]>
Commit | Line | Data |
---|---|---|
355885d5 MK |
1 | #include "cache.h" |
2 | #include "object.h" | |
3 | #include "blob.h" | |
4 | #include "tree.h" | |
5 | #include "tree-walk.h" | |
6 | #include "commit.h" | |
7 | #include "tag.h" | |
8 | #include "fsck.h" | |
cec097be | 9 | #include "refs.h" |
a2b26ffb | 10 | #include "url.h" |
a18fcc9f | 11 | #include "utf8.h" |
cd94c6f9 | 12 | #include "sha1-array.h" |
7b35efd7 | 13 | #include "decorate.h" |
159e7b08 | 14 | #include "oidset.h" |
27387444 | 15 | #include "packfile.h" |
ed8b10f6 JK |
16 | #include "submodule-config.h" |
17 | #include "config.h" | |
07259e74 | 18 | #include "credential.h" |
159e7b08 JK |
19 | |
20 | static struct oidset gitmodules_found = OIDSET_INIT; | |
21 | static struct oidset gitmodules_done = OIDSET_INIT; | |
355885d5 | 22 | |
f50c4407 | 23 | #define FSCK_FATAL -1 |
f27d05b1 | 24 | #define FSCK_INFO -2 |
f50c4407 | 25 | |
c99ba492 | 26 | #define FOREACH_MSG_ID(FUNC) \ |
f50c4407 JS |
27 | /* fatal errors */ \ |
28 | FUNC(NUL_IN_HEADER, FATAL) \ | |
29 | FUNC(UNTERMINATED_HEADER, FATAL) \ | |
c99ba492 JS |
30 | /* errors */ \ |
31 | FUNC(BAD_DATE, ERROR) \ | |
32 | FUNC(BAD_DATE_OVERFLOW, ERROR) \ | |
33 | FUNC(BAD_EMAIL, ERROR) \ | |
34 | FUNC(BAD_NAME, ERROR) \ | |
35 | FUNC(BAD_OBJECT_SHA1, ERROR) \ | |
36 | FUNC(BAD_PARENT_SHA1, ERROR) \ | |
37 | FUNC(BAD_TAG_OBJECT, ERROR) \ | |
38 | FUNC(BAD_TIMEZONE, ERROR) \ | |
39 | FUNC(BAD_TREE, ERROR) \ | |
40 | FUNC(BAD_TREE_SHA1, ERROR) \ | |
41 | FUNC(BAD_TYPE, ERROR) \ | |
42 | FUNC(DUPLICATE_ENTRIES, ERROR) \ | |
43 | FUNC(MISSING_AUTHOR, ERROR) \ | |
44 | FUNC(MISSING_COMMITTER, ERROR) \ | |
45 | FUNC(MISSING_EMAIL, ERROR) \ | |
46 | FUNC(MISSING_GRAFT, ERROR) \ | |
47 | FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \ | |
48 | FUNC(MISSING_OBJECT, ERROR) \ | |
49 | FUNC(MISSING_PARENT, ERROR) \ | |
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) \ | |
54 | FUNC(MISSING_TAG_OBJECT, ERROR) \ | |
55 | FUNC(MISSING_TREE, ERROR) \ | |
159e7b08 | 56 | FUNC(MISSING_TREE_OBJECT, ERROR) \ |
c99ba492 JS |
57 | FUNC(MISSING_TYPE, ERROR) \ |
58 | FUNC(MISSING_TYPE_ENTRY, ERROR) \ | |
c9ad147f | 59 | FUNC(MULTIPLE_AUTHORS, ERROR) \ |
c99ba492 JS |
60 | FUNC(TAG_OBJECT_NOT_TAG, ERROR) \ |
61 | FUNC(TREE_NOT_SORTED, ERROR) \ | |
62 | FUNC(UNKNOWN_TYPE, ERROR) \ | |
c99ba492 | 63 | FUNC(ZERO_PADDED_DATE, ERROR) \ |
159e7b08 JK |
64 | FUNC(GITMODULES_MISSING, ERROR) \ |
65 | FUNC(GITMODULES_BLOB, ERROR) \ | |
ed8b10f6 JK |
66 | FUNC(GITMODULES_PARSE, ERROR) \ |
67 | FUNC(GITMODULES_NAME, ERROR) \ | |
b7b1fca1 | 68 | FUNC(GITMODULES_SYMLINK, ERROR) \ |
a124133e | 69 | FUNC(GITMODULES_URL, ERROR) \ |
1a7fd1fb | 70 | FUNC(GITMODULES_PATH, ERROR) \ |
bb92255e | 71 | FUNC(GITMODULES_UPDATE, ERROR) \ |
c99ba492 JS |
72 | /* warnings */ \ |
73 | FUNC(BAD_FILEMODE, WARN) \ | |
c99ba492 JS |
74 | FUNC(EMPTY_NAME, WARN) \ |
75 | FUNC(FULL_PATHNAME, WARN) \ | |
76 | FUNC(HAS_DOT, WARN) \ | |
77 | FUNC(HAS_DOTDOT, WARN) \ | |
78 | FUNC(HAS_DOTGIT, WARN) \ | |
c99ba492 | 79 | FUNC(NULL_SHA1, WARN) \ |
f27d05b1 | 80 | FUNC(ZERO_PADDED_FILEMODE, WARN) \ |
6d2d780f | 81 | FUNC(NUL_IN_COMMIT, WARN) \ |
f27d05b1 JS |
82 | /* infos (reported as warnings, but ignored by default) */ \ |
83 | FUNC(BAD_TAG_NAME, INFO) \ | |
84 | FUNC(MISSING_TAGGER_ENTRY, INFO) | |
c99ba492 JS |
85 | |
86 | #define MSG_ID(id, msg_type) FSCK_MSG_##id, | |
87 | enum fsck_msg_id { | |
88 | FOREACH_MSG_ID(MSG_ID) | |
89 | FSCK_MSG_MAX | |
90 | }; | |
91 | #undef MSG_ID | |
92 | ||
f417eed8 JS |
93 | #define STR(x) #x |
94 | #define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type }, | |
c99ba492 | 95 | static struct { |
f417eed8 JS |
96 | const char *id_string; |
97 | const char *downcased; | |
c99ba492 JS |
98 | int msg_type; |
99 | } msg_id_info[FSCK_MSG_MAX + 1] = { | |
100 | FOREACH_MSG_ID(MSG_ID) | |
f417eed8 | 101 | { NULL, NULL, -1 } |
c99ba492 JS |
102 | }; |
103 | #undef MSG_ID | |
104 | ||
f417eed8 JS |
105 | static int parse_msg_id(const char *text) |
106 | { | |
107 | int i; | |
108 | ||
109 | if (!msg_id_info[0].downcased) { | |
110 | /* convert id_string to lower case, without underscores. */ | |
111 | for (i = 0; i < FSCK_MSG_MAX; i++) { | |
112 | const char *p = msg_id_info[i].id_string; | |
113 | int len = strlen(p); | |
114 | char *q = xmalloc(len); | |
115 | ||
116 | msg_id_info[i].downcased = q; | |
117 | while (*p) | |
118 | if (*p == '_') | |
119 | p++; | |
120 | else | |
121 | *(q)++ = tolower(*(p)++); | |
122 | *q = '\0'; | |
123 | } | |
124 | } | |
125 | ||
126 | for (i = 0; i < FSCK_MSG_MAX; i++) | |
127 | if (!strcmp(text, msg_id_info[i].downcased)) | |
128 | return i; | |
129 | ||
130 | return -1; | |
131 | } | |
132 | ||
c99ba492 JS |
133 | static int fsck_msg_type(enum fsck_msg_id msg_id, |
134 | struct fsck_options *options) | |
135 | { | |
136 | int msg_type; | |
137 | ||
0282f4dc JS |
138 | assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX); |
139 | ||
140 | if (options->msg_type) | |
141 | msg_type = options->msg_type[msg_id]; | |
142 | else { | |
143 | msg_type = msg_id_info[msg_id].msg_type; | |
144 | if (options->strict && msg_type == FSCK_WARN) | |
145 | msg_type = FSCK_ERROR; | |
146 | } | |
c99ba492 JS |
147 | |
148 | return msg_type; | |
149 | } | |
150 | ||
cd94c6f9 JS |
151 | static void init_skiplist(struct fsck_options *options, const char *path) |
152 | { | |
910650d2 | 153 | static struct oid_array skiplist = OID_ARRAY_INIT; |
cd94c6f9 | 154 | int sorted, fd; |
