Make on-disk index representation separate from in-core one
[git/git.git] / builtin-ls-files.c
1 /*
2 * This merges the file listing in the directory cache index
3 * with the actual working directory list, and shows different
4 * combinations of the two.
5 *
6 * Copyright (C) Linus Torvalds, 2005
7 */
8 #include "cache.h"
9 #include "quote.h"
10 #include "dir.h"
11 #include "builtin.h"
12 #include "tree.h"
13
14 static int abbrev;
15 static int show_deleted;
16 static int show_cached;
17 static int show_others;
18 static int show_stage;
19 static int show_unmerged;
20 static int show_modified;
21 static int show_killed;
22 static int show_valid_bit;
23 static int line_terminator = '\n';
24
25 static int prefix_len;
26 static int prefix_offset;
27 static const char **pathspec;
28 static int error_unmatch;
29 static char *ps_matched;
30 static const char *with_tree;
31
32 static const char *tag_cached = "";
33 static const char *tag_unmerged = "";
34 static const char *tag_removed = "";
35 static const char *tag_other = "";
36 static const char *tag_killed = "";
37 static const char *tag_modified = "";
38
39
40 /*
41 * Match a pathspec against a filename. The first "skiplen" characters
42 * are the common prefix
43 */
44 int pathspec_match(const char **spec, char *ps_matched,
45 const char *filename, int skiplen)
46 {
47 const char *m;
48
49 while ((m = *spec++) != NULL) {
50 int matchlen = strlen(m + skiplen);
51
52 if (!matchlen)
53 goto matched;
54 if (!strncmp(m + skiplen, filename + skiplen, matchlen)) {
55 if (m[skiplen + matchlen - 1] == '/')
56 goto matched;
57 switch (filename[skiplen + matchlen]) {
58 case '/': case '\0':
59 goto matched;
60 }
61 }
62 if (!fnmatch(m + skiplen, filename + skiplen, 0))
63 goto matched;
64 if (ps_matched)
65 ps_matched++;
66 continue;
67 matched:
68 if (ps_matched)
69 *ps_matched = 1;
70 return 1;
71 }
72 return 0;
73 }
74
75 static void show_dir_entry(const char *tag, struct dir_entry *ent)
76 {
77 int len = prefix_len;
78 int offset = prefix_offset;
79
80 if (len >= ent->len)
81 die("git-ls-files: internal error - directory entry not superset of prefix");
82
83 if (pathspec && !pathspec_match(pathspec, ps_matched, ent->name, len))
84 return;
85
86 fputs(tag, stdout);
87 write_name_quoted(ent->name + offset, stdout, line_terminator);
88 }
89
90 static void show_other_files(struct dir_struct *dir)
91 {
92 int i;
93
94
95 /*
96 * Skip matching and unmerged entries for the paths,
97 * since we want just "others".
98 *
99 * (Matching entries are normally pruned during
100 * the directory tree walk, but will show up for
101 * gitlinks because we don't necessarily have
102 * dir->show_other_directories set to suppress
103 * them).
104 */
105 for (i = 0; i < dir->nr; i++) {
106 struct dir_entry *ent = dir->entries[i];
107 int len, pos;
108 struct cache_entry *ce;
109
110 /*
111 * Remove the '/' at the end that directory
112 * walking adds for directory entries.
113 */
114 len = ent->len;
115 if (len && ent->name[len-1] == '/')
116 len--;
117 pos = cache_name_pos(ent->name, len);
118 if (0 <= pos)
119 continue; /* exact match */
120 pos = -pos - 1;
121 if (pos < active_nr) {
122 ce = active_cache[pos];
123 if (ce_namelen(ce) == len &&
124 !memcmp(ce->name, ent->name, len))
125 continue; /* Yup, this one exists unmerged */
126 }
127 show_dir_entry(tag_other, ent);
128 }
129 }
130
131 static void show_killed_files(struct dir_struct *dir)
132 {
133 int i;
134 for (i = 0; i < dir->nr; i++) {
135 struct dir_entry *ent = dir->entries[i];
136 char *cp, *sp;
137 int pos, len, killed = 0;
138
139 for (cp = ent->name; cp - ent->name < ent->len; cp = sp + 1) {
140 sp = strchr(cp, '/');
141 if (!sp) {
142 /* If ent->name is prefix of an entry in the
143 * cache, it will be killed.
