]>
Commit | Line | Data |
---|---|---|
8bc9a0c7 LT |
1 | /* |
2 | * GIT - The information manager from hell | |
3 | * | |
4 | * Copyright (C) Linus Torvalds, 2005 | |
5 | */ | |
4aab5b46 | 6 | #define NO_THE_INDEX_COMPATIBILITY_MACROS |
e83c5163 | 7 | #include "cache.h" |
bad68ec9 | 8 | #include "cache-tree.h" |
f35a6d3b | 9 | #include "refs.h" |
d616813d | 10 | #include "dir.h" |
bad68ec9 JH |
11 | |
12 | /* Index extensions. | |
13 | * | |
14 | * The first letter should be 'A'..'Z' for extensions that are not | |
15 | * necessary for a correct operation (i.e. optimization data). | |
16 | * When new extensions are added that _needs_ to be understood in | |
17 | * order to correctly interpret the index file, pick character that | |
18 | * is outside the range, to cause the reader to abort. | |
19 | */ | |
20 | ||
21 | #define CACHE_EXT(s) ( (s[0]<<24)|(s[1]<<16)|(s[2]<<8)|(s[3]) ) | |
22 | #define CACHE_EXT_TREE 0x54524545 /* "TREE" */ | |
e83c5163 | 23 | |
228e94f9 | 24 | struct index_state the_index; |
8fd2cb40 | 25 | |
cf558704 LT |
26 | static unsigned int hash_name(const char *name, int namelen) |
27 | { | |
28 | unsigned int hash = 0x123; | |
29 | ||
30 | do { | |
31 | unsigned char c = *name++; | |
32 | hash = hash*101 + c; | |
33 | } while (--namelen); | |
34 | return hash; | |
35 | } | |
36 | ||
9cb76b8c | 37 | static void hash_index_entry(struct index_state *istate, struct cache_entry *ce) |
cf558704 LT |
38 | { |
39 | void **pos; | |
40 | unsigned int hash = hash_name(ce->name, ce_namelen(ce)); | |
41 | ||
cf558704 LT |
42 | pos = insert_hash(hash, ce, &istate->name_hash); |
43 | if (pos) { | |
44 | ce->next = *pos; | |
45 | *pos = ce; | |
46 | } | |
47 | } | |
48 | ||
9cb76b8c JH |
49 | static void lazy_init_name_hash(struct index_state *istate) |
50 | { | |
51 | int nr; | |
52 | ||
53 | if (istate->name_hash_initialized) | |
54 | return; | |
55 | for (nr = 0; nr < istate->cache_nr; nr++) | |
56 | hash_index_entry(istate, istate->cache[nr]); | |
57 | istate->name_hash_initialized = 1; | |
58 | } | |
59 | ||
60 | static void set_index_entry(struct index_state *istate, int nr, struct cache_entry *ce) | |
61 | { | |
62 | istate->cache[nr] = ce; | |
63 | if (istate->name_hash_initialized) | |
64 | hash_index_entry(istate, ce); | |
65 | } | |
66 | ||
cf558704 LT |
67 | /* |
68 | * We don't actually *remove* it, we can just mark it invalid so that | |
69 | * we won't find it in lookups. | |
70 | * | |
71 | * Not only would we have to search the lists (simple enough), but | |
72 | * we'd also have to rehash other hash buckets in case this makes the | |
73 | * hash bucket empty (common). So it's much better to just mark | |
74 | * it. | |
75 | */ | |
76 | static void remove_hash_entry(struct index_state *istate, struct cache_entry *ce) | |
77 | { | |
78 | ce->ce_flags |= CE_UNHASHED; | |
79 | } | |
80 | ||
81 | static void replace_index_entry(struct index_state *istate, int nr, struct cache_entry *ce) | |
82 | { | |
83 | struct cache_entry *old = istate->cache[nr]; | |
84 | ||
85 | if (ce != old) { | |
86 | remove_hash_entry(istate, old); | |
87 | set_index_entry(istate, nr, ce); | |
88 | } | |
89 | istate->cache_changed = 1; | |
90 | } | |
91 | ||
92 | int index_name_exists(struct index_state *istate, const char *name, int namelen) | |
93 | { | |
94 | unsigned int hash = hash_name(name, namelen); | |
9cb76b8c JH |
95 | struct cache_entry *ce; |
96 | ||
97 | lazy_init_name_hash(istate); | |
98 | ce = lookup_hash(hash, &istate->name_hash); | |
cf558704 LT |
99 | |
100 | while (ce) { | |
101 | if (!(ce->ce_flags & CE_UNHASHED)) { | |
102 | if (!cache_name_compare(name, namelen, ce->name, ce->ce_flags)) | |
103 | return 1; | |
104 | } | |
105 | ce = ce->next; | |
106 | } | |
107 | return 0; | |
108 | } | |
109 | ||
415e96c8 JH |
110 | /* |
111 | * This only updates the "non-critical" parts of the directory | |
112 | * cache, ie the parts that aren't tracked by GIT, and only used | |
113 | * to validate the cache. | |
114 | */ | |
115 | void fill_stat_cache_info(struct cache_entry *ce, struct stat *st) | |
116 | { | |
7a51ed66 LT |
117 | ce->ce_ctime = st->st_ctime; |
118 | ce->ce_mtime = st->st_mtime; | |
119 | ce->ce_dev = st->st_dev; | |
120 | ce->ce_ino = st->st_ino; | |
121 | ce->ce_uid = st->st_uid; | |
122 | ce->ce_gid = st->st_gid; | |
123 | ce->ce_size = st->st_size; | |
5f73076c JH |
124 | |
125 | if (assume_unchanged) | |
7a51ed66 | 126 | ce->ce_flags |= CE_VALID; |
eadb5831 JH |
127 | |
128 | if (S_ISREG(st->st_mode)) | |
129 | ce_mark_uptodate(ce); | |
415e96c8 JH |
130 | } |
131 | ||
29e4d363 JH |
132 | static int ce_compare_data(struct cache_entry *ce, struct stat *st) |
133 | { | |
134 | int match = -1; | |
135 | int fd = open(ce->name, O_RDONLY); | |
136 | ||
137 | if (fd >= 0) { | |
138 | unsigned char sha1[20]; | |
53bca91a | 139 | if (!index_fd(sha1, fd, st, 0, OBJ_BLOB, ce->name)) |
a89fccd2 | 140 | match = hashcmp(sha1, ce->sha1); |
7f8508e8 | 141 | /* index_fd() closed the file descriptor already */ |
29e4d363 JH |
142 | } |
143 | return match; | |
144 | } | |
145 | ||
dc49cd76 | 146 | static int ce_compare_link(struct cache_entry *ce, size_t expected_size) |
29e4d363 JH |
147 | { |
148 | int match = -1; | |
149 | char *target; | |
150 | void *buffer; | |
151 | unsigned long size; | |
21666f1a | 152 | enum object_type type; |
29e4d363 JH |
153 | int len; |
154 | ||
155 | target = xmalloc(expected_size); | |
156 | len = readlink(ce->name, target, expected_size); | |
157 | if (len != expected_size) { | |
158 | free(target); | |
159 | return -1; | |
160 | } | |
21666f1a | 161 | buffer = read_sha1_file(ce->sha1, &type, &size); |
29e4d363 JH |
162 | if (!buffer) { |
163 | free(target); | |
164 | return -1; | |
165 | } | |
166 | if (size == expected_size) | |
167 | match = memcmp(buffer, target, size); | |
168 | free(buffer); | |
169 | free(target); | |
170 | return match; | |
171 | } | |
172 | ||
f35a6d3b LT |
173 | static int ce_compare_gitlink(struct cache_entry *ce) |
174 | { | |
175 | unsigned char sha1[20]; | |
176 | ||
177 | /* | |
178 | * We don't actually require that the .git directory | |
302b9282 | 179 | * under GITLINK directory be a valid git directory. It |
f35a6d3b LT |
180 | * might even be missing (in case nobody populated that |
181 | * sub-project). | |
182 | * | |
183 | * If so, we consider it always to match. | |
184 | */ | |
185 | if (resolve_gitlink_ref(ce->name, "HEAD", sha1) < 0) | |
