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 #define GOT_OBJECT_ID_HEX_MAXLEN SHA1_DIGEST_STRING_LENGTH
18 
19 enum got_hash_algorithm {
20 	GOT_HASH_SHA1,
21 	GOT_HASH_SHA256,
22 };
23 
24 struct got_object_id {
25 	u_int8_t sha1[SHA1_DIGEST_LENGTH];
26 };
27 
28 struct got_blob_object;
29 struct got_tree_object;
30 struct got_tree_entry;
31 struct got_tag_object;
32 struct got_commit_object;
33 
34 struct got_object_qid {
35 	STAILQ_ENTRY(got_object_qid) entry;
36 	struct got_object_id id;
37 	void *data; /* managed by API user */
38 };
39 
40 STAILQ_HEAD(got_object_id_queue, got_object_qid);
41 
42 const struct got_error *got_object_qid_alloc(struct got_object_qid **,
43     struct got_object_id *);
44 void got_object_qid_free(struct got_object_qid *);
45 void got_object_id_queue_free(struct got_object_id_queue *);
46 
47 /*
48  * Deep-copy elements from ID queue src to ID queue dest. Do not copy any
49  * qid->data pointers! This is the caller's responsibility if needed.
50  */
51 const struct got_error *got_object_id_queue_copy(
52     const struct got_object_id_queue *src, struct got_object_id_queue *dest);
53 
54 /* Object types. */
55 #define GOT_OBJ_TYPE_ANY		0 /* wildcard value at run-time */
56 #define GOT_OBJ_TYPE_COMMIT		1
57 #define GOT_OBJ_TYPE_TREE		2
58 #define GOT_OBJ_TYPE_BLOB		3
59 #define GOT_OBJ_TYPE_TAG		4
60 /* 5 is reserved */
61 #define GOT_OBJ_TYPE_OFFSET_DELTA	6
62 #define GOT_OBJ_TYPE_REF_DELTA		7
63 
64 /*
65  * Labels used in object data.
66  */
67 
68 #define GOT_OBJ_LABEL_COMMIT	"commit"
69 #define GOT_OBJ_LABEL_TREE	"tree"
70 #define GOT_OBJ_LABEL_BLOB	"blob"
71 #define GOT_OBJ_LABEL_TAG	"tag"
72 
73 #define GOT_COMMIT_LABEL_TREE		"tree "
74 #define GOT_COMMIT_LABEL_PARENT		"parent "
75 #define GOT_COMMIT_LABEL_AUTHOR		"author "
76 #define GOT_COMMIT_LABEL_COMMITTER	"committer "
77 
78 #define GOT_TAG_LABEL_OBJECT		"object "
79 #define GOT_TAG_LABEL_TYPE		"type "
80 #define GOT_TAG_LABEL_TAG		"tag "
81 #define GOT_TAG_LABEL_TAGGER		"tagger "
82 
83 struct got_repository;
84 
85 /*
86  * Obtain a string representation of an object ID. The output depends on
87  * the hash function used by the repository format (currently SHA1).
88  */
89 const struct got_error *got_object_id_str(char **, struct got_object_id *);
90 
91 /*
92  * Compare two object IDs. Return value behaves like memcmp(3).
93  */
94 int got_object_id_cmp(const struct got_object_id *,
95     const struct got_object_id *);
96 
97 /*
98  * Created a newly allocated copy of an object ID.
99  * The caller should dispose of it with free(3).
100  */
101 struct got_object_id *got_object_id_dup(struct got_object_id *);
102 
103 /*
104  * Get a newly allocated ID of the object which resides at the specified
105  * path in the specified tree.
106  * The caller should dispose of it with free(3).
107  */
108 const struct got_error *got_object_tree_find_path(struct got_object_id **id,
109     mode_t *mode, struct got_repository *repo, struct got_tree_object *tree,
110     const char *path);
111 
112 /*
113  * Get a newly allocated ID of the object which resides at the specified
114  * path in the tree of the specified commit.
115  * The caller should dispose of it with free(3).
116  */
117 const struct got_error *got_object_id_by_path(struct got_object_id **,
118     struct got_repository *, struct got_commit_object *, const char *);
119 
120 /*
121  * Obtain the type of an object.
122  * Returns one of the GOT_OBJ_TYPE_x values (see above).
123  */
124 const struct got_error *got_object_get_type(int *, struct got_repository *,
125     struct got_object_id *);
126 
127 /*
128  * Attempt to resolve the textual representation of an object ID
129  * to the ID of an existing object in the repository.
130  * The caller should dispose of the ID with free(3).
131  */
132 const struct got_error *got_object_resolve_id_str(struct got_object_id **,
133     struct got_repository *, const char *);
134 
135 /*
136  * Attempt to open a commit object in a repository.
137  * The caller must dispose of the commit with got_object_commit_close().
