]>
Commit | Line | Data |
---|---|---|
d812c3b6 | 1 | #include "git-compat-util.h" |
d4a4f929 | 2 | #include "date.h" |
a034e910 | 3 | #include "dir.h" |
41771fa4 | 4 | #include "hex.h" |
a034e910 | 5 | #include "object-store-ll.h" |
c339932b | 6 | #include "path.h" |
109cd76d | 7 | #include "repository.h" |
355885d5 | 8 | #include "object.h" |
27ab4784 | 9 | #include "attr.h" |
355885d5 MK |
10 | #include "blob.h" |
11 | #include "tree.h" | |
12 | #include "tree-walk.h" | |
13 | #include "commit.h" | |
14 | #include "tag.h" | |
15 | #include "fsck.h" | |
cec097be | 16 | #include "refs.h" |
a2b26ffb | 17 | #include "url.h" |
a18fcc9f | 18 | #include "utf8.h" |
159e7b08 | 19 | #include "oidset.h" |
27387444 | 20 | #include "packfile.h" |
ed8b10f6 JK |
21 | #include "submodule-config.h" |
22 | #include "config.h" | |
3ac68a93 | 23 | #include "help.h" |
159e7b08 | 24 | |
0fbcaef6 JK |
25 | static ssize_t max_tree_entry_len = 4096; |
26 | ||
f417eed8 | 27 | #define STR(x) #x |
a4a9cc19 | 28 | #define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type }, |
c99ba492 | 29 | static struct { |
f417eed8 JS |
30 | const char *id_string; |
31 | const char *downcased; | |
a4a9cc19 | 32 | const char *camelcased; |
1b32b59f | 33 | enum fsck_msg_type msg_type; |
c99ba492 | 34 | } msg_id_info[FSCK_MSG_MAX + 1] = { |
901f2f67 | 35 | FOREACH_FSCK_MSG_ID(MSG_ID) |
a4a9cc19 | 36 | { NULL, NULL, NULL, -1 } |
c99ba492 JS |
37 | }; |
38 | #undef MSG_ID | |
b5495024 | 39 | #undef STR |
c99ba492 | 40 | |
a46baac6 | 41 | static void prepare_msg_ids(void) |
f417eed8 JS |
42 | { |
43 | int i; | |
44 | ||
a46baac6 NTND |
45 | if (msg_id_info[0].downcased) |
46 | return; | |
47 | ||
48 | /* convert id_string to lower case, without underscores. */ | |
49 | for (i = 0; i < FSCK_MSG_MAX; i++) { | |
50 | const char *p = msg_id_info[i].id_string; | |
51 | int len = strlen(p); | |
52 | char *q = xmalloc(len); | |
53 | ||
54 | msg_id_info[i].downcased = q; | |
55 | while (*p) | |
56 | if (*p == '_') | |
57 | p++; | |
58 | else | |
59 | *(q)++ = tolower(*(p)++); | |
60 | *q = '\0'; | |
a4a9cc19 NTND |
61 | |
62 | p = msg_id_info[i].id_string; | |
63 | q = xmalloc(len); | |
64 | msg_id_info[i].camelcased = q; | |
65 | while (*p) { | |
66 | if (*p == '_') { | |
67 | p++; | |
68 | if (*p) | |
69 | *q++ = *p++; | |
70 | } else { | |
71 | *q++ = tolower(*p++); | |
72 | } | |
f417eed8 | 73 | } |
a4a9cc19 | 74 | *q = '\0'; |
f417eed8 | 75 | } |
a46baac6 NTND |
76 | } |
77 | ||
78 | static int parse_msg_id(const char *text) | |
79 | { | |
80 | int i; | |
81 | ||
82 | prepare_msg_ids(); | |
f417eed8 JS |
83 | |
84 | for (i = 0; i < FSCK_MSG_MAX; i++) | |
85 | if (!strcmp(text, msg_id_info[i].downcased)) | |
86 | return i; | |
87 | ||
88 | return -1; | |
89 | } | |
90 | ||
3ac68a93 NTND |
91 | void list_config_fsck_msg_ids(struct string_list *list, const char *prefix) |
92 | { | |
93 | int i; | |
94 | ||
95 | prepare_msg_ids(); | |
96 | ||
3ac68a93 | 97 | for (i = 0; i < FSCK_MSG_MAX; i++) |
a4a9cc19 | 98 | list_config_item(list, prefix, msg_id_info[i].camelcased); |
3ac68a93 NTND |
99 | } |
100 | ||
1b32b59f | 101 | static enum fsck_msg_type fsck_msg_type(enum fsck_msg_id msg_id, |
c99ba492 JS |
102 | struct fsck_options *options) |
103 | { | |
0282f4dc JS |
104 | assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX); |
105 | ||
e35d65a7 | 106 | if (!options->msg_type) { |
1b32b59f | 107 | enum fsck_msg_type msg_type = msg_id_info[msg_id].msg_type; |
e35d65a7 | 108 | |
0282f4dc JS |
109 | if (options->strict && msg_type == FSCK_WARN) |
110 | msg_type = FSCK_ERROR; | |
e35d65a7 | 111 | return msg_type; |
0282f4dc | 112 | } |
c99ba492 | 113 | |
e35d65a7 | 114 | return options->msg_type[msg_id]; |
c99ba492 JS |
115 | } |
116 | ||
1b32b59f | 117 | static enum fsck_msg_type parse_msg_type(const char *str) |
0282f4dc JS |
118 | { |
119 | if (!strcmp(str, "error")) | |
120 | return FSCK_ERROR; | |
121 | else if (!strcmp(str, "warn")) | |
122 | return FSCK_WARN; | |
efaba7cc JS |
123 | else if (!strcmp(str, "ignore")) |
124 | return FSCK_IGNORE; | |
0282f4dc JS |
125 | else |
126 | die("Unknown fsck message type: '%s'", str); | |
127 | } | |
128 | ||
5d477a33 JS |
129 | int is_valid_msg_type(const char *msg_id, const char *msg_type) |
130 | { | |
131 | if (parse_msg_id(msg_id) < 0) | |
132 | return 0; | |
133 | parse_msg_type(msg_type); | |
134 | return 1; | |
135 | } | |
136 | ||
53692df2 ÆAB |
137 | void fsck_set_msg_type_from_ids(struct fsck_options *options, |
138 | enum fsck_msg_id msg_id, | |
139 | enum fsck_msg_type msg_type) | |
140 | { | |
141 | if (!options->msg_type) { | |
142 | int i; | |
143 | enum fsck_msg_type *severity; | |
144 | ALLOC_ARRAY(severity, FSCK_MSG_MAX); | |
145 | for (i = 0; i < FSCK_MSG_MAX; i++) | |
146 | severity[i] = fsck_msg_type(i, options); | |
147 | options->msg_type = severity; | |
148 | } | |
149 | ||
150 | options->msg_type[msg_id] = msg_type; | |
151 | } | |
152 | ||
0282f4dc | 153 | void fsck_set_msg_type(struct fsck_options *options, |
f1abc2d0 | 154 | const char *msg_id_str, const char *msg_type_str) |
0282f4dc | 155 | { |
1b32b59f | 156 | int msg_id = parse_msg_id(msg_id_str); |
0fbcaef6 JK |
157 | char *to_free = NULL; |
158 | enum fsck_msg_type msg_type; | |
0282f4dc | 159 | |
f1abc2d0 ÆAB |
160 | if (msg_id < 0) |
161 | die("Unhandled message id: %s", msg_id_str); | |
0282f4dc | 162 | |
0fbcaef6 JK |
163 | if (msg_id == FSCK_MSG_LARGE_PATHNAME) { |
164 | const char *colon = strchr(msg_type_str, ':'); | |
165 | if (colon) { | |
166 | msg_type_str = to_free = | |
167 | xmemdupz(msg_type_str, colon - msg_type_str); | |
168 | colon++; | |
169 | if (!git_parse_ssize_t(colon, &max_tree_entry_len)) | |
170 | die("unable to parse max tree entry len: %s", colon); | |
171 | } | |
172 | } | |
173 | msg_type = parse_msg_type(msg_type_str); | |
174 | ||
f1abc2d0 ÆAB |
175 | if (msg_type != FSCK_ERROR && msg_id_info[msg_id].msg_type == FSCK_FATAL) |
176 | die("Cannot demote %s to %s", msg_id_str, msg_type_str); | |
f50c4407 | 177 | |
53692df2 | 178 | fsck_set_msg_type_from_ids(options, msg_id, msg_type); |
0fbcaef6 | 179 | free(to_free); |
0282f4dc JS |
180 | } |
181 | ||
182 | void fsck_set_msg_types(struct fsck_options *options, const char *values) | |