186 | return 0; | |
187 | return hashcmp(sha1, ce->sha1); | |
188 | } | |
189 | ||
29e4d363 JH |
190 | static int ce_modified_check_fs(struct cache_entry *ce, struct stat *st) |
191 | { | |
192 | switch (st->st_mode & S_IFMT) { | |
193 | case S_IFREG: | |
194 | if (ce_compare_data(ce, st)) | |
195 | return DATA_CHANGED; | |
196 | break; | |
197 | case S_IFLNK: | |
dc49cd76 | 198 | if (ce_compare_link(ce, xsize_t(st->st_size))) |
29e4d363 JH |
199 | return DATA_CHANGED; |
200 | break; | |
a8ee75bc | 201 | case S_IFDIR: |
7a51ed66 | 202 | if (S_ISGITLINK(ce->ce_mode)) |
a8ee75bc | 203 | return 0; |
29e4d363 JH |
204 | default: |
205 | return TYPE_CHANGED; | |
206 | } | |
207 | return 0; | |
208 | } | |
209 | ||
407c8eb0 | 210 | static int ce_match_stat_basic(struct cache_entry *ce, struct stat *st) |
734aab75 LT |
211 | { |
212 | unsigned int changed = 0; | |
213 | ||
7a51ed66 LT |
214 | if (ce->ce_flags & CE_REMOVE) |
215 | return MODE_CHANGED | DATA_CHANGED | TYPE_CHANGED; | |
216 | ||
217 | switch (ce->ce_mode & S_IFMT) { | |
8ae0a8c5 KS |
218 | case S_IFREG: |
219 | changed |= !S_ISREG(st->st_mode) ? TYPE_CHANGED : 0; | |
3e09cdfd JH |
220 | /* We consider only the owner x bit to be relevant for |
221 | * "mode changes" | |
222 | */ | |
223 | if (trust_executable_bit && | |
7a51ed66 | 224 | (0100 & (ce->ce_mode ^ st->st_mode))) |
ffbe1add | 225 | changed |= MODE_CHANGED; |
8ae0a8c5 KS |
226 | break; |
227 | case S_IFLNK: | |
78a8d641 JS |
228 | if (!S_ISLNK(st->st_mode) && |
229 | (has_symlinks || !S_ISREG(st->st_mode))) | |
230 | changed |= TYPE_CHANGED; | |
8ae0a8c5 | 231 | break; |
302b9282 | 232 | case S_IFGITLINK: |
f35a6d3b LT |
233 | if (!S_ISDIR(st->st_mode)) |
234 | changed |= TYPE_CHANGED; | |
235 | else if (ce_compare_gitlink(ce)) | |
236 | changed |= DATA_CHANGED; | |
a8ee75bc | 237 | return changed; |
8ae0a8c5 | 238 | default: |
7a51ed66 | 239 | die("internal error: ce_mode is %o", ce->ce_mode); |
8ae0a8c5 | 240 | } |
7a51ed66 | 241 | if (ce->ce_mtime != (unsigned int) st->st_mtime) |
ccc4feb5 | 242 | changed |= MTIME_CHANGED; |
7a51ed66 | 243 | if (ce->ce_ctime != (unsigned int) st->st_ctime) |
734aab75 | 244 | changed |= CTIME_CHANGED; |
ccc4feb5 | 245 | |
7a51ed66 LT |
246 | if (ce->ce_uid != (unsigned int) st->st_uid || |
247 | ce->ce_gid != (unsigned int) st->st_gid) | |
734aab75 | 248 | changed |= OWNER_CHANGED; |
7a51ed66 | 249 | if (ce->ce_ino != (unsigned int) st->st_ino) |
734aab75 | 250 | changed |= INODE_CHANGED; |
2cb45e95 LT |
251 | |
252 | #ifdef USE_STDEV | |
253 | /* | |
254 | * st_dev breaks on network filesystems where different | |
255 | * clients will have different views of what "device" | |
256 | * the filesystem is on | |
257 | */ | |
7a51ed66 | 258 | if (ce->ce_dev != (unsigned int) st->st_dev) |
2cb45e95 LT |
259 | changed |= INODE_CHANGED; |
260 | #endif | |
261 | ||
7a51ed66 | 262 | if (ce->ce_size != (unsigned int) st->st_size) |
734aab75 | 263 | changed |= DATA_CHANGED; |
b0391890 | 264 | |
407c8eb0 JH |
265 | return changed; |
266 | } | |
267 | ||
6d91da6d JH |
268 | static int is_racy_timestamp(struct index_state *istate, struct cache_entry *ce) |
269 | { | |
270 | return (istate->timestamp && | |
271 | ((unsigned int)istate->timestamp) <= ce->ce_mtime); | |
272 | } | |
273 | ||
4aab5b46 | 274 | int ie_match_stat(struct index_state *istate, |
4bd5b7da JH |
275 | struct cache_entry *ce, struct stat *st, |
276 | unsigned int options) | |
407c8eb0 | 277 | { |
5f73076c | 278 | unsigned int changed; |
4bd5b7da JH |
279 | int ignore_valid = options & CE_MATCH_IGNORE_VALID; |
280 | int assume_racy_is_modified = options & CE_MATCH_RACY_IS_DIRTY; | |
5f73076c JH |
281 | |
282 | /* | |
283 | * If it's marked as always valid in the index, it's | |
284 | * valid whatever the checked-out copy says. | |
285 | */ | |
7a51ed66 | 286 | if (!ignore_valid && (ce->ce_flags & CE_VALID)) |
5f73076c JH |
287 | return 0; |
288 | ||
289 | changed = ce_match_stat_basic(ce, st); | |
407c8eb0 | 290 | |
29e4d363 JH |
291 | /* |
292 | * Within 1 second of this sequence: | |
293 | * echo xyzzy >file && git-update-index --add file | |
294 | * running this command: | |
295 | * echo frotz >file | |
296 | * would give a falsely clean cache entry. The mtime and | |
297 | * length match the cache, and other stat fields do not change. | |
298 | * | |
299 | * We could detect this at update-index time (the cache entry | |
300 | * being registered/updated records the same time as "now") | |
301 | * and delay the return from git-update-index, but that would | |
302 | * effectively mean we can make at most one commit per second, | |
303 | * which is not acceptable. Instead, we check cache entries | |
304 | * whose mtime are the same as the index file timestamp more | |
5f73076c | 305 | * carefully than others. |
29e4d363 | 306 | */ |
6d91da6d | 307 | if (!changed && is_racy_timestamp(istate, ce)) { |
42f77406 JH |
308 | if (assume_racy_is_modified) |
309 | changed |= DATA_CHANGED; | |
310 | else | |
311 | changed |= ce_modified_check_fs(ce, st); | |
312 | } | |
b0391890 | 313 | |
29e4d363 | 314 | return changed; |
b0391890 JH |
315 | } |
316 | ||
4aab5b46 | 317 | int ie_modified(struct index_state *istate, |
4bd5b7da | 318 | struct cache_entry *ce, struct stat *st, unsigned int options) |
b0391890 | 319 | { |
29e4d363 | 320 | int changed, changed_fs; |
4bd5b7da JH |
321 | |
322 | changed = ie_match_stat(istate, ce, st, options); | |
b0391890 JH |
323 | if (!changed) |
324 | return 0; | |
b0391890 JH |
325 | /* |
326 | * If the mode or type has changed, there's no point in trying | |
327 | * to refresh the entry - it's not going to match | |
328 | */ | |
329 | if (changed & (MODE_CHANGED | TYPE_CHANGED)) | |
330 | return changed; | |
331 | ||
332 | /* Immediately after read-tree or update-index --cacheinfo, | |
333 | * the length field is zero. For other cases the ce_size | |
334 | * should match the SHA1 recorded in the index entry. | |
335 | */ | |
7a51ed66 | 336 | if ((changed & DATA_CHANGED) && ce->ce_size != 0) |
b0391890 JH |
337 | return changed; |
338 | ||
29e4d363 JH |
339 | changed_fs = ce_modified_check_fs(ce, st); |
340 | if (changed_fs) | |
341 | return changed | changed_fs; | |
b0391890 JH |
342 | return 0; |
343 | } | |
344 | ||
958ba6c9 LT |
345 | int base_name_compare(const char *name1, int len1, int mode1, |
346 | const char *name2, int len2, int mode2) | |
347 | { | |
348 | unsigned char c1, c2; | |
