1 /*
2  * Copyright (c) 2018 Stefan Sperling <stsp@openbsd.org>
3  *
4  * Permission to use, copy, modify, and distribute this software for any
5  * purpose with or without fee is hereby granted, provided that the above
6  * copyright notice and this permission notice appear in all copies.
7  *
8  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15  */
16 
17 #include <sys/queue.h>
18 #include <sys/stat.h>
19 
20 #include <limits.h>
21 #include <stdlib.h>
22 #include <stdio.h>
23 #include <sha1.h>
24 #include <string.h>
25 #include <zlib.h>
26 
27 #include "got_error.h"
28 #include "got_reference.h"
29 #include "got_repository.h"
30 #include "got_worktree.h"
31 #include "got_object.h"
32 
33 #include "got_lib_path.h"
34 #include "got_lib_delta.h"
35 #include "got_lib_inflate.h"
36 #include "got_lib_object.h"
37 #include "got_lib_pack.h"
38 #include "got_lib_repository.h"
39 #include "got_lib_worktree.h"
40 #include "got_lib_object_idcache.h"
41 
42 #ifndef nitems
43 #define nitems(_a) (sizeof(_a) / sizeof((_a)[0]))
44 #endif
45 
46 #define GOT_GIT_DIR	".git"
47 
48 /* Mandatory files and directories inside the git directory. */
49 #define GOT_OBJECTS_DIR		"objects"
50 #define GOT_REFS_DIR		"refs"
51 #define GOT_HEAD_FILE		"HEAD"
52 
53 /* Other files and directories inside the git directory. */
54 #define GOT_FETCH_HEAD_FILE	"FETCH_HEAD"
55 #define GOT_ORIG_HEAD_FILE	"ORIG_HEAD"
56 #define GOT_OBJECTS_PACK_DIR	"objects/pack"
57 
58 char *
59 got_repo_get_path(struct got_repository *repo)
60 {
61 	return strdup(repo->path);
62 }
63 
64 char *
65 got_repo_get_path_git_dir(struct got_repository *repo)
66 {
67 	return strdup(repo->path_git_dir);
68 }
69 
70 static char *
71 get_path_git_child(struct got_repository *repo, const char *basename)
72 {
73 	char *path_child;
74 	
75 	if (asprintf(&path_child, "%s/%s", repo->path_git_dir,
76 	    basename) == -1)
77 		return NULL;
78 
79 	return path_child;
80 }
81 
82 char *
83 got_repo_get_path_objects(struct got_repository *repo)
84 {
85 	return get_path_git_child(repo, GOT_OBJECTS_DIR);
86 }
87 
88 char *
89 got_repo_get_path_objects_pack(struct got_repository *repo)
90 {
91 	return get_path_git_child(repo, GOT_OBJECTS_PACK_DIR);
92 }
93 
94 char *
95 got_repo_get_path_refs(struct got_repository *repo)
96 {
97 	return get_path_git_child(repo, GOT_REFS_DIR);
98 }
99 
100 static char *
101 get_path_head(struct got_repository *repo)
102 {
103 	return get_path_git_child(repo, GOT_HEAD_FILE);
104 }
105 
106 static int
107 is_git_repo(struct got_repository *repo)
108 {
109 	char *path_git = got_repo_get_path_git_dir(repo);
110 	char *path_objects = got_repo_get_path_objects(repo);
111 	char *path_refs = got_repo_get_path_refs(repo);
112 	char *path_head = get_path_head(repo);
113 	int ret = 0;
114 	struct stat sb;
115 	struct got_reference *head_ref;
116 
117 	if (lstat(path_git, &sb) == -1)
118 		goto done;
119 	if (!S_ISDIR(sb.st_mode))
120 		goto done;
121 
122 	if (lstat(path_objects, &sb) == -1)
123 		goto done;
124 	if (!S_ISDIR(sb.st_mode))
125 		goto done;
126 
127 	if (lstat(path_refs, &sb) == -1)
128 		goto done;
129 	if (!S_ISDIR(sb.st_mode))
130 		goto done;
131 
132 	if (lstat(path_head, &sb) == -1)
133 		goto done;