365c27fb | 155 | char buffer[GIT_MAX_HEXSZ + 1]; |
156 | struct object_id oid; | |
cd94c6f9 JS |
157 | |
158 | if (options->skiplist) | |
159 | sorted = options->skiplist->sorted; | |
160 | else { | |
161 | sorted = 1; | |
162 | options->skiplist = &skiplist; | |
163 | } | |
164 | ||
165 | fd = open(path, O_RDONLY); | |
166 | if (fd < 0) | |
167 | die("Could not open skip list: %s", path); | |
168 | for (;;) { | |
365c27fb | 169 | const char *p; |
cd94c6f9 JS |
170 | int result = read_in_full(fd, buffer, sizeof(buffer)); |
171 | if (result < 0) | |
172 | die_errno("Could not read '%s'", path); | |
173 | if (!result) | |
174 | break; | |
365c27fb | 175 | if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') |
cd94c6f9 | 176 | die("Invalid SHA-1: %s", buffer); |
910650d2 | 177 | oid_array_append(&skiplist, &oid); |
cd94c6f9 | 178 | if (sorted && skiplist.nr > 1 && |
ee3051bd | 179 | oidcmp(&skiplist.oid[skiplist.nr - 2], |
180 | &oid) > 0) | |
cd94c6f9 JS |
181 | sorted = 0; |
182 | } | |
183 | close(fd); | |
184 | ||
185 | if (sorted) | |
186 | skiplist.sorted = 1; | |
187 | } | |
188 | ||
0282f4dc JS |
189 | static int parse_msg_type(const char *str) |
190 | { | |
191 | if (!strcmp(str, "error")) | |
192 | return FSCK_ERROR; | |
193 | else if (!strcmp(str, "warn")) | |
194 | return FSCK_WARN; | |
efaba7cc JS |
195 | else if (!strcmp(str, "ignore")) |
196 | return FSCK_IGNORE; | |
0282f4dc JS |
197 | else |
198 | die("Unknown fsck message type: '%s'", str); | |
199 | } | |
200 | ||
5d477a33 JS |
201 | int is_valid_msg_type(const char *msg_id, const char *msg_type) |
202 | { | |
203 | if (parse_msg_id(msg_id) < 0) | |
204 | return 0; | |
205 | parse_msg_type(msg_type); | |
206 | return 1; | |
207 | } | |
208 | ||
0282f4dc JS |
209 | void fsck_set_msg_type(struct fsck_options *options, |
210 | const char *msg_id, const char *msg_type) | |
211 | { | |
212 | int id = parse_msg_id(msg_id), type; | |
213 | ||
214 | if (id < 0) | |
215 | die("Unhandled message id: %s", msg_id); | |
216 | type = parse_msg_type(msg_type); | |
217 | ||
f50c4407 JS |
218 | if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL) |
219 | die("Cannot demote %s to %s", msg_id, msg_type); | |
220 | ||
0282f4dc JS |
221 | if (!options->msg_type) { |
222 | int i; | |
b32fa95f JK |
223 | int *msg_type; |
224 | ALLOC_ARRAY(msg_type, FSCK_MSG_MAX); | |
0282f4dc JS |
225 | for (i = 0; i < FSCK_MSG_MAX; i++) |
226 | msg_type[i] = fsck_msg_type(i, options); | |
227 | options->msg_type = msg_type; | |
228 | } | |
229 | ||
230 | options->msg_type[id] = type; | |
231 | } | |
232 | ||
233 | void fsck_set_msg_types(struct fsck_options *options, const char *values) | |
234 | { | |
235 | char *buf = xstrdup(values), *to_free = buf; | |
236 | int done = 0; | |
237 | ||
238 | while (!done) { | |
239 | int len = strcspn(buf, " ,|"), equal; | |
240 | ||
241 | done = !buf[len]; | |
242 | if (!len) { | |
243 | buf++; | |
244 | continue; | |
245 | } | |
246 | buf[len] = '\0'; | |
247 | ||
248 | for (equal = 0; | |
249 | equal < len && buf[equal] != '=' && buf[equal] != ':'; | |
250 | equal++) | |
251 | buf[equal] = tolower(buf[equal]); | |
252 | buf[equal] = '\0'; | |
253 | ||
cd94c6f9 JS |
254 | if (!strcmp(buf, "skiplist")) { |
255 | if (equal == len) | |
256 | die("skiplist requires a path"); | |
257 | init_skiplist(options, buf + equal + 1); | |
258 | buf += len + 1; | |
259 | continue; | |
260 | } | |
261 | ||
0282f4dc JS |
262 | if (equal == len) |
263 | die("Missing '=': '%s'", buf); | |
355885d5 | 264 | |
0282f4dc JS |
265 | fsck_set_msg_type(options, buf, buf + equal + 1); |
266 | buf += len + 1; | |
267 | } | |
268 | free(to_free); | |
269 | } | |
270 | ||
71ab8fa8 JS |
271 | static void append_msg_id(struct strbuf *sb, const char *msg_id) |
272 | { | |
273 | for (;;) { | |
274 | char c = *(msg_id)++; | |
275 | ||
276 | if (!c) | |
277 | break; | |
278 | if (c != '_') | |
279 | strbuf_addch(sb, tolower(c)); | |
280 | else { | |
281 | assert(*msg_id); | |
282 | strbuf_addch(sb, *(msg_id)++); | |
283 | } | |
284 | } | |
285 | ||
286 | strbuf_addstr(sb, ": "); | |
287 | } | |
288 | ||
c99ba492 JS |
289 | __attribute__((format (printf, 4, 5))) |
290 | static int report(struct fsck_options *options, struct object *object, | |
291 | enum fsck_msg_id id, const char *fmt, ...) | |
292 | { | |
293 | va_list ap; | |
294 | struct strbuf sb = STRBUF_INIT; | |
295 | int msg_type = fsck_msg_type(id, options), result; | |
296 | ||
efaba7cc JS |
297 | if (msg_type == FSCK_IGNORE) |
298 | return 0; | |
299 | ||
cd94c6f9 | 300 | if (options->skiplist && object && |
910650d2 | 301 | oid_array_lookup(options->skiplist, &object->oid) >= 0) |
cd94c6f9 JS |
302 | return 0; |
303 | ||
f50c4407 JS |
304 | if (msg_type == FSCK_FATAL) |
305 | msg_type = FSCK_ERROR; | |
f27d05b1 JS |
306 | else if (msg_type == FSCK_INFO) |
307 | msg_type = FSCK_WARN; | |
f50c4407 | 308 | |
71ab8fa8 JS |
309 | append_msg_id(&sb, msg_id_info[id].id_string); |
310 | ||
c99ba492 JS |
311 | va_start(ap, fmt); |
312 | strbuf_vaddf(&sb, fmt, ap); | |
1cd772cc | 313 | result = options->error_func(options, object, msg_type, sb.buf); |
c99ba492 JS |
314 | strbuf_release(&sb); |
315 | va_end(ap); | |
316 | ||
317 | return result; | |
318 | } | |
319 | ||
7b35efd7 JS |
320 | static char *get_object_name(struct fsck_options *options, struct object *obj) |
321 | { | |
322 | if (!options->object_names) | |
323 | return NULL; | |
324 | return lookup_decoration(options->object_names, obj); | |