144 */
145 pos = cache_name_pos(ent->name, ent->len);
146 if (0 <= pos)
147 die("bug in show-killed-files");
148 pos = -pos - 1;
149 while (pos < active_nr &&
150 ce_stage(active_cache[pos]))
151 pos++; /* skip unmerged */
152 if (active_nr <= pos)
153 break;
154 /* pos points at a name immediately after
155 * ent->name in the cache. Does it expect
156 * ent->name to be a directory?
157 */
158 len = ce_namelen(active_cache[pos]);
159 if ((ent->len < len) &&
160 !strncmp(active_cache[pos]->name,
161 ent->name, ent->len) &&
162 active_cache[pos]->name[ent->len] == '/')
163 killed = 1;
164 break;
165 }
166 if (0 <= cache_name_pos(ent->name, sp - ent->name)) {
167 /* If any of the leading directories in
168 * ent->name is registered in the cache,
169 * ent->name will be killed.
170 */
171 killed = 1;
172 break;
173 }
174 }
175 if (killed)
176 show_dir_entry(tag_killed, dir->entries[i]);
177 }
178 }
179
180 static void show_ce_entry(const char *tag, struct cache_entry *ce)
181 {
182 int len = prefix_len;
183 int offset = prefix_offset;
184
185 if (len >= ce_namelen(ce))
186 die("git-ls-files: internal error - cache entry not superset of prefix");
187
188 if (pathspec && !pathspec_match(pathspec, ps_matched, ce->name, len))
189 return;
190
191 if (tag && *tag && show_valid_bit &&
192 (ce->ce_flags & CE_VALID)) {
193 static char alttag[4];
194 memcpy(alttag, tag, 3);
195 if (isalpha(tag[0]))
196 alttag[0] = tolower(tag[0]);
197 else if (tag[0] == '?')
198 alttag[0] = '!';
199 else {
200 alttag[0] = 'v';
201 alttag[1] = tag[0];
202 alttag[2] = ' ';
203 alttag[3] = 0;
204 }
205 tag = alttag;
206 }
207
208 if (!show_stage) {
209 fputs(tag, stdout);
210 } else {
211 printf("%s%06o %s %d\t",
212 tag,
213 ce->ce_mode,
214 abbrev ? find_unique_abbrev(ce->sha1,abbrev)
215 : sha1_to_hex(ce->sha1),
216 ce_stage(ce));
217 }
218 write_name_quoted(ce->name + offset, stdout, line_terminator);
219 }
220
221 static void show_files(struct dir_struct *dir, const char *prefix)
222 {
223 int i;
224
225 /* For cached/deleted files we don't need to even do the readdir */
226 if (show_others || show_killed) {
227 const char *path = ".", *base = "";
228 int baselen = prefix_len;
229
230 if (baselen)
231 path = base = prefix;
232 read_directory(dir, path, base, baselen, pathspec);
233 if (show_others)
234 show_other_files(dir);
235 if (show_killed)
236 show_killed_files(dir);
237 }
238 if (show_cached | show_stage) {
239 for (i = 0; i < active_nr; i++) {
240 struct cache_entry *ce = active_cache[i];
241 if (excluded(dir, ce->name) != dir->show_ignored)
242 continue;
243 if (show_unmerged && !ce_stage(ce))
244 continue;
245 if (ce->ce_flags & CE_UPDATE)
246 continue;
247 show_ce_entry(ce_stage(ce) ? tag_unmerged : tag_cached, ce);
248 }
249 }
250 if (show_deleted | show_modified) {
251 for (i = 0; i < active_nr; i++) {
252 struct cache_entry *ce = active_cache[i];
253 struct stat st;
254 int err;
255 if (excluded(dir, ce->name) != dir->show_ignored)
256 continue;
257 err = lstat(ce->name, &st);
258 if (show_deleted && err)
