]>
Commit | Line | Data |
---|---|---|
355885d5 | 1 | #include "cache.h" |
cbd53a21 | 2 | #include "object-store.h" |
109cd76d | 3 | #include "repository.h" |
355885d5 MK |
4 | #include "object.h" |
5 | #include "blob.h" | |
6 | #include "tree.h" | |
7 | #include "tree-walk.h" | |
8 | #include "commit.h" | |
9 | #include "tag.h" | |
10 | #include "fsck.h" | |
cec097be | 11 | #include "refs.h" |
a2b26ffb | 12 | #include "url.h" |
a18fcc9f | 13 | #include "utf8.h" |
7b35efd7 | 14 | #include "decorate.h" |
159e7b08 | 15 | #include "oidset.h" |
27387444 | 16 | #include "packfile.h" |
ed8b10f6 JK |
17 | #include "submodule-config.h" |
18 | #include "config.h" | |
07259e74 | 19 | #include "credential.h" |
3ac68a93 | 20 | #include "help.h" |
159e7b08 JK |
21 | |
22 | static struct oidset gitmodules_found = OIDSET_INIT; | |
23 | static struct oidset gitmodules_done = OIDSET_INIT; | |
355885d5 | 24 | |
f50c4407 | 25 | #define FSCK_FATAL -1 |
f27d05b1 | 26 | #define FSCK_INFO -2 |
f50c4407 | 27 | |
c99ba492 | 28 | #define FOREACH_MSG_ID(FUNC) \ |
f50c4407 JS |
29 | /* fatal errors */ \ |
30 | FUNC(NUL_IN_HEADER, FATAL) \ | |
31 | FUNC(UNTERMINATED_HEADER, FATAL) \ | |
c99ba492 JS |
32 | /* errors */ \ |
33 | FUNC(BAD_DATE, ERROR) \ | |
34 | FUNC(BAD_DATE_OVERFLOW, ERROR) \ | |
35 | FUNC(BAD_EMAIL, ERROR) \ | |
36 | FUNC(BAD_NAME, ERROR) \ | |
37 | FUNC(BAD_OBJECT_SHA1, ERROR) \ | |
38 | FUNC(BAD_PARENT_SHA1, ERROR) \ | |
39 | FUNC(BAD_TAG_OBJECT, ERROR) \ | |
40 | FUNC(BAD_TIMEZONE, ERROR) \ | |
41 | FUNC(BAD_TREE, ERROR) \ | |
42 | FUNC(BAD_TREE_SHA1, ERROR) \ | |
43 | FUNC(BAD_TYPE, ERROR) \ | |
44 | FUNC(DUPLICATE_ENTRIES, ERROR) \ | |
45 | FUNC(MISSING_AUTHOR, ERROR) \ | |
46 | FUNC(MISSING_COMMITTER, ERROR) \ | |
47 | FUNC(MISSING_EMAIL, ERROR) \ | |
c99ba492 JS |
48 | FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \ |
49 | FUNC(MISSING_OBJECT, ERROR) \ | |
c99ba492 JS |
50 | FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \ |
51 | FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \ | |
52 | FUNC(MISSING_TAG, ERROR) \ | |
53 | FUNC(MISSING_TAG_ENTRY, ERROR) \ | |
c99ba492 | 54 | FUNC(MISSING_TREE, ERROR) \ |
159e7b08 | 55 | FUNC(MISSING_TREE_OBJECT, ERROR) \ |
c99ba492 JS |
56 | FUNC(MISSING_TYPE, ERROR) \ |
57 | FUNC(MISSING_TYPE_ENTRY, ERROR) \ | |
c9ad147f | 58 | FUNC(MULTIPLE_AUTHORS, ERROR) \ |
c99ba492 JS |
59 | FUNC(TREE_NOT_SORTED, ERROR) \ |
60 | FUNC(UNKNOWN_TYPE, ERROR) \ | |
c99ba492 | 61 | FUNC(ZERO_PADDED_DATE, ERROR) \ |
159e7b08 JK |
62 | FUNC(GITMODULES_MISSING, ERROR) \ |
63 | FUNC(GITMODULES_BLOB, ERROR) \ | |
0d68764d | 64 | FUNC(GITMODULES_LARGE, ERROR) \ |
ed8b10f6 | 65 | FUNC(GITMODULES_NAME, ERROR) \ |
b7b1fca1 | 66 | FUNC(GITMODULES_SYMLINK, ERROR) \ |
a124133e | 67 | FUNC(GITMODULES_URL, ERROR) \ |
1a7fd1fb | 68 | FUNC(GITMODULES_PATH, ERROR) \ |
bb92255e | 69 | FUNC(GITMODULES_UPDATE, ERROR) \ |
c99ba492 JS |
70 | /* warnings */ \ |
71 | FUNC(BAD_FILEMODE, WARN) \ | |
c99ba492 JS |
72 | FUNC(EMPTY_NAME, WARN) \ |
73 | FUNC(FULL_PATHNAME, WARN) \ | |
74 | FUNC(HAS_DOT, WARN) \ | |
75 | FUNC(HAS_DOTDOT, WARN) \ | |
76 | FUNC(HAS_DOTGIT, WARN) \ | |
c99ba492 | 77 | FUNC(NULL_SHA1, WARN) \ |
f27d05b1 | 78 | FUNC(ZERO_PADDED_FILEMODE, WARN) \ |
6d2d780f | 79 | FUNC(NUL_IN_COMMIT, WARN) \ |
f27d05b1 | 80 | /* infos (reported as warnings, but ignored by default) */ \ |
64eb14d3 | 81 | FUNC(GITMODULES_PARSE, INFO) \ |
f27d05b1 | 82 | FUNC(BAD_TAG_NAME, INFO) \ |
acf9de4c ÆAB |
83 | FUNC(MISSING_TAGGER_ENTRY, INFO) \ |
84 | /* ignored (elevated when requested) */ \ | |
85 | FUNC(EXTRA_HEADER_ENTRY, IGNORE) | |
c99ba492 JS |
86 | |
87 | #define MSG_ID(id, msg_type) FSCK_MSG_##id, | |
88 | enum fsck_msg_id { | |
89 | FOREACH_MSG_ID(MSG_ID) | |
90 | FSCK_MSG_MAX | |
91 | }; | |
92 | #undef MSG_ID | |
93 | ||
f417eed8 | 94 | #define STR(x) #x |
a4a9cc19 | 95 | #define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type }, |
c99ba492 | 96 | static struct { |
f417eed8 JS |
97 | const char *id_string; |
98 | const char *downcased; | |
a4a9cc19 | 99 | const char *camelcased; |
c99ba492 JS |
100 | int msg_type; |
101 | } msg_id_info[FSCK_MSG_MAX + 1] = { | |
102 | FOREACH_MSG_ID(MSG_ID) | |
a4a9cc19 | 103 | { NULL, NULL, NULL, -1 } |
c99ba492 JS |
104 | }; |
105 | #undef MSG_ID | |
106 | ||
a46baac6 | 107 | static void prepare_msg_ids(void) |
f417eed8 JS |
108 | { |
109 | int i; | |
110 | ||
a46baac6 NTND |
111 | if (msg_id_info[0].downcased) |
112 | return; | |
113 | ||
114 | /* convert id_string to lower case, without underscores. */ | |
115 | for (i = 0; i < FSCK_MSG_MAX; i++) { | |
116 | const char *p = msg_id_info[i].id_string; | |
117 | int len = strlen(p); | |
118 | char *q = xmalloc(len); | |
119 | ||
120 | msg_id_info[i].downcased = q; | |
121 | while (*p) | |
122 | if (*p == '_') | |
123 | p++; | |
124 | else | |
125 | *(q)++ = tolower(*(p)++); | |
126 | *q = '\0'; | |
a4a9cc19 NTND |
127 | |
128 | p = msg_id_info[i].id_string; | |
129 | q = xmalloc(len); | |
130 | msg_id_info[i].camelcased = q; | |
131 | while (*p) { | |
132 | if (*p == '_') { | |
133 | p++; | |
134 | if (*p) | |
135 | *q++ = *p++; | |
136 | } else { | |
137 | *q++ = tolower(*p++); | |
138 | } | |
f417eed8 | 139 | } |
a4a9cc19 | 140 | *q = '\0'; |
f417eed8 | 141 | } |
a46baac6 NTND |
142 | } |
143 | ||
144 | static int parse_msg_id(const char *text) | |
145 | { | |
146 | int i; | |
147 | ||
148 | prepare_msg_ids(); | |
f417eed8 JS |
149 | |
150 | for (i = 0; i < FSCK_MSG_MAX; i++) | |
151 | if (!strcmp(text, msg_id_info[i].downcased)) | |
152 | return i; | |
153 | ||
154 | return -1; | |
155 | } | |
156 | ||
3ac68a93 NTND |
157 | void list_config_fsck_msg_ids(struct string_list *list, const char *prefix) |
158 | { | |
159 | int i; | |
160 | ||
161 | prepare_msg_ids(); | |
162 | ||
3ac68a93 | 163 | for (i = 0; i < FSCK_MSG_MAX; i++) |
a4a9cc19 | 164 | list_config_item(list, prefix, msg_id_info[i].camelcased); |
3ac68a93 NTND |
165 | } |
166 | ||
c99ba492 JS |
167 | static int fsck_msg_type(enum fsck_msg_id msg_id, |
168 | struct fsck_options *options) | |
169 | { | |
170 | int msg_type; | |
171 | ||
0282f4dc JS |
172 | assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX); |
173 | ||
174 | if (options->msg_type) | |
175 | msg_type = options->msg_type[msg_id]; | |
176 | else { | |
177 | msg_type = msg_id_info[msg_id].msg_type; | |