325 | } | |
326 | ||
327 | static void put_object_name(struct fsck_options *options, struct object *obj, | |
328 | const char *fmt, ...) | |
329 | { | |
330 | va_list ap; | |
331 | struct strbuf buf = STRBUF_INIT; | |
332 | char *existing; | |
333 | ||
334 | if (!options->object_names) | |
335 | return; | |
336 | existing = lookup_decoration(options->object_names, obj); | |
337 | if (existing) | |
338 | return; | |
339 | va_start(ap, fmt); | |
340 | strbuf_vaddf(&buf, fmt, ap); | |
341 | add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL)); | |
342 | va_end(ap); | |
343 | } | |
344 | ||
90cf590f JS |
345 | static const char *describe_object(struct fsck_options *o, struct object *obj) |
346 | { | |
347 | static struct strbuf buf = STRBUF_INIT; | |
348 | char *name; | |
349 | ||
350 | strbuf_reset(&buf); | |
351 | strbuf_addstr(&buf, oid_to_hex(&obj->oid)); | |
352 | if (o->object_names && (name = lookup_decoration(o->object_names, obj))) | |
353 | strbuf_addf(&buf, " (%s)", name); | |
354 | ||
355 | return buf.buf; | |
356 | } | |
357 | ||
22410549 | 358 | static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options) |
355885d5 MK |
359 | { |
360 | struct tree_desc desc; | |
361 | struct name_entry entry; | |
362 | int res = 0; | |
7b35efd7 | 363 | const char *name; |
355885d5 MK |
364 | |
365 | if (parse_tree(tree)) | |
366 | return -1; | |
367 | ||
7b35efd7 | 368 | name = get_object_name(options, &tree->object); |
8354fa3d DT |
369 | if (init_tree_desc_gently(&desc, tree->buffer, tree->size)) |
370 | return -1; | |
371 | while (tree_entry_gently(&desc, &entry)) { | |
7b35efd7 | 372 | struct object *obj; |
355885d5 MK |
373 | int result; |
374 | ||
375 | if (S_ISGITLINK(entry.mode)) | |
376 | continue; | |
7b35efd7 JS |
377 | |
378 | if (S_ISDIR(entry.mode)) { | |
2720f6db RS |
379 | obj = (struct object *)lookup_tree(entry.oid); |
380 | if (name && obj) | |
7b35efd7 JS |
381 | put_object_name(options, obj, "%s%s/", name, |
382 | entry.path); | |
383 | result = options->walk(obj, OBJ_TREE, data, options); | |
384 | } | |
385 | else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) { | |
2720f6db RS |
386 | obj = (struct object *)lookup_blob(entry.oid); |
387 | if (name && obj) | |
7b35efd7 JS |
388 | put_object_name(options, obj, "%s%s", name, |
389 | entry.path); | |
390 | result = options->walk(obj, OBJ_BLOB, data, options); | |
391 | } | |
355885d5 | 392 | else { |
82247e9b | 393 | result = error("in tree %s: entry %s has bad mode %.6o", |
90cf590f | 394 | describe_object(options, &tree->object), entry.path, entry.mode); |
355885d5 MK |
395 | } |
396 | if (result < 0) | |
397 | return result; | |
398 | if (!res) | |
399 | res = result; | |
400 | } | |
401 | return res; | |
402 | } | |
403 | ||
22410549 | 404 | static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options) |
355885d5 | 405 | { |
7b35efd7 | 406 | int counter = 0, generation = 0, name_prefix_len = 0; |
355885d5 MK |
407 | struct commit_list *parents; |
408 | int res; | |
409 | int result; | |
7b35efd7 | 410 | const char *name; |
355885d5 MK |
411 | |
412 | if (parse_commit(commit)) | |
413 | return -1; | |
414 | ||
7b35efd7 JS |
415 | name = get_object_name(options, &commit->object); |
416 | if (name) | |
417 | put_object_name(options, &commit->tree->object, "%s:", name); | |
418 | ||
22410549 | 419 | result = options->walk((struct object *)commit->tree, OBJ_TREE, data, options); |
355885d5 MK |
420 | if (result < 0) |
421 | return result; | |
422 | res = result; | |
423 | ||
424 | parents = commit->parents; | |
7b35efd7 JS |
425 | if (name && parents) { |
426 | int len = strlen(name), power; | |
427 | ||
428 | if (len && name[len - 1] == '^') { | |
429 | generation = 1; | |
430 | name_prefix_len = len - 1; | |
431 | } | |
432 | else { /* parse ~<generation> suffix */ | |
433 | for (generation = 0, power = 1; | |
434 | len && isdigit(name[len - 1]); | |
435 | power *= 10) | |
436 | generation += power * (name[--len] - '0'); | |
437 | if (power > 1 && len && name[len - 1] == '~') | |
438 | name_prefix_len = len - 1; | |
439 | } | |
440 | } | |
441 | ||
355885d5 | 442 | while (parents) { |
7b35efd7 JS |
443 | if (name) { |
444 | struct object *obj = &parents->item->object; | |
445 | ||
446 | if (++counter > 1) | |
447 | put_object_name(options, obj, "%s^%d", | |
448 | name, counter); | |
449 | else if (generation > 0) | |
450 | put_object_name(options, obj, "%.*s~%d", | |
451 | name_prefix_len, name, generation + 1); | |
452 | else | |
453 | put_object_name(options, obj, "%s^", name); | |
454 | } | |
22410549 | 455 | result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options); |
355885d5 MK |
456 | if (result < 0) |
457 | return result; | |
458 | if (!res) | |
459 | res = result; | |
460 | parents = parents->next; | |
461 | } | |
462 | return res; | |
463 | } | |
464 | ||
22410549 | 465 | static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options) |
355885d5 | 466 | { |
7b35efd7 JS |
467 | char *name = get_object_name(options, &tag->object); |
468 | ||
355885d5 MK |
469 | if (parse_tag(tag)) |
470 | return -1; | |
7b35efd7 JS |
471 | if (name) |
472 | put_object_name(options, tag->tagged, "%s", name); | |
22410549 | 473 | return options->walk(tag->tagged, OBJ_ANY, data, options); |
355885d5 MK |
474 | } |
475 | ||
22410549 | 476 | int fsck_walk(struct object *obj, void *data, struct fsck_options *options) |
355885d5 MK |
477 | { |
478 | if (!obj) | |
479 | return -1; | |
a2b22854 JK |
480 | |
