Merge branch 'ps/pathspec-empty-prefix-origin'
[git/git.git] / pathspec.c
1 #include "cache.h"
2 #include "dir.h"
3 #include "pathspec.h"
4 #include "attr.h"
5
6 /*
7 * Finds which of the given pathspecs match items in the index.
8 *
9 * For each pathspec, sets the corresponding entry in the seen[] array
10 * (which should be specs items long, i.e. the same size as pathspec)
11 * to the nature of the "closest" (i.e. most specific) match found for
12 * that pathspec in the index, if it was a closer type of match than
13 * the existing entry. As an optimization, matching is skipped
14 * altogether if seen[] already only contains non-zero entries.
15 *
16 * If seen[] has not already been written to, it may make sense
17 * to use find_pathspecs_matching_against_index() instead.
18 */
19 void add_pathspec_matches_against_index(const struct pathspec *pathspec,
20 char *seen)
21 {
22 int num_unmatched = 0, i;
23
24 /*
25 * Since we are walking the index as if we were walking the directory,
26 * we have to mark the matched pathspec as seen; otherwise we will
27 * mistakenly think that the user gave a pathspec that did not match
28 * anything.
29 */
30 for (i = 0; i < pathspec->nr; i++)
31 if (!seen[i])
32 num_unmatched++;
33 if (!num_unmatched)
34 return;
35 for (i = 0; i < active_nr; i++) {
36 const struct cache_entry *ce = active_cache[i];
37 ce_path_match(ce, pathspec, seen);
38 }
39 }
40
41 /*
42 * Finds which of the given pathspecs match items in the index.
43 *
44 * This is a one-shot wrapper around add_pathspec_matches_against_index()
45 * which allocates, populates, and returns a seen[] array indicating the
46 * nature of the "closest" (i.e. most specific) matches which each of the
47 * given pathspecs achieves against all items in the index.
48 */
49 char *find_pathspecs_matching_against_index(const struct pathspec *pathspec)
50 {
51 char *seen = xcalloc(pathspec->nr, 1);
52 add_pathspec_matches_against_index(pathspec, seen);
53 return seen;
54 }
55
56 /*
57 * Magic pathspec
58 *
59 * Possible future magic semantics include stuff like:
60 *
61 * { PATHSPEC_RECURSIVE, '*', "recursive" },
62 * { PATHSPEC_REGEXP, '\0', "regexp" },
63 *
64 */
65
66 static struct pathspec_magic {
67 unsigned bit;
68 char mnemonic; /* this cannot be ':'! */
69 const char *name;
70 } pathspec_magic[] = {
71 { PATHSPEC_FROMTOP, '/', "top" },
72 { PATHSPEC_LITERAL, '\0', "literal" },
73 { PATHSPEC_GLOB, '\0', "glob" },
74 { PATHSPEC_ICASE, '\0', "icase" },
75 { PATHSPEC_EXCLUDE, '!', "exclude" },
76 { PATHSPEC_ATTR, '\0', "attr" },
77 };
78
79 static void prefix_magic(struct strbuf *sb, int prefixlen, unsigned magic)
80 {
81 int i;
82 strbuf_addstr(sb, ":(");
83 for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++)
84 if (magic & pathspec_magic[i].bit) {
85 if (sb->buf[sb->len - 1] != '(')
86 strbuf_addch(sb, ',');
87 strbuf_addstr(sb, pathspec_magic[i].name);
88 }
89 strbuf_addf(sb, ",prefix:%d)", prefixlen);
90 }
91
92 static size_t strcspn_escaped(const char *s, const char *stop)
93 {
94 const char *i;
95
96 for (i = s; *i; i++) {
97 /* skip the escaped character */
98 if (i[0] == '\\' && i[1]) {
99 i++;
100 continue;
101 }
102
103 if (strchr(stop, *i))
104 break;
105 }
106 return i - s;
107 }
108
109 static inline int invalid_value_char(const char ch)
110 {
111 if (isalnum(ch) || strchr(",-_", ch))
112 return 0;
