i]; i++) {
2325
		if (!is_dir_sep(p[i]))
2326
			continue;
2327

2328
		sep = p[i];
2329
		p[i] = '\0';
2330
		/* allow missing components, but no symlinks */
2331
		ret = lstat(p, &st) || !S_ISLNK(st.st_mode) ? 0 : -1;
2332
		p[i] = sep;
2333
		if (ret)
2334
			error(_("expected '%.*s' in submodule path '%s' not to "
2335
				"be a symbolic link"), i, p, p);
2336
	}
2337
	if (!lstat(p, &st) && S_ISLNK(st.st_mode))
2338
		ret = error(_("expected submodule path '%s' not to be a "
2339
			      "symbolic link"), p);
2340
	free(p);
2341
	return ret;
2342
}
2343

2344

2345
/*
2346
 * Embeds a single submodules git directory into the superprojects git dir,
2347
 * non recursively.
2348
 */
2349
static void relocate_single_git_dir_into_superproject(const char *path,
2350
						      const char *super_prefix)
2351
{
2352
	char *old_git_dir = NULL, *real_old_git_dir = NULL, *real_new_git_dir = NULL;
2353
	struct strbuf new_gitdir = STRBUF_INIT;
2354
	const struct submodule *sub;
2355

2356
	if (validate_submodule_path(path) < 0)
2357
		exit(128);
2358

2359
	if (submodule_uses_worktrees(path))
2360
		die(_("relocate_gitdir for submodule '%s' with "
2361
		      "more than one worktree not supported"), path);
2362

2363
	old_git_dir = xstrfmt("%s/.git", path);
2364
	if (read_gitfile(old_git_dir))
2365
		/* If it is an actual gitfile, it doesn't need migration. */
2366
		return;
2367

2368
	real_old_git_dir = real_pathdup(old_git_dir, 1);
2369

2370
	sub = submodule_from_path(the_repository, null_oid(), path);
2371
	if (!sub)
2372
		die(_("could not lookup name for submodule '%s'"), path);
2373

2374
	submodule_name_to_gitdir(&new_gitdir, the_repository, sub->name);
2375
	if (validate_submodule_git_dir(new_gitdir.buf, sub->name) < 0)
2376
		die(_("refusing to move '%s' into an existing git dir"),
2377
		    real_old_git_dir);
2378
	if (safe_create_leading_directories_const(new_gitdir.buf) < 0)
2379
		die(_("could not create directory '%s'"), new_gitdir.buf);
2380
	real_new_git_dir = real_pathdup(new_gitdir.buf, 1);
2381

2382
	fprintf(stderr, _("Migrating git directory of '%s%s' from\n'%s' to\n'%s'\n"),
2383
		super_prefix ? super_prefix : "", path,
2384
		real_old_git_dir, real_new_git_dir);
2385

2386
	relocate_gitdir(path, real_old_git_dir, real_new_git_dir);
2387

2388
	free(old_git_dir);
2389
	free(real_old_git_dir);
2390
	free(real_new_git_dir);
2391
	strbuf_release(&new_gitdir);
2392
}
2393

2394
static void absorb_git_dir_into_superproject_recurse(const char *path,
2395
						     const char *super_prefix)
2396
{
2397

2398
	struct child_process cp = CHILD_PROCESS_INIT;
2399

2400
	if (validate_submodule_path(path) < 0)
2401
		exit(128);
2402

2403
	cp.dir = path;
2404
	cp.git_cmd = 1;
2405
	cp.no_stdin = 1;
2406
	strvec_pushl(&cp.args, "submodule--helper",
2407
		     "absorbgitdirs", NULL);
2408
	strvec_pushf(&cp.args, "--super-prefix=%s%s/", super_prefix ?
2409
		     super_prefix : "", path);
2410

2411
	prepare_submodule_repo_env(&cp.env);
2412
	if (run_command(&cp))
2413
		die(_("could not recurse into submodule '%s'"), path);
2414
}
2415