481 | if (obj->type == OBJ_NONE) | |
c251c83d | 482 | parse_object(&obj->oid); |
a2b22854 | 483 | |
355885d5 MK |
484 | switch (obj->type) { |
485 | case OBJ_BLOB: | |
486 | return 0; | |
487 | case OBJ_TREE: | |
22410549 | 488 | return fsck_walk_tree((struct tree *)obj, data, options); |
355885d5 | 489 | case OBJ_COMMIT: |
22410549 | 490 | return fsck_walk_commit((struct commit *)obj, data, options); |
355885d5 | 491 | case OBJ_TAG: |
22410549 | 492 | return fsck_walk_tag((struct tag *)obj, data, options); |
355885d5 | 493 | default: |
90cf590f | 494 | error("Unknown object type for %s", describe_object(options, obj)); |
355885d5 MK |
495 | return -1; |
496 | } | |
497 | } | |
ba002f3b MK |
498 | |
499 | /* | |
500 | * The entries in a tree are ordered in the _path_ order, | |
501 | * which means that a directory entry is ordered by adding | |
502 | * a slash to the end of it. | |
503 | * | |
504 | * So a directory called "a" is ordered _after_ a file | |
505 | * called "a.c", because "a/" sorts after "a.c". | |
506 | */ | |
507 | #define TREE_UNORDERED (-1) | |
508 | #define TREE_HAS_DUPS (-2) | |
509 | ||
510 | static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2) | |
511 | { | |
512 | int len1 = strlen(name1); | |
513 | int len2 = strlen(name2); | |
514 | int len = len1 < len2 ? len1 : len2; | |
515 | unsigned char c1, c2; | |
516 | int cmp; | |
517 | ||
518 | cmp = memcmp(name1, name2, len); | |
519 | if (cmp < 0) | |
520 | return 0; | |
521 | if (cmp > 0) | |
522 | return TREE_UNORDERED; | |
523 | ||
524 | /* | |
525 | * Ok, the first <len> characters are the same. | |
526 | * Now we need to order the next one, but turn | |
527 | * a '\0' into a '/' for a directory entry. | |
528 | */ | |
529 | c1 = name1[len]; | |
530 | c2 = name2[len]; | |
531 | if (!c1 && !c2) | |
532 | /* | |
533 | * git-write-tree used to write out a nonsense tree that has | |
534 | * entries with the same name, one blob and one tree. Make | |
535 | * sure we do not have duplicate entries. | |
536 | */ | |
537 | return TREE_HAS_DUPS; | |
538 | if (!c1 && S_ISDIR(mode1)) | |
539 | c1 = '/'; | |
540 | if (!c2 && S_ISDIR(mode2)) | |
541 | c2 = '/'; | |
542 | return c1 < c2 ? 0 : TREE_UNORDERED; | |
543 | } | |
544 | ||
22410549 | 545 | static int fsck_tree(struct tree *item, struct fsck_options *options) |
ba002f3b | 546 | { |
8354fa3d | 547 | int retval = 0; |
c479d14a | 548 | int has_null_sha1 = 0; |
ba002f3b MK |
549 | int has_full_path = 0; |
550 | int has_empty_name = 0; | |
5d34a435 JK |
551 | int has_dot = 0; |
552 | int has_dotdot = 0; | |
5c17f512 | 553 | int has_dotgit = 0; |
ba002f3b MK |
554 | int has_zero_pad = 0; |
555 | int has_bad_modes = 0; | |
556 | int has_dup_entries = 0; | |
557 | int not_properly_sorted = 0; | |
558 | struct tree_desc desc; | |
559 | unsigned o_mode; | |
560 | const char *o_name; | |
ba002f3b | 561 | |
8354fa3d DT |
562 | if (init_tree_desc_gently(&desc, item->buffer, item->size)) { |
563 | retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree"); | |
564 | return retval; | |
565 | } | |
ba002f3b MK |
566 | |
567 | o_mode = 0; | |
568 | o_name = NULL; | |
ba002f3b MK |
569 | |
570 | while (desc.size) { | |
571 | unsigned mode; | |
288a74bc | 572 | const char *name, *backslash; |
ce6663a9 | 573 | const struct object_id *oid; |
ba002f3b | 574 | |
ce6663a9 | 575 | oid = tree_entry_extract(&desc, &name, &mode); |
ba002f3b | 576 | |
ce6663a9 | 577 | has_null_sha1 |= is_null_oid(oid); |
effd12ec HS |
578 | has_full_path |= !!strchr(name, '/'); |
579 | has_empty_name |= !*name; | |
580 | has_dot |= !strcmp(name, "."); | |
581 | has_dotdot |= !strcmp(name, ".."); | |
ed9c3220 | 582 | has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name); |
ba002f3b | 583 | has_zero_pad |= *(char *)desc.buffer == '0'; |
159e7b08 | 584 | |
b7b1fca1 JK |
585 | if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) { |
586 | if (!S_ISLNK(mode)) | |
587 | oidset_insert(&gitmodules_found, oid); | |
588 | else | |
589 | retval += report(options, &item->object, | |
590 | FSCK_MSG_GITMODULES_SYMLINK, | |
591 | ".gitmodules is a symbolic link"); | |
592 | } | |
159e7b08 | 593 | |
288a74bc JS |
594 | if ((backslash = strchr(name, '\\'))) { |
595 | while (backslash) { | |
596 | backslash++; | |
597 | has_dotgit |= is_ntfs_dotgit(backslash); | |
bdfef049 JS |
598 | if (is_ntfs_dotgitmodules(backslash)) { |
599 | if (!S_ISLNK(mode)) | |
600 | oidset_insert(&gitmodules_found, oid); | |
601 | else | |
602 | retval += report(options, &item->object, | |
603 | FSCK_MSG_GITMODULES_SYMLINK, | |
604 | ".gitmodules is a symbolic link"); | |
605 | } | |
288a74bc JS |
606 | backslash = strchr(backslash, '\\'); |
607 | } | |
608 | } | |
609 | ||
8354fa3d DT |
610 | if (update_tree_entry_gently(&desc)) { |
611 | retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree"); | |
612 | break; | |
613 | } | |
ba002f3b MK |
614 | |
615 | switch (mode) { | |
616 | /* | |
617 | * Standard modes.. | |
618 | */ | |
619 | case S_IFREG | 0755: | |
620 | case S_IFREG | 0644: | |
621 | case S_IFLNK: | |
622 | case S_IFDIR: | |
623 | case S_IFGITLINK: | |
624 | break; | |
625 | /* | |
626 | * This is nonstandard, but we had a few of these | |
627 | * early on when we honored the full set of mode | |
628 | * bits.. | |
629 | */ | |
630 | case S_IFREG | 0664: | |
22410549 | 631 | if (!options->strict) |
ba002f3b | 632 | break; |
1cf01a34 | 633 | /* fallthrough */ |
ba002f3b MK |
634 | default: |
635 | has_bad_modes = 1; | |
636 | } | |
637 | ||