349 | int len = len1 < len2 ? len1 : len2; | |
350 | int cmp; | |
351 | ||
352 | cmp = memcmp(name1, name2, len); | |
353 | if (cmp) | |
354 | return cmp; | |
355 | c1 = name1[len]; | |
356 | c2 = name2[len]; | |
1833a925 | 357 | if (!c1 && S_ISDIR(mode1)) |
958ba6c9 | 358 | c1 = '/'; |
1833a925 | 359 | if (!c2 && S_ISDIR(mode2)) |
958ba6c9 LT |
360 | c2 = '/'; |
361 | return (c1 < c2) ? -1 : (c1 > c2) ? 1 : 0; | |
362 | } | |
363 | ||
95fd5bf8 | 364 | int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2) |
eb38c22f | 365 | { |
95fd5bf8 LT |
366 | int len1 = flags1 & CE_NAMEMASK; |
367 | int len2 = flags2 & CE_NAMEMASK; | |
eb38c22f LT |
368 | int len = len1 < len2 ? len1 : len2; |
369 | int cmp; | |
370 | ||
371 | cmp = memcmp(name1, name2, len); | |
372 | if (cmp) | |
373 | return cmp; | |
374 | if (len1 < len2) | |
375 | return -1; | |
376 | if (len1 > len2) | |
377 | return 1; | |
5f73076c | 378 | |
7b80be15 JH |
379 | /* Compare stages */ |
380 | flags1 &= CE_STAGEMASK; | |
381 | flags2 &= CE_STAGEMASK; | |
5f73076c | 382 | |
95fd5bf8 LT |
383 | if (flags1 < flags2) |
384 | return -1; | |
385 | if (flags1 > flags2) | |
386 | return 1; | |
eb38c22f LT |
387 | return 0; |
388 | } | |
389 | ||
4aab5b46 | 390 | int index_name_pos(struct index_state *istate, const char *name, int namelen) |
eb38c22f LT |
391 | { |
392 | int first, last; | |
393 | ||
394 | first = 0; | |
4aab5b46 | 395 | last = istate->cache_nr; |
eb38c22f LT |
396 | while (last > first) { |
397 | int next = (last + first) >> 1; | |
4aab5b46 | 398 | struct cache_entry *ce = istate->cache[next]; |
7a51ed66 | 399 | int cmp = cache_name_compare(name, namelen, ce->name, ce->ce_flags); |
eb38c22f | 400 | if (!cmp) |
76e7f4ec | 401 | return next; |
eb38c22f LT |
402 | if (cmp < 0) { |
403 | last = next; | |
404 | continue; | |
405 | } | |
406 | first = next+1; | |
407 | } | |
76e7f4ec | 408 | return -first-1; |
eb38c22f LT |
409 | } |
410 | ||
7b937ca3 | 411 | /* Remove entry, return true if there are more entries to go.. */ |
4aab5b46 | 412 | int remove_index_entry_at(struct index_state *istate, int pos) |
7b937ca3 | 413 | { |
cf558704 LT |
414 | struct cache_entry *ce = istate->cache[pos]; |
415 | ||
416 | remove_hash_entry(istate, ce); | |
4aab5b46 JH |
417 | istate->cache_changed = 1; |
418 | istate->cache_nr--; | |
419 | if (pos >= istate->cache_nr) | |
7b937ca3 | 420 | return 0; |
4aab5b46 JH |
421 | memmove(istate->cache + pos, |
422 | istate->cache + pos + 1, | |
423 | (istate->cache_nr - pos) * sizeof(struct cache_entry *)); | |
7b937ca3 LT |
424 | return 1; |
425 | } | |
426 | ||
4aab5b46 | 427 | int remove_file_from_index(struct index_state *istate, const char *path) |
197ee8c9 | 428 | { |
4aab5b46 | 429 | int pos = index_name_pos(istate, path, strlen(path)); |
c4e3cca1 JH |
430 | if (pos < 0) |
431 | pos = -pos-1; | |
09d5dc32 | 432 | cache_tree_invalidate_path(istate->cache_tree, path); |
4aab5b46 JH |
433 | while (pos < istate->cache_nr && !strcmp(istate->cache[pos]->name, path)) |
434 | remove_index_entry_at(istate, pos); | |
197ee8c9 LT |
435 | return 0; |
436 | } | |
437 | ||
20314271 JS |
438 | static int compare_name(struct cache_entry *ce, const char *path, int namelen) |
439 | { | |
440 | return namelen != ce_namelen(ce) || memcmp(path, ce->name, namelen); | |
441 | } | |
442 | ||
443 | static int index_name_pos_also_unmerged(struct index_state *istate, | |
444 | const char *path, int namelen) | |
445 | { | |
446 | int pos = index_name_pos(istate, path, namelen); | |
447 | struct cache_entry *ce; | |
448 | ||
449 | if (pos >= 0) | |
450 | return pos; | |
451 | ||
452 | /* maybe unmerged? */ | |
453 | pos = -1 - pos; | |
454 | if (pos >= istate->cache_nr || | |
455 | compare_name((ce = istate->cache[pos]), path, namelen)) | |
456 | return -1; | |
457 | ||
458 | /* order of preference: stage 2, 1, 3 */ | |
459 | if (ce_stage(ce) == 1 && pos + 1 < istate->cache_nr && | |
460 | ce_stage((ce = istate->cache[pos + 1])) == 2 && | |
461 | !compare_name(ce, path, namelen)) | |
462 | pos++; | |
463 | return pos; | |
464 | } | |
465 | ||
4aab5b46 | 466 | int add_file_to_index(struct index_state *istate, const char *path, int verbose) |
11be42a4 | 467 | { |
0781b8a9 | 468 | int size, namelen, pos; |
11be42a4 JS |
469 | struct stat st; |
470 | struct cache_entry *ce; | |
fb63d7f8 | 471 | unsigned ce_option = CE_MATCH_IGNORE_VALID|CE_MATCH_RACY_IS_DIRTY; |
11be42a4 JS |
472 | |
473 | if (lstat(path, &st)) | |
474 | die("%s: unable to stat (%s)", path, strerror(errno)); | |
475 | ||
f35a6d3b LT |
476 | if (!S_ISREG(st.st_mode) && !S_ISLNK(st.st_mode) && !S_ISDIR(st.st_mode)) |
477 | die("%s: can only add regular files, symbolic links or git-directories", path); | |
11be42a4 JS |
478 | |
479 | namelen = strlen(path); | |
09595258 LT |
480 | if (S_ISDIR(st.st_mode)) { |
481 | while (namelen && path[namelen-1] == '/') | |
482 | namelen--; | |
483 | } | |
11be42a4 JS |
484 | size = cache_entry_size(namelen); |
485 | ce = xcalloc(1, size); | |
486 | memcpy(ce->name, path, namelen); | |
7a51ed66 | 487 | ce->ce_flags = namelen; |
11be42a4 JS |
488 | fill_stat_cache_info(ce, &st); |
489 | ||
78a8d641 | 490 | if (trust_executable_bit && has_symlinks) |
185c975f JH |
491 | ce->ce_mode = create_ce_mode(st.st_mode); |
492 | else { | |
78a8d641 JS |
493 | /* If there is an existing entry, pick the mode bits and type |
494 | * from it, otherwise assume unexecutable regular file. | |
11be42a4 | 495 | */ |
185c975f | 496 | struct cache_entry *ent; |
20314271 | 497 | int pos = index_name_pos_also_unmerged(istate, path, namelen); |
185c975f | 498 | |
4aab5b46 | 499 | ent = (0 <= pos) ? istate->cache[pos] : NULL; |
185c975f | 500 | ce->ce_mode = ce_mode_from_stat(ent, st.st_mode); |
11be42a4 JS |
501 | } |
502 | ||
0781b8a9 JH |
503 | pos = index_name_pos(istate, ce->name, namelen); |
504 | if (0 <= pos && | |
505 | !ce_stage(istate->cache[pos]) && | |
fb63d7f8 | 506 | !ie_match_stat(istate, istate->cache[pos], &st, ce_option)) { |
0781b8a9 JH |
507 | /* Nothing changed, really */ |
508 | free(ce); | |
eadb5831 | 509 | ce_mark_uptodate(istate->cache[pos]); |
0781b8a9 JH |
510 | return 0; |
511 | } | |
512 | ||
11be42a4 JS |
513 | if (index_path(ce->sha1, path, &st, 1)) |
514 | die("unable to index file %s", path); | |
4aab5b46 | 515 | if (add_index_entry(istate, ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE)) |
11be42a4 JS |