259 show_ce_entry(tag_removed, ce);
260 if (show_modified && ce_modified(ce, &st, 0))
261 show_ce_entry(tag_modified, ce);
262 }
263 }
264 }
265
266 /*
267 * Prune the index to only contain stuff starting with "prefix"
268 */
269 static void prune_cache(const char *prefix)
270 {
271 int pos = cache_name_pos(prefix, prefix_len);
272 unsigned int first, last;
273
274 if (pos < 0)
275 pos = -pos-1;
276 memmove(active_cache, active_cache + pos,
277 (active_nr - pos) * sizeof(struct cache_entry *));
278 active_nr -= pos;
279 first = 0;
280 last = active_nr;
281 while (last > first) {
282 int next = (last + first) >> 1;
283 struct cache_entry *ce = active_cache[next];
284 if (!strncmp(ce->name, prefix, prefix_len)) {
285 first = next+1;
286 continue;
287 }
288 last = next;
289 }
290 active_nr = last;
291 }
292
293 static const char *verify_pathspec(const char *prefix)
294 {
295 const char **p, *n, *prev;
296 unsigned long max;
297
298 prev = NULL;
299 max = PATH_MAX;
300 for (p = pathspec; (n = *p) != NULL; p++) {
301 int i, len = 0;
302 for (i = 0; i < max; i++) {
303 char c = n[i];
304 if (prev && prev[i] != c)
305 break;
306 if (!c || c == '*' || c == '?')
307 break;
308 if (c == '/')
309 len = i+1;
310 }
311 prev = n;
312 if (len < max) {
313 max = len;
314 if (!max)
315 break;
316 }
317 }
318
319 if (prefix_offset > max || memcmp(prev, prefix, prefix_offset))
320 die("git-ls-files: cannot generate relative filenames containing '..'");
321
322 prefix_len = max;
323 return max ? xmemdupz(prev, max) : NULL;
324 }
325
326 /*
327 * Read the tree specified with --with-tree option
328 * (typically, HEAD) into stage #1 and then
329 * squash them down to stage #0. This is used for
330 * --error-unmatch to list and check the path patterns
331 * that were given from the command line. We are not
332 * going to write this index out.
333 */
334 void overlay_tree_on_cache(const char *tree_name, const char *prefix)
335 {
336 struct tree *tree;
337 unsigned char sha1[20];
338 const char **match;
339 struct cache_entry *last_stage0 = NULL;
340 int i;
341
342 if (get_sha1(tree_name, sha1))
343 die("tree-ish %s not found.", tree_name);
344 tree = parse_tree_indirect(sha1);
345 if (!tree)
346 die("bad tree-ish %s", tree_name);
347
348 /* Hoist the unmerged entries up to stage #3 to make room */
349 for (i = 0; i < active_nr; i++) {
350 struct cache_entry *ce = active_cache[i];
351 if (!ce_stage(ce))
352 continue;
353 ce->ce_flags |= CE_STAGEMASK;
354 }
355
356 if (prefix) {
357 static const char *(matchbuf[2]);
358 matchbuf[0] = prefix;
359 matchbuf [1] = NULL;
360 match = matchbuf;
361 } else
362 match = NULL;
363 if (read_tree(tree, 1, match))
364 die("unable to read tree entries %s", tree_name);
365
366 for (i = 0; i < active_nr; i++) {
367 struct cache_entry *ce = active_cache[i];
368 switch (ce_stage(ce)) {
369 case 0:
370 last_stage0 = ce;
371 /* fallthru */
372 default:
373 continue;
374 case 1:
375 /*
376 * If there is stage #0 entry for this, we do not
377 * need to show it. We use CE_UPDATE bit to mark
378 * such an entry.