183 | { | |
184 | char *buf = xstrdup(values), *to_free = buf; | |
185 | int done = 0; | |
186 | ||
187 | while (!done) { | |
188 | int len = strcspn(buf, " ,|"), equal; | |
189 | ||
190 | done = !buf[len]; | |
191 | if (!len) { | |
192 | buf++; | |
193 | continue; | |
194 | } | |
195 | buf[len] = '\0'; | |
196 | ||
197 | for (equal = 0; | |
198 | equal < len && buf[equal] != '=' && buf[equal] != ':'; | |
199 | equal++) | |
200 | buf[equal] = tolower(buf[equal]); | |
201 | buf[equal] = '\0'; | |
202 | ||
cd94c6f9 JS |
203 | if (!strcmp(buf, "skiplist")) { |
204 | if (equal == len) | |
205 | die("skiplist requires a path"); | |
24eb33eb | 206 | oidset_parse_file(&options->skiplist, buf + equal + 1); |
cd94c6f9 JS |
207 | buf += len + 1; |
208 | continue; | |
209 | } | |
210 | ||
0282f4dc JS |
211 | if (equal == len) |
212 | die("Missing '=': '%s'", buf); | |
355885d5 | 213 | |
0282f4dc JS |
214 | fsck_set_msg_type(options, buf, buf + equal + 1); |
215 | buf += len + 1; | |
216 | } | |
217 | free(to_free); | |
218 | } | |
219 | ||
f5979376 JK |
220 | static int object_on_skiplist(struct fsck_options *opts, |
221 | const struct object_id *oid) | |
fb162877 | 222 | { |
f5979376 | 223 | return opts && oid && oidset_contains(&opts->skiplist, oid); |
fb162877 RJ |
224 | } |
225 | ||
38370253 JK |
226 | __attribute__((format (printf, 5, 6))) |
227 | static int report(struct fsck_options *options, | |
228 | const struct object_id *oid, enum object_type object_type, | |
35af754b | 229 | enum fsck_msg_id msg_id, const char *fmt, ...) |
c99ba492 JS |
230 | { |
231 | va_list ap; | |
232 | struct strbuf sb = STRBUF_INIT; | |
1b32b59f ÆAB |
233 | enum fsck_msg_type msg_type = fsck_msg_type(msg_id, options); |
234 | int result; | |
c99ba492 | 235 | |
efaba7cc JS |
236 | if (msg_type == FSCK_IGNORE) |
237 | return 0; | |
238 | ||
38370253 | 239 | if (object_on_skiplist(options, oid)) |
cd94c6f9 JS |
240 | return 0; |
241 | ||
f50c4407 JS |
242 | if (msg_type == FSCK_FATAL) |
243 | msg_type = FSCK_ERROR; | |
f27d05b1 JS |
244 | else if (msg_type == FSCK_INFO) |
245 | msg_type = FSCK_WARN; | |
f50c4407 | 246 | |
034a7b7b | 247 | prepare_msg_ids(); |
35af754b | 248 | strbuf_addf(&sb, "%s: ", msg_id_info[msg_id].camelcased); |
71ab8fa8 | 249 | |
c99ba492 JS |
250 | va_start(ap, fmt); |
251 | strbuf_vaddf(&sb, fmt, ap); | |
38370253 | 252 | result = options->error_func(options, oid, object_type, |
394d5d31 | 253 | msg_type, msg_id, sb.buf); |
c99ba492 JS |
254 | strbuf_release(&sb); |
255 | va_end(ap); | |
256 | ||
257 | return result; | |
258 | } | |
259 | ||
a59cfb32 JK |
260 | void fsck_enable_object_names(struct fsck_options *options) |
261 | { | |
262 | if (!options->object_names) | |
73390290 | 263 | options->object_names = kh_init_oid_map(); |
a59cfb32 JK |
264 | } |
265 | ||
73390290 JK |
266 | const char *fsck_get_object_name(struct fsck_options *options, |
267 | const struct object_id *oid) | |
7b35efd7 | 268 | { |
73390290 | 269 | khiter_t pos; |
7b35efd7 JS |
270 | if (!options->object_names) |
271 | return NULL; | |
73390290 JK |
272 | pos = kh_get_oid_map(options->object_names, *oid); |
273 | if (pos >= kh_end(options->object_names)) | |
274 | return NULL; | |
275 | return kh_value(options->object_names, pos); | |
7b35efd7 JS |
276 | } |
277 | ||
73390290 JK |
278 | void fsck_put_object_name(struct fsck_options *options, |
279 | const struct object_id *oid, | |
a59cfb32 | 280 | const char *fmt, ...) |
7b35efd7 JS |
281 | { |
282 | va_list ap; | |
283 | struct strbuf buf = STRBUF_INIT; | |
73390290 JK |
284 | khiter_t pos; |
285 | int hashret; | |
7b35efd7 JS |
286 | |
287 | if (!options->object_names) | |
288 | return; | |
73390290 JK |
289 | |
290 | pos = kh_put_oid_map(options->object_names, *oid, &hashret); | |
291 | if (!hashret) | |
7b35efd7 JS |
292 | return; |
293 | va_start(ap, fmt); | |
294 | strbuf_vaddf(&buf, fmt, ap); | |
73390290 | 295 | kh_value(options->object_names, pos) = strbuf_detach(&buf, NULL); |
7b35efd7 JS |
296 | va_end(ap); |
297 | } | |
298 | ||
a59cfb32 | 299 | const char *fsck_describe_object(struct fsck_options *options, |
73390290 | 300 | const struct object_id *oid) |
90cf590f | 301 | { |
a59cfb32 JK |
302 | static struct strbuf bufs[] = { |
303 | STRBUF_INIT, STRBUF_INIT, STRBUF_INIT, STRBUF_INIT | |
304 | }; | |
305 | static int b = 0; | |
306 | struct strbuf *buf; | |
73390290 | 307 | const char *name = fsck_get_object_name(options, oid); |
a59cfb32 JK |
308 | |
309 | buf = bufs + b; | |
310 | b = (b + 1) % ARRAY_SIZE(bufs); | |
311 | strbuf_reset(buf); | |
73390290 | 312 | strbuf_addstr(buf, oid_to_hex(oid)); |
a59cfb32 JK |
313 | if (name) |
314 | strbuf_addf(buf, " (%s)", name); | |
90cf590f | 315 | |
a59cfb32 | 316 | return buf->buf; |
90cf590f JS |
317 | } |
318 | ||
22410549 | 319 | static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options) |
355885d5 MK |
320 | { |
321 | struct tree_desc desc; | |
322 | struct name_entry entry; | |
323 | int res = 0; | |
7b35efd7 | 324 | const char *name; |
355885d5 MK |
325 | |
326 | if (parse_tree(tree)) | |
327 | return -1; | |
328 | ||
73390290 | 329 | name = fsck_get_object_name(options, &tree->object.oid); |
ec18b10b | 330 | if (init_tree_desc_gently(&desc, tree->buffer, tree->size, 0)) |
8354fa3d DT |
331 | return -1; |
332 | while (tree_entry_gently(&desc, &entry)) { | |
7b35efd7 | 333 | struct object *obj; |
355885d5 MK |
334 | int result; |
335 | ||
336 | if (S_ISGITLINK(entry.mode)) | |
337 | continue; | |
7b35efd7 JS |
338 | |
339 | if (S_ISDIR(entry.mode)) { | |
ea82b2a0 | 340 | obj = (struct object *)lookup_tree(the_repository, &entry.oid); |
2720f6db | 341 | if (name && obj) |
73390290 | 342 | fsck_put_object_name(options, &entry.oid, "%s%s/", |
a59cfb32 | 343 | name, entry.path); |
7b35efd7 JS |
344 | result = options->walk(obj, OBJ_TREE, data, options); |
345 | } | |
346 | else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) { | |
ea82b2a0 | 347 | obj = (struct object *)lookup_blob(the_repository, &entry.oid); |
2720f6db | 348 | if (name && obj) |
73390290 | 349 | fsck_put_object_name(options, &entry.oid, "%s%s", |
a59cfb32 | 350 | name, entry.path); |
7b35efd7 JS |