516 | die("unable to add %s to index",path); |
517 | if (verbose) | |
518 | printf("add '%s'\n", path); | |
11be42a4 JS |
519 | return 0; |
520 | } | |
521 | ||
6640f881 CR |
522 | struct cache_entry *make_cache_entry(unsigned int mode, |
523 | const unsigned char *sha1, const char *path, int stage, | |
524 | int refresh) | |
525 | { | |
526 | int size, len; | |
527 | struct cache_entry *ce; | |
528 | ||
529 | if (!verify_path(path)) | |
530 | return NULL; | |
531 | ||
532 | len = strlen(path); | |
533 | size = cache_entry_size(len); | |
534 | ce = xcalloc(1, size); | |
535 | ||
536 | hashcpy(ce->sha1, sha1); | |
537 | memcpy(ce->name, path, len); | |
538 | ce->ce_flags = create_ce_flags(len, stage); | |
539 | ce->ce_mode = create_ce_mode(mode); | |
540 | ||
541 | if (refresh) | |
542 | return refresh_cache_entry(ce, 0); | |
543 | ||
544 | return ce; | |
545 | } | |
546 | ||
dbbce55b | 547 | int ce_same_name(struct cache_entry *a, struct cache_entry *b) |
7b937ca3 LT |
548 | { |
549 | int len = ce_namelen(a); | |
550 | return ce_namelen(b) == len && !memcmp(a->name, b->name, len); | |
551 | } | |
552 | ||
c0fd1f51 LT |
553 | int ce_path_match(const struct cache_entry *ce, const char **pathspec) |
554 | { | |
555 | const char *match, *name; | |
556 | int len; | |
557 | ||
558 | if (!pathspec) | |
559 | return 1; | |
560 | ||
561 | len = ce_namelen(ce); | |
562 | name = ce->name; | |
563 | while ((match = *pathspec++) != NULL) { | |
564 | int matchlen = strlen(match); | |
565 | if (matchlen > len) | |
566 | continue; | |
567 | if (memcmp(name, match, matchlen)) | |
568 | continue; | |
569 | if (matchlen && name[matchlen-1] == '/') | |
570 | return 1; | |
571 | if (name[matchlen] == '/' || !name[matchlen]) | |
572 | return 1; | |
f332726e LT |
573 | if (!matchlen) |
574 | return 1; | |
c0fd1f51 LT |
575 | } |
576 | return 0; | |
577 | } | |
578 | ||
8dcf39c4 LT |
579 | /* |
580 | * We fundamentally don't like some paths: we don't want | |
581 | * dot or dot-dot anywhere, and for obvious reasons don't | |
582 | * want to recurse into ".git" either. | |
583 | * | |
584 | * Also, we don't want double slashes or slashes at the | |
585 | * end that can make pathnames ambiguous. | |
586 | */ | |
587 | static int verify_dotfile(const char *rest) | |
588 | { | |
589 | /* | |
590 | * The first character was '.', but that | |
591 | * has already been discarded, we now test | |
592 | * the rest. | |
593 | */ | |
594 | switch (*rest) { | |
595 | /* "." is not allowed */ | |
596 | case '\0': case '/': | |
597 | return 0; | |
598 | ||
599 | /* | |
600 | * ".git" followed by NUL or slash is bad. This | |
601 | * shares the path end test with the ".." case. | |
602 | */ | |
603 | case 'g': | |
604 | if (rest[1] != 'i') | |
605 | break; | |
606 | if (rest[2] != 't') | |
607 | break; | |
608 | rest += 2; | |
609 | /* fallthrough */ | |
610 | case '.': | |
611 | if (rest[1] == '\0' || rest[1] == '/') | |
612 | return 0; | |
613 | } | |
614 | return 1; | |
615 | } | |
616 | ||
617 | int verify_path(const char *path) | |
618 | { | |
619 | char c; | |
620 | ||
621 | goto inside; | |
622 | for (;;) { | |
623 | if (!c) | |
624 | return 1; | |
625 | if (c == '/') { | |
626 | inside: | |
627 | c = *path++; | |
628 | switch (c) { | |
629 | default: | |
630 | continue; | |
631 | case '/': case '\0': | |
632 | break; | |
633 | case '.': | |
634 | if (verify_dotfile(path)) | |
635 | continue; | |
636 | } | |
637 | return 0; | |
638 | } | |
639 | c = *path++; | |
640 | } | |
641 | } | |
642 | ||
12676608 LT |
643 | /* |
644 | * Do we have another file that has the beginning components being a | |
645 | * proper superset of the name we're trying to add? | |
0f1e4f04 | 646 | */ |
4aab5b46 JH |
647 | static int has_file_name(struct index_state *istate, |
648 | const struct cache_entry *ce, int pos, int ok_to_replace) | |
0f1e4f04 | 649 | { |
12676608 LT |
650 | int retval = 0; |
651 | int len = ce_namelen(ce); | |
b155725d | 652 | int stage = ce_stage(ce); |
12676608 | 653 | const char *name = ce->name; |
0f1e4f04 | 654 | |
4aab5b46 JH |
655 | while (pos < istate->cache_nr) { |
656 | struct cache_entry *p = istate->cache[pos++]; | |
0f1e4f04 | 657 | |
12676608 | 658 | if (len >= ce_namelen(p)) |
0f1e4f04 | 659 | break; |
12676608 LT |
660 | if (memcmp(name, p->name, len)) |
661 | break; | |
b155725d JH |
662 | if (ce_stage(p) != stage) |
663 | continue; | |
12676608 LT |
664 | if (p->name[len] != '/') |
665 | continue; | |
7a51ed66 | 666 | if (p->ce_flags & CE_REMOVE) |
21cd8d00 | 667 | continue; |
12676608 LT |
668 | retval = -1; |
669 | if (!ok_to_replace) | |
670 | break; | |
4aab5b46 | 671 | remove_index_entry_at(istate, --pos); |
0f1e4f04 | 672 | } |
12676608 LT |
673 | return retval; |
674 | } | |
0f1e4f04 | 675 | |
12676608 LT |
676 | /* |
677 | * Do we have another file with a pathname that is a proper | |
678 | * subset of the name we're trying to add? | |
679 | */ | |
4aab5b46 JH |
680 | static int has_dir_name(struct index_state *istate, |
681 | const struct cache_entry *ce, int pos, int ok_to_replace) | |
12676608 LT |
682 | { |
683 | int retval = 0; | |
b155725d | 684 | int stage = ce_stage(ce); |
12676608 LT |
685 | const char *name = ce->name; |
686 | const char *slash = name + ce_namelen(ce); | |
0f1e4f04 | 687 | |
12676608 LT |
688 | for (;;) { |
689 | int len; | |
0f1e4f04 | 690 | |
12676608 LT |
691 | for (;;) { |
692 | if (*--slash == '/') | |
693 | break; | |
694 | if (slash <= ce->name) | |
695 | return retval; | |
696 | } | |
697 | len = slash - name; | |
0f1e4f04 | 698 | |
7a51ed66 | 699 | pos = index_name_pos(istate, name, create_ce_flags(len, stage)); |
12676608 | 700 | if (pos >= 0) { |
21cd8d00 JH |
701 | /* |
702 | * Found one, but not so fast. This could | |
703 | * be a marker that says "I was here, but | |
704 | * I am being removed". Such an entry is | |
705 | * not a part of the resulting tree, and | |
706 | * it is Ok to have a directory at the same | |
707 | * path. | |
708 | */ | |
077c48df | 709 | if (!(istate->cache[pos]->ce_flags & CE_REMOVE)) { |
21cd8d00 JH |
710 | retval = -1; |
711 | if (!ok_to_replace) | |
712 | break; | |
4aab5b46 | 713 | remove_index_entry_at(istate, pos); |
21cd8d00 JH |
714 | continue; |
715 | } | |
12676608 | 716 | } |
21cd8d00 JH |
717 | else |
718 | pos = -pos-1; | |
12676608 LT |
719 | |
720 | /* | |
721 | * Trivial optimization: if we find an entry that | |
722 | * already matches the sub-directory, then we know | |