138  */
139 const struct got_error *got_object_open_as_commit(struct got_commit_object **,
140     struct got_repository *, struct got_object_id *);
141 
142 /* Dispose of a commit object. */
143 void got_object_commit_close(struct got_commit_object *);
144 
145 /* Obtain the ID of the tree created in a commit. */
146 struct got_object_id *got_object_commit_get_tree_id(struct got_commit_object *);
147 
148 /* Obtain the number of parent commits of a commit. */
149 int got_object_commit_get_nparents(struct got_commit_object *);
150 
151 /* Obtain the list of parent commits of a commit. */
152 const struct got_object_id_queue *got_object_commit_get_parent_ids(
153     struct got_commit_object *);
154 
155 /* Get the author's name and email address. */
156 const char *got_object_commit_get_author(struct got_commit_object *);
157 
158 /* Get an author's commit timestamp in UTC. */
159 time_t got_object_commit_get_author_time(struct got_commit_object *);
160 
161 /* Get an author's timezone offset. */
162 time_t got_object_commit_get_author_gmtoff(struct got_commit_object *);
163 
164 /* Get the committer's name and email address. */
165 const char *got_object_commit_get_committer(struct got_commit_object *);
166 
167 /* Get a committer's commit timestamp in UTC. */
168 time_t got_object_commit_get_committer_time(struct got_commit_object *);
169 
170 /* Get a committer's timezone offset. */
171 time_t got_object_commit_get_committer_gmtoff(struct got_commit_object *);
172 
173 /*
174  * Get the commit log message.
175  * PGP-signatures contained in the log message will be stripped.
176  * The caller must dispose of it with free(3).
177  */
178 const struct got_error *got_object_commit_get_logmsg(char **,
179     struct got_commit_object *);
180 
181 /* Get the raw commit log message.*/
182 const char *got_object_commit_get_logmsg_raw(struct got_commit_object *);
183 
184 /*
185  * Attempt to open a tree object in a repository.
186  * The caller must dispose of the tree with got_object_tree_close().
187  */
188 const struct got_error *got_object_open_as_tree(struct got_tree_object **,
189     struct got_repository *, struct got_object_id *);
190 
191 /* Dispose of a tree object. */
192 void got_object_tree_close(struct got_tree_object *);
193 
194 /* Get the number of entries in this tree object. */
195 int got_object_tree_get_nentries(struct got_tree_object *);
196 
197 /* Get the first tree entry from a tree, or NULL if there is none. */
198 struct got_tree_entry *got_object_tree_get_first_entry(
199     struct got_tree_object *);
200 
201 /* Get the last tree entry from a tree, or NULL if there is none. */
202 struct got_tree_entry *got_object_tree_get_last_entry(struct got_tree_object *);
203 
204 /* Get the entry with the specified index from a tree object. */
205 struct got_tree_entry *got_object_tree_get_entry(
206     struct got_tree_object *, int);
207 
208 /* Find a particular entry in a tree by name. */
209 struct got_tree_entry *got_object_tree_find_entry(
210     struct got_tree_object *, const char *);
211 
212 /* Get the file permission mode of a tree entry. */
213 mode_t got_tree_entry_get_mode(struct got_tree_entry *);
214 
215 /* Get the name of a tree entry. */
216 const char *got_tree_entry_get_name(struct got_tree_entry *);
217 
218 /* Get the object ID of a tree entry. */
219 struct got_object_id *got_tree_entry_get_id(struct got_tree_entry *);
220 
221 /*
222  * Get a string containing the target path of a given a symlink tree entry.
223  * The caller should dispose of it with free(3).
224  */
225 const struct got_error *got_tree_entry_get_symlink_target(char **,
226     struct got_tree_entry *, struct got_repository *);
227 
228 /* Get the index of a tree entry. */
229 int got_tree_entry_get_index(struct got_tree_entry *);
230 
231 /* Get the next tree entry from a tree, or NULL if there is none. */
232 struct got_tree_entry *got_tree_entry_get_next(struct got_tree_object *,
233     struct got_tree_entry *);
234 
235 /* Get the previous tree entry from a tree, or NULL if there is none. */
236 struct got_tree_entry *got_tree_entry_get_prev(struct got_tree_object *,
237     struct got_tree_entry *);
238 
239 /* Return non-zero if the specified tree entry is a Git submodule. */
240 int got_object_tree_entry_is_submodule(struct got_tree_entry *);
241 
242 /* Return non-zero if the specified tree entry is a symbolic link. */
243 int got_object_tree_entry_is_symlink(struct got_tree_entry *);
244 
245 /*
246  * Resolve an in-repository symlink at the specified path in the tree
247  * corresponding to the specified commit. If the specified path is not
248  * a symlink then set *link_target to NULL.
249  * Otherwise, resolve symlinks recursively and return the final link
250  * target path. The caller must dispose of it with free(3).
251  */
252 const struct got_error *got_object_resolve_symlinks(char **, const char *,
253     struct got_commit_object *, struct got_repository *);
254 
255 /*
256  * Compare two trees and indicate whether the entry at the specified path
257  * differs between them. The path must not be the root path "/"; the function
258  * got_object_id_cmp() should be used instead to compare the tree roots.