638 | if (o_name) { | |
639 | switch (verify_ordered(o_mode, o_name, mode, name)) { | |
640 | case TREE_UNORDERED: | |
641 | not_properly_sorted = 1; | |
642 | break; | |
643 | case TREE_HAS_DUPS: | |
644 | has_dup_entries = 1; | |
645 | break; | |
646 | default: | |
647 | break; | |
648 | } | |
649 | } | |
650 | ||
651 | o_mode = mode; | |
652 | o_name = name; | |
ba002f3b MK |
653 | } |
654 | ||
c479d14a | 655 | if (has_null_sha1) |
c99ba492 | 656 | retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1"); |
ba002f3b | 657 | if (has_full_path) |
c99ba492 | 658 | retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames"); |
ba002f3b | 659 | if (has_empty_name) |
c99ba492 | 660 | retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname"); |
5d34a435 | 661 | if (has_dot) |
c99ba492 | 662 | retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'"); |
5d34a435 | 663 | if (has_dotdot) |
c99ba492 | 664 | retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'"); |
5c17f512 | 665 | if (has_dotgit) |
c99ba492 | 666 | retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'"); |
ba002f3b | 667 | if (has_zero_pad) |
c99ba492 | 668 | retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes"); |
ba002f3b | 669 | if (has_bad_modes) |
c99ba492 | 670 | retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes"); |
ba002f3b | 671 | if (has_dup_entries) |
c99ba492 | 672 | retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries"); |
ba002f3b | 673 | if (not_properly_sorted) |
c99ba492 | 674 | retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted"); |
ba002f3b MK |
675 | return retval; |
676 | } | |
677 | ||
84d18c0b | 678 | static int verify_headers(const void *data, unsigned long size, |
b2f44feb | 679 | struct object *obj, struct fsck_options *options) |
4d0d8975 JS |
680 | { |
681 | const char *buffer = (const char *)data; | |
682 | unsigned long i; | |
683 | ||
684 | for (i = 0; i < size; i++) { | |
685 | switch (buffer[i]) { | |
686 | case '\0': | |
c99ba492 JS |
687 | return report(options, obj, |
688 | FSCK_MSG_NUL_IN_HEADER, | |
689 | "unterminated header: NUL at offset %ld", i); | |
4d0d8975 JS |
690 | case '\n': |
691 | if (i + 1 < size && buffer[i + 1] == '\n') | |
692 | return 0; | |
693 | } | |
694 | } | |
695 | ||
84d18c0b JH |
696 | /* |
697 | * We did not find double-LF that separates the header | |
698 | * and the body. Not having a body is not a crime but | |
699 | * we do want to see the terminating LF for the last header | |
700 | * line. | |
701 | */ | |
702 | if (size && buffer[size - 1] == '\n') | |
703 | return 0; | |
704 | ||
c99ba492 JS |
705 | return report(options, obj, |
706 | FSCK_MSG_UNTERMINATED_HEADER, "unterminated header"); | |
4d0d8975 JS |
707 | } |
708 | ||
22410549 | 709 | static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options) |
daae1922 | 710 | { |
e6826e33 | 711 | const char *p = *ident; |
d4b8de04 JK |
712 | char *end; |
713 | ||
e6826e33 JS |
714 | *ident = strchrnul(*ident, '\n'); |
715 | if (**ident == '\n') | |
716 | (*ident)++; | |
717 | ||
718 | if (*p == '<') | |
c99ba492 | 719 | return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email"); |
e6826e33 JS |
720 | p += strcspn(p, "<>\n"); |
721 | if (*p == '>') | |
c99ba492 | 722 | return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name"); |
e6826e33 | 723 | if (*p != '<') |
c99ba492 | 724 | return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email"); |
e6826e33 | 725 | if (p[-1] != ' ') |
c99ba492 | 726 | return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email"); |
e6826e33 JS |
727 | p++; |
728 | p += strcspn(p, "<>\n"); | |
729 | if (*p != '>') | |
c99ba492 | 730 | return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email"); |
e6826e33 JS |
731 | p++; |
732 | if (*p != ' ') | |
c99ba492 | 733 | return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date"); |
e6826e33 JS |
734 | p++; |
735 | if (*p == '0' && p[1] != ' ') | |
c99ba492 | 736 | return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date"); |
1aeb7e75 | 737 | if (date_overflows(parse_timestamp(p, &end, 10))) |
c99ba492 | 738 | return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow"); |
e6826e33 | 739 | if ((end == p || *end != ' ')) |
c99ba492 | 740 | return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date"); |
e6826e33 JS |
741 | p = end + 1; |
742 | if ((*p != '+' && *p != '-') || | |
743 | !isdigit(p[1]) || | |
744 | !isdigit(p[2]) || | |
745 | !isdigit(p[3]) || | |
746 | !isdigit(p[4]) || | |
747 | (p[5] != '\n')) | |
c99ba492 | 748 | return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone"); |
e6826e33 | 749 | p += 6; |
daae1922 JN |
750 | return 0; |
751 | } | |
752 | ||
bc6b8fc1 | 753 | static int fsck_commit_buffer(struct commit *commit, const char *buffer, |
22410549 | 754 | unsigned long size, struct fsck_options *options) |
ba002f3b | 755 | { |
ba002f3b MK |
756 | unsigned char tree_sha1[20], sha1[20]; |
757 | struct commit_graft *graft; | |
c9ad147f | 758 | unsigned parent_count, parent_line_count = 0, author_count; |
daae1922 | 759 | int err; |
6d2d780f | 760 | const char *buffer_begin = buffer; |
ba002f3b | 761 | |
b2f44feb | 762 | if (verify_headers(buffer, size, &commit->object, options)) |
4d0d8975 JS |
763 | return -1; |
764 | ||
cf4fff57 | 765 | if (!skip_prefix(buffer, "tree ", &buffer)) |
c99ba492 | 766 | return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line"); |
b3584761 JS |
767 | if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n') { |