b155725d | 723 | * we're ok, and we can exit. |
12676608 | 724 | */ |
4aab5b46 JH |
725 | while (pos < istate->cache_nr) { |
726 | struct cache_entry *p = istate->cache[pos]; | |
b155725d JH |
727 | if ((ce_namelen(p) <= len) || |
728 | (p->name[len] != '/') || | |
729 | memcmp(p->name, name, len)) | |
730 | break; /* not our subdirectory */ | |
077c48df JH |
731 | if (ce_stage(p) == stage && !(p->ce_flags & CE_REMOVE)) |
732 | /* | |
733 | * p is at the same stage as our entry, and | |
b155725d JH |
734 | * is a subdirectory of what we are looking |
735 | * at, so we cannot have conflicts at our | |
736 | * level or anything shorter. | |
737 | */ | |
738 | return retval; | |
739 | pos++; | |
192268c1 | 740 | } |
0f1e4f04 | 741 | } |
12676608 LT |
742 | return retval; |
743 | } | |
744 | ||
745 | /* We may be in a situation where we already have path/file and path | |
746 | * is being added, or we already have path and path/file is being | |
747 | * added. Either one would result in a nonsense tree that has path | |
748 | * twice when git-write-tree tries to write it out. Prevent it. | |
a6080a0a | 749 | * |
12676608 LT |
750 | * If ok-to-replace is specified, we remove the conflicting entries |
751 | * from the cache so the caller should recompute the insert position. | |
752 | * When this happens, we return non-zero. | |
753 | */ | |
4aab5b46 JH |
754 | static int check_file_directory_conflict(struct index_state *istate, |
755 | const struct cache_entry *ce, | |
756 | int pos, int ok_to_replace) | |
12676608 | 757 | { |
21cd8d00 JH |
758 | int retval; |
759 | ||
760 | /* | |
761 | * When ce is an "I am going away" entry, we allow it to be added | |
762 | */ | |
7a51ed66 | 763 | if (ce->ce_flags & CE_REMOVE) |
21cd8d00 JH |
764 | return 0; |
765 | ||
12676608 LT |
766 | /* |
767 | * We check if the path is a sub-path of a subsequent pathname | |
768 | * first, since removing those will not change the position | |
21cd8d00 | 769 | * in the array. |
12676608 | 770 | */ |
4aab5b46 | 771 | retval = has_file_name(istate, ce, pos, ok_to_replace); |
21cd8d00 | 772 | |
12676608 LT |
773 | /* |
774 | * Then check if the path might have a clashing sub-directory | |
775 | * before it. | |
776 | */ | |
4aab5b46 | 777 | return retval + has_dir_name(istate, ce, pos, ok_to_replace); |
0f1e4f04 JH |
778 | } |
779 | ||
af3785dc | 780 | static int add_index_entry_with_check(struct index_state *istate, struct cache_entry *ce, int option) |
197ee8c9 LT |
781 | { |
782 | int pos; | |
192268c1 JH |
783 | int ok_to_add = option & ADD_CACHE_OK_TO_ADD; |
784 | int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE; | |
b155725d | 785 | int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK; |
5f73076c | 786 | |
09d5dc32 | 787 | cache_tree_invalidate_path(istate->cache_tree, ce->name); |
7a51ed66 | 788 | pos = index_name_pos(istate, ce->name, ce->ce_flags); |
197ee8c9 | 789 | |
3e09cdfd | 790 | /* existing match? Just replace it. */ |
76e7f4ec | 791 | if (pos >= 0) { |
cf558704 | 792 | replace_index_entry(istate, pos, ce); |
197ee8c9 LT |
793 | return 0; |
794 | } | |
76e7f4ec | 795 | pos = -pos-1; |
197ee8c9 | 796 | |
7b937ca3 LT |
797 | /* |
798 | * Inserting a merged entry ("stage 0") into the index | |
799 | * will always replace all non-merged entries.. | |
800 | */ | |
4aab5b46 JH |
801 | if (pos < istate->cache_nr && ce_stage(ce) == 0) { |
802 | while (ce_same_name(istate->cache[pos], ce)) { | |
7b937ca3 | 803 | ok_to_add = 1; |
4aab5b46 | 804 | if (!remove_index_entry_at(istate, pos)) |
7b937ca3 LT |
805 | break; |
806 | } | |
807 | } | |
808 | ||
121481ab LT |
809 | if (!ok_to_add) |
810 | return -1; | |
8dcf39c4 LT |
811 | if (!verify_path(ce->name)) |
812 | return -1; | |
121481ab | 813 | |
3e09cdfd | 814 | if (!skip_df_check && |
4aab5b46 | 815 | check_file_directory_conflict(istate, ce, pos, ok_to_replace)) { |
192268c1 | 816 | if (!ok_to_replace) |
4aab5b46 JH |
817 | return error("'%s' appears as both a file and as a directory", |
818 | ce->name); | |
7a51ed66 | 819 | pos = index_name_pos(istate, ce->name, ce->ce_flags); |
192268c1 JH |
820 | pos = -pos-1; |
821 | } | |
af3785dc JH |
822 | return pos + 1; |
823 | } | |
824 | ||
825 | int add_index_entry(struct index_state *istate, struct cache_entry *ce, int option) | |
826 | { | |
827 | int pos; | |
828 | ||
829 | if (option & ADD_CACHE_JUST_APPEND) | |
830 | pos = istate->cache_nr; | |
831 | else { | |
832 | int ret; | |
833 | ret = add_index_entry_with_check(istate, ce, option); | |
834 | if (ret <= 0) | |
835 | return ret; | |
836 | pos = ret - 1; | |
837 | } | |
0f1e4f04 | 838 | |
197ee8c9 | 839 | /* Make sure the array is big enough .. */ |
4aab5b46 JH |
840 | if (istate->cache_nr == istate->cache_alloc) { |
841 | istate->cache_alloc = alloc_nr(istate->cache_alloc); | |
842 | istate->cache = xrealloc(istate->cache, | |
843 | istate->cache_alloc * sizeof(struct cache_entry *)); | |
197ee8c9 LT |
844 | } |
845 | ||
846 | /* Add it in.. */ | |
4aab5b46 | 847 | istate->cache_nr++; |
af3785dc | 848 | if (istate->cache_nr > pos + 1) |
4aab5b46 JH |
849 | memmove(istate->cache + pos + 1, |
850 | istate->cache + pos, | |
851 | (istate->cache_nr - pos - 1) * sizeof(ce)); | |
cf558704 | 852 | set_index_entry(istate, pos, ce); |
4aab5b46 | 853 | istate->cache_changed = 1; |
197ee8c9 LT |
854 | return 0; |
855 | } | |
856 | ||
405e5b2f LT |
857 | /* |
858 | * "refresh" does not calculate a new sha1 file or bring the | |
859 | * cache up-to-date for mode/content changes. But what it | |
860 | * _does_ do is to "re-match" the stat information of a file | |
861 | * with the cache, so that you can refresh the cache for a | |
862 | * file that hasn't been changed but where the stat entry is | |
863 | * out of date. | |
864 | * | |
865 | * For example, you'd want to do this after doing a "git-read-tree", | |
866 | * to link up the stat cache details with the proper files. | |
867 | */ | |
4aab5b46 | 868 | static struct cache_entry *refresh_cache_ent(struct index_state *istate, |
4bd5b7da JH |
869 | struct cache_entry *ce, |
870 | unsigned int options, int *err) | |
405e5b2f LT |
871 | { |
872 | struct stat st; | |
873 | struct cache_entry *updated; | |
874 | int changed, size; | |
4bd5b7da | 875 | int ignore_valid = options & CE_MATCH_IGNORE_VALID; |
405e5b2f | 876 | |
eadb5831 JH |
877 | if (ce_uptodate(ce)) |
878 | return ce; | |
879 | ||
8fd2cb40 | 880 | if (lstat(ce->name, &st) < 0) { |