113 return -1;
114 }
115
116 static char *attr_value_unescape(const char *value)
117 {
118 const char *src;
119 char *dst, *ret;
120
121 ret = xmallocz(strlen(value));
122 for (src = value, dst = ret; *src; src++, dst++) {
123 if (*src == '\\') {
124 if (!src[1])
125 die(_("Escape character '\\' not allowed as "
126 "last character in attr value"));
127 src++;
128 }
129 if (invalid_value_char(*src))
130 die("cannot use '%c' for value matching", *src);
131 *dst = *src;
132 }
133 *dst = '\0';
134 return ret;
135 }
136
137 static void parse_pathspec_attr_match(struct pathspec_item *item, const char *value)
138 {
139 struct string_list_item *si;
140 struct string_list list = STRING_LIST_INIT_DUP;
141
142 if (item->attr_check || item->attr_match)
143 die(_("Only one 'attr:' specification is allowed."));
144
145 if (!value || !*value)
146 die(_("attr spec must not be empty"));
147
148 string_list_split(&list, value, ' ', -1);
149 string_list_remove_empty_items(&list, 0);
150
151 item->attr_check = attr_check_alloc();
152 item->attr_match = xcalloc(list.nr, sizeof(struct attr_match));
153
154 for_each_string_list_item(si, &list) {
155 size_t attr_len;
156 char *attr_name;
157 const struct git_attr *a;
158
159 int j = item->attr_match_nr++;
160 const char *attr = si->string;
161 struct attr_match *am = &item->attr_match[j];
162
163 switch (*attr) {
164 case '!':
165 am->match_mode = MATCH_UNSPECIFIED;
166 attr++;
167 attr_len = strlen(attr);
168 break;
169 case '-':
170 am->match_mode = MATCH_UNSET;
171 attr++;
172 attr_len = strlen(attr);
173 break;
174 default:
175 attr_len = strcspn(attr, "=");
176 if (attr[attr_len] != '=')
177 am->match_mode = MATCH_SET;
178 else {
179 const char *v = &attr[attr_len + 1];
180 am->match_mode = MATCH_VALUE;
181 am->value = attr_value_unescape(v);
182 }
183 break;
184 }
185
186 attr_name = xmemdupz(attr, attr_len);
187 a = git_attr(attr_name);
188 if (!a)
189 die(_("invalid attribute name %s"), attr_name);
190
191 attr_check_append(item->attr_check, a);
192
193 free(attr_name);
194 }
195
196 if (item->attr_check->nr != item->attr_match_nr)
197 die("BUG: should have same number of entries");
198
199 string_list_clear(&list, 0);
200 }
201
202 static inline int get_literal_global(void)
203 {
204 static int literal = -1;
205
206 if (literal < 0)
207 literal = git_env_bool(GIT_LITERAL_PATHSPECS_ENVIRONMENT, 0);
208
209 return literal;
210 }
211
212 static inline int get_glob_global(void)
213 {
214 static int glob = -1;
215
216 if (glob < 0)
217 glob = git_env_bool(GIT_GLOB_PATHSPECS_ENVIRONMENT, 0);
218
219 return glob;
220 }
221
222 static inline int get_noglob_global(void)
223 {
224 static int noglob = -1;
225
226 if (noglob < 0)
227 noglob = git_env_bool(GIT_NOGLOB_PATHSPECS_ENVIRONMENT, 0);
228
229 return noglob;
230 }
231
232 static inline int get_icase_global(void)
233 {
234 static int icase = -1;
235
236 if (icase < 0)
237 icase = git_env_bool(GIT_ICASE_PATHSPECS_ENVIRONMENT, 0);
238
239 return icase;
240 }
241
242 static int get_global_magic(int element_magic)
243 {
244 int global_magic = 0;
245
246 if (get_literal_global())
247 global_magic |= PATHSPEC_LITERAL;
248
249 /* --glob-pathspec is overridden by :(literal) */
250 if (get_glob_global() && !(element_magic & PATHSPEC_LITERAL))
251 global_magic |= PATHSPEC_GLOB;
252
253 if (get_glob_global() && get_noglob_global())