351 | result = options->walk(obj, OBJ_BLOB, data, options); |
352 | } | |
355885d5 | 353 | else { |
82247e9b | 354 | result = error("in tree %s: entry %s has bad mode %.6o", |
73390290 | 355 | fsck_describe_object(options, &tree->object.oid), |
a59cfb32 | 356 | entry.path, entry.mode); |
355885d5 MK |
357 | } |
358 | if (result < 0) | |
359 | return result; | |
360 | if (!res) | |
361 | res = result; | |
362 | } | |
363 | return res; | |
364 | } | |
365 | ||
22410549 | 366 | static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options) |
355885d5 | 367 | { |
7b35efd7 | 368 | int counter = 0, generation = 0, name_prefix_len = 0; |
355885d5 MK |
369 | struct commit_list *parents; |
370 | int res; | |
371 | int result; | |
7b35efd7 | 372 | const char *name; |
355885d5 | 373 | |
ecb5091f | 374 | if (repo_parse_commit(the_repository, commit)) |
355885d5 MK |
375 | return -1; |
376 | ||
73390290 | 377 | name = fsck_get_object_name(options, &commit->object.oid); |
7b35efd7 | 378 | if (name) |
73390290 | 379 | fsck_put_object_name(options, get_commit_tree_oid(commit), |
a59cfb32 | 380 | "%s:", name); |
7b35efd7 | 381 | |
ecb5091f | 382 | result = options->walk((struct object *) repo_get_commit_tree(the_repository, commit), |
2e27bd77 | 383 | OBJ_TREE, data, options); |
355885d5 MK |
384 | if (result < 0) |
385 | return result; | |
386 | res = result; | |
387 | ||
388 | parents = commit->parents; | |
7b35efd7 JS |
389 | if (name && parents) { |
390 | int len = strlen(name), power; | |
391 | ||
392 | if (len && name[len - 1] == '^') { | |
393 | generation = 1; | |
394 | name_prefix_len = len - 1; | |
395 | } | |
396 | else { /* parse ~<generation> suffix */ | |
397 | for (generation = 0, power = 1; | |
398 | len && isdigit(name[len - 1]); | |
399 | power *= 10) | |
400 | generation += power * (name[--len] - '0'); | |
401 | if (power > 1 && len && name[len - 1] == '~') | |
402 | name_prefix_len = len - 1; | |
e89f8936 JS |
403 | else { |
404 | /* Maybe a non-first parent, e.g. HEAD^2 */ | |
405 | generation = 0; | |
406 | name_prefix_len = len; | |
407 | } | |
7b35efd7 JS |
408 | } |
409 | } | |
410 | ||
355885d5 | 411 | while (parents) { |
7b35efd7 | 412 | if (name) { |
73390290 | 413 | struct object_id *oid = &parents->item->object.oid; |
7b35efd7 | 414 | |
b84c7838 | 415 | if (counter++) |
73390290 | 416 | fsck_put_object_name(options, oid, "%s^%d", |
a59cfb32 | 417 | name, counter); |
7b35efd7 | 418 | else if (generation > 0) |
73390290 | 419 | fsck_put_object_name(options, oid, "%.*s~%d", |
a59cfb32 JK |
420 | name_prefix_len, name, |
421 | generation + 1); | |
7b35efd7 | 422 | else |
73390290 | 423 | fsck_put_object_name(options, oid, "%s^", name); |
7b35efd7 | 424 | } |
22410549 | 425 | result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options); |
355885d5 MK |
426 | if (result < 0) |
427 | return result; | |
428 | if (!res) | |
429 | res = result; | |
430 | parents = parents->next; | |
431 | } | |
432 | return res; | |
433 | } | |
434 | ||
22410549 | 435 | static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options) |
355885d5 | 436 | { |
73390290 | 437 | const char *name = fsck_get_object_name(options, &tag->object.oid); |
7b35efd7 | 438 | |
355885d5 MK |
439 | if (parse_tag(tag)) |
440 | return -1; | |
7b35efd7 | 441 | if (name) |
73390290 | 442 | fsck_put_object_name(options, &tag->tagged->oid, "%s", name); |
22410549 | 443 | return options->walk(tag->tagged, OBJ_ANY, data, options); |
355885d5 MK |
444 | } |
445 | ||
22410549 | 446 | int fsck_walk(struct object *obj, void *data, struct fsck_options *options) |
355885d5 MK |
447 | { |
448 | if (!obj) | |
449 | return -1; | |
a2b22854 JK |
450 | |
451 | if (obj->type == OBJ_NONE) | |
109cd76d | 452 | parse_object(the_repository, &obj->oid); |
a2b22854 | 453 | |
355885d5 MK |
454 | switch (obj->type) { |
455 | case OBJ_BLOB: | |
456 | return 0; | |
457 | case OBJ_TREE: | |
22410549 | 458 | return fsck_walk_tree((struct tree *)obj, data, options); |
355885d5 | 459 | case OBJ_COMMIT: |
22410549 | 460 | return fsck_walk_commit((struct commit *)obj, data, options); |
355885d5 | 461 | case OBJ_TAG: |
22410549 | 462 | return fsck_walk_tag((struct tag *)obj, data, options); |
355885d5 | 463 | default: |
a59cfb32 | 464 | error("Unknown object type for %s", |
73390290 | 465 | fsck_describe_object(options, &obj->oid)); |
355885d5 MK |
466 | return -1; |
467 | } | |
468 | } | |
ba002f3b | 469 | |
9068cfb2 RS |
470 | struct name_stack { |
471 | const char **names; | |
472 | size_t nr, alloc; | |
473 | }; | |
474 | ||
475 | static void name_stack_push(struct name_stack *stack, const char *name) | |
476 | { | |
477 | ALLOC_GROW(stack->names, stack->nr + 1, stack->alloc); | |
478 | stack->names[stack->nr++] = name; | |
479 | } | |
480 | ||
481 | static const char *name_stack_pop(struct name_stack *stack) | |
482 | { | |
483 | return stack->nr ? stack->names[--stack->nr] : NULL; | |
484 | } | |
485 | ||
486 | static void name_stack_clear(struct name_stack *stack) | |
487 | { | |
488 | FREE_AND_NULL(stack->names); | |
489 | stack->nr = stack->alloc = 0; | |
490 | } | |
491 | ||
ba002f3b MK |
492 | /* |
493 | * The entries in a tree are ordered in the _path_ order, | |
494 | * which means that a directory entry is ordered by adding | |
495 | * a slash to the end of it. | |
496 | * | |
497 | * So a directory called "a" is ordered _after_ a file | |
498 | * called "a.c", because "a/" sorts after "a.c". | |
499 | */ | |
500 | #define TREE_UNORDERED (-1) | |
501 | #define TREE_HAS_DUPS (-2) | |
502 | ||
9068cfb2 RS |
503 | static int is_less_than_slash(unsigned char c) |
504 | { | |
505 | return '\0' < c && c < '/'; | |
506 | } | |
507 | ||
508 | static int verify_ordered(unsigned mode1, const char *name1, | |
509 | unsigned mode2, const char *name2, | |
510 | struct name_stack *candidates) | |
ba002f3b MK |
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 = '/'; | |
9068cfb2 RS |
542 | |
543 | /* | |
544 | * There can be non-consecutive duplicates due to the implicitly | |
86715592 | 545 | * added slash, e.g.: |
9068cfb2 RS |
546 | * |
547 | * foo | |
548 | * foo.bar | |
549 | * foo.bar.baz | |
550 | * foo.bar/ | |
551 | * foo/ | |
552 | * | |