379 */
380 if (last_stage0 &&
381 !strcmp(last_stage0->name, ce->name))
382 ce->ce_flags |= CE_UPDATE;
383 }
384 }
385 }
386
387 int report_path_error(const char *ps_matched, const char **pathspec, int prefix_offset)
388 {
389 /*
390 * Make sure all pathspec matched; otherwise it is an error.
391 */
392 int num, errors = 0;
393 for (num = 0; pathspec[num]; num++) {
394 int other, found_dup;
395
396 if (ps_matched[num])
397 continue;
398 /*
399 * The caller might have fed identical pathspec
400 * twice. Do not barf on such a mistake.
401 */
402 for (found_dup = other = 0;
403 !found_dup && pathspec[other];
404 other++) {
405 if (other == num || !ps_matched[other])
406 continue;
407 if (!strcmp(pathspec[other], pathspec[num]))
408 /*
409 * Ok, we have a match already.
410 */
411 found_dup = 1;
412 }
413 if (found_dup)
414 continue;
415
416 error("pathspec '%s' did not match any file(s) known to git.",
417 pathspec[num] + prefix_offset);
418 errors++;
419 }
420 return errors;
421 }
422
423 static const char ls_files_usage[] =
424 "git-ls-files [-z] [-t] [-v] (--[cached|deleted|others|stage|unmerged|killed|modified])* "
425 "[ --ignored ] [--exclude=<pattern>] [--exclude-from=<file>] "
426 "[ --exclude-per-directory=<filename> ] [--exclude-standard] "
427 "[--full-name] [--abbrev] [--] [<file>]*";
428
429 int cmd_ls_files(int argc, const char **argv, const char *prefix)
430 {
431 int i;
432 int exc_given = 0, require_work_tree = 0;
433 struct dir_struct dir;
434
435 memset(&dir, 0, sizeof(dir));
436 if (prefix)
437 prefix_offset = strlen(prefix);
438 git_config(git_default_config);
439
440 for (i = 1; i < argc; i++) {
441 const char *arg = argv[i];
442
443 if (!strcmp(arg, "--")) {
444 i++;
445 break;
446 }
447 if (!strcmp(arg, "-z")) {
448 line_terminator = 0;
449 continue;
450 }
451 if (!strcmp(arg, "-t") || !strcmp(arg, "-v")) {
452 tag_cached = "H ";
453 tag_unmerged = "M ";
454 tag_removed = "R ";
455 tag_modified = "C ";
456 tag_other = "? ";
457 tag_killed = "K ";
458 if (arg[1] == 'v')
459 show_valid_bit = 1;
460 continue;
461 }
462 if (!strcmp(arg, "-c") || !strcmp(arg, "--cached")) {
463 show_cached = 1;
464 continue;
465 }
466 if (!strcmp(arg, "-d") || !strcmp(arg, "--deleted")) {
467 show_deleted = 1;
468 continue;
469 }
470 if (!strcmp(arg, "-m") || !strcmp(arg, "--modified")) {
471 show_modified = 1;
472 require_work_tree = 1;
473 continue;
474 }
475 if (!strcmp(arg, "-o") || !strcmp(arg, "--others")) {
476 show_others = 1;
477 require_work_tree = 1;
478 continue;
479 }
480 if (!strcmp(arg, "-i") || !strcmp(arg, "--ignored")) {
481 dir.show_ignored = 1;
482 require_work_tree = 1;
483 continue;
484 }
485 if (!strcmp(arg, "-s") || !strcmp(arg, "--stage")) {
486 show_stage = 1;
487 continue;
488 }
489 if (!strcmp(arg, "-k") || !strcmp(arg, "--killed")) {
490 show_killed = 1;
491 require_work_tree = 1;
492 continue;
493 }
494 if (!strcmp(arg, "--directory")) {
495 dir.show_other_directories = 1;
496 continue;
497 }
498 if (!strcmp(arg, "--no-empty-directory")) {
499 dir.hide_empty_directories = 1;
500 continue;
501 }
502 if (!strcmp(arg, "-u") || !strcmp(arg, "--unmerged")) {
503 /* There's no point in showing unmerged unless
504 * you also show the stage information.