254 die(_("global 'glob' and 'noglob' pathspec settings are incompatible"));
255
256 if (get_icase_global())
257 global_magic |= PATHSPEC_ICASE;
258
259 if ((global_magic & PATHSPEC_LITERAL) &&
260 (global_magic & ~PATHSPEC_LITERAL))
261 die(_("global 'literal' pathspec setting is incompatible "
262 "with all other global pathspec settings"));
263
264 /* --noglob-pathspec adds :(literal) _unless_ :(glob) is specified */
265 if (get_noglob_global() && !(element_magic & PATHSPEC_GLOB))
266 global_magic |= PATHSPEC_LITERAL;
267
268 return global_magic;
269 }
270
271 /*
272 * Parse the pathspec element looking for long magic
273 *
274 * saves all magic in 'magic'
275 * if prefix magic is used, save the prefix length in 'prefix_len'
276 * returns the position in 'elem' after all magic has been parsed
277 */
278 static const char *parse_long_magic(unsigned *magic, int *prefix_len,
279 struct pathspec_item *item,
280 const char *elem)
281 {
282 const char *pos;
283 const char *nextat;
284
285 for (pos = elem + 2; *pos && *pos != ')'; pos = nextat) {
286 size_t len = strcspn_escaped(pos, ",)");
287 int i;
288
289 if (pos[len] == ',')
290 nextat = pos + len + 1; /* handle ',' */
291 else
292 nextat = pos + len; /* handle ')' and '\0' */
293
294 if (!len)
295 continue;
296
297 if (starts_with(pos, "prefix:")) {
298 char *endptr;
299 *prefix_len = strtol(pos + 7, &endptr, 10);
300 if (endptr - pos != len)
301 die(_("invalid parameter for pathspec magic 'prefix'"));
302 continue;
303 }
304
305 if (starts_with(pos, "attr:")) {
306 char *attr_body = xmemdupz(pos + 5, len - 5);
307 parse_pathspec_attr_match(item, attr_body);
308 *magic |= PATHSPEC_ATTR;
309 free(attr_body);
310 continue;
311 }
312
313 for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
314 if (strlen(pathspec_magic[i].name) == len &&
315 !strncmp(pathspec_magic[i].name, pos, len)) {
316 *magic |= pathspec_magic[i].bit;
317 break;
318 }
319 }
320
321 if (ARRAY_SIZE(pathspec_magic) <= i)
322 die(_("Invalid pathspec magic '%.*s' in '%s'"),
323 (int) len, pos, elem);
324 }
325
326 if (*pos != ')')
327 die(_("Missing ')' at the end of pathspec magic in '%s'"),
328 elem);
329 pos++;
330
331 return pos;
332 }
333
334 /*
335 * Parse the pathspec element looking for short magic
336 *
337 * saves all magic in 'magic'
338 * returns the position in 'elem' after all magic has been parsed
339 */
340 static const char *parse_short_magic(unsigned *magic, const char *elem)
341 {
342 const char *pos;
343
344 for (pos = elem + 1; *pos && *pos != ':'; pos++) {
345 char ch = *pos;
346 int i;
347
348 /* Special case alias for '!' */
349 if (ch == '^') {
350 *magic |= PATHSPEC_EXCLUDE;
351 continue;
352 }
353
354 if (!is_pathspec_magic(ch))
355 break;
356
357 for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
358 if (pathspec_magic[i].mnemonic == ch) {
359 *magic |= pathspec_magic[i].bit;
360 break;
361 }
362 }
363
364 if (ARRAY_SIZE(pathspec_magic) <= i)
365 die(_("Unimplemented pathspec magic '%c' in '%s'"),
366 ch, elem);
367 }
368
369 if (*pos == ':')
370 pos++;
371
372 return pos;
373 }
374
375 static const char *parse_element_magic(unsigned *magic, int *prefix_len,
376 struct pathspec_item *item,
377 const char *elem)
378 {
379 if (elem[0] != ':' || get_literal_global())
380 return elem; /* nothing to do */
381 else if (elem[1] == '(')
382 /* longhand */