553 | * Record non-directory candidates (like "foo" and "foo.bar" in | |
554 | * the example) on a stack and check directory candidates (like | |
555 | * foo/" and "foo.bar/") against that stack. | |
556 | */ | |
557 | if (!c1 && is_less_than_slash(c2)) { | |
558 | name_stack_push(candidates, name1); | |
559 | } else if (c2 == '/' && is_less_than_slash(c1)) { | |
560 | for (;;) { | |
561 | const char *p; | |
562 | const char *f_name = name_stack_pop(candidates); | |
563 | ||
564 | if (!f_name) | |
565 | break; | |
566 | if (!skip_prefix(name2, f_name, &p)) | |
fe747043 | 567 | continue; |
9068cfb2 RS |
568 | if (!*p) |
569 | return TREE_HAS_DUPS; | |
570 | if (is_less_than_slash(*p)) { | |
571 | name_stack_push(candidates, f_name); | |
572 | break; | |
573 | } | |
574 | } | |
575 | } | |
576 | ||
ba002f3b MK |
577 | return c1 < c2 ? 0 : TREE_UNORDERED; |
578 | } | |
579 | ||
9e1947cb | 580 | static int fsck_tree(const struct object_id *tree_oid, |
23a173a7 JK |
581 | const char *buffer, unsigned long size, |
582 | struct fsck_options *options) | |
ba002f3b | 583 | { |
8354fa3d | 584 | int retval = 0; |
c479d14a | 585 | int has_null_sha1 = 0; |
ba002f3b MK |
586 | int has_full_path = 0; |
587 | int has_empty_name = 0; | |
5d34a435 JK |
588 | int has_dot = 0; |
589 | int has_dotdot = 0; | |
5c17f512 | 590 | int has_dotgit = 0; |
ba002f3b MK |
591 | int has_zero_pad = 0; |
592 | int has_bad_modes = 0; | |
593 | int has_dup_entries = 0; | |
594 | int not_properly_sorted = 0; | |
0fbcaef6 | 595 | int has_large_name = 0; |
ba002f3b MK |
596 | struct tree_desc desc; |
597 | unsigned o_mode; | |
598 | const char *o_name; | |
9068cfb2 | 599 | struct name_stack df_dup_candidates = { NULL }; |
ba002f3b | 600 | |
53602a93 | 601 | if (init_tree_desc_gently(&desc, buffer, size, TREE_DESC_RAW_MODES)) { |
0282f679 JK |
602 | retval += report(options, tree_oid, OBJ_TREE, |
603 | FSCK_MSG_BAD_TREE, | |
604 | "cannot be parsed as a tree"); | |
8354fa3d DT |
605 | return retval; |
606 | } | |
ba002f3b MK |
607 | |
608 | o_mode = 0; | |
609 | o_name = NULL; | |
ba002f3b MK |
610 | |
611 | while (desc.size) { | |
5ec1e728 | 612 | unsigned short mode; |
288a74bc | 613 | const char *name, *backslash; |
9e1947cb | 614 | const struct object_id *entry_oid; |
ba002f3b | 615 | |
9e1947cb | 616 | entry_oid = tree_entry_extract(&desc, &name, &mode); |
ba002f3b | 617 | |
9e1947cb | 618 | has_null_sha1 |= is_null_oid(entry_oid); |
effd12ec HS |
619 | has_full_path |= !!strchr(name, '/'); |
620 | has_empty_name |= !*name; | |
621 | has_dot |= !strcmp(name, "."); | |
622 | has_dotdot |= !strcmp(name, ".."); | |
ed9c3220 | 623 | has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name); |
ba002f3b | 624 | has_zero_pad |= *(char *)desc.buffer == '0'; |
0fbcaef6 | 625 | has_large_name |= tree_entry_len(&desc.entry) > max_tree_entry_len; |
159e7b08 | 626 | |
b7b1fca1 JK |
627 | if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) { |
628 | if (!S_ISLNK(mode)) | |
9e1947cb JK |
629 | oidset_insert(&options->gitmodules_found, |
630 | entry_oid); | |
b7b1fca1 | 631 | else |
38370253 | 632 | retval += report(options, |
9e1947cb | 633 | tree_oid, OBJ_TREE, |
b7b1fca1 JK |
634 | FSCK_MSG_GITMODULES_SYMLINK, |
635 | ".gitmodules is a symbolic link"); | |
636 | } | |
159e7b08 | 637 | |
f8587c31 | 638 | if (is_hfs_dotgitattributes(name) || is_ntfs_dotgitattributes(name)) { |
27ab4784 PS |
639 | if (!S_ISLNK(mode)) |
640 | oidset_insert(&options->gitattributes_found, | |
641 | entry_oid); | |
642 | else | |
f8587c31 PS |
643 | retval += report(options, tree_oid, OBJ_TREE, |
644 | FSCK_MSG_GITATTRIBUTES_SYMLINK, | |
645 | ".gitattributes is a symlink"); | |
646 | } | |
647 | ||
bb6832d5 JK |
648 | if (S_ISLNK(mode)) { |
649 | if (is_hfs_dotgitignore(name) || | |
650 | is_ntfs_dotgitignore(name)) | |
651 | retval += report(options, tree_oid, OBJ_TREE, | |
652 | FSCK_MSG_GITIGNORE_SYMLINK, | |
653 | ".gitignore is a symlink"); | |
bb6832d5 JK |
654 | if (is_hfs_dotmailmap(name) || |
655 | is_ntfs_dotmailmap(name)) | |
656 | retval += report(options, tree_oid, OBJ_TREE, | |
657 | FSCK_MSG_MAILMAP_SYMLINK, | |
658 | ".mailmap is a symlink"); | |
659 | } | |
660 | ||
288a74bc JS |
661 | if ((backslash = strchr(name, '\\'))) { |
662 | while (backslash) { | |
663 | backslash++; | |
664 | has_dotgit |= is_ntfs_dotgit(backslash); | |
bdfef049 JS |
665 | if (is_ntfs_dotgitmodules(backslash)) { |
666 | if (!S_ISLNK(mode)) | |
9e1947cb JK |
667 | oidset_insert(&options->gitmodules_found, |
668 | entry_oid); | |
bdfef049 | 669 | else |
9e1947cb | 670 | retval += report(options, tree_oid, OBJ_TREE, |
bdfef049 JS |
671 | FSCK_MSG_GITMODULES_SYMLINK, |
672 | ".gitmodules is a symbolic link"); | |
673 | } | |
288a74bc JS |
674 | backslash = strchr(backslash, '\\'); |
675 | } | |
676 | } | |
677 | ||
8354fa3d | 678 | if (update_tree_entry_gently(&desc)) { |
0282f679 JK |
679 | retval += report(options, tree_oid, OBJ_TREE, |
680 | FSCK_MSG_BAD_TREE, | |
681 | "cannot be parsed as a tree"); | |
8354fa3d DT |
682 | break; |
683 | } | |
ba002f3b MK |
684 | |
685 | switch (mode) { | |
686 | /* | |
687 | * Standard modes.. | |
688 | */ | |
689 | case S_IFREG | 0755: | |
690 | case S_IFREG | 0644: | |
691 | case S_IFLNK: | |
692 | case S_IFDIR: | |
693 | case S_IFGITLINK: | |
694 | break; | |
695 | /* | |
696 | * This is nonstandard, but we had a few of these | |
697 | * early on when we honored the full set of mode | |
698 | * bits.. | |
699 | */ | |
700 | case S_IFREG | 0664: | |
22410549 | 701 | if (!options->strict) |
ba002f3b | 702 | break; |
1cf01a34 | 703 | /* fallthrough */ |
ba002f3b MK |
704 | default: |
705 | has_bad_modes = 1; | |
706 | } | |
707 | ||
708 | if (o_name) { | |
9068cfb2 RS |
709 | switch (verify_ordered(o_mode, o_name, mode, name, |
710 | &df_dup_candidates)) { | |
ba002f3b MK |
711 | case TREE_UNORDERED: |
712 | not_properly_sorted = 1; | |
713 | break; | |
714 | case TREE_HAS_DUPS: | |
715 | has_dup_entries = 1; | |
716 | break; | |
717 | default: | |
718 | break; | |
719 | } | |
720 | } | |
721 | ||
722 | o_mode = mode; | |
723 | o_name = name; | |
ba002f3b MK |
724 | } |
725 | ||
9068cfb2 RS |
726 | name_stack_clear(&df_dup_candidates); |
727 | ||