2416
/*
2417
 * Migrate the git directory of the submodule given by path from
2418
 * having its git directory within the working tree to the git dir nested
2419
 * in its superprojects git dir under modules/.
2420
 */
2421
void absorb_git_dir_into_superproject(const char *path,
2422
				      const char *super_prefix)
2423
{
2424
	int err_code;
2425
	const char *sub_git_dir;
2426
	struct strbuf gitdir = STRBUF_INIT;
2427

2428
	if (validate_submodule_path(path) < 0)
2429
		exit(128);
2430

2431
	strbuf_addf(&gitdir, "%s/.git", path);
2432
	sub_git_dir = resolve_gitdir_gently(gitdir.buf, &err_code);
2433

2434
	/* Not populated? */
2435
	if (!sub_git_dir) {
2436
		const struct submodule *sub;
2437
		struct strbuf sub_gitdir = STRBUF_INIT;
2438

2439
		if (err_code == READ_GITFILE_ERR_STAT_FAILED) {
2440
			/* unpopulated as expected */
2441
			strbuf_release(&gitdir);
2442
			return;
2443
		}
2444

2445
		if (err_code != READ_GITFILE_ERR_NOT_A_REPO)
2446
			/* We don't know what broke here. */
2447
			read_gitfile_error_die(err_code, path, NULL);
2448

2449
		/*
2450
		* Maybe populated, but no git directory was found?
2451
		* This can happen if the superproject is a submodule
2452
		* itself and was just absorbed. The absorption of the
2453
		* superproject did not rewrite the git file links yet,
2454
		* fix it now.
2455
		*/
2456
		sub = submodule_from_path(the_repository, null_oid(), path);
2457
		if (!sub)
2458
			die(_("could not lookup name for submodule '%s'"), path);
2459
		submodule_name_to_gitdir(&sub_gitdir, the_repository, sub->name);
2460
		connect_work_tree_and_git_dir(path, sub_gitdir.buf, 0);
2461
		strbuf_release(&sub_gitdir);
2462
	} else {
2463
		/* Is it already absorbed into the superprojects git dir? */
2464
		char *real_sub_git_dir = real_pathdup(sub_git_dir, 1);
2465
		char *real_common_git_dir = real_pathdup(get_git_common_dir(), 1);
2466

2467
		if (!starts_with(real_sub_git_dir, real_common_git_dir))
2468
			relocate_single_git_dir_into_superproject(path, super_prefix);
2469

2470
		free(real_sub_git_dir);
2471
		free(real_common_git_dir);
2472
	}
2473
	strbuf_release(&gitdir);
2474

2475
	absorb_git_dir_into_superproject_recurse(path, super_prefix);
2476
}
2477

2478
int get_superproject_working_tree(struct strbuf *buf)
2479
{
2480
	struct child_process cp = CHILD_PROCESS_INIT;
2481
	struct strbuf sb = STRBUF_INIT;
2482
	struct strbuf one_up = STRBUF_INIT;
2483
	char *cwd = xgetcwd();
2484
	int ret = 0;
2485
	const char *subpath;
2486
	int code;
2487
	ssize_t len;
2488

2489
	if (!is_inside_work_tree())
2490
		/*
2491
		 * FIXME:
2492
		 * We might have a superproject, but it is harder
2493
		 * to determine.
2494
		 */
2495
		return 0;
2496

2497
	if (!strbuf_realpath(&one_up, "../", 0))
2498
		return 0;
2499

2500
	subpath = relative_path(cwd, one_up.buf, &sb);
2501
	strbuf_release(&one_up);
2502

2503
	prepare_submodule_repo_env(&cp.env);
2504
	strvec_pop(&cp.env);
2505

2506
	strvec_pushl(&cp.args, "--literal-pathspecs", "-C", "..",
2507
		     "ls-files", "-z", "--stage", "--full-name", "--",
2508
		     subpath, NULL);
2509
	strbuf_reset(&sb);
2510

2511
	cp.no_stdin = 1;
2512
	cp.no_stderr = 1;
2513
	cp.out = -1;
2514
	cp.git_cmd = 1;
2515

2516
	if (start_command(&cp))
2517
		die(_("could not start ls-files in .."));
2518

2519
	len = strbuf_read(&sb, cp.out, PATH_MAX);
2520
	close(cp.out);
2521

2522
	if (starts_with(sb.buf, "160000")) {
2523
		int super_sub_len;
2524
		int cwd_len = strlen(cwd);
2525
		char *super_sub, *super_wt;
2526