768 | err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1"); | |
769 | if (err) | |
770 | return err; | |
771 | } | |
2d820a61 | 772 | buffer += 41; |
cf4fff57 | 773 | while (skip_prefix(buffer, "parent ", &buffer)) { |
b3584761 JS |
774 | if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') { |
775 | err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1"); | |
776 | if (err) | |
777 | return err; | |
778 | } | |
2d820a61 | 779 | buffer += 41; |
9d02150c | 780 | parent_line_count++; |
ba002f3b | 781 | } |
8b65a34c | 782 | graft = lookup_commit_graft(&commit->object.oid); |
9d02150c | 783 | parent_count = commit_list_count(commit->parents); |
ba002f3b | 784 | if (graft) { |
9d02150c | 785 | if (graft->nr_parent == -1 && !parent_count) |
ba002f3b | 786 | ; /* shallow commit */ |
b3584761 JS |
787 | else if (graft->nr_parent != parent_count) { |
788 | err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing"); | |
789 | if (err) | |
790 | return err; | |
791 | } | |
ba002f3b | 792 | } else { |
b3584761 JS |
793 | if (parent_count != parent_line_count) { |
794 | err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing"); | |
795 | if (err) | |
796 | return err; | |
797 | } | |
ba002f3b | 798 | } |
c9ad147f JS |
799 | author_count = 0; |
800 | while (skip_prefix(buffer, "author ", &buffer)) { | |
801 | author_count++; | |
802 | err = fsck_ident(&buffer, &commit->object, options); | |
803 | if (err) | |
804 | return err; | |
ba002f3b | 805 | } |
c9ad147f JS |
806 | if (author_count < 1) |
807 | err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line"); | |
808 | else if (author_count > 1) | |
809 | err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines"); | |
daae1922 JN |
810 | if (err) |
811 | return err; | |
cf4fff57 | 812 | if (!skip_prefix(buffer, "committer ", &buffer)) |
c99ba492 | 813 | return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line"); |
22410549 | 814 | err = fsck_ident(&buffer, &commit->object, options); |
daae1922 JN |
815 | if (err) |
816 | return err; | |
5af29718 JH |
817 | if (!commit->tree) { |
818 | err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", sha1_to_hex(tree_sha1)); | |
819 | if (err) | |
820 | return err; | |
821 | } | |
6d2d780f JH |
822 | if (memchr(buffer_begin, '\0', size)) { |
823 | err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT, | |
824 | "NUL byte in the commit object body"); | |
825 | if (err) | |
826 | return err; | |
827 | } | |
ba002f3b MK |
828 | return 0; |
829 | } | |
830 | ||
90a398bb | 831 | static int fsck_commit(struct commit *commit, const char *data, |
22410549 | 832 | unsigned long size, struct fsck_options *options) |
bc6b8fc1 | 833 | { |
90a398bb | 834 | const char *buffer = data ? data : get_commit_buffer(commit, &size); |
22410549 | 835 | int ret = fsck_commit_buffer(commit, buffer, size, options); |
90a398bb JS |
836 | if (!data) |
837 | unuse_commit_buffer(commit, buffer); | |
bc6b8fc1 JK |
838 | return ret; |
839 | } | |
840 | ||
cec097be | 841 | static int fsck_tag_buffer(struct tag *tag, const char *data, |
22410549 | 842 | unsigned long size, struct fsck_options *options) |
cec097be JS |
843 | { |
844 | unsigned char sha1[20]; | |
845 | int ret = 0; | |
846 | const char *buffer; | |
847 | char *to_free = NULL, *eol; | |
848 | struct strbuf sb = STRBUF_INIT; | |
849 | ||
850 | if (data) | |
851 | buffer = data; | |
852 | else { | |
853 | enum object_type type; | |
854 | ||
855 | buffer = to_free = | |
ed1c9977 | 856 | read_sha1_file(tag->object.oid.hash, &type, &size); |
cec097be | 857 | if (!buffer) |
c99ba492 JS |
858 | return report(options, &tag->object, |
859 | FSCK_MSG_MISSING_TAG_OBJECT, | |
cec097be JS |
860 | "cannot read tag object"); |
861 | ||
862 | if (type != OBJ_TAG) { | |
c99ba492 JS |
863 | ret = report(options, &tag->object, |
864 | FSCK_MSG_TAG_OBJECT_NOT_TAG, | |
cec097be | 865 | "expected tag got %s", |
debca9d2 | 866 | type_name(type)); |
cec097be JS |
867 | goto done; |
868 | } | |
869 | } | |
870 | ||
8a272f29 RS |
871 | ret = verify_headers(buffer, size, &tag->object, options); |
872 | if (ret) | |
cec097be JS |
873 | goto done; |
874 | ||
875 | if (!skip_prefix(buffer, "object ", &buffer)) { | |
c99ba492 | 876 | ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line"); |
cec097be JS |
877 | goto done; |
878 | } | |
879 | if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') { | |
c99ba492 | 880 | ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1"); |
7d7d5b05 JS |
881 | if (ret) |
882 | goto done; | |
cec097be JS |
883 | } |
884 | buffer += 41; | |
885 | ||
886 | if (!skip_prefix(buffer, "type ", &buffer)) { | |
c99ba492 | 887 | ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line"); |
cec097be JS |
888 | goto done; |
889 | } | |
890 | eol = strchr(buffer, '\n'); | |
891 | if (!eol) { | |
c99ba492 | 892 | ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line"); |
cec097be JS |
893 | goto done; |
894 | } | |
895 | if (type_from_string_gently(buffer, eol - buffer, 1) < 0) | |
c99ba492 | 896 | ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value"); |
cec097be JS |
897 | if (ret) |
898 | goto done; | |
899 | buffer = eol + 1; | |
900 | ||
901 | if (!skip_prefix(buffer, "tag ", &buffer)) { | |
c99ba492 | 902 | ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line"); |
cec097be JS |
903 | goto done; |
904 | } | |
905 | eol = strchr(buffer, '\n'); | |
906 | if (!eol) { | |
c99ba492 | 907 | ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line"); |