c479d14a | 728 | if (has_null_sha1) |
0282f679 JK |
729 | retval += report(options, tree_oid, OBJ_TREE, |
730 | FSCK_MSG_NULL_SHA1, | |
731 | "contains entries pointing to null sha1"); | |
ba002f3b | 732 | if (has_full_path) |
0282f679 JK |
733 | retval += report(options, tree_oid, OBJ_TREE, |
734 | FSCK_MSG_FULL_PATHNAME, | |
735 | "contains full pathnames"); | |
ba002f3b | 736 | if (has_empty_name) |
0282f679 JK |
737 | retval += report(options, tree_oid, OBJ_TREE, |
738 | FSCK_MSG_EMPTY_NAME, | |
739 | "contains empty pathname"); | |
5d34a435 | 740 | if (has_dot) |
0282f679 JK |
741 | retval += report(options, tree_oid, OBJ_TREE, |
742 | FSCK_MSG_HAS_DOT, | |
743 | "contains '.'"); | |
5d34a435 | 744 | if (has_dotdot) |
0282f679 JK |
745 | retval += report(options, tree_oid, OBJ_TREE, |
746 | FSCK_MSG_HAS_DOTDOT, | |
747 | "contains '..'"); | |
5c17f512 | 748 | if (has_dotgit) |
0282f679 JK |
749 | retval += report(options, tree_oid, OBJ_TREE, |
750 | FSCK_MSG_HAS_DOTGIT, | |
751 | "contains '.git'"); | |
ba002f3b | 752 | if (has_zero_pad) |
0282f679 JK |
753 | retval += report(options, tree_oid, OBJ_TREE, |
754 | FSCK_MSG_ZERO_PADDED_FILEMODE, | |
755 | "contains zero-padded file modes"); | |
ba002f3b | 756 | if (has_bad_modes) |
0282f679 JK |
757 | retval += report(options, tree_oid, OBJ_TREE, |
758 | FSCK_MSG_BAD_FILEMODE, | |
759 | "contains bad file modes"); | |
ba002f3b | 760 | if (has_dup_entries) |
0282f679 JK |
761 | retval += report(options, tree_oid, OBJ_TREE, |
762 | FSCK_MSG_DUPLICATE_ENTRIES, | |
763 | "contains duplicate file entries"); | |
ba002f3b | 764 | if (not_properly_sorted) |
0282f679 JK |
765 | retval += report(options, tree_oid, OBJ_TREE, |
766 | FSCK_MSG_TREE_NOT_SORTED, | |
767 | "not properly sorted"); | |
0fbcaef6 JK |
768 | if (has_large_name) |
769 | retval += report(options, tree_oid, OBJ_TREE, | |
770 | FSCK_MSG_LARGE_PATHNAME, | |
771 | "contains excessively large pathname"); | |
ba002f3b MK |
772 | return retval; |
773 | } | |
774 | ||
8e430903 JK |
775 | /* |
776 | * Confirm that the headers of a commit or tag object end in a reasonable way, | |
777 | * either with the usual "\n\n" separator, or at least with a trailing newline | |
778 | * on the final header line. | |
779 | * | |
780 | * This property is important for the memory safety of our callers. It allows | |
781 | * them to scan the buffer linewise without constantly checking the remaining | |
782 | * size as long as: | |
783 | * | |
784 | * - they check that there are bytes left in the buffer at the start of any | |
785 | * line (i.e., that the last newline they saw was not the final one we | |
786 | * found here) | |
787 | * | |
788 | * - any intra-line scanning they do will stop at a newline, which will worst | |
789 | * case hit the newline we found here as the end-of-header. This makes it | |
790 | * OK for them to use helpers like parse_oid_hex(), or even skip_prefix(). | |
791 | */ | |
84d18c0b | 792 | static int verify_headers(const void *data, unsigned long size, |
cc579000 JK |
793 | const struct object_id *oid, enum object_type type, |
794 | struct fsck_options *options) | |
4d0d8975 JS |
795 | { |
796 | const char *buffer = (const char *)data; | |
797 | unsigned long i; | |
798 | ||
799 | for (i = 0; i < size; i++) { | |
800 | switch (buffer[i]) { | |
801 | case '\0': | |
cc579000 | 802 | return report(options, oid, type, |
c99ba492 JS |
803 | FSCK_MSG_NUL_IN_HEADER, |
804 | "unterminated header: NUL at offset %ld", i); | |
4d0d8975 JS |
805 | case '\n': |
806 | if (i + 1 < size && buffer[i + 1] == '\n') | |
807 | return 0; | |
808 | } | |
809 | } | |
810 | ||
84d18c0b JH |
811 | /* |
812 | * We did not find double-LF that separates the header | |
813 | * and the body. Not having a body is not a crime but | |
814 | * we do want to see the terminating LF for the last header | |
815 | * line. | |
816 | */ | |
817 | if (size && buffer[size - 1] == '\n') | |
818 | return 0; | |
819 | ||
cc579000 | 820 | return report(options, oid, type, |
c99ba492 | 821 | FSCK_MSG_UNTERMINATED_HEADER, "unterminated header"); |
4d0d8975 JS |
822 | } |
823 | ||
78543993 JK |
824 | static int fsck_ident(const char **ident, |
825 | const struct object_id *oid, enum object_type type, | |
826 | struct fsck_options *options) | |
daae1922 | 827 | { |
e6826e33 | 828 | const char *p = *ident; |
d4b8de04 JK |
829 | char *end; |
830 | ||
e6826e33 JS |
831 | *ident = strchrnul(*ident, '\n'); |
832 | if (**ident == '\n') | |
833 | (*ident)++; | |
834 | ||
835 | if (*p == '<') | |
78543993 | 836 | return report(options, oid, type, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email"); |
e6826e33 JS |
837 | p += strcspn(p, "<>\n"); |
838 | if (*p == '>') | |
78543993 | 839 | return report(options, oid, type, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name"); |
e6826e33 | 840 | if (*p != '<') |
78543993 | 841 | return report(options, oid, type, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email"); |
e6826e33 | 842 | if (p[-1] != ' ') |
78543993 | 843 | return report(options, oid, type, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email"); |
e6826e33 JS |
844 | p++; |
845 | p += strcspn(p, "<>\n"); | |
846 | if (*p != '>') | |
78543993 | 847 | return report(options, oid, type, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email"); |
e6826e33 JS |
848 | p++; |
849 | if (*p != ' ') | |
78543993 | 850 | return report(options, oid, type, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date"); |
e6826e33 | 851 | p++; |
8e430903 JK |
852 | /* |
853 | * Our timestamp parser is based on the C strto*() functions, which | |
854 | * will happily eat whitespace, including the newline that is supposed | |
855 | * to prevent us walking past the end of the buffer. So do our own | |
856 | * scan, skipping linear whitespace but not newlines, and then | |
857 | * confirming we found a digit. We _could_ be even more strict here, | |
858 | * as we really expect only a single space, but since we have | |
859 | * traditionally allowed extra whitespace, we'll continue to do so. | |
860 | */ | |
861 | while (*p == ' ' || *p == '\t') | |
862 | p++; | |
863 | if (!isdigit(*p)) | |
864 | return report(options, oid, type, FSCK_MSG_BAD_DATE, | |