259  */
260 const struct got_error *got_object_tree_path_changed(int *,
261     struct got_tree_object *, struct got_tree_object *, const char *,
262     struct got_repository *);
263 
264 /*
265  * Attempt to open a blob object in a repository.
266  * The size_t argument specifies the block size of an associated read buffer.
267  * The caller must dispose of the blob with got_object_blob_close().
268  */
269 const struct got_error *got_object_open_as_blob(struct got_blob_object **,
270     struct got_repository *, struct got_object_id *, size_t, int);
271 
272 /* Dispose of a blob object. */
273 const struct got_error *got_object_blob_close(struct got_blob_object *);
274 
275 /*
276  * Get the length of header data at the beginning of the blob's read buffer.
277  * Note that header data is only present upon the first invocation of
278  * got_object_blob_read_block() after the blob is opened.
279  */
280 size_t got_object_blob_get_hdrlen(struct got_blob_object *);
281 
282 /*
283  * Get a pointer to the blob's read buffer.
284  * The read buffer is filled by got_object_blob_read_block().
285  */
286 const uint8_t *got_object_blob_get_read_buf(struct got_blob_object *);
287 
288 /*
289  * Read the next chunk of data from a blob, up to the blob's read buffer
290  * block size. The size_t output argument indicates how many bytes have
291  * been read into the blob's read buffer. Zero bytes will be reported if
292  * all data in the blob has been read.
293  */
294 const struct got_error *got_object_blob_read_block(size_t *,
295     struct got_blob_object *);
296 
297 /* Rewind an open blob's data stream back to the beginning. */
298 void got_object_blob_rewind(struct got_blob_object *);
299 
300 /*
301  * Heuristic to check whether the blob contains binary data.  Rewinds
302  * the blob's data stream back after the header.
303  */
304 const struct got_error *got_object_blob_is_binary(int *,
305     struct got_blob_object *);
306 
307 /*
308  * getline(3) for blobs.
309  */
310 const struct got_error *got_object_blob_getline(char **, ssize_t *,
311     size_t *, struct got_blob_object *);
312 
313 /*
314  * Read the entire content of a blob and write it to the specified file.
315  * Flush and rewind the file as well. Indicate the amount of bytes
316  * written in the size_t output argument, and the number of lines in the
317  * file in the int argument, and line offsets in the off_t argument
318  * (NULL can be passed for any output argument).
319  */
320 const struct got_error *got_object_blob_dump_to_file(off_t *, int *,
321     off_t **, FILE *, struct got_blob_object *);
322 
323 /*
324  * Read the entire content of a blob into a newly allocated string buffer
325  * and terminate it with '\0'. This is intended for blobs which contain a
326  * symlink target path. It should not be used to process arbitrary blobs.
327  * Use got_object_blob_dump_to_file() or got_tree_entry_get_symlink_target()
328  * instead if possible. The caller must dispose of the string with free(3).
329  */
330 const struct got_error *got_object_blob_read_to_str(char **,
331     struct got_blob_object *);
332 
333 /*
334  * Attempt to open a tag object in a repository.
335  * The caller must dispose of the tree with got_tag_object_close().
336  */
337 const struct got_error *got_object_open_as_tag(struct got_tag_object **,
338     struct got_repository *, struct got_object_id *);
339 
340 /* Dispose of a tag object. */
341 void got_object_tag_close(struct got_tag_object *);
342 
343 /* Get the name of a tag. */
344 const char *got_object_tag_get_name(struct got_tag_object *);
345 
346 /* Get type of the object a tag points to. */
347 int got_object_tag_get_object_type(struct got_tag_object *);
348 
349 /*
350  * Get ID of the object a tag points to.
351  * This must not be freed by the caller. Use got_object_id_dup() if needed.
352  */
353 struct got_object_id *got_object_tag_get_object_id(struct got_tag_object *);
354 
355 
356 /* Get the timestamp of the tag. */
357 time_t got_object_tag_get_tagger_time(struct got_tag_object *);
358 
359 /* Get the tag's timestamp's GMT offset.  */
360 time_t got_object_tag_get_tagger_gmtoff(struct got_tag_object *);
361 
362 /* Get the author of the tag. */
363 const char *got_object_tag_get_tagger(struct got_tag_object *);
364 
365 /* Get the tag message associated with the tag. */
366 const char *got_object_tag_get_message(struct got_tag_object *);
367 
368 const struct got_error *got_object_commit_add_parent(struct got_commit_object *,
369     const char *);
370 
371 /* Create a new tag object in the repository. */
372 const struct got_error *got_object_tag_create(struct got_object_id **,
373     const char *, struct got_object_id *, const char *,
374     time_t, const char *, const char *, struct got_repository *, int verbosity);
375 
376 /* Increment commit object reference counter. */
377 void got_object_commit_retain(struct got_commit_object *);