cec097be JS |
908 | goto done; |
909 | } | |
910 | strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer); | |
f27d05b1 JS |
911 | if (check_refname_format(sb.buf, 0)) { |
912 | ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME, | |
c99ba492 | 913 | "invalid 'tag' name: %.*s", |
7add4419 | 914 | (int)(eol - buffer), buffer); |
f27d05b1 JS |
915 | if (ret) |
916 | goto done; | |
917 | } | |
cec097be JS |
918 | buffer = eol + 1; |
919 | ||
f27d05b1 | 920 | if (!skip_prefix(buffer, "tagger ", &buffer)) { |
cec097be | 921 | /* early tags do not contain 'tagger' lines; warn only */ |
f27d05b1 JS |
922 | ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line"); |
923 | if (ret) | |
924 | goto done; | |
925 | } | |
cec097be | 926 | else |
22410549 | 927 | ret = fsck_ident(&buffer, &tag->object, options); |
cec097be JS |
928 | |
929 | done: | |
930 | strbuf_release(&sb); | |
931 | free(to_free); | |
932 | return ret; | |
933 | } | |
934 | ||
90a398bb | 935 | static int fsck_tag(struct tag *tag, const char *data, |
22410549 | 936 | unsigned long size, struct fsck_options *options) |
ba002f3b MK |
937 | { |
938 | struct object *tagged = tag->tagged; | |
939 | ||
940 | if (!tagged) | |
c99ba492 | 941 | return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object"); |
cec097be | 942 | |
22410549 | 943 | return fsck_tag_buffer(tag, data, size, options); |
ba002f3b MK |
944 | } |
945 | ||
a2b26ffb JN |
946 | /* |
947 | * Like builtin/submodule--helper.c's starts_with_dot_slash, but without | |
948 | * relying on the platform-dependent is_dir_sep helper. | |
949 | * | |
950 | * This is for use in checking whether a submodule URL is interpreted as | |
951 | * relative to the current directory on any platform, since \ is a | |
952 | * directory separator on Windows but not on other platforms. | |
953 | */ | |
954 | static int starts_with_dot_slash(const char *str) | |
955 | { | |
956 | return str[0] == '.' && (str[1] == '/' || str[1] == '\\'); | |
957 | } | |
958 | ||
959 | /* | |
960 | * Like starts_with_dot_slash, this is a variant of submodule--helper's | |
961 | * helper of the same name with the twist that it accepts backslash as a | |
962 | * directory separator even on non-Windows platforms. | |
963 | */ | |
964 | static int starts_with_dot_dot_slash(const char *str) | |
965 | { | |
966 | return str[0] == '.' && starts_with_dot_slash(str + 1); | |
967 | } | |
968 | ||
969 | static int submodule_url_is_relative(const char *url) | |
970 | { | |
971 | return starts_with_dot_slash(url) || starts_with_dot_dot_slash(url); | |
972 | } | |
973 | ||
974 | /* | |
975 | * Check whether a transport is implemented by git-remote-curl. | |
976 | * | |
977 | * If it is, returns 1 and writes the URL that would be passed to | |
978 | * git-remote-curl to the "out" parameter. | |
979 | * | |
980 | * Otherwise, returns 0 and leaves "out" untouched. | |
981 | * | |
982 | * Examples: | |
983 | * http::https://example.com/repo.git -> 1, https://example.com/repo.git | |
984 | * https://example.com/repo.git -> 1, https://example.com/repo.git | |
985 | * git://example.com/repo.git -> 0 | |
986 | * | |
987 | * This is for use in checking for previously exploitable bugs that | |
988 | * required a submodule URL to be passed to git-remote-curl. | |
989 | */ | |
990 | static int url_to_curl_url(const char *url, const char **out) | |
991 | { | |
992 | /* | |
993 | * We don't need to check for case-aliases, "http.exe", and so | |
994 | * on because in the default configuration, is_transport_allowed | |
995 | * prevents URLs with those schemes from being cloned | |
996 | * automatically. | |
997 | */ | |
998 | if (skip_prefix(url, "http::", out) || | |
999 | skip_prefix(url, "https::", out) || | |
1000 | skip_prefix(url, "ftp::", out) || | |
1001 | skip_prefix(url, "ftps::", out)) | |
1002 | return 1; | |
1003 | if (starts_with(url, "http://") || | |
1004 | starts_with(url, "https://") || | |
1005 | starts_with(url, "ftp://") || | |
1006 | starts_with(url, "ftps://")) { | |
1007 | *out = url; | |
1008 | return 1; | |
1009 | } | |
1010 | return 0; | |
1011 | } | |
1012 | ||
07259e74 JK |
1013 | static int check_submodule_url(const char *url) |
1014 | { | |
a2b26ffb | 1015 | const char *curl_url; |
07259e74 JK |
1016 | |
1017 | if (looks_like_command_line_option(url)) | |
1018 | return -1; | |
1019 | ||
a2b26ffb JN |
1020 | if (submodule_url_is_relative(url)) { |
1021 | /* | |
1022 | * This could be appended to an http URL and url-decoded; | |
1023 | * check for malicious characters. | |
1024 | */ | |
1025 | char *decoded = url_decode(url); | |
1026 | int has_nl = !!strchr(decoded, '\n'); | |
1027 | free(decoded); | |
1028 | if (has_nl) | |
1029 | return -1; | |
1030 | } | |
1031 | ||
1032 | else if (url_to_curl_url(url, &curl_url)) { | |
1033 | struct credential c = CREDENTIAL_INIT; | |
1034 | int ret = credential_from_url_gently(&c, curl_url, 1); | |
1035 | credential_clear(&c); | |
1036 | return ret; | |
1037 | } | |
1038 | ||
1039 | return 0; | |
07259e74 JK |
1040 | } |
1041 | ||
ed8b10f6 JK |
1042 | struct fsck_gitmodules_data { |
1043 | struct object *obj; | |
1044 | struct fsck_options *options; | |
1045 | int ret; | |
1046 | }; | |
1047 | ||
1048 | static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata) | |
1049 | { | |
1050 | struct fsck_gitmodules_data *data = vdata; | |
1051 | const char *subsection, *key; | |
1052 | int subsection_len; | |
1053 | char *name; | |
1054 | ||