178 | if (options->strict && msg_type == FSCK_WARN) | |
179 | msg_type = FSCK_ERROR; | |
180 | } | |
c99ba492 JS |
181 | |
182 | return msg_type; | |
183 | } | |
184 | ||
0282f4dc JS |
185 | static int parse_msg_type(const char *str) |
186 | { | |
187 | if (!strcmp(str, "error")) | |
188 | return FSCK_ERROR; | |
189 | else if (!strcmp(str, "warn")) | |
190 | return FSCK_WARN; | |
efaba7cc JS |
191 | else if (!strcmp(str, "ignore")) |
192 | return FSCK_IGNORE; | |
0282f4dc JS |
193 | else |
194 | die("Unknown fsck message type: '%s'", str); | |
195 | } | |
196 | ||
5d477a33 JS |
197 | int is_valid_msg_type(const char *msg_id, const char *msg_type) |
198 | { | |
199 | if (parse_msg_id(msg_id) < 0) | |
200 | return 0; | |
201 | parse_msg_type(msg_type); | |
202 | return 1; | |
203 | } | |
204 | ||
0282f4dc JS |
205 | void fsck_set_msg_type(struct fsck_options *options, |
206 | const char *msg_id, const char *msg_type) | |
207 | { | |
208 | int id = parse_msg_id(msg_id), type; | |
209 | ||
210 | if (id < 0) | |
211 | die("Unhandled message id: %s", msg_id); | |
212 | type = parse_msg_type(msg_type); | |
213 | ||
f50c4407 JS |
214 | if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL) |
215 | die("Cannot demote %s to %s", msg_id, msg_type); | |
216 | ||
0282f4dc JS |
217 | if (!options->msg_type) { |
218 | int i; | |
b32fa95f JK |
219 | int *msg_type; |
220 | ALLOC_ARRAY(msg_type, FSCK_MSG_MAX); | |
0282f4dc JS |
221 | for (i = 0; i < FSCK_MSG_MAX; i++) |
222 | msg_type[i] = fsck_msg_type(i, options); | |
223 | options->msg_type = msg_type; | |
224 | } | |
225 | ||
226 | options->msg_type[id] = type; | |
227 | } | |
228 | ||
229 | void fsck_set_msg_types(struct fsck_options *options, const char *values) | |
230 | { | |
231 | char *buf = xstrdup(values), *to_free = buf; | |
232 | int done = 0; | |
233 | ||
234 | while (!done) { | |
235 | int len = strcspn(buf, " ,|"), equal; | |
236 | ||
237 | done = !buf[len]; | |
238 | if (!len) { | |
239 | buf++; | |
240 | continue; | |
241 | } | |
242 | buf[len] = '\0'; | |
243 | ||
244 | for (equal = 0; | |
245 | equal < len && buf[equal] != '=' && buf[equal] != ':'; | |
246 | equal++) | |
247 | buf[equal] = tolower(buf[equal]); | |
248 | buf[equal] = '\0'; | |
249 | ||
cd94c6f9 JS |
250 | if (!strcmp(buf, "skiplist")) { |
251 | if (equal == len) | |
252 | die("skiplist requires a path"); | |
24eb33eb | 253 | oidset_parse_file(&options->skiplist, buf + equal + 1); |
cd94c6f9 JS |
254 | buf += len + 1; |
255 | continue; | |
256 | } | |
257 | ||
0282f4dc JS |
258 | if (equal == len) |
259 | die("Missing '=': '%s'", buf); | |
355885d5 | 260 | |
0282f4dc JS |
261 | fsck_set_msg_type(options, buf, buf + equal + 1); |
262 | buf += len + 1; | |
263 | } | |
264 | free(to_free); | |
265 | } | |
266 | ||
71ab8fa8 JS |
267 | static void append_msg_id(struct strbuf *sb, const char *msg_id) |
268 | { | |
269 | for (;;) { | |
270 | char c = *(msg_id)++; | |
271 | ||
272 | if (!c) | |
273 | break; | |
274 | if (c != '_') | |
275 | strbuf_addch(sb, tolower(c)); | |
276 | else { | |
277 | assert(*msg_id); | |
278 | strbuf_addch(sb, *(msg_id)++); | |
279 | } | |
280 | } | |
281 | ||
282 | strbuf_addstr(sb, ": "); | |
283 | } | |
284 | ||
f5979376 JK |
285 | static int object_on_skiplist(struct fsck_options *opts, |
286 | const struct object_id *oid) | |
fb162877 | 287 | { |
f5979376 | 288 | return opts && oid && oidset_contains(&opts->skiplist, oid); |
fb162877 RJ |
289 | } |
290 | ||
38370253 JK |
291 | __attribute__((format (printf, 5, 6))) |
292 | static int report(struct fsck_options *options, | |
293 | const struct object_id *oid, enum object_type object_type, | |
294 | enum fsck_msg_id id, const char *fmt, ...) | |
c99ba492 JS |
295 | { |
296 | va_list ap; | |
297 | struct strbuf sb = STRBUF_INIT; | |
298 | int msg_type = fsck_msg_type(id, options), result; | |
299 | ||
efaba7cc JS |
300 | if (msg_type == FSCK_IGNORE) |
301 | return 0; | |
302 | ||
38370253 | 303 | if (object_on_skiplist(options, oid)) |
cd94c6f9 JS |
304 | return 0; |
305 | ||
f50c4407 JS |
306 | if (msg_type == FSCK_FATAL) |
307 | msg_type = FSCK_ERROR; | |
f27d05b1 JS |
308 | else if (msg_type == FSCK_INFO) |
309 | msg_type = FSCK_WARN; | |
f50c4407 | 310 | |
71ab8fa8 JS |
311 | append_msg_id(&sb, msg_id_info[id].id_string); |
312 | ||
c99ba492 JS |
313 | va_start(ap, fmt); |
314 | strbuf_vaddf(&sb, fmt, ap); | |
38370253 | 315 | result = options->error_func(options, oid, object_type, |
5afc4b1d | 316 | msg_type, sb.buf); |
c99ba492 JS |
317 | strbuf_release(&sb); |
318 | va_end(ap); | |
319 | ||
320 | return result; | |
321 | } | |
322 | ||
a59cfb32 JK |
323 | void fsck_enable_object_names(struct fsck_options *options) |
324 | { | |
325 | if (!options->object_names) | |
73390290 | 326 | options->object_names = kh_init_oid_map(); |
a59cfb32 JK |
327 | } |
328 | ||
73390290 JK |
329 | const char *fsck_get_object_name(struct fsck_options *options, |
330 | const struct object_id *oid) | |
7b35efd7 | 331 | { |
73390290 | 332 | khiter_t pos; |
7b35efd7 JS |
333 | if (!options->object_names) |
334 | return NULL; | |
73390290 JK |
335 | pos = kh_get_oid_map(options->object_names, *oid); |
336 | if (pos >= kh_end(options->object_names)) | |
337 | return NULL; | |
338 | return kh_value(options->object_names, pos); | |
7b35efd7 JS |
339 | } |
340 | ||
73390290 JK |
341 | void fsck_put_object_name(struct fsck_options *options, |
342 | const struct object_id *oid, | |
a59cfb32 | 343 | const char *fmt, ...) |
7b35efd7 JS |
344 | { |
345 | va_list ap; | |
346 | struct strbuf buf = STRBUF_INIT; | |
73390290 JK |
347 | khiter_t pos; |
348 | int hashret; | |
7b35efd7 JS |
349 | |
350 | if (!options->object_names) | |
351 | return; | |
73390290 JK |
352 | |
353 | pos = kh_put_oid_map(options->object_names, *oid, &hashret); | |
354 | if (!hashret) | |
7b35efd7 JS |
355 | return; |
356 | va_start(ap, fmt); | |
357 | strbuf_vaddf(&buf, fmt, ap); | |
73390290 | 358 | kh_value(options->object_names, pos) = strbuf_detach(&buf, NULL); |
7b35efd7 JS |
359 | va_end(ap); |
360 | } | |
361 | ||
a59cfb32 | 362 | const char *fsck_describe_object(struct fsck_options *options, |
73390290 | 363 | const struct object_id *oid) |
90cf590f | 364 | { |
a59cfb32 JK |
365 | static struct strbuf bufs[] = { |
366 | STRBUF_INIT, STRBUF_INIT, STRBUF_INIT, STRBUF_INIT | |