ec0cc704 JH |
881 | if (err) |
882 | *err = errno; | |
8fd2cb40 JS |
883 | return NULL; |
884 | } | |
405e5b2f | 885 | |
4bd5b7da | 886 | changed = ie_match_stat(istate, ce, &st, options); |
405e5b2f | 887 | if (!changed) { |
4bd5b7da JH |
888 | /* |
889 | * The path is unchanged. If we were told to ignore | |
890 | * valid bit, then we did the actual stat check and | |
891 | * found that the entry is unmodified. If the entry | |
892 | * is not marked VALID, this is the place to mark it | |
893 | * valid again, under "assume unchanged" mode. | |
894 | */ | |
895 | if (ignore_valid && assume_unchanged && | |
7a51ed66 | 896 | !(ce->ce_flags & CE_VALID)) |
405e5b2f | 897 | ; /* mark this one VALID again */ |
eadb5831 JH |
898 | else { |
899 | /* | |
900 | * We do not mark the index itself "modified" | |
901 | * because CE_UPTODATE flag is in-core only; | |
902 | * we are not going to write this change out. | |
903 | */ | |
904 | ce_mark_uptodate(ce); | |
8fd2cb40 | 905 | return ce; |
eadb5831 | 906 | } |
405e5b2f LT |
907 | } |
908 | ||
4bd5b7da | 909 | if (ie_modified(istate, ce, &st, options)) { |
ec0cc704 JH |
910 | if (err) |
911 | *err = EINVAL; | |
8fd2cb40 JS |
912 | return NULL; |
913 | } | |
405e5b2f LT |
914 | |
915 | size = ce_size(ce); | |
916 | updated = xmalloc(size); | |
917 | memcpy(updated, ce, size); | |
918 | fill_stat_cache_info(updated, &st); | |
4bd5b7da JH |
919 | /* |
920 | * If ignore_valid is not set, we should leave CE_VALID bit | |
921 | * alone. Otherwise, paths marked with --no-assume-unchanged | |
922 | * (i.e. things to be edited) will reacquire CE_VALID bit | |
923 | * automatically, which is not really what we want. | |
405e5b2f | 924 | */ |
4bd5b7da | 925 | if (!ignore_valid && assume_unchanged && |
7a51ed66 LT |
926 | !(ce->ce_flags & CE_VALID)) |
927 | updated->ce_flags &= ~CE_VALID; | |
405e5b2f LT |
928 | |
929 | return updated; | |
930 | } | |
931 | ||
d616813d | 932 | int refresh_index(struct index_state *istate, unsigned int flags, const char **pathspec, char *seen) |
405e5b2f LT |
933 | { |
934 | int i; | |
935 | int has_errors = 0; | |
936 | int really = (flags & REFRESH_REALLY) != 0; | |
937 | int allow_unmerged = (flags & REFRESH_UNMERGED) != 0; | |
938 | int quiet = (flags & REFRESH_QUIET) != 0; | |
939 | int not_new = (flags & REFRESH_IGNORE_MISSING) != 0; | |
4bd5b7da | 940 | unsigned int options = really ? CE_MATCH_IGNORE_VALID : 0; |
405e5b2f | 941 | |
4aab5b46 | 942 | for (i = 0; i < istate->cache_nr; i++) { |
405e5b2f | 943 | struct cache_entry *ce, *new; |
ec0cc704 JH |
944 | int cache_errno = 0; |
945 | ||
4aab5b46 | 946 | ce = istate->cache[i]; |
405e5b2f | 947 | if (ce_stage(ce)) { |
4aab5b46 JH |
948 | while ((i < istate->cache_nr) && |
949 | ! strcmp(istate->cache[i]->name, ce->name)) | |
405e5b2f LT |
950 | i++; |
951 | i--; | |
952 | if (allow_unmerged) | |
953 | continue; | |
954 | printf("%s: needs merge\n", ce->name); | |
955 | has_errors = 1; | |
956 | continue; | |
957 | } | |
958 | ||
d616813d AJ |
959 | if (pathspec && !match_pathspec(pathspec, ce->name, strlen(ce->name), 0, seen)) |
960 | continue; | |
961 | ||
4bd5b7da | 962 | new = refresh_cache_ent(istate, ce, options, &cache_errno); |
8fd2cb40 | 963 | if (new == ce) |
405e5b2f | 964 | continue; |
8fd2cb40 JS |
965 | if (!new) { |
966 | if (not_new && cache_errno == ENOENT) | |
405e5b2f | 967 | continue; |
8fd2cb40 | 968 | if (really && cache_errno == EINVAL) { |
405e5b2f LT |
969 | /* If we are doing --really-refresh that |
970 | * means the index is not valid anymore. | |
971 | */ | |
7a51ed66 | 972 | ce->ce_flags &= ~CE_VALID; |
4aab5b46 | 973 | istate->cache_changed = 1; |
405e5b2f LT |
974 | } |
975 | if (quiet) | |
976 | continue; | |
977 | printf("%s: needs update\n", ce->name); | |
978 | has_errors = 1; | |
979 | continue; | |
980 | } | |
cf558704 LT |
981 | |
982 | replace_index_entry(istate, i, new); | |
405e5b2f LT |
983 | } |
984 | return has_errors; | |
985 | } | |
986 | ||
ec0cc704 JH |
987 | struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really) |
988 | { | |
4aab5b46 | 989 | return refresh_cache_ent(&the_index, ce, really, NULL); |
ec0cc704 JH |
990 | } |
991 | ||
bad68ec9 | 992 | static int verify_hdr(struct cache_header *hdr, unsigned long size) |
e83c5163 LT |
993 | { |
994 | SHA_CTX c; | |
bad68ec9 | 995 | unsigned char sha1[20]; |
e83c5163 | 996 | |
ccc4feb5 | 997 | if (hdr->hdr_signature != htonl(CACHE_SIGNATURE)) |
e83c5163 | 998 | return error("bad signature"); |
ca9be054 LT |
999 | if (hdr->hdr_version != htonl(2)) |
1000 | return error("bad index version"); | |
e83c5163 | 1001 | SHA1_Init(&c); |
ca9be054 | 1002 | SHA1_Update(&c, hdr, size - 20); |
e83c5163 | 1003 | SHA1_Final(sha1, &c); |
a89fccd2 | 1004 | if (hashcmp(sha1, (unsigned char *)hdr + size - 20)) |
ca9be054 | 1005 | return error("bad index file sha1 signature"); |
e83c5163 LT |
1006 | return 0; |
1007 | } | |
1008 | ||
4aab5b46 JH |
1009 | static int read_index_extension(struct index_state *istate, |
1010 | const char *ext, void *data, unsigned long sz) | |
bad68ec9 JH |
1011 | { |
1012 | switch (CACHE_EXT(ext)) { | |
1013 | case CACHE_EXT_TREE: | |
4aab5b46 | 1014 | istate->cache_tree = cache_tree_read(data, sz); |
bad68ec9 JH |
1015 | break; |
1016 | default: | |
1017 | if (*ext < 'A' || 'Z' < *ext) | |
1018 | return error("index uses %.4s extension, which we do not understand", | |
1019 | ext); | |
1020 | fprintf(stderr, "ignoring %.4s extension\n", ext); | |
1021 | break; | |
1022 | } | |
1023 | return 0; | |
1024 | } | |
1025 | ||
4aab5b46 | 1026 | int read_index(struct index_state *istate) |
8fd2cb40 | 1027 | { |
4aab5b46 | 1028 | return read_index_from(istate, get_index_file()); |
8fd2cb40 JS |
1029 | } |
1030 | ||
7a51ed66 LT |
1031 | static void convert_from_disk(struct ondisk_cache_entry *ondisk, struct cache_entry *ce) |
1032 | { | |
7fec10b7 JH |
1033 | size_t len; |
1034 | ||
7a51ed66 LT |
1035 | ce->ce_ctime = ntohl(ondisk->ctime.sec); |
1036 | ce->ce_mtime = ntohl(ondisk->mtime.sec); | |
1037 | ce->ce_dev = ntohl(ondisk->dev); | |
1038 | ce->ce_ino = ntohl(ondisk->ino); | |
1039 | ce->ce_mode = ntohl(ondisk->mode); | |
1040 | ce->ce_uid = ntohl(ondisk->uid); | |
1041 | ce->ce_gid = ntohl(ondisk->gid); | |
1042 | ce->ce_size = ntohl(ondisk->size); | |
1043 | /* On-disk flags are just 16 bits */ | |
1044 | ce->ce_flags = ntohs(ondisk->flags); | |