383 return parse_long_magic(magic, prefix_len, item, elem);
384 else
385 /* shorthand */
386 return parse_short_magic(magic, elem);
387 }
388
389 static void strip_submodule_slash_cheap(struct pathspec_item *item)
390 {
391 if (item->len >= 1 && item->match[item->len - 1] == '/') {
392 int i = cache_name_pos(item->match, item->len - 1);
393
394 if (i >= 0 && S_ISGITLINK(active_cache[i]->ce_mode)) {
395 item->len--;
396 item->match[item->len] = '\0';
397 }
398 }
399 }
400
401 static void strip_submodule_slash_expensive(struct pathspec_item *item)
402 {
403 int i;
404
405 for (i = 0; i < active_nr; i++) {
406 struct cache_entry *ce = active_cache[i];
407 int ce_len = ce_namelen(ce);
408
409 if (!S_ISGITLINK(ce->ce_mode))
410 continue;
411
412 if (item->len <= ce_len || item->match[ce_len] != '/' ||
413 memcmp(ce->name, item->match, ce_len))
414 continue;
415
416 if (item->len == ce_len + 1) {
417 /* strip trailing slash */
418 item->len--;
419 item->match[item->len] = '\0';
420 } else {
421 die(_("Pathspec '%s' is in submodule '%.*s'"),
422 item->original, ce_len, ce->name);
423 }
424 }
425 }
426
427 static void die_inside_submodule_path(struct pathspec_item *item)
428 {
429 int i;
430
431 for (i = 0; i < active_nr; i++) {
432 struct cache_entry *ce = active_cache[i];
433 int ce_len = ce_namelen(ce);
434
435 if (!S_ISGITLINK(ce->ce_mode))
436 continue;
437
438 if (item->len < ce_len ||
439 !(item->match[ce_len] == '/' || item->match[ce_len] == '\0') ||
440 memcmp(ce->name, item->match, ce_len))
441 continue;
442
443 die(_("Pathspec '%s' is in submodule '%.*s'"),
444 item->original, ce_len, ce->name);
445 }
446 }
447
448 /*
449 * Perform the initialization of a pathspec_item based on a pathspec element.
450 */
451 static void init_pathspec_item(struct pathspec_item *item, unsigned flags,
452 const char *prefix, int prefixlen,
453 const char *elt)
454 {
455 unsigned magic = 0, element_magic = 0;
456 const char *copyfrom = elt;
457 char *match;
458 int pathspec_prefix = -1;
459
460 item->attr_check = NULL;
461 item->attr_match = NULL;
462 item->attr_match_nr = 0;
463
464 /* PATHSPEC_LITERAL_PATH ignores magic */
465 if (flags & PATHSPEC_LITERAL_PATH) {
466 magic = PATHSPEC_LITERAL;
467 } else {
468 copyfrom = parse_element_magic(&element_magic,
469 &pathspec_prefix,
470 item,
471 elt);
472 magic |= element_magic;
473 magic |= get_global_magic(element_magic);
474 }
475
476 item->magic = magic;
477
478 if (pathspec_prefix >= 0 &&
479 (prefixlen || (prefix && *prefix)))
480 die("BUG: 'prefix' magic is supposed to be used at worktree's root");
481
482 if ((magic & PATHSPEC_LITERAL) && (magic & PATHSPEC_GLOB))
483 die(_("%s: 'literal' and 'glob' are incompatible"), elt);
484
485 /* Create match string which will be used for pathspec matching */
486 if (pathspec_prefix >= 0) {
487 match = xstrdup(copyfrom);
488 prefixlen = pathspec_prefix;
489 } else if (magic & PATHSPEC_FROMTOP) {
490 match = xstrdup(copyfrom);
491 prefixlen = 0;
492 } else {
493 match = prefix_path_gently(prefix, prefixlen,
494 &prefixlen, copyfrom);
495 if (!match)
496 die(_("%s: '%s' is outside repository"), elt, copyfrom);
497 }
498
499 item->match = match;
500 item->len = strlen(item->match);
501 item->prefix = prefixlen;
502
503 /*
504 * Prefix the pathspec (keep all magic) and assign to
505 * original. Useful for passing to another command.