367 | }; | |
368 | static int b = 0; | |
369 | struct strbuf *buf; | |
73390290 | 370 | const char *name = fsck_get_object_name(options, oid); |
a59cfb32 JK |
371 | |
372 | buf = bufs + b; | |
373 | b = (b + 1) % ARRAY_SIZE(bufs); | |
374 | strbuf_reset(buf); | |
73390290 | 375 | strbuf_addstr(buf, oid_to_hex(oid)); |
a59cfb32 JK |
376 | if (name) |
377 | strbuf_addf(buf, " (%s)", name); | |
90cf590f | 378 | |
a59cfb32 | 379 | return buf->buf; |
90cf590f JS |
380 | } |
381 | ||
22410549 | 382 | static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options) |
355885d5 MK |
383 | { |
384 | struct tree_desc desc; | |
385 | struct name_entry entry; | |
386 | int res = 0; | |
7b35efd7 | 387 | const char *name; |
355885d5 MK |
388 | |
389 | if (parse_tree(tree)) | |
390 | return -1; | |
391 | ||
73390290 | 392 | name = fsck_get_object_name(options, &tree->object.oid); |
8354fa3d DT |
393 | if (init_tree_desc_gently(&desc, tree->buffer, tree->size)) |
394 | return -1; | |
395 | while (tree_entry_gently(&desc, &entry)) { | |
7b35efd7 | 396 | struct object *obj; |
355885d5 MK |
397 | int result; |
398 | ||
399 | if (S_ISGITLINK(entry.mode)) | |
400 | continue; | |
7b35efd7 JS |
401 | |
402 | if (S_ISDIR(entry.mode)) { | |
ea82b2a0 | 403 | obj = (struct object *)lookup_tree(the_repository, &entry.oid); |
2720f6db | 404 | if (name && obj) |
73390290 | 405 | fsck_put_object_name(options, &entry.oid, "%s%s/", |
a59cfb32 | 406 | name, entry.path); |
7b35efd7 JS |
407 | result = options->walk(obj, OBJ_TREE, data, options); |
408 | } | |
409 | else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) { | |
ea82b2a0 | 410 | obj = (struct object *)lookup_blob(the_repository, &entry.oid); |
2720f6db | 411 | if (name && obj) |
73390290 | 412 | fsck_put_object_name(options, &entry.oid, "%s%s", |
a59cfb32 | 413 | name, entry.path); |
7b35efd7 JS |
414 | result = options->walk(obj, OBJ_BLOB, data, options); |
415 | } | |
355885d5 | 416 | else { |
82247e9b | 417 | result = error("in tree %s: entry %s has bad mode %.6o", |
73390290 | 418 | fsck_describe_object(options, &tree->object.oid), |
a59cfb32 | 419 | entry.path, entry.mode); |
355885d5 MK |
420 | } |
421 | if (result < 0) | |
422 | return result; | |
423 | if (!res) | |
424 | res = result; | |
425 | } | |
426 | return res; | |
427 | } | |
428 | ||
22410549 | 429 | static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options) |
355885d5 | 430 | { |
7b35efd7 | 431 | int counter = 0, generation = 0, name_prefix_len = 0; |
355885d5 MK |
432 | struct commit_list *parents; |
433 | int res; | |
434 | int result; | |
7b35efd7 | 435 | const char *name; |
355885d5 MK |
436 | |
437 | if (parse_commit(commit)) | |
438 | return -1; | |
439 | ||
73390290 | 440 | name = fsck_get_object_name(options, &commit->object.oid); |
7b35efd7 | 441 | if (name) |
73390290 | 442 | fsck_put_object_name(options, get_commit_tree_oid(commit), |
a59cfb32 | 443 | "%s:", name); |
7b35efd7 | 444 | |
2e27bd77 DS |
445 | result = options->walk((struct object *)get_commit_tree(commit), |
446 | OBJ_TREE, data, options); | |
355885d5 MK |
447 | if (result < 0) |
448 | return result; | |
449 | res = result; | |
450 | ||
451 | parents = commit->parents; | |
7b35efd7 JS |
452 | if (name && parents) { |
453 | int len = strlen(name), power; | |
454 | ||
455 | if (len && name[len - 1] == '^') { | |
456 | generation = 1; | |
457 | name_prefix_len = len - 1; | |
458 | } | |
459 | else { /* parse ~<generation> suffix */ | |
460 | for (generation = 0, power = 1; | |
461 | len && isdigit(name[len - 1]); | |
462 | power *= 10) | |
463 | generation += power * (name[--len] - '0'); | |
464 | if (power > 1 && len && name[len - 1] == '~') | |
465 | name_prefix_len = len - 1; | |
e89f8936 JS |
466 | else { |
467 | /* Maybe a non-first parent, e.g. HEAD^2 */ | |
468 | generation = 0; | |
469 | name_prefix_len = len; | |
470 | } | |
7b35efd7 JS |
471 | } |
472 | } | |
473 | ||
355885d5 | 474 | while (parents) { |
7b35efd7 | 475 | if (name) { |
73390290 | 476 | struct object_id *oid = &parents->item->object.oid; |
7b35efd7 | 477 | |
b84c7838 | 478 | if (counter++) |
73390290 | 479 | fsck_put_object_name(options, oid, "%s^%d", |
a59cfb32 | 480 | name, counter); |
7b35efd7 | 481 | else if (generation > 0) |
73390290 | 482 | fsck_put_object_name(options, oid, "%.*s~%d", |
a59cfb32 JK |
483 | name_prefix_len, name, |
484 | generation + 1); | |
7b35efd7 | 485 | else |
73390290 | 486 | fsck_put_object_name(options, oid, "%s^", name); |
7b35efd7 | 487 | } |
22410549 | 488 | result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options); |
355885d5 MK |
489 | if (result < 0) |
490 | return result; | |
491 | if (!res) | |
492 | res = result; | |
493 | parents = parents->next; | |
494 | } | |
495 | return res; | |
496 | } | |
497 | ||
22410549 | 498 | static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options) |
355885d5 | 499 | { |
73390290 | 500 | const char *name = fsck_get_object_name(options, &tag->object.oid); |
7b35efd7 | 501 | |
355885d5 MK |
502 | if (parse_tag(tag)) |
503 | return -1; | |
7b35efd7 | 504 | if (name) |
73390290 | 505 | fsck_put_object_name(options, &tag->tagged->oid, "%s", name); |
22410549 | 506 | return options->walk(tag->tagged, OBJ_ANY, data, options); |
355885d5 MK |
507 | } |
508 | ||
22410549 | 509 | int fsck_walk(struct object *obj, void *data, struct fsck_options *options) |
355885d5 MK |
510 | { |
511 | if (!obj) | |
512 | return -1; | |
a2b22854 JK |
513 | |
514 | if (obj->type == OBJ_NONE) | |
109cd76d | 515 | parse_object(the_repository, &obj->oid); |
a2b22854 | 516 | |
355885d5 MK |
517 | switch (obj->type) { |
518 | case OBJ_BLOB: | |
519 | return 0; | |
520 | case OBJ_TREE: | |
22410549 | 521 | return fsck_walk_tree((struct tree *)obj, data, options); |
355885d5 | 522 | case OBJ_COMMIT: |
22410549 | 523 | return fsck_walk_commit((struct commit *)obj, data, options); |
355885d5 | 524 | case OBJ_TAG: |
22410549 | 525 | return fsck_walk_tag((struct tag *)obj, data, options); |
355885d5 | 526 | default: |
a59cfb32 | 527 | error("Unknown object type for %s", |
73390290 | 528 | fsck_describe_object(options, &obj->oid)); |
355885d5 MK |
529 | return -1; |
530 | } | |
531 | } | |
ba002f3b | 532 | |
9068cfb2 RS |
533 | struct name_stack { |
534 | const char **names; | |
535 | size_t nr, alloc; | |
536 | }; | |