505 */
506 show_stage = 1;
507 show_unmerged = 1;
508 continue;
509 }
510 if (!strcmp(arg, "-x") && i+1 < argc) {
511 exc_given = 1;
512 add_exclude(argv[++i], "", 0, &dir.exclude_list[EXC_CMDL]);
513 continue;
514 }
515 if (!prefixcmp(arg, "--exclude=")) {
516 exc_given = 1;
517 add_exclude(arg+10, "", 0, &dir.exclude_list[EXC_CMDL]);
518 continue;
519 }
520 if (!strcmp(arg, "-X") && i+1 < argc) {
521 exc_given = 1;
522 add_excludes_from_file(&dir, argv[++i]);
523 continue;
524 }
525 if (!prefixcmp(arg, "--exclude-from=")) {
526 exc_given = 1;
527 add_excludes_from_file(&dir, arg+15);
528 continue;
529 }
530 if (!prefixcmp(arg, "--exclude-per-directory=")) {
531 exc_given = 1;
532 dir.exclude_per_dir = arg + 24;
533 continue;
534 }
535 if (!strcmp(arg, "--exclude-standard")) {
536 exc_given = 1;
537 setup_standard_excludes(&dir);
538 continue;
539 }
540 if (!strcmp(arg, "--full-name")) {
541 prefix_offset = 0;
542 continue;
543 }
544 if (!strcmp(arg, "--error-unmatch")) {
545 error_unmatch = 1;
546 continue;
547 }
548 if (!prefixcmp(arg, "--with-tree=")) {
549 with_tree = arg + 12;
550 continue;
551 }
552 if (!prefixcmp(arg, "--abbrev=")) {
553 abbrev = strtoul(arg+9, NULL, 10);
554 if (abbrev && abbrev < MINIMUM_ABBREV)
555 abbrev = MINIMUM_ABBREV;
556 else if (abbrev > 40)
557 abbrev = 40;
558 continue;
559 }
560 if (!strcmp(arg, "--abbrev")) {
561 abbrev = DEFAULT_ABBREV;
562 continue;
563 }
564 if (*arg == '-')
565 usage(ls_files_usage);
566 break;
567 }
568
569 if (require_work_tree && !is_inside_work_tree())
570 setup_work_tree();
571
572 pathspec = get_pathspec(prefix, argv + i);
573
574 /* Verify that the pathspec matches the prefix */
575 if (pathspec)
576 prefix = verify_pathspec(prefix);
577
578 /* Treat unmatching pathspec elements as errors */
579 if (pathspec && error_unmatch) {
580 int num;
581 for (num = 0; pathspec[num]; num++)
582 ;
583 ps_matched = xcalloc(1, num);
584 }
585
586 if (dir.show_ignored && !exc_given) {
587 fprintf(stderr, "%s: --ignored needs some exclude pattern\n",
588 argv[0]);
589 exit(1);
590 }
591
592 /* With no flags, we default to showing the cached files */
593 if (!(show_stage | show_deleted | show_others | show_unmerged |
594 show_killed | show_modified))
595 show_cached = 1;
596
597 read_cache();
598 if (prefix)
599 prune_cache(prefix);
600 if (with_tree) {
601 /*
602 * Basic sanity check; show-stages and show-unmerged
603 * would not make any sense with this option.
604 */
605 if (show_stage || show_unmerged)
606 die("ls-files --with-tree is incompatible with -s or -u");
607 overlay_tree_on_cache(with_tree, prefix);
608 }
609 show_files(&dir, prefix);
610
611 if (ps_matched) {
612 int bad;
613 bad = report_path_error(ps_matched, pathspec, prefix_offset);
614 if (bad)
615 fprintf(stderr, "Did you forget to 'git add'?\n");
616
617 return bad ? 1 : 0;
618 }
619
620 return 0;
621 }