1045 | hashcpy(ce->sha1, ondisk->sha1); | |
7fec10b7 JH |
1046 | |
1047 | len = ce->ce_flags & CE_NAMEMASK; | |
1048 | if (len == CE_NAMEMASK) | |
1049 | len = strlen(ondisk->name); | |
1050 | /* | |
1051 | * NEEDSWORK: If the original index is crafted, this copy could | |
1052 | * go unchecked. | |
1053 | */ | |
1054 | memcpy(ce->name, ondisk->name, len + 1); | |
7a51ed66 LT |
1055 | } |
1056 | ||
cf558704 LT |
1057 | static inline size_t estimate_cache_size(size_t ondisk_size, unsigned int entries) |
1058 | { | |
1059 | long per_entry; | |
1060 | ||
1061 | per_entry = sizeof(struct cache_entry) - sizeof(struct ondisk_cache_entry); | |
1062 | ||
1063 | /* | |
1064 | * Alignment can cause differences. This should be "alignof", but | |
1065 | * since that's a gcc'ism, just use the size of a pointer. | |
1066 | */ | |
1067 | per_entry += sizeof(void *); | |
1068 | return ondisk_size + entries*per_entry; | |
1069 | } | |
1070 | ||
8fd2cb40 | 1071 | /* remember to discard_cache() before reading a different cache! */ |
4aab5b46 | 1072 | int read_index_from(struct index_state *istate, const char *path) |
e83c5163 LT |
1073 | { |
1074 | int fd, i; | |
1075 | struct stat st; | |
7a51ed66 | 1076 | unsigned long src_offset, dst_offset; |
e83c5163 | 1077 | struct cache_header *hdr; |
7a51ed66 LT |
1078 | void *mmap; |
1079 | size_t mmap_size; | |
e83c5163 LT |
1080 | |
1081 | errno = EBUSY; | |
7a51ed66 | 1082 | if (istate->alloc) |
4aab5b46 | 1083 | return istate->cache_nr; |
5d1a5c02 | 1084 | |
e83c5163 | 1085 | errno = ENOENT; |
4aab5b46 | 1086 | istate->timestamp = 0; |
8fd2cb40 | 1087 | fd = open(path, O_RDONLY); |
5d1a5c02 LT |
1088 | if (fd < 0) { |
1089 | if (errno == ENOENT) | |
1090 | return 0; | |
1091 | die("index file open failed (%s)", strerror(errno)); | |
1092 | } | |
e83c5163 | 1093 | |
3511a377 | 1094 | if (fstat(fd, &st)) |
5fe5c830 | 1095 | die("cannot stat the open index (%s)", strerror(errno)); |
3511a377 LFC |
1096 | |
1097 | errno = EINVAL; | |
7a51ed66 LT |
1098 | mmap_size = xsize_t(st.st_size); |
1099 | if (mmap_size < sizeof(struct cache_header) + 20) | |
3511a377 LFC |
1100 | die("index file smaller than expected"); |
1101 | ||
7a51ed66 | 1102 | mmap = xmmap(NULL, mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0); |
e83c5163 | 1103 | close(fd); |
7a51ed66 LT |
1104 | if (mmap == MAP_FAILED) |
1105 | die("unable to map index file"); | |
e83c5163 | 1106 | |
7a51ed66 LT |
1107 | hdr = mmap; |
1108 | if (verify_hdr(hdr, mmap_size) < 0) | |
e83c5163 LT |
1109 | goto unmap; |
1110 | ||
4aab5b46 JH |
1111 | istate->cache_nr = ntohl(hdr->hdr_entries); |
1112 | istate->cache_alloc = alloc_nr(istate->cache_nr); | |
1113 | istate->cache = xcalloc(istate->cache_alloc, sizeof(struct cache_entry *)); | |
e83c5163 | 1114 | |
7a51ed66 LT |
1115 | /* |
1116 | * The disk format is actually larger than the in-memory format, | |
1117 | * due to space for nsec etc, so even though the in-memory one | |
1118 | * has room for a few more flags, we can allocate using the same | |
1119 | * index size | |
1120 | */ | |
cf558704 | 1121 | istate->alloc = xmalloc(estimate_cache_size(mmap_size, istate->cache_nr)); |
7a51ed66 LT |
1122 | |
1123 | src_offset = sizeof(*hdr); | |
1124 | dst_offset = 0; | |
4aab5b46 | 1125 | for (i = 0; i < istate->cache_nr; i++) { |
7a51ed66 | 1126 | struct ondisk_cache_entry *disk_ce; |
4aab5b46 JH |
1127 | struct cache_entry *ce; |
1128 | ||
7a51ed66 LT |
1129 | disk_ce = (struct ondisk_cache_entry *)((char *)mmap + src_offset); |
1130 | ce = (struct cache_entry *)((char *)istate->alloc + dst_offset); | |
1131 | convert_from_disk(disk_ce, ce); | |
cf558704 | 1132 | set_index_entry(istate, i, ce); |
7a51ed66 LT |
1133 | |
1134 | src_offset += ondisk_ce_size(ce); | |
1135 | dst_offset += ce_size(ce); | |
e83c5163 | 1136 | } |
4aab5b46 | 1137 | istate->timestamp = st.st_mtime; |
7a51ed66 | 1138 | while (src_offset <= mmap_size - 20 - 8) { |
bad68ec9 JH |
1139 | /* After an array of active_nr index entries, |
1140 | * there can be arbitrary number of extended | |
1141 | * sections, each of which is prefixed with | |
1142 | * extension name (4-byte) and section length | |
1143 | * in 4-byte network byte order. | |
1144 | */ | |
1145 | unsigned long extsize; | |
7a51ed66 | 1146 | memcpy(&extsize, (char *)mmap + src_offset + 4, 4); |
bad68ec9 | 1147 | extsize = ntohl(extsize); |
4aab5b46 | 1148 | if (read_index_extension(istate, |
7a51ed66 LT |
1149 | (const char *) mmap + src_offset, |
1150 | (char *) mmap + src_offset + 8, | |
1d7f171c | 1151 | extsize) < 0) |
bad68ec9 | 1152 | goto unmap; |
7a51ed66 LT |
1153 | src_offset += 8; |
1154 | src_offset += extsize; | |
bad68ec9 | 1155 | } |
7a51ed66 | 1156 | munmap(mmap, mmap_size); |
4aab5b46 | 1157 | return istate->cache_nr; |
e83c5163 LT |
1158 | |
1159 | unmap: | |
7a51ed66 | 1160 | munmap(mmap, mmap_size); |
e83c5163 | 1161 | errno = EINVAL; |
5d1a5c02 | 1162 | die("index file corrupt"); |
e83c5163 LT |
1163 | } |
1164 | ||
4aab5b46 | 1165 | int discard_index(struct index_state *istate) |
6d297f81 | 1166 | { |
4aab5b46 JH |
1167 | istate->cache_nr = 0; |
1168 | istate->cache_changed = 0; | |
1169 | istate->timestamp = 0; | |
cf558704 | 1170 | free_hash(&istate->name_hash); |
4aab5b46 | 1171 | cache_tree_free(&(istate->cache_tree)); |
7a51ed66 LT |
1172 | free(istate->alloc); |
1173 | istate->alloc = NULL; | |
6d297f81 JS |
1174 | |
1175 | /* no need to throw away allocated active_cache */ | |
7a51ed66 | 1176 | return 0; |
6d297f81 JS |
1177 | } |
1178 | ||
4990aadc | 1179 | #define WRITE_BUFFER_SIZE 8192 |
bf0f910d | 1180 | static unsigned char write_buffer[WRITE_BUFFER_SIZE]; |
4990aadc LT |
1181 | static unsigned long write_buffer_len; |
1182 | ||
6015c28b JH |
1183 | static int ce_write_flush(SHA_CTX *context, int fd) |
1184 | { | |
1185 | unsigned int buffered = write_buffer_len; | |
1186 | if (buffered) { | |
1187 | SHA1_Update(context, write_buffer, buffered); | |
93822c22 | 1188 | if (write_in_full(fd, write_buffer, buffered) != buffered) |
6015c28b JH |
1189 | return -1; |
1190 | write_buffer_len = 0; | |
1191 | } | |
1192 | return 0; | |
1193 | } | |
1194 | ||
ca9be054 | 1195 | static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len) |
4990aadc LT |
1196 | { |
1197 | while (len) { | |
1198 | unsigned int buffered = write_buffer_len; | |
1199 | unsigned int partial = WRITE_BUFFER_SIZE - buffered; | |
1200 | if (partial > len) | |
1201 | partial = len; | |