537 | ||
538 | static void name_stack_push(struct name_stack *stack, const char *name) | |
539 | { | |
540 | ALLOC_GROW(stack->names, stack->nr + 1, stack->alloc); | |
541 | stack->names[stack->nr++] = name; | |
542 | } | |
543 | ||
544 | static const char *name_stack_pop(struct name_stack *stack) | |
545 | { | |
546 | return stack->nr ? stack->names[--stack->nr] : NULL; | |
547 | } | |
548 | ||
549 | static void name_stack_clear(struct name_stack *stack) | |
550 | { | |
551 | FREE_AND_NULL(stack->names); | |
552 | stack->nr = stack->alloc = 0; | |
553 | } | |
554 | ||
ba002f3b MK |
555 | /* |
556 | * The entries in a tree are ordered in the _path_ order, | |
557 | * which means that a directory entry is ordered by adding | |
558 | * a slash to the end of it. | |
559 | * | |
560 | * So a directory called "a" is ordered _after_ a file | |
561 | * called "a.c", because "a/" sorts after "a.c". | |
562 | */ | |
563 | #define TREE_UNORDERED (-1) | |
564 | #define TREE_HAS_DUPS (-2) | |
565 | ||
9068cfb2 RS |
566 | static int is_less_than_slash(unsigned char c) |
567 | { | |
568 | return '\0' < c && c < '/'; | |
569 | } | |
570 | ||
571 | static int verify_ordered(unsigned mode1, const char *name1, | |
572 | unsigned mode2, const char *name2, | |
573 | struct name_stack *candidates) | |
ba002f3b MK |
574 | { |
575 | int len1 = strlen(name1); | |
576 | int len2 = strlen(name2); | |
577 | int len = len1 < len2 ? len1 : len2; | |
578 | unsigned char c1, c2; | |
579 | int cmp; | |
580 | ||
581 | cmp = memcmp(name1, name2, len); | |
582 | if (cmp < 0) | |
583 | return 0; | |
584 | if (cmp > 0) | |
585 | return TREE_UNORDERED; | |
586 | ||
587 | /* | |
588 | * Ok, the first <len> characters are the same. | |
589 | * Now we need to order the next one, but turn | |
590 | * a '\0' into a '/' for a directory entry. | |
591 | */ | |
592 | c1 = name1[len]; | |
593 | c2 = name2[len]; | |
594 | if (!c1 && !c2) | |
595 | /* | |
596 | * git-write-tree used to write out a nonsense tree that has | |
597 | * entries with the same name, one blob and one tree. Make | |
598 | * sure we do not have duplicate entries. | |
599 | */ | |
600 | return TREE_HAS_DUPS; | |
601 | if (!c1 && S_ISDIR(mode1)) | |
602 | c1 = '/'; | |
603 | if (!c2 && S_ISDIR(mode2)) | |
604 | c2 = '/'; | |
9068cfb2 RS |
605 | |
606 | /* | |
607 | * There can be non-consecutive duplicates due to the implicitly | |
86715592 | 608 | * added slash, e.g.: |
9068cfb2 RS |
609 | * |
610 | * foo | |
611 | * foo.bar | |
612 | * foo.bar.baz | |
613 | * foo.bar/ | |
614 | * foo/ | |
615 | * | |
616 | * Record non-directory candidates (like "foo" and "foo.bar" in | |
617 | * the example) on a stack and check directory candidates (like | |
618 | * foo/" and "foo.bar/") against that stack. | |
619 | */ | |
620 | if (!c1 && is_less_than_slash(c2)) { | |
621 | name_stack_push(candidates, name1); | |
622 | } else if (c2 == '/' && is_less_than_slash(c1)) { | |
623 | for (;;) { | |
624 | const char *p; | |
625 | const char *f_name = name_stack_pop(candidates); | |
626 | ||
627 | if (!f_name) | |
628 | break; | |
629 | if (!skip_prefix(name2, f_name, &p)) | |
fe747043 | 630 | continue; |
9068cfb2 RS |
631 | if (!*p) |
632 | return TREE_HAS_DUPS; | |
633 | if (is_less_than_slash(*p)) { | |
634 | name_stack_push(candidates, f_name); | |
635 | break; | |
636 | } | |
637 | } | |
638 | } | |
639 | ||
ba002f3b MK |
640 | return c1 < c2 ? 0 : TREE_UNORDERED; |
641 | } | |
642 | ||
b2f2039c | 643 | static int fsck_tree(const struct object_id *oid, |
23a173a7 JK |
644 | const char *buffer, unsigned long size, |
645 | struct fsck_options *options) | |
ba002f3b | 646 | { |
8354fa3d | 647 | int retval = 0; |
c479d14a | 648 | int has_null_sha1 = 0; |
ba002f3b MK |
649 | int has_full_path = 0; |
650 | int has_empty_name = 0; | |
5d34a435 JK |
651 | int has_dot = 0; |
652 | int has_dotdot = 0; | |
5c17f512 | 653 | int has_dotgit = 0; |
ba002f3b MK |
654 | int has_zero_pad = 0; |
655 | int has_bad_modes = 0; | |
656 | int has_dup_entries = 0; | |
657 | int not_properly_sorted = 0; | |
658 | struct tree_desc desc; | |
659 | unsigned o_mode; | |
660 | const char *o_name; | |
9068cfb2 | 661 | struct name_stack df_dup_candidates = { NULL }; |
ba002f3b | 662 | |
23a173a7 | 663 | if (init_tree_desc_gently(&desc, buffer, size)) { |
b2f2039c | 664 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree"); |
8354fa3d DT |
665 | return retval; |
666 | } | |
ba002f3b MK |
667 | |
668 | o_mode = 0; | |
669 | o_name = NULL; | |
ba002f3b MK |
670 | |
671 | while (desc.size) { | |
5ec1e728 | 672 | unsigned short mode; |
288a74bc | 673 | const char *name, *backslash; |
ce6663a9 | 674 | const struct object_id *oid; |
ba002f3b | 675 | |
ce6663a9 | 676 | oid = tree_entry_extract(&desc, &name, &mode); |
ba002f3b | 677 | |
ce6663a9 | 678 | has_null_sha1 |= is_null_oid(oid); |
effd12ec HS |
679 | has_full_path |= !!strchr(name, '/'); |
680 | has_empty_name |= !*name; | |
681 | has_dot |= !strcmp(name, "."); | |
682 | has_dotdot |= !strcmp(name, ".."); | |
ed9c3220 | 683 | has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name); |
ba002f3b | 684 | has_zero_pad |= *(char *)desc.buffer == '0'; |
159e7b08 | 685 | |
b7b1fca1 JK |
686 | if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) { |
687 | if (!S_ISLNK(mode)) | |
688 | oidset_insert(&gitmodules_found, oid); | |
689 | else | |
38370253 | 690 | retval += report(options, |
b2f2039c | 691 | oid, OBJ_TREE, |
b7b1fca1 JK |
692 | FSCK_MSG_GITMODULES_SYMLINK, |
693 | ".gitmodules is a symbolic link"); | |
694 | } | |
159e7b08 | 695 | |
288a74bc JS |
696 | if ((backslash = strchr(name, '\\'))) { |
697 | while (backslash) { | |
698 | backslash++; | |
699 | has_dotgit |= is_ntfs_dotgit(backslash); | |
bdfef049 JS |
700 | if (is_ntfs_dotgitmodules(backslash)) { |
701 | if (!S_ISLNK(mode)) | |
702 | oidset_insert(&gitmodules_found, oid); | |
703 | else | |
7034cd09 | 704 | retval += report(options, oid, OBJ_TREE, |
bdfef049 JS |
705 | FSCK_MSG_GITMODULES_SYMLINK, |
706 | ".gitmodules is a symbolic link"); | |
707 | } | |
288a74bc JS |
708 | backslash = strchr(backslash, '\\'); |
709 | } | |
710 | } | |
711 | ||
8354fa3d | 712 | if (update_tree_entry_gently(&desc)) { |
b2f2039c | 713 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree"); |
8354fa3d DT |
714 | break; |
715 | } | |
