1 /*
2  * Copyright (c) 2017 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 
19 #include <stdio.h>
20 #include <stdlib.h>
21 #include <string.h>
22 #include <sha1.h>
23 #include <zlib.h>
24 #include <ctype.h>
25 #include <limits.h>
26 
27 #include "got_error.h"
28 #include "got_object.h"
29 #include "got_repository.h"
30 #include "got_sha1.h"
31 
32 #ifndef MIN
33 #define	MIN(_a,_b) ((_a) < (_b) ? (_a) : (_b))
34 #endif
35 
36 #ifndef nitems
37 #define nitems(_a) (sizeof(_a) / sizeof((_a)[0]))
38 #endif
39 
40 #define GOT_OBJ_TAG_COMMIT	"commit"
41 #define GOT_OBJ_TAG_TREE	"tree"
42 #define GOT_OBJ_TAG_BLOB	"blob"
43 
44 #define GOT_COMMIT_TAG_TREE		"tree "
45 #define GOT_COMMIT_TAG_PARENT		"parent "
46 #define GOT_COMMIT_TAG_AUTHOR		"author "
47 #define GOT_COMMIT_TAG_COMMITTER	"committer "
48 
49 const char *
50 got_object_id_str(struct got_object_id *id, char *buf, size_t size)
51 {
52 	char *p = buf;
53 	char hex[3];
54 	int i;
55 
56 	if (size < SHA1_DIGEST_STRING_LENGTH)
57 		return NULL;
58 
59 	for (i = 0; i < SHA1_DIGEST_LENGTH; i++) {
60 		snprintf(hex, sizeof(hex), "%.2x", id->sha1[i]);
61 		p[0] = hex[0];
62 		p[1] = hex[1];
63 		p += 2;
64 	}
65 	p[0] = '\0';
66 
67 	return buf;
68 }
69 
70 struct got_zstream_buf {
71 	z_stream z;
72 	char *inbuf;
73 	size_t inlen;
74 	char *outbuf;
75 	size_t outlen;
76 	int flags;
77 #define GOT_ZSTREAM_F_HAVE_MORE 0x01
78 };
79 
80 static void
81 inflate_end(struct got_zstream_buf *zb)
82 {
83 	free(zb->inbuf);
84 	free(zb->outbuf);
85 	inflateEnd(&zb->z);
86 }
87 
88 static const struct got_error *
89 inflate_init(struct got_zstream_buf *zb, size_t bufsize)
90 {
91 	const struct got_error *err = NULL;
92 
93 	memset(zb, 0, sizeof(*zb));
94 
95 	zb->z.zalloc = Z_NULL;
96 	zb->z.zfree = Z_NULL;
97 	if (inflateInit(&zb->z) != Z_OK) {
98 		err = got_error(GOT_ERR_IO);
99 		goto done;
100 	}
101 
102 	zb->inlen = zb->outlen = bufsize;
103 
104 	zb->inbuf = calloc(1, zb->inlen);
105 	if (zb->inbuf == NULL) {
106 		err = got_error(GOT_ERR_NO_MEM);
107 		goto done;
108 	}
109 
110 	zb->outbuf = calloc(1, zb->outlen);
111 	if (zb->outbuf == NULL) {
112 		err = got_error(GOT_ERR_NO_MEM);
113 		goto done;
114 	}
115 
116 done:
117 	if (err)
118 		inflate_end(zb);
119 	return err;
120 }
121 
122 static const struct got_error *
123 inflate_read(struct got_zstream_buf *zb, FILE *f, size_t *outlenp)
124 {
125 	size_t last_total_out = zb->z.total_out;
126 	z_stream *z = &zb->z;
127 	int n, ret;
128 
129 	z->next_out = zb->outbuf;
130 	z->avail_out = zb->outlen;
131 
132 	if (z->avail_in == 0 && (zb->flags & GOT_ZSTREAM_F_HAVE_MORE) == 0) {
133 		int i;
134 		n = fread(zb->inbuf, 1, zb->inlen, f);
135 		if (n == 0) {
136 			if (ferror(f))
137 				return got_error(GOT_ERR_IO);
138 			*outlenp = 0;
139 			return NULL;
140 		}
141 		z->next_in = zb->inbuf;
142 		z->avail_in = n;
143 	}
144 
145 	ret = inflate(z, Z_SYNC_FLUSH);
146 	if (ret == Z_OK) {
147 		if (z->avail_out == 0)
148 			zb->flags |= GOT_ZSTREAM_F_HAVE_MORE;
149 		else
150 			zb->flags &= ~GOT_ZSTREAM_F_HAVE_MORE;
151 	} else if (ret != Z_STREAM_END)
152 		return got_error(GOT_ERR_DECOMPRESSION);
153 
154 	*outlenp = z->total_out - last_total_out;
155 	return NULL;
156 }
157 
158 static const struct got_error *
159 parse_object_header(struct got_object **obj, char *buf, size_t len)
160 {
161 	const char *obj_tags[] = {
162 		GOT_OBJ_TAG_COMMIT,
163 		GOT_OBJ_TAG_TREE,
164 		GOT_OBJ_TAG_BLOB
165 	};
166 	const int obj_types[] = {
167 		GOT_OBJ_TYPE_COMMIT,
168 		GOT_OBJ_TYPE_TREE,
169 		GOT_OBJ_TYPE_BLOB,
170 	};
171 	int type = 0;
172 	size_t size = 0, hdrlen = 0;
173 	int i;
174 	char *p = strchr(buf, '\0');
175 
176 	if (p == NULL)
177 		return got_error(GOT_ERR_BAD_OBJ_HDR);
178 
179 	hdrlen = strlen(buf) + 1 /* '\0' */;
180 
181 	for (i = 0; i < nitems(obj_tags); i++) {
182 		const char *tag = obj_tags[i];
183 		size_t tlen = strlen(tag);
184 		const char *errstr;