1202 | memcpy(write_buffer + buffered, data, partial); | |
1203 | buffered += partial; | |
1204 | if (buffered == WRITE_BUFFER_SIZE) { | |
6015c28b JH |
1205 | write_buffer_len = buffered; |
1206 | if (ce_write_flush(context, fd)) | |
4990aadc LT |
1207 | return -1; |
1208 | buffered = 0; | |
1209 | } | |
1210 | write_buffer_len = buffered; | |
1211 | len -= partial; | |
1d7f171c | 1212 | data = (char *) data + partial; |
a6080a0a JH |
1213 | } |
1214 | return 0; | |
4990aadc LT |
1215 | } |
1216 | ||
bad68ec9 | 1217 | static int write_index_ext_header(SHA_CTX *context, int fd, |
ac58c7b1 | 1218 | unsigned int ext, unsigned int sz) |
bad68ec9 JH |
1219 | { |
1220 | ext = htonl(ext); | |
1221 | sz = htonl(sz); | |
968a1d65 DR |
1222 | return ((ce_write(context, fd, &ext, 4) < 0) || |
1223 | (ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0; | |
bad68ec9 JH |
1224 | } |
1225 | ||
1226 | static int ce_flush(SHA_CTX *context, int fd) | |
4990aadc LT |
1227 | { |
1228 | unsigned int left = write_buffer_len; | |
ca9be054 | 1229 | |
4990aadc LT |
1230 | if (left) { |
1231 | write_buffer_len = 0; | |
ca9be054 | 1232 | SHA1_Update(context, write_buffer, left); |
4990aadc | 1233 | } |
ca9be054 | 1234 | |
2c865d9a QH |
1235 | /* Flush first if not enough space for SHA1 signature */ |
1236 | if (left + 20 > WRITE_BUFFER_SIZE) { | |
93822c22 | 1237 | if (write_in_full(fd, write_buffer, left) != left) |
2c865d9a QH |
1238 | return -1; |
1239 | left = 0; | |
1240 | } | |
1241 | ||
ca9be054 | 1242 | /* Append the SHA1 signature at the end */ |
bad68ec9 | 1243 | SHA1_Final(write_buffer + left, context); |
ca9be054 | 1244 | left += 20; |
93822c22 | 1245 | return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0; |
4990aadc LT |
1246 | } |
1247 | ||
407c8eb0 JH |
1248 | static void ce_smudge_racily_clean_entry(struct cache_entry *ce) |
1249 | { | |
1250 | /* | |
1251 | * The only thing we care about in this function is to smudge the | |
1252 | * falsely clean entry due to touch-update-touch race, so we leave | |
1253 | * everything else as they are. We are called for entries whose | |
1254 | * ce_mtime match the index file mtime. | |
1255 | */ | |
1256 | struct stat st; | |
1257 | ||
1258 | if (lstat(ce->name, &st) < 0) | |
1259 | return; | |
1260 | if (ce_match_stat_basic(ce, &st)) | |
1261 | return; | |
1262 | if (ce_modified_check_fs(ce, &st)) { | |
4b3511b0 JH |
1263 | /* This is "racily clean"; smudge it. Note that this |
1264 | * is a tricky code. At first glance, it may appear | |
1265 | * that it can break with this sequence: | |
1266 | * | |
1267 | * $ echo xyzzy >frotz | |
1268 | * $ git-update-index --add frotz | |
1269 | * $ : >frotz | |
1270 | * $ sleep 3 | |
1271 | * $ echo filfre >nitfol | |
1272 | * $ git-update-index --add nitfol | |
1273 | * | |
b7e58b17 | 1274 | * but it does not. When the second update-index runs, |
4b3511b0 JH |
1275 | * it notices that the entry "frotz" has the same timestamp |
1276 | * as index, and if we were to smudge it by resetting its | |
1277 | * size to zero here, then the object name recorded | |
1278 | * in index is the 6-byte file but the cached stat information | |
1279 | * becomes zero --- which would then match what we would | |
a6080a0a | 1280 | * obtain from the filesystem next time we stat("frotz"). |
4b3511b0 JH |
1281 | * |
1282 | * However, the second update-index, before calling | |
1283 | * this function, notices that the cached size is 6 | |
1284 | * bytes and what is on the filesystem is an empty | |
1285 | * file, and never calls us, so the cached size information | |
1286 | * for "frotz" stays 6 which does not match the filesystem. | |
1287 | */ | |
7a51ed66 | 1288 | ce->ce_size = 0; |
407c8eb0 JH |
1289 | } |
1290 | } | |
1291 | ||
7a51ed66 LT |
1292 | static int ce_write_entry(SHA_CTX *c, int fd, struct cache_entry *ce) |
1293 | { | |
1294 | int size = ondisk_ce_size(ce); | |
1295 | struct ondisk_cache_entry *ondisk = xcalloc(1, size); | |
1296 | ||
1297 | ondisk->ctime.sec = htonl(ce->ce_ctime); | |
1298 | ondisk->ctime.nsec = 0; | |
1299 | ondisk->mtime.sec = htonl(ce->ce_mtime); | |
1300 | ondisk->mtime.nsec = 0; | |
1301 | ondisk->dev = htonl(ce->ce_dev); | |
1302 | ondisk->ino = htonl(ce->ce_ino); | |
1303 | ondisk->mode = htonl(ce->ce_mode); | |
1304 | ondisk->uid = htonl(ce->ce_uid); | |
1305 | ondisk->gid = htonl(ce->ce_gid); | |
1306 | ondisk->size = htonl(ce->ce_size); | |
1307 | hashcpy(ondisk->sha1, ce->sha1); | |
1308 | ondisk->flags = htons(ce->ce_flags); | |
1309 | memcpy(ondisk->name, ce->name, ce_namelen(ce)); | |
1310 | ||
1311 | return ce_write(c, fd, ondisk, size); | |
1312 | } | |
1313 | ||
4aab5b46 | 1314 | int write_index(struct index_state *istate, int newfd) |
197ee8c9 LT |
1315 | { |
1316 | SHA_CTX c; | |
1317 | struct cache_header hdr; | |
1dffb8fa | 1318 | int i, err, removed; |
4aab5b46 JH |
1319 | struct cache_entry **cache = istate->cache; |
1320 | int entries = istate->cache_nr; | |
025a0709 JH |
1321 | |
1322 | for (i = removed = 0; i < entries; i++) | |
7a51ed66 | 1323 | if (cache[i]->ce_flags & CE_REMOVE) |
025a0709 | 1324 | removed++; |
197ee8c9 | 1325 | |
ccc4feb5 | 1326 | hdr.hdr_signature = htonl(CACHE_SIGNATURE); |
ca9be054 | 1327 | hdr.hdr_version = htonl(2); |
025a0709 | 1328 | hdr.hdr_entries = htonl(entries - removed); |
197ee8c9 LT |
1329 | |
1330 | SHA1_Init(&c); | |
ca9be054 | 1331 | if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0) |
197ee8c9 LT |
1332 | return -1; |
1333 | ||
1334 | for (i = 0; i < entries; i++) { | |
1335 | struct cache_entry *ce = cache[i]; | |
7a51ed66 | 1336 | if (ce->ce_flags & CE_REMOVE) |
aa16021e | 1337 | continue; |
6d91da6d | 1338 | if (is_racy_timestamp(istate, ce)) |
407c8eb0 | 1339 | ce_smudge_racily_clean_entry(ce); |
7a51ed66 | 1340 | if (ce_write_entry(&c, newfd, ce) < 0) |
197ee8c9 LT |
1341 | return -1; |
1342 | } | |
1af1c2b6 | 1343 | |
bad68ec9 | 1344 | /* Write extension data here */ |
4aab5b46 | 1345 | if (istate->cache_tree) { |
1dffb8fa PH |
1346 | struct strbuf sb; |
1347 | ||
1348 | strbuf_init(&sb, 0); | |
1349 | cache_tree_write(&sb, istate->cache_tree); | |
1350 | err = write_index_ext_header(&c, newfd, CACHE_EXT_TREE, sb.len) < 0 | |
1351 | || ce_write(&c, newfd, sb.buf, sb.len) < 0; | |
1352 | strbuf_release(&sb); | |
1353 | if (err) | |
bad68ec9 | 1354 | return -1; |
bad68ec9 JH |
1355 | } |
1356 | return ce_flush(&c, newfd); | |
197ee8c9 | 1357 | } |