134 	if (!S_ISREG(sb.st_mode))
135 		goto done;
136 
137 	/* Check if the HEAD reference can be opened. */
138 	if (got_ref_open(&head_ref, repo, GOT_REF_HEAD) != NULL)
139 		goto done;
140 	got_ref_close(head_ref);
141 
142 	ret = 1;
143 done:
144 	free(path_git);
145 	free(path_objects);
146 	free(path_refs);
147 	free(path_head);
148 	return ret;
149 
150 }
151 
152 #ifndef GOT_NO_OBJ_CACHE
153 static const struct got_error *
154 cache_add(struct got_object_cache *cache, struct got_object_id *id, void *item)
155 {
156 	const struct got_error *err = NULL;
157 	struct got_object_cache_entry *ce;
158 	int nelem;
159 
160 	nelem = got_object_idcache_num_elements(cache->idcache);
161 	if (nelem >= cache->size) {
162 		err = got_object_idcache_remove_least_used((void **)&ce,
163 		    cache->idcache);
164 		if (err)
165 			return err;
166 		switch (cache->type) {
167 		case GOT_OBJECT_CACHE_TYPE_OBJ:
168 			got_object_close(ce->data.obj);
169 			break;
170 		case GOT_OBJECT_CACHE_TYPE_TREE:
171 			got_object_tree_close(ce->data.tree);
172 			break;
173 		case GOT_OBJECT_CACHE_TYPE_COMMIT:
174 			got_object_commit_close(ce->data.commit);
175 			break;
176 		}
177 		free(ce);
178 	}
179 
180 	ce = calloc(1, sizeof(*ce));
181 	if (ce == NULL)
182 		return got_error_from_errno();
183 	memcpy(&ce->id, id, sizeof(ce->id));
184 	switch (cache->type) {
185 	case GOT_OBJECT_CACHE_TYPE_OBJ:
186 		ce->data.obj = (struct got_object *)item;
187 		break;
188 	case GOT_OBJECT_CACHE_TYPE_TREE:
189 		ce->data.tree = (struct got_tree_object *)item;
190 		break;
191 	case GOT_OBJECT_CACHE_TYPE_COMMIT:
192 		ce->data.commit = (struct got_commit_object *)item;
193 		break;
194 	}
195 
196 	err = got_object_idcache_add(cache->idcache, id, ce);
197 	if (err) {
198 		if (err->code == GOT_ERR_OBJ_EXISTS) {
199 			free(ce);
200 			err = NULL;
201 		}
202 	}
203 	return err;
204 }
205 #endif
206 
207 const struct got_error *
208 got_repo_cache_object(struct got_repository *repo, struct got_object_id *id,
209     struct got_object *obj)
210 {
211 #ifndef GOT_NO_OBJ_CACHE
212 	const struct got_error *err = NULL;
213 	err = cache_add(&repo->objcache, id, obj);
214 	if (err)
215 		return err;
216 	obj->refcnt++;
217 #endif
218 	return NULL;
219 }
220 
221 struct got_object *
222 got_repo_get_cached_object(struct got_repository *repo,
223     struct got_object_id *id)
224 {
225 	struct got_object_cache_entry *ce;
226 
227 	ce = got_object_idcache_get(repo->objcache.idcache, id);
228 	if (ce) {
229 		repo->objcache.cache_hit++;
230 		return ce->data.obj;
231 	}
232 
233 	repo->objcache.cache_miss++;
234 	return NULL;
235 }
236 
237 const struct got_error *
238 got_repo_cache_tree(struct got_repository *repo, struct got_object_id *id,
239     struct got_tree_object *tree)
240 {
241 #ifndef GOT_NO_OBJ_CACHE
242 	const struct got_error *err = NULL;
243 	err = cache_add(&repo->treecache, id, tree);
244 	if (err)
245 		return err;
246 	tree->refcnt++;
247 #endif
248 	return NULL;
249 }
250 
251 struct got_tree_object *
252 got_repo_get_cached_tree(struct got_repository *repo,
253     struct got_object_id *id)
254 {
255 	struct got_object_cache_entry *ce;
256 
257 	ce = got_object_idcache_get(repo->treecache.idcache, id);