ba002f3b MK |
716 | |
717 | switch (mode) { | |
718 | /* | |
719 | * Standard modes.. | |
720 | */ | |
721 | case S_IFREG | 0755: | |
722 | case S_IFREG | 0644: | |
723 | case S_IFLNK: | |
724 | case S_IFDIR: | |
725 | case S_IFGITLINK: | |
726 | break; | |
727 | /* | |
728 | * This is nonstandard, but we had a few of these | |
729 | * early on when we honored the full set of mode | |
730 | * bits.. | |
731 | */ | |
732 | case S_IFREG | 0664: | |
22410549 | 733 | if (!options->strict) |
ba002f3b | 734 | break; |
1cf01a34 | 735 | /* fallthrough */ |
ba002f3b MK |
736 | default: |
737 | has_bad_modes = 1; | |
738 | } | |
739 | ||
740 | if (o_name) { | |
9068cfb2 RS |
741 | switch (verify_ordered(o_mode, o_name, mode, name, |
742 | &df_dup_candidates)) { | |
ba002f3b MK |
743 | case TREE_UNORDERED: |
744 | not_properly_sorted = 1; | |
745 | break; | |
746 | case TREE_HAS_DUPS: | |
747 | has_dup_entries = 1; | |
748 | break; | |
749 | default: | |
750 | break; | |
751 | } | |
752 | } | |
753 | ||
754 | o_mode = mode; | |
755 | o_name = name; | |
ba002f3b MK |
756 | } |
757 | ||
9068cfb2 RS |
758 | name_stack_clear(&df_dup_candidates); |
759 | ||
c479d14a | 760 | if (has_null_sha1) |
b2f2039c | 761 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1"); |
ba002f3b | 762 | if (has_full_path) |
b2f2039c | 763 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_FULL_PATHNAME, "contains full pathnames"); |
ba002f3b | 764 | if (has_empty_name) |
b2f2039c | 765 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_EMPTY_NAME, "contains empty pathname"); |
5d34a435 | 766 | if (has_dot) |
b2f2039c | 767 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_HAS_DOT, "contains '.'"); |
5d34a435 | 768 | if (has_dotdot) |
b2f2039c | 769 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_HAS_DOTDOT, "contains '..'"); |
5c17f512 | 770 | if (has_dotgit) |
b2f2039c | 771 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_HAS_DOTGIT, "contains '.git'"); |
ba002f3b | 772 | if (has_zero_pad) |
b2f2039c | 773 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes"); |
ba002f3b | 774 | if (has_bad_modes) |
b2f2039c | 775 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_BAD_FILEMODE, "contains bad file modes"); |
ba002f3b | 776 | if (has_dup_entries) |
b2f2039c | 777 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries"); |
ba002f3b | 778 | if (not_properly_sorted) |
b2f2039c | 779 | retval += report(options, oid, OBJ_TREE, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted"); |
ba002f3b MK |
780 | return retval; |
781 | } | |
782 | ||
84d18c0b | 783 | static int verify_headers(const void *data, unsigned long size, |
cc579000 JK |
784 | const struct object_id *oid, enum object_type type, |
785 | struct fsck_options *options) | |
4d0d8975 JS |
786 | { |
787 | const char *buffer = (const char *)data; | |
788 | unsigned long i; | |
789 | ||
790 | for (i = 0; i < size; i++) { | |
791 | switch (buffer[i]) { | |
792 | case '\0': | |
cc579000 | 793 | return report(options, oid, type, |
c99ba492 JS |
794 | FSCK_MSG_NUL_IN_HEADER, |
795 | "unterminated header: NUL at offset %ld", i); | |
4d0d8975 JS |
796 | case '\n': |
797 | if (i + 1 < size && buffer[i + 1] == '\n') | |
798 | return 0; | |
799 | } | |
800 | } | |
801 | ||
84d18c0b JH |
802 | /* |
803 | * We did not find double-LF that separates the header | |
804 | * and the body. Not having a body is not a crime but | |
805 | * we do want to see the terminating LF for the last header | |
806 | * line. | |
807 | */ | |
808 | if (size && buffer[size - 1] == '\n') | |
809 | return 0; | |
810 | ||
cc579000 | 811 | return report(options, oid, type, |
c99ba492 | 812 | FSCK_MSG_UNTERMINATED_HEADER, "unterminated header"); |
4d0d8975 JS |
813 | } |
814 | ||
78543993 JK |
815 | static int fsck_ident(const char **ident, |
816 | const struct object_id *oid, enum object_type type, | |
817 | struct fsck_options *options) | |
daae1922 | 818 | { |
e6826e33 | 819 | const char *p = *ident; |
d4b8de04 JK |
820 | char *end; |
821 | ||
e6826e33 JS |
822 | *ident = strchrnul(*ident, '\n'); |
823 | if (**ident == '\n') | |
824 | (*ident)++; | |
825 | ||
826 | if (*p == '<') | |
78543993 | 827 | return report(options, oid, type, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email"); |
e6826e33 JS |
828 | p += strcspn(p, "<>\n"); |
829 | if (*p == '>') | |
78543993 | 830 | return report(options, oid, type, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name"); |
e6826e33 | 831 | if (*p != '<') |
78543993 | 832 | return report(options, oid, type, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email"); |
e6826e33 | 833 | if (p[-1] != ' ') |
78543993 | 834 | return report(options, oid, type, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email"); |
e6826e33 JS |
835 | p++; |
836 | p += strcspn(p, "<>\n"); | |
837 | if (*p != '>') | |
78543993 | 838 | return report(options, oid, type, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email"); |
e6826e33 JS |
839 | p++; |
840 | if (*p != ' ') | |
78543993 | 841 | return report(options, oid, type, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date"); |
e6826e33 JS |
842 | p++; |
843 | if (*p == '0' && p[1] != ' ') | |
78543993 | 844 | return report(options, oid, type, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date"); |
1aeb7e75 | 845 | if (date_overflows(parse_timestamp(p, &end, 10))) |
78543993 | 846 | return report(options, oid, type, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow"); |
e6826e33 | 847 | if ((end == p || *end != ' ')) |
78543993 | 848 | return report(options, oid, type, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date"); |
e6826e33 JS |
849 | p = end + 1; |
850 | if ((*p != '+' && *p != '-') || | |
851 | !isdigit(p[1]) || | |
852 | !isdigit(p[2]) || | |
853 | !isdigit(p[3]) || | |
854 | !isdigit(p[4]) || | |
855 | (p[5] != '\n')) | |
78543993 | 856 | return report(options, oid, type, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone"); |
e6826e33 | 857 | p += 6; |
daae1922 JN |
858 | return 0; |
859 | } | |
860 | ||
c5b4269b JK |
861 | static int fsck_commit(const struct object_id *oid, |
862 | const char *buffer, unsigned long size, | |
863 | struct fsck_options *options) | |
ba002f3b | 864 | { |
f648ee70 | 865 | struct object_id tree_oid, parent_oid; |
ec652315 | 866 | unsigned author_count; |