865 | "invalid author/committer line - bad date"); | |
e6826e33 | 866 | if (*p == '0' && p[1] != ' ') |
78543993 | 867 | return report(options, oid, type, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date"); |
1aeb7e75 | 868 | if (date_overflows(parse_timestamp(p, &end, 10))) |
78543993 | 869 | return report(options, oid, type, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow"); |
e6826e33 | 870 | if ((end == p || *end != ' ')) |
78543993 | 871 | return report(options, oid, type, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date"); |
e6826e33 JS |
872 | p = end + 1; |
873 | if ((*p != '+' && *p != '-') || | |
874 | !isdigit(p[1]) || | |
875 | !isdigit(p[2]) || | |
876 | !isdigit(p[3]) || | |
877 | !isdigit(p[4]) || | |
878 | (p[5] != '\n')) | |
78543993 | 879 | return report(options, oid, type, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone"); |
e6826e33 | 880 | p += 6; |
daae1922 JN |
881 | return 0; |
882 | } | |
883 | ||
c5b4269b JK |
884 | static int fsck_commit(const struct object_id *oid, |
885 | const char *buffer, unsigned long size, | |
886 | struct fsck_options *options) | |
ba002f3b | 887 | { |
f648ee70 | 888 | struct object_id tree_oid, parent_oid; |
ec652315 | 889 | unsigned author_count; |
daae1922 | 890 | int err; |
6d2d780f | 891 | const char *buffer_begin = buffer; |
8e430903 | 892 | const char *buffer_end = buffer + size; |
c54f5ca9 | 893 | const char *p; |
ba002f3b | 894 | |
8e430903 JK |
895 | /* |
896 | * We _must_ stop parsing immediately if this reports failure, as the | |
897 | * memory safety of the rest of the function depends on it. See the | |
898 | * comment above the definition of verify_headers() for more details. | |
899 | */ | |
c5b4269b | 900 | if (verify_headers(buffer, size, oid, OBJ_COMMIT, options)) |
4d0d8975 JS |
901 | return -1; |
902 | ||
8e430903 | 903 | if (buffer >= buffer_end || !skip_prefix(buffer, "tree ", &buffer)) |
c5b4269b | 904 | return report(options, oid, OBJ_COMMIT, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line"); |
c54f5ca9 | 905 | if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') { |
c5b4269b | 906 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1"); |
b3584761 JS |
907 | if (err) |
908 | return err; | |
909 | } | |
c54f5ca9 | 910 | buffer = p + 1; |
8e430903 | 911 | while (buffer < buffer_end && skip_prefix(buffer, "parent ", &buffer)) { |
f648ee70 | 912 | if (parse_oid_hex(buffer, &parent_oid, &p) || *p != '\n') { |
c5b4269b | 913 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1"); |
b3584761 JS |
914 | if (err) |
915 | return err; | |
916 | } | |
c54f5ca9 | 917 | buffer = p + 1; |
ba002f3b | 918 | } |
c9ad147f | 919 | author_count = 0; |
8e430903 | 920 | while (buffer < buffer_end && skip_prefix(buffer, "author ", &buffer)) { |
c9ad147f | 921 | author_count++; |
c5b4269b | 922 | err = fsck_ident(&buffer, oid, OBJ_COMMIT, options); |
c9ad147f JS |
923 | if (err) |
924 | return err; | |
ba002f3b | 925 | } |
c9ad147f | 926 | if (author_count < 1) |
c5b4269b | 927 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line"); |
c9ad147f | 928 | else if (author_count > 1) |
c5b4269b | 929 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines"); |
daae1922 JN |
930 | if (err) |
931 | return err; | |
8e430903 | 932 | if (buffer >= buffer_end || !skip_prefix(buffer, "committer ", &buffer)) |
c5b4269b JK |
933 | return report(options, oid, OBJ_COMMIT, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line"); |
934 | err = fsck_ident(&buffer, oid, OBJ_COMMIT, options); | |
daae1922 JN |
935 | if (err) |
936 | return err; | |
6d2d780f | 937 | if (memchr(buffer_begin, '\0', size)) { |
c5b4269b | 938 | err = report(options, oid, OBJ_COMMIT, FSCK_MSG_NUL_IN_COMMIT, |
6d2d780f JH |
939 | "NUL byte in the commit object body"); |
940 | if (err) | |
941 | return err; | |
942 | } | |
ba002f3b MK |
943 | return 0; |
944 | } | |
945 | ||
103fb6d4 | 946 | static int fsck_tag(const struct object_id *oid, const char *buffer, |
2175a0c6 | 947 | unsigned long size, struct fsck_options *options) |
cec097be | 948 | { |
f648ee70 | 949 | struct object_id tagged_oid; |
acf9de4c ÆAB |
950 | int tagged_type; |
951 | return fsck_tag_standalone(oid, buffer, size, options, &tagged_oid, | |
952 | &tagged_type); | |
953 | } | |
954 | ||
955 | int fsck_tag_standalone(const struct object_id *oid, const char *buffer, | |
956 | unsigned long size, struct fsck_options *options, | |
957 | struct object_id *tagged_oid, | |
958 | int *tagged_type) | |
959 | { | |
cec097be | 960 | int ret = 0; |
23a173a7 | 961 | char *eol; |
cec097be | 962 | struct strbuf sb = STRBUF_INIT; |
8e430903 | 963 | const char *buffer_end = buffer + size; |
c54f5ca9 | 964 | const char *p; |
cec097be | 965 | |
8e430903 JK |
966 | /* |
967 | * We _must_ stop parsing immediately if this reports failure, as the | |
968 | * memory safety of the rest of the function depends on it. See the | |
969 | * comment above the definition of verify_headers() for more details. | |
970 | */ | |
103fb6d4 | 971 | ret = verify_headers(buffer, size, oid, OBJ_TAG, options); |
8a272f29 | 972 | if (ret) |
cec097be JS |
973 | goto done; |
974 | ||
8e430903 | 975 | if (buffer >= buffer_end || !skip_prefix(buffer, "object ", &buffer)) { |
103fb6d4 | 976 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line"); |
cec097be JS |
977 | goto done; |
978 | } | |
acf9de4c | 979 | if (parse_oid_hex(buffer, tagged_oid, &p) || *p != '\n') { |
103fb6d4 | 980 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1"); |
7d7d5b05 JS |
981 | if (ret) |
982 | goto done; | |
cec097be | 983 | } |
c54f5ca9 | 984 | buffer = p + 1; |
cec097be | 985 | |
8e430903 | 986 | if (buffer >= buffer_end || !skip_prefix(buffer, "type ", &buffer)) { |
103fb6d4 | 987 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line"); |
cec097be JS |
988 | goto done; |
989 | } | |
8e430903 | 990 | eol = memchr(buffer, '\n', buffer_end - buffer); |
cec097be | 991 | if (!eol) { |
103fb6d4 | 992 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line"); |
cec097be JS |
993 | goto done; |
994 | } | |
acf9de4c ÆAB |
995 | *tagged_type = type_from_string_gently(buffer, eol - buffer, 1); |
996 | if (*tagged_type < 0) | |