2527
		/*
2528
		 * There is a superproject having this repo as a submodule.
2529
		 * The format is <mode> SP <hash> SP <stage> TAB <full name> \0,
2530
		 * We're only interested in the name after the tab.
2531
		 */
2532
		super_sub = strchr(sb.buf, '\t') + 1;
2533
		super_sub_len = strlen(super_sub);
2534

2535
		if (super_sub_len > cwd_len ||
2536
		    strcmp(&cwd[cwd_len - super_sub_len], super_sub))
2537
			BUG("returned path string doesn't match cwd?");
2538

2539
		super_wt = xstrdup(cwd);
2540
		super_wt[cwd_len - super_sub_len] = '\0';
2541

2542
		strbuf_realpath(buf, super_wt, 1);
2543
		ret = 1;
2544
		free(super_wt);
2545
	}
2546
	free(cwd);
2547
	strbuf_release(&sb);
2548

2549
	code = finish_command(&cp);
2550

2551
	if (code == 128)
2552
		/* '../' is not a git repository */
2553
		return 0;
2554
	if (code == 0 && len == 0)
2555
		/* There is an unrelated git repository at '../' */
2556
		return 0;
2557
	if (code)
2558
		die(_("ls-tree returned unexpected return code %d"), code);
2559

2560
	return ret;
2561
}
2562

2563
/*
2564
 * Put the gitdir for a submodule (given relative to the main
2565
 * repository worktree) into `buf`, or return -1 on error.
2566
 */
2567
int submodule_to_gitdir(struct strbuf *buf, const char *submodule)
2568
{
2569
	const struct submodule *sub;
2570
	const char *git_dir;
2571
	int ret = 0;
2572

2573
	if (validate_submodule_path(submodule) < 0)
2574
		exit(128);
2575

2576
	strbuf_reset(buf);
2577
	strbuf_addstr(buf, submodule);
2578
	strbuf_complete(buf, '/');
2579
	strbuf_addstr(buf, ".git");
2580

2581
	git_dir = read_gitfile(buf->buf);
2582
	if (git_dir) {
2583
		strbuf_reset(buf);
2584
		strbuf_addstr(buf, git_dir);
2585
	}
2586
	if (!is_git_directory(buf->buf)) {
2587
		sub = submodule_from_path(the_repository, null_oid(),
2588
					  submodule);
2589
		if (!sub) {
2590
			ret = -1;
2591
			goto cleanup;
2592
		}
2593
		strbuf_reset(buf);
2594
		submodule_name_to_gitdir(buf, the_repository, sub->name);
2595
	}
2596

2597
cleanup:
2598
	return ret;
2599
}
2600

2601
void submodule_name_to_gitdir(struct strbuf *buf, struct repository *r,
2602
			      const char *submodule_name)
2603
{
2604
	/*
2605
	 * NEEDSWORK: The current way of mapping a submodule's name to
2606
	 * its location in .git/modules/ has problems with some naming
2607
	 * schemes. For example, if a submodule is named "foo" and
2608
	 * another is named "foo/bar" (whether present in the same
2609
	 * superproject commit or not - the problem will arise if both
2610
	 * superproject commits have been checked out at any point in
2611
	 * time), or if two submodule names only have different cases in
2612
	 * a case-insensitive filesystem.
2613
	 *
2614
	 * There are several solutions, including encoding the path in
2615
	 * some way, introducing a submodule.<name>.gitdir config in
2616
	 * .git/config (not .gitmodules) that allows overriding what the
2617
	 * gitdir of a submodule would be (and teach Git, upon noticing
2618
	 * a clash, to automatically determine a non-clashing name and
2619
	 * to write such a config), or introducing a
2620
	 * submodule.<name>.gitdir config in .gitmodules that repo
2621
	 * administrators can explicitly set. Nothing has been decided,
2622
	 * so for now, just append the name at the end of the path.
2623
	 */
2624
	strbuf_repo_git_path(buf, r, "modules/");
2625
	strbuf_addstr(buf, submodule_name);
2626
}
2627