daae1922 | 867 | int err; |
6d2d780f | 868 | const char *buffer_begin = buffer; |
c54f5ca9 | 869 | const char *p; |
ba002f3b | 870 | |
c5b4269b | 871 | if (verify_headers(buffer, size, oid, OBJ_COMMIT, options)) |
4d0d8975 JS |
872 | return -1; |
873 | ||
cf4fff57 | 874 | if (!skip_prefix(buffer, "tree ", &buffer)) |
c5b4269b | 875 | return report(options, oid, OBJ_COMMIT, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line"); |
c54f5ca9 | 876 | if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') { |
c5b4269b | 877 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1"); |
b3584761 JS |
878 | if (err) |
879 | return err; | |
880 | } | |
c54f5ca9 | 881 | buffer = p + 1; |
cf4fff57 | 882 | while (skip_prefix(buffer, "parent ", &buffer)) { |
f648ee70 | 883 | if (parse_oid_hex(buffer, &parent_oid, &p) || *p != '\n') { |
c5b4269b | 884 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1"); |
b3584761 JS |
885 | if (err) |
886 | return err; | |
887 | } | |
c54f5ca9 | 888 | buffer = p + 1; |
ba002f3b | 889 | } |
c9ad147f JS |
890 | author_count = 0; |
891 | while (skip_prefix(buffer, "author ", &buffer)) { | |
892 | author_count++; | |
c5b4269b | 893 | err = fsck_ident(&buffer, oid, OBJ_COMMIT, options); |
c9ad147f JS |
894 | if (err) |
895 | return err; | |
ba002f3b | 896 | } |
c9ad147f | 897 | if (author_count < 1) |
c5b4269b | 898 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line"); |
c9ad147f | 899 | else if (author_count > 1) |
c5b4269b | 900 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines"); |
daae1922 JN |
901 | if (err) |
902 | return err; | |
cf4fff57 | 903 | if (!skip_prefix(buffer, "committer ", &buffer)) |
c5b4269b JK |
904 | return report(options, oid, OBJ_COMMIT, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line"); |
905 | err = fsck_ident(&buffer, oid, OBJ_COMMIT, options); | |
daae1922 JN |
906 | if (err) |
907 | return err; | |
6d2d780f | 908 | if (memchr(buffer_begin, '\0', size)) { |
c5b4269b | 909 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_NUL_IN_COMMIT, |
6d2d780f JH |
910 | "NUL byte in the commit object body"); |
911 | if (err) | |
912 | return err; | |
913 | } | |
ba002f3b MK |
914 | return 0; |
915 | } | |
916 | ||
103fb6d4 | 917 | static int fsck_tag(const struct object_id *oid, const char *buffer, |
2175a0c6 | 918 | unsigned long size, struct fsck_options *options) |
cec097be | 919 | { |
f648ee70 | 920 | struct object_id tagged_oid; |
acf9de4c ÆAB |
921 | int tagged_type; |
922 | return fsck_tag_standalone(oid, buffer, size, options, &tagged_oid, | |
923 | &tagged_type); | |
924 | } | |
925 | ||
926 | int fsck_tag_standalone(const struct object_id *oid, const char *buffer, | |
927 | unsigned long size, struct fsck_options *options, | |
928 | struct object_id *tagged_oid, | |
929 | int *tagged_type) | |
930 | { | |
cec097be | 931 | int ret = 0; |
23a173a7 | 932 | char *eol; |
cec097be | 933 | struct strbuf sb = STRBUF_INIT; |
c54f5ca9 | 934 | const char *p; |
cec097be | 935 | |
103fb6d4 | 936 | ret = verify_headers(buffer, size, oid, OBJ_TAG, options); |
8a272f29 | 937 | if (ret) |
cec097be JS |
938 | goto done; |
939 | ||
940 | if (!skip_prefix(buffer, "object ", &buffer)) { | |
103fb6d4 | 941 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line"); |
cec097be JS |
942 | goto done; |
943 | } | |
acf9de4c | 944 | if (parse_oid_hex(buffer, tagged_oid, &p) || *p != '\n') { |
103fb6d4 | 945 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1"); |
7d7d5b05 JS |
946 | if (ret) |
947 | goto done; | |
cec097be | 948 | } |
c54f5ca9 | 949 | buffer = p + 1; |
cec097be JS |
950 | |
951 | if (!skip_prefix(buffer, "type ", &buffer)) { | |
103fb6d4 | 952 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line"); |
cec097be JS |
953 | goto done; |
954 | } | |
955 | eol = strchr(buffer, '\n'); | |
956 | if (!eol) { | |
103fb6d4 | 957 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line"); |
cec097be JS |
958 | goto done; |
959 | } | |
acf9de4c ÆAB |
960 | *tagged_type = type_from_string_gently(buffer, eol - buffer, 1); |
961 | if (*tagged_type < 0) | |
103fb6d4 | 962 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_BAD_TYPE, "invalid 'type' value"); |
cec097be JS |
963 | if (ret) |
964 | goto done; | |
965 | buffer = eol + 1; | |
966 | ||
967 | if (!skip_prefix(buffer, "tag ", &buffer)) { | |
103fb6d4 | 968 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line"); |
cec097be JS |
969 | goto done; |
970 | } | |
971 | eol = strchr(buffer, '\n'); | |
972 | if (!eol) { | |
103fb6d4 | 973 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line"); |
cec097be JS |
974 | goto done; |
975 | } | |
976 | strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer); | |
f27d05b1 | 977 | if (check_refname_format(sb.buf, 0)) { |
103fb6d4 | 978 | ret = report(options, oid, OBJ_TAG, |
38370253 JK |
979 | FSCK_MSG_BAD_TAG_NAME, |
980 | "invalid 'tag' name: %.*s", | |
981 | (int)(eol - buffer), buffer); | |
f27d05b1 JS |
982 | if (ret) |
983 | goto done; | |
984 | } | |
cec097be JS |
985 | buffer = eol + 1; |
986 | ||
f27d05b1 | 987 | if (!skip_prefix(buffer, "tagger ", &buffer)) { |
cec097be | 988 | /* early tags do not contain 'tagger' lines; warn only */ |
103fb6d4 | 989 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line"); |
f27d05b1 JS |
990 | if (ret) |
991 | goto done; | |
992 | } | |
cec097be | 993 | else |
103fb6d4 | 994 | ret = fsck_ident(&buffer, oid, OBJ_TAG, options); |
9a1a3a4d ÆAB |
995 | if (!*buffer) |
996 | goto done; | |
cec097be | 997 | |
acf9de4c ÆAB |
998 | if (!starts_with(buffer, "\n")) { |
999 | /* | |
1000 | * The verify_headers() check will allow | |
1001 | * e.g. "[...]tagger <tagger>\nsome | |
1002 | * garbage\n\nmessage" to pass, thinking "some | |
1003 | * garbage" could be a custom header. E.g. "mktag" | |
1004 | * doesn't want any unknown headers. | |
1005 | */ | |
1006 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_EXTRA_HEADER_ENTRY, "invalid format - extra header(s) after 'tagger'"); | |
1007 | if (ret) | |
1008 | goto done; | |
1009 | } | |
1010 | ||
cec097be JS |
1011 | done: |
1012 | strbuf_release(&sb); | |
cec097be JS |
1013 | return ret; |
1014 | } | |
1015 | ||
a2b26ffb JN |
1016 | /* |
1017 | * Like builtin/submodule--helper.c's starts_with_dot_slash, but without | |
1018 | * relying on the platform-dependent is_dir_sep helper. | |