506 */
507 if ((flags & PATHSPEC_PREFIX_ORIGIN) &&
508 !get_literal_global()) {
509 struct strbuf sb = STRBUF_INIT;
510
511 /* Preserve the actual prefix length of each pattern */
512 prefix_magic(&sb, prefixlen, element_magic);
513
514 strbuf_addstr(&sb, match);
515 item->original = strbuf_detach(&sb, NULL);
516 } else {
517 item->original = xstrdup(elt);
518 }
519
520 if (flags & PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAP)
521 strip_submodule_slash_cheap(item);
522
523 if (flags & PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE)
524 strip_submodule_slash_expensive(item);
525
526 if (magic & PATHSPEC_LITERAL) {
527 item->nowildcard_len = item->len;
528 } else {
529 item->nowildcard_len = simple_length(item->match);
530 if (item->nowildcard_len < prefixlen)
531 item->nowildcard_len = prefixlen;
532 }
533
534 item->flags = 0;
535 if (magic & PATHSPEC_GLOB) {
536 /*
537 * FIXME: should we enable ONESTAR in _GLOB for
538 * pattern "* * / * . c"?
539 */
540 } else {
541 if (item->nowildcard_len < item->len &&
542 item->match[item->nowildcard_len] == '*' &&
543 no_wildcard(item->match + item->nowildcard_len + 1))
544 item->flags |= PATHSPEC_ONESTAR;
545 }
546
547 /* sanity checks, pathspec matchers assume these are sane */
548 if (item->nowildcard_len > item->len ||
549 item->prefix > item->len) {
550 /*
551 * This case can be triggered by the user pointing us to a
552 * pathspec inside a submodule, which is an input error.
553 * Detect that here and complain, but fallback in the
554 * non-submodule case to a BUG, as we have no idea what
555 * would trigger that.
556 */
557 die_inside_submodule_path(item);
558 die ("BUG: item->nowildcard_len > item->len || item->prefix > item->len)");
559 }
560 }
561
562 static int pathspec_item_cmp(const void *a_, const void *b_)
563 {
564 struct pathspec_item *a, *b;
565
566 a = (struct pathspec_item *)a_;
567 b = (struct pathspec_item *)b_;
568 return strcmp(a->match, b->match);
569 }
570
571 static void NORETURN unsupported_magic(const char *pattern,
572 unsigned magic)
573 {
574 struct strbuf sb = STRBUF_INIT;
575 int i;
576 for (i = 0; i < ARRAY_SIZE(pathspec_magic); i++) {
577 const struct pathspec_magic *m = pathspec_magic + i;
578 if (!(magic & m->bit))
579 continue;
580 if (sb.len)
581 strbuf_addstr(&sb, ", ");
582
583 if (m->mnemonic)
584 strbuf_addf(&sb, _("'%s' (mnemonic: '%c')"),
585 m->name, m->mnemonic);
586 else
587 strbuf_addf(&sb, "'%s'", m->name);
588 }
589 /*
590 * We may want to substitute "this command" with a command
591 * name. E.g. when add--interactive dies when running
592 * "checkout -p"
593 */
594 die(_("%s: pathspec magic not supported by this command: %s"),
595 pattern, sb.buf);
596 }
597
598 /*
599 * Given command line arguments and a prefix, convert the input to
600 * pathspec. die() if any magic in magic_mask is used.
