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