1019 | * | |
1020 | * This is for use in checking whether a submodule URL is interpreted as | |
1021 | * relative to the current directory on any platform, since \ is a | |
1022 | * directory separator on Windows but not on other platforms. | |
1023 | */ | |
1024 | static int starts_with_dot_slash(const char *str) | |
1025 | { | |
1026 | return str[0] == '.' && (str[1] == '/' || str[1] == '\\'); | |
1027 | } | |
1028 | ||
1029 | /* | |
1030 | * Like starts_with_dot_slash, this is a variant of submodule--helper's | |
1031 | * helper of the same name with the twist that it accepts backslash as a | |
1032 | * directory separator even on non-Windows platforms. | |
1033 | */ | |
1034 | static int starts_with_dot_dot_slash(const char *str) | |
1035 | { | |
1036 | return str[0] == '.' && starts_with_dot_slash(str + 1); | |
1037 | } | |
1038 | ||
1039 | static int submodule_url_is_relative(const char *url) | |
1040 | { | |
1041 | return starts_with_dot_slash(url) || starts_with_dot_dot_slash(url); | |
1042 | } | |
1043 | ||
c44088ec JN |
1044 | /* |
1045 | * Count directory components that a relative submodule URL should chop | |
1046 | * from the remote_url it is to be resolved against. | |
1047 | * | |
1048 | * In other words, this counts "../" components at the start of a | |
1049 | * submodule URL. | |
1050 | * | |
1051 | * Returns the number of directory components to chop and writes a | |
1052 | * pointer to the next character of url after all leading "./" and | |
1053 | * "../" components to out. | |
1054 | */ | |
1055 | static int count_leading_dotdots(const char *url, const char **out) | |
1056 | { | |
1057 | int result = 0; | |
1058 | while (1) { | |
1059 | if (starts_with_dot_dot_slash(url)) { | |
1060 | result++; | |
1061 | url += strlen("../"); | |
1062 | continue; | |
1063 | } | |
1064 | if (starts_with_dot_slash(url)) { | |
1065 | url += strlen("./"); | |
1066 | continue; | |
1067 | } | |
1068 | *out = url; | |
1069 | return result; | |
1070 | } | |
1071 | } | |
a2b26ffb JN |
1072 | /* |
1073 | * Check whether a transport is implemented by git-remote-curl. | |
1074 | * | |
1075 | * If it is, returns 1 and writes the URL that would be passed to | |
1076 | * git-remote-curl to the "out" parameter. | |
1077 | * | |
1078 | * Otherwise, returns 0 and leaves "out" untouched. | |
1079 | * | |
1080 | * Examples: | |
1081 | * http::https://example.com/repo.git -> 1, https://example.com/repo.git | |
1082 | * https://example.com/repo.git -> 1, https://example.com/repo.git | |
1083 | * git://example.com/repo.git -> 0 | |
1084 | * | |
1085 | * This is for use in checking for previously exploitable bugs that | |
1086 | * required a submodule URL to be passed to git-remote-curl. | |
1087 | */ | |
1088 | static int url_to_curl_url(const char *url, const char **out) | |
1089 | { | |
1090 | /* | |
1091 | * We don't need to check for case-aliases, "http.exe", and so | |
1092 | * on because in the default configuration, is_transport_allowed | |
1093 | * prevents URLs with those schemes from being cloned | |
1094 | * automatically. | |
1095 | */ | |
1096 | if (skip_prefix(url, "http::", out) || | |
1097 | skip_prefix(url, "https::", out) || | |
1098 | skip_prefix(url, "ftp::", out) || | |
1099 | skip_prefix(url, "ftps::", out)) | |
1100 | return 1; | |
1101 | if (starts_with(url, "http://") || | |
1102 | starts_with(url, "https://") || | |
1103 | starts_with(url, "ftp://") || | |
1104 | starts_with(url, "ftps://")) { | |
1105 | *out = url; | |
1106 | return 1; | |
1107 | } | |
1108 | return 0; | |
1109 | } | |
1110 | ||
07259e74 JK |
1111 | static int check_submodule_url(const char *url) |
1112 | { | |
a2b26ffb | 1113 | const char *curl_url; |
07259e74 JK |
1114 | |
1115 | if (looks_like_command_line_option(url)) | |
1116 | return -1; | |
1117 | ||
6aed5673 | 1118 | if (submodule_url_is_relative(url) || starts_with(url, "git://")) { |
c44088ec JN |
1119 | char *decoded; |
1120 | const char *next; | |
1121 | int has_nl; | |
1122 | ||
a2b26ffb JN |
1123 | /* |
1124 | * This could be appended to an http URL and url-decoded; | |
1125 | * check for malicious characters. | |
1126 | */ | |
c44088ec JN |
1127 | decoded = url_decode(url); |
1128 | has_nl = !!strchr(decoded, '\n'); | |
1129 | ||
a2b26ffb JN |
1130 | free(decoded); |
1131 | if (has_nl) | |
1132 | return -1; | |
c44088ec JN |
1133 | |
1134 | /* | |
1135 | * URLs which escape their root via "../" can overwrite | |
1136 | * the host field and previous components, resolving to | |
1a3609e4 JN |
1137 | * URLs like https::example.com/submodule.git and |
1138 | * https:///example.com/submodule.git that were | |
c44088ec JN |
1139 | * susceptible to CVE-2020-11008. |
1140 | */ | |
1141 | if (count_leading_dotdots(url, &next) > 0 && | |
1a3609e4 | 1142 | (*next == ':' || *next == '/')) |
c44088ec | 1143 | return -1; |
a2b26ffb JN |
1144 | } |
1145 | ||
1146 | else if (url_to_curl_url(url, &curl_url)) { | |
1147 | struct credential c = CREDENTIAL_INIT; | |
1a3609e4 JN |
1148 | int ret = 0; |
1149 | if (credential_from_url_gently(&c, curl_url, 1) || | |
1150 | !*c.host) | |
1151 | ret = -1; | |
a2b26ffb JN |
1152 | credential_clear(&c); |
1153 | return ret; | |
1154 | } | |
1155 | ||
1156 | return 0; | |
07259e74 JK |
1157 | } |
1158 | ||
ed8b10f6 | 1159 | struct fsck_gitmodules_data { |
6da40b22 | 1160 | const struct object_id *oid; |
ed8b10f6 JK |
1161 | struct fsck_options *options; |
1162 | int ret; | |
1163 | }; | |
1164 | ||
1165 | static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata) | |
1166 | { | |
1167 | struct fsck_gitmodules_data *data = vdata; | |
1168 | const char *subsection, *key; | |
f5914f4b | 1169 | size_t subsection_len; |
ed8b10f6 JK |
1170 | char *name; |
1171 | ||
1172 | if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 || | |
1173 | !subsection) | |
1174 | return 0; | |
1175 | ||
1176 | name = xmemdupz(subsection, subsection_len); | |
1177 | if (check_submodule_name(name) < 0) | |
38370253 | 1178 | data->ret |= report(data->options, |
6da40b22 | 1179 | data->oid, OBJ_BLOB, |
ed8b10f6 JK |
1180 | FSCK_MSG_GITMODULES_NAME, |
1181 | "disallowed submodule name: %s", | |
1182 | name); | |
a124133e | 1183 | if (!strcmp(key, "url") && value && |
07259e74 | 1184 | check_submodule_url(value) < 0) |
38370253 | 1185 | data->ret |= report(data->options, |