601 */
602 void parse_pathspec(struct pathspec *pathspec,
603 unsigned magic_mask, unsigned flags,
604 const char *prefix, const char **argv)
605 {
606 struct pathspec_item *item;
607 const char *entry = argv ? *argv : NULL;
608 int i, n, prefixlen, warn_empty_string, nr_exclude = 0;
609
610 memset(pathspec, 0, sizeof(*pathspec));
611
612 if (flags & PATHSPEC_MAXDEPTH_VALID)
613 pathspec->magic |= PATHSPEC_MAXDEPTH;
614
615 /* No arguments, no prefix -> no pathspec */
616 if (!entry && !prefix)
617 return;
618
619 if ((flags & PATHSPEC_PREFER_CWD) &&
620 (flags & PATHSPEC_PREFER_FULL))
621 die("BUG: PATHSPEC_PREFER_CWD and PATHSPEC_PREFER_FULL are incompatible");
622
623 /* No arguments with prefix -> prefix pathspec */
624 if (!entry) {
625 if (flags & PATHSPEC_PREFER_FULL)
626 return;
627
628 if (!(flags & PATHSPEC_PREFER_CWD))
629 die("BUG: PATHSPEC_PREFER_CWD requires arguments");
630
631 pathspec->items = item = xcalloc(1, sizeof(*item));
632 item->match = xstrdup(prefix);
633 item->original = xstrdup(prefix);
634 item->nowildcard_len = item->len = strlen(prefix);
635 item->prefix = item->len;
636 pathspec->nr = 1;
637 return;
638 }
639
640 n = 0;
641 warn_empty_string = 1;
642 while (argv[n]) {
643 if (*argv[n] == '\0' && warn_empty_string) {
644 warning(_("empty strings as pathspecs will be made invalid in upcoming releases. "
645 "please use . instead if you meant to match all paths"));
646 warn_empty_string = 0;
647 }
648 n++;
649 }
650
651 pathspec->nr = n;
652 ALLOC_ARRAY(pathspec->items, n + 1);
653 item = pathspec->items;
654 prefixlen = prefix ? strlen(prefix) : 0;
655
656 for (i = 0; i < n; i++) {
657 entry = argv[i];
658
659 init_pathspec_item(item + i, flags, prefix, prefixlen, entry);
660
661 if (item[i].magic & PATHSPEC_EXCLUDE)
662 nr_exclude++;
663 if (item[i].magic & magic_mask)
664 unsupported_magic(entry, item[i].magic & magic_mask);
665
666 if ((flags & PATHSPEC_SYMLINK_LEADING_PATH) &&
667 has_symlink_leading_path(item[i].match, item[i].len)) {
668 die(_("pathspec '%s' is beyond a symbolic link"), entry);
669 }
670
671 if (item[i].nowildcard_len < item[i].len)
672 pathspec->has_wildcard = 1;
673 pathspec->magic |= item[i].magic;
674 }
675
676 /*
677 * If everything is an exclude pattern, add one positive pattern
678 * that matches everyting. We allocated an extra one for this.
679 */
680 if (nr_exclude == n) {
681 int plen = (!(flags & PATHSPEC_PREFER_CWD)) ? 0 : prefixlen;
682 init_pathspec_item(item + n, 0, prefix, plen, "");
683 pathspec->nr++;
684 }
685
686 if (pathspec->magic & PATHSPEC_MAXDEPTH) {
687 if (flags & PATHSPEC_KEEP_ORDER)
688 die("BUG: PATHSPEC_MAXDEPTH_VALID and PATHSPEC_KEEP_ORDER are incompatible");
689 QSORT(pathspec->items, pathspec->nr, pathspec_item_cmp);
690 }
691 }
692
693 void copy_pathspec(struct pathspec *dst, const struct pathspec *src)
694 {
695 int i, j;
696
697 *dst = *src;
698 ALLOC_ARRAY(dst->items, dst->nr);
699 COPY_ARRAY(dst->items, src->items, dst->nr);
700
701 for (i = 0; i < dst->nr; i++) {
702 struct pathspec_item *d = &dst->items[i];
703 struct pathspec_item *s = &src->items[i];
704
705 d->match = xstrdup(s->match);
706 d->original = xstrdup(s->original);
707
708 ALLOC_ARRAY(d->attr_match, d->attr_match_nr);
709 COPY_ARRAY(d->attr_match, s->attr_match, d->attr_match_nr);
710 for (j = 0; j < d->attr_match_nr; j++) {
711 const char *value = s->attr_match[j].value;
712 d->attr_match[j].value = xstrdup_or_null(value);
713 }
714
715 d->attr_check = attr_check_dup(s->attr_check);
716 }
717 }
718
719 void clear_pathspec(struct pathspec *pathspec)
720 {
721 int i, j;
722
723 for (i = 0; i < pathspec->nr; i++) {
724 free(pathspec->items[i].match);
725 free(pathspec->items[i].original);
726
727 for (j = 0; j < pathspec->items[i].attr_match_nr; j++)
728 free(pathspec->items[i].attr_match[j].value);
729 free(pathspec->items[i].attr_match);
730
731 if (pathspec->items[i].attr_check)
732 attr_check_free(pathspec->items[i].attr_check);
733 }
734
735 free(pathspec->items);
736 pathspec->items = NULL;
737 pathspec->nr = 0;
738 }