258 	if (ce) {
259 		repo->treecache.cache_hit++;
260 		return ce->data.tree;
261 	}
262 
263 	repo->treecache.cache_miss++;
264 	return NULL;
265 }
266 
267 const struct got_error *
268 got_repo_cache_commit(struct got_repository *repo, struct got_object_id *id,
269     struct got_commit_object *commit)
270 {
271 #ifndef GOT_NO_OBJ_CACHE
272 	const struct got_error *err = NULL;
273 	err = cache_add(&repo->commitcache, id, commit);
274 	if (err)
275 		return err;
276 
277 	commit->refcnt++;
278 #endif
279 	return NULL;
280 }
281 
282 struct got_commit_object *
283 got_repo_get_cached_commit(struct got_repository *repo,
284     struct got_object_id *id)
285 {
286 	struct got_object_cache_entry *ce;
287 
288 	ce = got_object_idcache_get(repo->commitcache.idcache, id);
289 	if (ce) {
290 		repo->commitcache.cache_hit++;
291 		return ce->data.commit;
292 	}
293 
294 	repo->commitcache.cache_miss++;
295 	return NULL;
296 }
297 
298 const struct got_error *
299 got_repo_open(struct got_repository **ret, const char *path)
300 {
301 	struct got_repository *repo = NULL;
302 	const struct got_error *err = NULL;
303 	char *abspath;
304 
305 	if (got_path_is_absolute(path))
306 		abspath = strdup(path);
307 	else
308 		abspath = got_path_get_absolute(path);
309 	if (abspath == NULL)
310 		return got_error(GOT_ERR_BAD_PATH);
311 
312 	repo = calloc(1, sizeof(*repo));
313 	if (repo == NULL) {
314 		err = got_error_from_errno();
315 		goto done;
316 	}
317 
318 	repo->objcache.type = GOT_OBJECT_CACHE_TYPE_OBJ;
319 	repo->objcache.size = GOT_OBJECT_CACHE_SIZE_OBJ;
320 	repo->objcache.idcache = got_object_idcache_alloc(repo->objcache.size);
321 	if (repo->objcache.idcache == NULL) {
322 		err = got_error_from_errno();
323 		goto done;
324 	}
325 
326 	repo->treecache.type = GOT_OBJECT_CACHE_TYPE_TREE;
327 	repo->treecache.size = GOT_OBJECT_CACHE_SIZE_TREE;
328 	repo->treecache.idcache =
329 	    got_object_idcache_alloc(repo->treecache.size);
330 	if (repo->treecache.idcache == NULL) {
331 		err = got_error_from_errno();
332 		goto done;
333 	}
334 
335 	repo->commitcache.type = GOT_OBJECT_CACHE_TYPE_COMMIT;
336 	repo->commitcache.size = GOT_OBJECT_CACHE_SIZE_COMMIT;
337 	repo->commitcache.idcache =
338 	    got_object_idcache_alloc(repo->commitcache.size);
339 	if (repo->commitcache.idcache == NULL) {
340 		err = got_error_from_errno();
341 		goto done;
342 	}
343 
344 	repo->path = got_path_normalize(abspath);
345 	if (repo->path == NULL) {
346 		err = got_error(GOT_ERR_BAD_PATH);
347 		goto done;
348 	}
349 
350 	repo->path_git_dir = strdup(repo->path);
351 	if (repo->path_git_dir == NULL) {
352 		err = got_error_from_errno();
353 		goto done;
354 	}
355 	if (!is_git_repo(repo)) {
356 		free(repo->path_git_dir);
357 		if (asprintf(&repo->path_git_dir, "%s/%s", repo->path,
358 		    GOT_GIT_DIR) == -1) {
359 			err = got_error_from_errno();
360 			goto done;
361 		}
362 		if (!is_git_repo(repo)) {
363 			struct got_worktree *worktree;
364 			if (got_worktree_open(&worktree, repo->path) == NULL) {
365 				free(repo->path_git_dir);
366 				repo->path_git_dir =
367 				    strdup(worktree->repo_path);
368 				if (repo->path_git_dir == NULL) {
369 					err = got_error_from_errno();
370 					goto done;