103fb6d4 | 997 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_BAD_TYPE, "invalid 'type' value"); |
cec097be JS |
998 | if (ret) |
999 | goto done; | |
1000 | buffer = eol + 1; | |
1001 | ||
8e430903 | 1002 | if (buffer >= buffer_end || !skip_prefix(buffer, "tag ", &buffer)) { |
103fb6d4 | 1003 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line"); |
cec097be JS |
1004 | goto done; |
1005 | } | |
8e430903 | 1006 | eol = memchr(buffer, '\n', buffer_end - buffer); |
cec097be | 1007 | if (!eol) { |
103fb6d4 | 1008 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line"); |
cec097be JS |
1009 | goto done; |
1010 | } | |
1011 | strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer); | |
f27d05b1 | 1012 | if (check_refname_format(sb.buf, 0)) { |
103fb6d4 | 1013 | ret = report(options, oid, OBJ_TAG, |
38370253 JK |
1014 | FSCK_MSG_BAD_TAG_NAME, |
1015 | "invalid 'tag' name: %.*s", | |
1016 | (int)(eol - buffer), buffer); | |
f27d05b1 JS |
1017 | if (ret) |
1018 | goto done; | |
1019 | } | |
cec097be JS |
1020 | buffer = eol + 1; |
1021 | ||
8e430903 | 1022 | if (buffer >= buffer_end || !skip_prefix(buffer, "tagger ", &buffer)) { |
cec097be | 1023 | /* early tags do not contain 'tagger' lines; warn only */ |
103fb6d4 | 1024 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line"); |
f27d05b1 JS |
1025 | if (ret) |
1026 | goto done; | |
1027 | } | |
cec097be | 1028 | else |
103fb6d4 | 1029 | ret = fsck_ident(&buffer, oid, OBJ_TAG, options); |
cec097be | 1030 | |
8e430903 | 1031 | if (buffer < buffer_end && !starts_with(buffer, "\n")) { |
acf9de4c ÆAB |
1032 | /* |
1033 | * The verify_headers() check will allow | |
1034 | * e.g. "[...]tagger <tagger>\nsome | |
1035 | * garbage\n\nmessage" to pass, thinking "some | |
1036 | * garbage" could be a custom header. E.g. "mktag" | |
1037 | * doesn't want any unknown headers. | |
1038 | */ | |
1039 | ret = report(options, oid, OBJ_TAG, FSCK_MSG_EXTRA_HEADER_ENTRY, "invalid format - extra header(s) after 'tagger'"); | |
1040 | if (ret) | |
1041 | goto done; | |
1042 | } | |
1043 | ||
cec097be JS |
1044 | done: |
1045 | strbuf_release(&sb); | |
cec097be JS |
1046 | return ret; |
1047 | } | |
1048 | ||
ed8b10f6 | 1049 | struct fsck_gitmodules_data { |
6da40b22 | 1050 | const struct object_id *oid; |
ed8b10f6 JK |
1051 | struct fsck_options *options; |
1052 | int ret; | |
1053 | }; | |
1054 | ||
a4e7e317 GC |
1055 | static int fsck_gitmodules_fn(const char *var, const char *value, |
1056 | const struct config_context *ctx UNUSED, | |
1057 | void *vdata) | |
ed8b10f6 JK |
1058 | { |
1059 | struct fsck_gitmodules_data *data = vdata; | |
1060 | const char *subsection, *key; | |
f5914f4b | 1061 | size_t subsection_len; |
ed8b10f6 JK |
1062 | char *name; |
1063 | ||
1064 | if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 || | |
1065 | !subsection) | |
1066 | return 0; | |
1067 | ||
1068 | name = xmemdupz(subsection, subsection_len); | |
1069 | if (check_submodule_name(name) < 0) | |
38370253 | 1070 | data->ret |= report(data->options, |
6da40b22 | 1071 | data->oid, OBJ_BLOB, |
ed8b10f6 JK |
1072 | FSCK_MSG_GITMODULES_NAME, |
1073 | "disallowed submodule name: %s", | |
1074 | name); | |
a124133e | 1075 | if (!strcmp(key, "url") && value && |
07259e74 | 1076 | check_submodule_url(value) < 0) |
38370253 | 1077 | data->ret |= report(data->options, |
6da40b22 | 1078 | data->oid, OBJ_BLOB, |
a124133e JK |
1079 | FSCK_MSG_GITMODULES_URL, |
1080 | "disallowed submodule url: %s", | |
1081 | value); | |
1a7fd1fb JK |
1082 | if (!strcmp(key, "path") && value && |
1083 | looks_like_command_line_option(value)) | |
38370253 | 1084 | data->ret |= report(data->options, |
6da40b22 | 1085 | data->oid, OBJ_BLOB, |
1a7fd1fb JK |
1086 | FSCK_MSG_GITMODULES_PATH, |
1087 | "disallowed submodule path: %s", | |
1088 | value); | |
bb92255e JN |
1089 | if (!strcmp(key, "update") && value && |
1090 | parse_submodule_update_type(value) == SM_UPDATE_COMMAND) | |
7034cd09 | 1091 | data->ret |= report(data->options, data->oid, OBJ_BLOB, |
bb92255e JN |
1092 | FSCK_MSG_GITMODULES_UPDATE, |
1093 | "disallowed submodule update setting: %s", | |
1094 | value); | |
ed8b10f6 JK |
1095 | free(name); |
1096 | ||
1097 | return 0; | |
1098 | } | |
1099 | ||
6da40b22 | 1100 | static int fsck_blob(const struct object_id *oid, const char *buf, |
7ac4f3a0 JK |
1101 | unsigned long size, struct fsck_options *options) |
1102 | { | |
bb3a9265 | 1103 | int ret = 0; |
ed8b10f6 | 1104 | |
6da40b22 | 1105 | if (object_on_skiplist(options, oid)) |
fb162877 RJ |
1106 | return 0; |
1107 | ||
bb3a9265 PS |
1108 | if (oidset_contains(&options->gitmodules_found, oid)) { |
1109 | struct config_options config_opts = { 0 }; | |
1110 | struct fsck_gitmodules_data data; | |
1111 | ||
1112 | oidset_insert(&options->gitmodules_done, oid); | |
1113 | ||
1114 | if (!buf) { | |
1115 | /* | |
1116 | * A missing buffer here is a sign that the caller found the | |
1117 | * blob too gigantic to load into memory. Let's just consider | |
1118 | * that an error. | |
1119 | */ | |
1120 | return report(options, oid, OBJ_BLOB, | |
1121 | FSCK_MSG_GITMODULES_LARGE, | |
1122 | ".gitmodules too large to parse"); | |
1123 | } | |
1124 | ||
1125 | data.oid = oid; | |
1126 | data.options = options; | |
1127 | data.ret = 0; | |
1128 | config_opts.error_action = CONFIG_ERROR_SILENT; | |
1129 | if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB, | |
809d8680 GC |
1130 | ".gitmodules", buf, size, &data, |
1131 | CONFIG_SCOPE_UNKNOWN, &config_opts)) | |
bb3a9265 PS |
1132 | data.ret |= report(options, oid, OBJ_BLOB, |
1133 | FSCK_MSG_GITMODULES_PARSE, | |
1134 | "could not parse gitmodules blob"); | |
1135 | ret |= data.ret; | |
ed8b10f6 JK |
1136 | } |
1137 | ||
27ab4784 PS |
1138 | if (oidset_contains(&options->gitattributes_found, oid)) { |
1139 | const char *ptr; | |
1140 | ||
1141 | oidset_insert(&options->gitattributes_done, oid); | |
1142 | ||
1143 | if (!buf || size > ATTR_MAX_FILE_SIZE) { | |
1144 | /* | |
1145 | * A missing buffer here is a sign that the caller found the | |
1146 | * blob too gigantic to load into memory. Let's just consider | |
1147 | * that an error. | |
1148 | */ | |
1149 | return report(options, oid, OBJ_BLOB, | |