1055 | if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 || | |
1056 | !subsection) | |
1057 | return 0; | |
1058 | ||
1059 | name = xmemdupz(subsection, subsection_len); | |
1060 | if (check_submodule_name(name) < 0) | |
1061 | data->ret |= report(data->options, data->obj, | |
1062 | FSCK_MSG_GITMODULES_NAME, | |
1063 | "disallowed submodule name: %s", | |
1064 | name); | |
a124133e | 1065 | if (!strcmp(key, "url") && value && |
07259e74 | 1066 | check_submodule_url(value) < 0) |
a124133e JK |
1067 | data->ret |= report(data->options, data->obj, |
1068 | FSCK_MSG_GITMODULES_URL, | |
1069 | "disallowed submodule url: %s", | |
1070 | value); | |
1a7fd1fb JK |
1071 | if (!strcmp(key, "path") && value && |
1072 | looks_like_command_line_option(value)) | |
1073 | data->ret |= report(data->options, data->obj, | |
1074 | FSCK_MSG_GITMODULES_PATH, | |
1075 | "disallowed submodule path: %s", | |
1076 | value); | |
bb92255e JN |
1077 | if (!strcmp(key, "update") && value && |
1078 | parse_submodule_update_type(value) == SM_UPDATE_COMMAND) | |
1079 | data->ret |= report(data->options, data->obj, | |
1080 | FSCK_MSG_GITMODULES_UPDATE, | |
1081 | "disallowed submodule update setting: %s", | |
1082 | value); | |
ed8b10f6 JK |
1083 | free(name); |
1084 | ||
1085 | return 0; | |
1086 | } | |
1087 | ||
7ac4f3a0 JK |
1088 | static int fsck_blob(struct blob *blob, const char *buf, |
1089 | unsigned long size, struct fsck_options *options) | |
1090 | { | |
ed8b10f6 JK |
1091 | struct fsck_gitmodules_data data; |
1092 | ||
1093 | if (!oidset_contains(&gitmodules_found, &blob->object.oid)) | |
1094 | return 0; | |
1095 | oidset_insert(&gitmodules_done, &blob->object.oid); | |
1096 | ||
1097 | if (!buf) { | |
1098 | /* | |
1099 | * A missing buffer here is a sign that the caller found the | |
1100 | * blob too gigantic to load into memory. Let's just consider | |
1101 | * that an error. | |
1102 | */ | |
1103 | return report(options, &blob->object, | |
1104 | FSCK_MSG_GITMODULES_PARSE, | |
1105 | ".gitmodules too large to parse"); | |
1106 | } | |
1107 | ||
1108 | data.obj = &blob->object; | |
1109 | data.options = options; | |
1110 | data.ret = 0; | |
1111 | if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB, | |
1112 | ".gitmodules", buf, size, &data)) | |
1113 | data.ret |= report(options, &blob->object, | |
1114 | FSCK_MSG_GITMODULES_PARSE, | |
1115 | "could not parse gitmodules blob"); | |
1116 | ||
1117 | return data.ret; | |
7ac4f3a0 JK |
1118 | } |
1119 | ||
90a398bb | 1120 | int fsck_object(struct object *obj, void *data, unsigned long size, |
22410549 | 1121 | struct fsck_options *options) |
ba002f3b MK |
1122 | { |
1123 | if (!obj) | |
c99ba492 | 1124 | return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck"); |
ba002f3b MK |
1125 | |
1126 | if (obj->type == OBJ_BLOB) | |
7ac4f3a0 | 1127 | return fsck_blob((struct blob *)obj, data, size, options); |
ba002f3b | 1128 | if (obj->type == OBJ_TREE) |
22410549 | 1129 | return fsck_tree((struct tree *) obj, options); |
ba002f3b | 1130 | if (obj->type == OBJ_COMMIT) |
90a398bb | 1131 | return fsck_commit((struct commit *) obj, (const char *) data, |
22410549 | 1132 | size, options); |
ba002f3b | 1133 | if (obj->type == OBJ_TAG) |
90a398bb | 1134 | return fsck_tag((struct tag *) obj, (const char *) data, |
22410549 | 1135 | size, options); |
ba002f3b | 1136 | |
c99ba492 | 1137 | return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)", |
ba002f3b MK |
1138 | obj->type); |
1139 | } | |
d6ffc8d7 | 1140 | |
1cd772cc JS |
1141 | int fsck_error_function(struct fsck_options *o, |
1142 | struct object *obj, int msg_type, const char *message) | |
d6ffc8d7 | 1143 | { |
0282f4dc | 1144 | if (msg_type == FSCK_WARN) { |
90cf590f | 1145 | warning("object %s: %s", describe_object(o, obj), message); |
0282f4dc JS |
1146 | return 0; |
1147 | } | |
90cf590f | 1148 | error("object %s: %s", describe_object(o, obj), message); |
d6ffc8d7 MK |
1149 | return 1; |
1150 | } | |
159e7b08 JK |
1151 | |
1152 | int fsck_finish(struct fsck_options *options) | |
1153 | { | |
1154 | int ret = 0; | |
1155 | struct oidset_iter iter; | |
1156 | const struct object_id *oid; | |
1157 | ||
1158 | oidset_iter_init(&gitmodules_found, &iter); | |
1159 | while ((oid = oidset_iter_next(&iter))) { | |
1160 | struct blob *blob; | |
1161 | enum object_type type; | |
1162 | unsigned long size; | |
1163 | char *buf; | |
1164 | ||
1165 | if (oidset_contains(&gitmodules_done, oid)) | |
1166 | continue; | |
1167 | ||
1168 | blob = lookup_blob(oid); | |
1169 | if (!blob) { | |
1170 | ret |= report(options, &blob->object, | |
1171 | FSCK_MSG_GITMODULES_BLOB, | |
1172 | "non-blob found at .gitmodules"); | |
1173 | continue; | |
1174 | } | |
1175 | ||
1176 | buf = read_sha1_file(oid->hash, &type, &size); | |
1177 | if (!buf) { | |
27387444 JK |
1178 | if (is_promisor_object(&blob->object.oid)) |
1179 | continue; | |
159e7b08 JK |
1180 | ret |= report(options, &blob->object, |
1181 | FSCK_MSG_GITMODULES_MISSING, | |
1182 | "unable to read .gitmodules blob"); | |
1183 | continue; | |
1184 | } | |
1185 | ||
1186 | if (type == OBJ_BLOB) | |
1187 | ret |= fsck_blob(blob, buf, size, options); | |
1188 | else | |
1189 | ret |= report(options, &blob->object, | |
1190 | FSCK_MSG_GITMODULES_BLOB, | |
1191 | "non-blob found at .gitmodules"); | |
1192 | free(buf); | |
1193 | } | |
1194 | ||
1195 | ||
1196 | oidset_clear(&gitmodules_found); | |
1197 | oidset_clear(&gitmodules_done); | |
1198 | return ret; | |
1199 | } |