371 				}
372 				if (!is_git_repo(repo)) {
373 					free(repo->path_git_dir);
374 					if (asprintf(&repo->path_git_dir,
375 					    "%s/%s", worktree->repo_path,
376 					    GOT_GIT_DIR) == -1) {
377 						err = got_error_from_errno();
378 						goto done;
379 					}
380 				}
381 				got_worktree_close(worktree);
382 			}
383 		}
384 		if (!is_git_repo(repo)) {
385 			err = got_error(GOT_ERR_NOT_GIT_REPO);
386 			goto done;
387 		}
388 	}
389 		
390 	*ret = repo;
391 done:
392 	if (err)
393 		got_repo_close(repo);
394 	free(abspath);
395 	return err;
396 }
397 
398 #if 0
399 static void
400 print_cache_stats(struct got_object_cache *cache, const char *name)
401 {
402 	fprintf(stderr, "%s cache: %d elements, %d hits, %d missed\n",
403 	    name, got_object_idcache_num_elements(cache->idcache),
404 	    cache->cache_hit, cache->cache_miss);
405 }
406 
407 void check_refcount(struct got_object_id *id, void *data, void *arg)
408 {
409 	struct got_object_cache *cache = arg;
410 	struct got_object_cache_entry *ce = data;
411 	struct got_object *obj;
412 	struct got_tree_object *tree;
413 	struct got_commit_object *commit;
414 	char *id_str;
415 
416 	if (got_object_id_str(&id_str, id) != NULL)
417 		return;
418 
419 	switch (cache->type) {
420 	case GOT_OBJECT_CACHE_TYPE_OBJ:
421 		obj = ce->data.obj;
422 		if (obj->refcnt == 1)
423 			break;
424 		fprintf(stderr, "object %s has %d unclaimed references\n",
425 		    id_str, obj->refcnt - 1);
426 		break;
427 	case GOT_OBJECT_CACHE_TYPE_TREE:
428 		tree = ce->data.tree;
429 		if (tree->refcnt == 1)
430 			break;
431 		fprintf(stderr, "tree %s has %d unclaimed references\n",
432 		    id_str, tree->refcnt - 1);
433 		break;
434 	case GOT_OBJECT_CACHE_TYPE_COMMIT:
435 		commit = ce->data.commit;
436 		if (commit->refcnt == 1)
437 			break;
438 		fprintf(stderr, "commit %s has %d unclaimed references\n",
439 		    id_str, commit->refcnt);
440 		break;
441 	}
442 	free(id_str);
443 }
444 #endif
445 
446 void
447 got_repo_close(struct got_repository *repo)
448 {
449 	int i;
450 
451 	for (i = 0; i < nitems(repo->packidx_cache); i++) {
452 		if (repo->packidx_cache[i] == NULL)
453 			break;
454 		got_packidx_close(repo->packidx_cache[i]);
455 	}
456 
457 	for (i = 0; i < nitems(repo->packs); i++) {
458 		if (repo->packs[i].path_packfile == NULL)
459 			break;
460 		got_pack_close(&repo->packs[i]);
461 	}
462 
463 	free(repo->path);
464 	free(repo->path_git_dir);
465 
466 #if 0
467 	print_cache_stats(&repo->objcache, "object");
468 	print_cache_stats(&repo->treecache, "tree");
469 	print_cache_stats(&repo->commitcache, "commit");
470 	got_object_idcache_for_each(repo->objcache.idcache, check_refcount,
471 	    &repo->objcache);
472 	got_object_idcache_for_each(repo->treecache.idcache, check_refcount,
473 	    &repo->treecache);
474 	got_object_idcache_for_each(repo->commitcache.idcache, check_refcount,
475 	    &repo->commitcache);
476 #endif
477 
478 	if (repo->objcache.idcache)
479 		got_object_idcache_free(repo->objcache.idcache);
480 	if (repo->treecache.idcache)
481 		got_object_idcache_free(repo->treecache.idcache);
482 	if (repo->commitcache.idcache)
483 		got_object_idcache_free(repo->commitcache.idcache);
484 	free(repo);
485 }