1150 | FSCK_MSG_GITATTRIBUTES_LARGE, | |
1151 | ".gitattributes too large to parse"); | |
1152 | } | |
1153 | ||
1154 | for (ptr = buf; *ptr; ) { | |
1155 | const char *eol = strchrnul(ptr, '\n'); | |
1156 | if (eol - ptr >= ATTR_MAX_LINE_LENGTH) { | |
1157 | ret |= report(options, oid, OBJ_BLOB, | |
1158 | FSCK_MSG_GITATTRIBUTES_LINE_LENGTH, | |
1159 | ".gitattributes has too long lines to parse"); | |
1160 | break; | |
1161 | } | |
1162 | ||
1163 | ptr = *eol ? eol + 1 : eol; | |
1164 | } | |
1165 | } | |
1166 | ||
bb3a9265 | 1167 | return ret; |
7ac4f3a0 JK |
1168 | } |
1169 | ||
90a398bb | 1170 | int fsck_object(struct object *obj, void *data, unsigned long size, |
22410549 | 1171 | struct fsck_options *options) |
ba002f3b MK |
1172 | { |
1173 | if (!obj) | |
38370253 | 1174 | return report(options, NULL, OBJ_NONE, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck"); |
ba002f3b | 1175 | |
35ff327e JK |
1176 | return fsck_buffer(&obj->oid, obj->type, data, size, options); |
1177 | } | |
1178 | ||
1179 | int fsck_buffer(const struct object_id *oid, enum object_type type, | |
1180 | void *data, unsigned long size, | |
1181 | struct fsck_options *options) | |
1182 | { | |
1183 | if (type == OBJ_BLOB) | |
1184 | return fsck_blob(oid, data, size, options); | |
1185 | if (type == OBJ_TREE) | |
1186 | return fsck_tree(oid, data, size, options); | |
1187 | if (type == OBJ_COMMIT) | |
1188 | return fsck_commit(oid, data, size, options); | |
1189 | if (type == OBJ_TAG) | |
1190 | return fsck_tag(oid, data, size, options); | |
1191 | ||
1192 | return report(options, oid, type, | |
38370253 JK |
1193 | FSCK_MSG_UNKNOWN_TYPE, |
1194 | "unknown type '%d' (internal fsck error)", | |
35ff327e | 1195 | type); |
ba002f3b | 1196 | } |
d6ffc8d7 | 1197 | |
1cd772cc | 1198 | int fsck_error_function(struct fsck_options *o, |
5afc4b1d | 1199 | const struct object_id *oid, |
0b4e9013 | 1200 | enum object_type object_type UNUSED, |
394d5d31 | 1201 | enum fsck_msg_type msg_type, |
0b4e9013 | 1202 | enum fsck_msg_id msg_id UNUSED, |
394d5d31 | 1203 | const char *message) |
d6ffc8d7 | 1204 | { |
0282f4dc | 1205 | if (msg_type == FSCK_WARN) { |
5afc4b1d | 1206 | warning("object %s: %s", fsck_describe_object(o, oid), message); |
0282f4dc JS |
1207 | return 0; |
1208 | } | |
5afc4b1d | 1209 | error("object %s: %s", fsck_describe_object(o, oid), message); |
d6ffc8d7 MK |
1210 | return 1; |
1211 | } | |
159e7b08 | 1212 | |
a59a8c68 PS |
1213 | static int fsck_blobs(struct oidset *blobs_found, struct oidset *blobs_done, |
1214 | enum fsck_msg_id msg_missing, enum fsck_msg_id msg_type, | |
1215 | struct fsck_options *options, const char *blob_type) | |
159e7b08 JK |
1216 | { |
1217 | int ret = 0; | |
1218 | struct oidset_iter iter; | |
1219 | const struct object_id *oid; | |
1220 | ||
a59a8c68 | 1221 | oidset_iter_init(blobs_found, &iter); |
159e7b08 | 1222 | while ((oid = oidset_iter_next(&iter))) { |
159e7b08 JK |
1223 | enum object_type type; |
1224 | unsigned long size; | |
1225 | char *buf; | |
1226 | ||
a59a8c68 | 1227 | if (oidset_contains(blobs_done, oid)) |
159e7b08 JK |
1228 | continue; |
1229 | ||
bc726bd0 | 1230 | buf = repo_read_object_file(the_repository, oid, &type, &size); |
159e7b08 | 1231 | if (!buf) { |
b8b00f16 | 1232 | if (is_promisor_object(oid)) |
27387444 | 1233 | continue; |
38370253 | 1234 | ret |= report(options, |
a59a8c68 PS |
1235 | oid, OBJ_BLOB, msg_missing, |
1236 | "unable to read %s blob", blob_type); | |
159e7b08 JK |
1237 | continue; |
1238 | } | |
1239 | ||
1240 | if (type == OBJ_BLOB) | |
b8b00f16 | 1241 | ret |= fsck_blob(oid, buf, size, options); |
159e7b08 | 1242 | else |
a59a8c68 PS |
1243 | ret |= report(options, oid, type, msg_type, |
1244 | "non-blob found at %s", blob_type); | |
159e7b08 JK |
1245 | free(buf); |
1246 | } | |
1247 | ||
a59a8c68 PS |
1248 | oidset_clear(blobs_found); |
1249 | oidset_clear(blobs_done); | |
1250 | ||
1251 | return ret; | |
1252 | } | |
1253 | ||
1254 | int fsck_finish(struct fsck_options *options) | |
1255 | { | |
1256 | int ret = 0; | |
1257 | ||
1258 | ret |= fsck_blobs(&options->gitmodules_found, &options->gitmodules_done, | |
1259 | FSCK_MSG_GITMODULES_MISSING, FSCK_MSG_GITMODULES_BLOB, | |
1260 | options, ".gitmodules"); | |
27ab4784 PS |
1261 | ret |= fsck_blobs(&options->gitattributes_found, &options->gitattributes_done, |
1262 | FSCK_MSG_GITATTRIBUTES_MISSING, FSCK_MSG_GITATTRIBUTES_BLOB, | |
1263 | options, ".gitattributes"); | |
159e7b08 | 1264 | |
159e7b08 JK |
1265 | return ret; |
1266 | } | |
1f3299fd | 1267 | |
a4e7e317 GC |
1268 | int git_fsck_config(const char *var, const char *value, |
1269 | const struct config_context *ctx, void *cb) | |
1f3299fd | 1270 | { |
fb79f5bf | 1271 | struct fsck_options *options = cb; |
d49cb162 JK |
1272 | const char *msg_id; |
1273 | ||
1f3299fd ÆAB |
1274 | if (strcmp(var, "fsck.skiplist") == 0) { |
1275 | const char *path; | |
1276 | struct strbuf sb = STRBUF_INIT; | |
1277 | ||
1278 | if (git_config_pathname(&path, var, value)) | |
1279 | return 1; | |
1280 | strbuf_addf(&sb, "skiplist=%s", path); | |
1281 | free((char *)path); | |
1282 | fsck_set_msg_types(options, sb.buf); | |
1283 | strbuf_release(&sb); | |
1284 | return 0; | |
1285 | } | |
1286 | ||
d49cb162 JK |
1287 | if (skip_prefix(var, "fsck.", &msg_id)) { |
1288 | if (!value) | |
1289 | return config_error_nonbool(var); | |
1290 | fsck_set_msg_type(options, msg_id, value); | |
1f3299fd ÆAB |
1291 | return 0; |
1292 | } | |
1293 | ||
a4e7e317 | 1294 | return git_default_config(var, value, ctx, cb); |
1f3299fd | 1295 | } |
3745e269 ÆAB |
1296 | |
1297 | /* | |
1298 | * Custom error callbacks that are used in more than one place. | |
1299 | */ | |
1300 | ||
1301 | int fsck_error_cb_print_missing_gitmodules(struct fsck_options *o, | |
1302 | const struct object_id *oid, | |
1303 | enum object_type object_type, | |
1304 | enum fsck_msg_type msg_type, | |
1305 | enum fsck_msg_id msg_id, | |
1306 | const char *message) | |
1307 | { | |
1308 | if (msg_id == FSCK_MSG_GITMODULES_MISSING) { | |
1309 | puts(oid_to_hex(oid)); | |
1310 | return 0; | |
1311 | } | |
1312 | return fsck_error_function(o, oid, object_type, msg_type, msg_id, message); | |
1313 | } |