6da40b22 | 1186 | data->oid, OBJ_BLOB, |
a124133e JK |
1187 | FSCK_MSG_GITMODULES_URL, |
1188 | "disallowed submodule url: %s", | |
1189 | value); | |
1a7fd1fb JK |
1190 | if (!strcmp(key, "path") && value && |
1191 | looks_like_command_line_option(value)) | |
38370253 | 1192 | data->ret |= report(data->options, |
6da40b22 | 1193 | data->oid, OBJ_BLOB, |
1a7fd1fb JK |
1194 | FSCK_MSG_GITMODULES_PATH, |
1195 | "disallowed submodule path: %s", | |
1196 | value); | |
bb92255e JN |
1197 | if (!strcmp(key, "update") && value && |
1198 | parse_submodule_update_type(value) == SM_UPDATE_COMMAND) | |
7034cd09 | 1199 | data->ret |= report(data->options, data->oid, OBJ_BLOB, |
bb92255e JN |
1200 | FSCK_MSG_GITMODULES_UPDATE, |
1201 | "disallowed submodule update setting: %s", | |
1202 | value); | |
ed8b10f6 JK |
1203 | free(name); |
1204 | ||
1205 | return 0; | |
1206 | } | |
1207 | ||
6da40b22 | 1208 | static int fsck_blob(const struct object_id *oid, const char *buf, |
7ac4f3a0 JK |
1209 | unsigned long size, struct fsck_options *options) |
1210 | { | |
ed8b10f6 | 1211 | struct fsck_gitmodules_data data; |
de6bd9e3 | 1212 | struct config_options config_opts = { 0 }; |
ed8b10f6 | 1213 | |
6da40b22 | 1214 | if (!oidset_contains(&gitmodules_found, oid)) |
ed8b10f6 | 1215 | return 0; |
6da40b22 | 1216 | oidset_insert(&gitmodules_done, oid); |
ed8b10f6 | 1217 | |
6da40b22 | 1218 | if (object_on_skiplist(options, oid)) |
fb162877 RJ |
1219 | return 0; |
1220 | ||
ed8b10f6 JK |
1221 | if (!buf) { |
1222 | /* | |
1223 | * A missing buffer here is a sign that the caller found the | |
1224 | * blob too gigantic to load into memory. Let's just consider | |
1225 | * that an error. | |
1226 | */ | |
6da40b22 | 1227 | return report(options, oid, OBJ_BLOB, |
0d68764d | 1228 | FSCK_MSG_GITMODULES_LARGE, |
ed8b10f6 JK |
1229 | ".gitmodules too large to parse"); |
1230 | } | |
1231 | ||
6da40b22 | 1232 | data.oid = oid; |
ed8b10f6 JK |
1233 | data.options = options; |
1234 | data.ret = 0; | |
de6bd9e3 | 1235 | config_opts.error_action = CONFIG_ERROR_SILENT; |
ed8b10f6 | 1236 | if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB, |
de6bd9e3 | 1237 | ".gitmodules", buf, size, &data, &config_opts)) |
6da40b22 | 1238 | data.ret |= report(options, oid, OBJ_BLOB, |
ed8b10f6 JK |
1239 | FSCK_MSG_GITMODULES_PARSE, |
1240 | "could not parse gitmodules blob"); | |
1241 | ||
1242 | return data.ret; | |
7ac4f3a0 JK |
1243 | } |
1244 | ||
90a398bb | 1245 | int fsck_object(struct object *obj, void *data, unsigned long size, |
22410549 | 1246 | struct fsck_options *options) |
ba002f3b MK |
1247 | { |
1248 | if (!obj) | |
38370253 | 1249 | return report(options, NULL, OBJ_NONE, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck"); |
ba002f3b MK |
1250 | |
1251 | if (obj->type == OBJ_BLOB) | |
6da40b22 | 1252 | return fsck_blob(&obj->oid, data, size, options); |
ba002f3b | 1253 | if (obj->type == OBJ_TREE) |
b2f2039c | 1254 | return fsck_tree(&obj->oid, data, size, options); |
ba002f3b | 1255 | if (obj->type == OBJ_COMMIT) |
c5b4269b | 1256 | return fsck_commit(&obj->oid, data, size, options); |
ba002f3b | 1257 | if (obj->type == OBJ_TAG) |
103fb6d4 | 1258 | return fsck_tag(&obj->oid, data, size, options); |
ba002f3b | 1259 | |
38370253 JK |
1260 | return report(options, &obj->oid, obj->type, |
1261 | FSCK_MSG_UNKNOWN_TYPE, | |
1262 | "unknown type '%d' (internal fsck error)", | |
1263 | obj->type); | |
ba002f3b | 1264 | } |
d6ffc8d7 | 1265 | |
1cd772cc | 1266 | int fsck_error_function(struct fsck_options *o, |
5afc4b1d JK |
1267 | const struct object_id *oid, |
1268 | enum object_type object_type, | |
1269 | int msg_type, const char *message) | |
d6ffc8d7 | 1270 | { |
0282f4dc | 1271 | if (msg_type == FSCK_WARN) { |
5afc4b1d | 1272 | warning("object %s: %s", fsck_describe_object(o, oid), message); |
0282f4dc JS |
1273 | return 0; |
1274 | } | |
5afc4b1d | 1275 | error("object %s: %s", fsck_describe_object(o, oid), message); |
d6ffc8d7 MK |
1276 | return 1; |
1277 | } | |
159e7b08 | 1278 | |
5476e1ef JT |
1279 | void register_found_gitmodules(const struct object_id *oid) |
1280 | { | |
1281 | oidset_insert(&gitmodules_found, oid); | |
1282 | } | |
1283 | ||
159e7b08 JK |
1284 | int fsck_finish(struct fsck_options *options) |
1285 | { | |
1286 | int ret = 0; | |
1287 | struct oidset_iter iter; | |
1288 | const struct object_id *oid; | |
1289 | ||
1290 | oidset_iter_init(&gitmodules_found, &iter); | |
1291 | while ((oid = oidset_iter_next(&iter))) { | |
159e7b08 JK |
1292 | enum object_type type; |
1293 | unsigned long size; | |
1294 | char *buf; | |
1295 | ||
1296 | if (oidset_contains(&gitmodules_done, oid)) | |
1297 | continue; | |
1298 | ||
7913f53b | 1299 | buf = read_object_file(oid, &type, &size); |
159e7b08 | 1300 | if (!buf) { |
b8b00f16 | 1301 | if (is_promisor_object(oid)) |
27387444 | 1302 | continue; |
38370253 | 1303 | ret |= report(options, |
b8b00f16 | 1304 | oid, OBJ_BLOB, |
159e7b08 JK |
1305 | FSCK_MSG_GITMODULES_MISSING, |
1306 | "unable to read .gitmodules blob"); | |
1307 | continue; | |
1308 | } | |
1309 | ||
1310 | if (type == OBJ_BLOB) | |
b8b00f16 | 1311 | ret |= fsck_blob(oid, buf, size, options); |
159e7b08 | 1312 | else |
38370253 | 1313 | ret |= report(options, |
b8b00f16 | 1314 | oid, type, |
159e7b08 JK |
1315 | FSCK_MSG_GITMODULES_BLOB, |
1316 | "non-blob found at .gitmodules"); | |
1317 | free(buf); | |
1318 | } | |
1319 | ||
1320 | ||
1321 | oidset_clear(&gitmodules_found); | |
1322 | oidset_clear(&gitmodules_done); | |
1323 | return ret; | |
1324 | } | |
1f3299fd ÆAB |
1325 | |
1326 | int fsck_config_internal(const char *var, const char *value, void *cb, | |
1327 | struct fsck_options *options) | |
1328 | { | |
1329 | if (strcmp(var, "fsck.skiplist") == 0) { | |
1330 | const char *path; | |
1331 | struct strbuf sb = STRBUF_INIT; | |
1332 | ||
1333 | if (git_config_pathname(&path, var, value)) | |
1334 | return 1; | |
1335 | strbuf_addf(&sb, "skiplist=%s", path); | |
1336 | free((char *)path); | |
1337 | fsck_set_msg_types(options, sb.buf); | |
1338 | strbuf_release(&sb); | |
1339 | return 0; | |
1340 | } | |
1341 | ||
1342 | if (skip_prefix(var, "fsck.", &var)) { | |
1343 | fsck_set_msg_type(options, var, value); | |
1344 | return 0; | |
1345 | } | |
1346 | ||
1347 | return git_default_config(var, value, cb); | |
1348 | } |