185 
186 		if (strncmp(buf, tag, tlen) != 0)
187 			continue;
188 
189 		type = obj_types[i];
190 		if (len <= tlen)
191 			return got_error(GOT_ERR_BAD_OBJ_HDR);
192 		size = strtonum(buf + tlen, 0, LONG_MAX, &errstr);
193 		if (errstr != NULL)
194 			return got_error(GOT_ERR_BAD_OBJ_HDR);
195 		break;
196 	}
197 
198 	if (type == 0)
199 		return got_error(GOT_ERR_BAD_OBJ_HDR);
200 
201 	*obj = calloc(1, sizeof(**obj));
202 	(*obj)->type = type;
203 	(*obj)->hdrlen = hdrlen;
204 	(*obj)->size = size;
205 	return NULL;
206 }
207 
208 static const struct got_error *
209 read_object_header(struct got_object **obj, struct got_repository *repo,
210     const char *path)
211 {
212 	const struct got_error *err;
213 	FILE *f;
214 	struct got_zstream_buf zb;
215 	size_t outlen;
216 	int i, ret;
217 
218 	f = fopen(path, "rb");
219 	if (f == NULL)
220 		return got_error(GOT_ERR_BAD_PATH);
221 
222 	err = inflate_init(&zb, 64);
223 	if (err) {
224 		fclose(f);
225 		return err;
226 	}
227 
228 	err = inflate_read(&zb, f, &outlen);
229 	if (err)
230 		goto done;
231 
232 	err = parse_object_header(obj, zb.outbuf, outlen);
233 done:
234 	inflate_end(&zb);
235 	fclose(f);
236 	return err;
237 }
238 
239 static const struct got_error *
240 object_path(char **path, struct got_object_id *id,
241     struct got_repository *repo)
242 {
243 	const struct got_error *err = NULL;
244 	char hex[SHA1_DIGEST_STRING_LENGTH];
245 	char *path_objects = got_repo_get_path_objects(repo);
246 
247 	if (path_objects == NULL)
248 		return got_error(GOT_ERR_NO_MEM);
249 
250 	got_object_id_str(id, hex, sizeof(hex));
251 
252 	if (asprintf(path, "%s/%.2x/%s", path_objects,
253 	    id->sha1[0], hex + 2) == -1)
254 		err = got_error(GOT_ERR_NO_MEM);
255 
256 	free(path_objects);
257 	return err;
258 }
259 
260 const struct got_error *
261 got_object_open(struct got_object **obj, struct got_repository *repo,
262     struct got_object_id *id)
263 {
264 	const struct got_error *err = NULL;
265 	char *path = NULL;
266 
267 	err = object_path(&path, id, repo);
268 	if (err)
269 		return err;
270 
271 	err = read_object_header(obj, repo, path);
272 	if (err == NULL)
273 		memcpy((*obj)->id.sha1, id->sha1, SHA1_DIGEST_LENGTH);
274 done:
275 	free(path);
276 	return err;
277 }
278 
279 void
280 got_object_close(struct got_object *obj)
281 {
282 	free(obj);
283 }
284 
285 static int
286 commit_object_valid(struct got_commit_object *commit)
287 {
288 	int i;
289 	int n;
290 
291 	if (commit == NULL)
292 		return 0;
293 
294 	n = 0;
295 	for (i = 0; i < SHA1_DIGEST_LENGTH; i++) {
296 		if (commit->tree_id.sha1[i] == 0)
297 			n++;
298 	}
299 	if (n == SHA1_DIGEST_LENGTH)
300 		return 0;
301 
302 	return 1;
303 }
304 
305 static const struct got_error *
306 parse_commit_object(struct got_commit_object **commit, char *buf, size_t len)
307 {
308 	const struct got_error *err = NULL;
309 	char *s = buf;
310 	size_t tlen;
311 	ssize_t remain = (ssize_t)len;
312  
313 	*commit = calloc(1, sizeof(**commit));
314 	if (*commit == NULL)
315 		return got_error(GOT_ERR_NO_MEM);
316 
317 	SIMPLEQ_INIT(&(*commit)->parent_ids);
318 
319 	tlen = strlen(GOT_COMMIT_TAG_TREE);
320 	if (strncmp(s, GOT_COMMIT_TAG_TREE, tlen) == 0) {
321 		remain -= tlen;
322 		if (remain < SHA1_DIGEST_STRING_LENGTH) {
323 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
324 			goto done;
325 		}
326 		s += tlen;
327 		if (!got_parse_sha1_digest((*commit)->tree_id.sha1, s)) {
328 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
329 			goto done;
330 		}
331 		remain -= SHA1_DIGEST_STRING_LENGTH;
332 		s += SHA1_DIGEST_STRING_LENGTH;
333 	} else {
334 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
335 		goto done;
336 	}
337 
338 	tlen = strlen(GOT_COMMIT_TAG_PARENT);
339 	while (strncmp(s, GOT_COMMIT_TAG_PARENT, tlen) == 0) {
340 		struct got_parent_id *pid;
341 
342 		remain -= tlen;
343 		if (remain < SHA1_DIGEST_STRING_LENGTH) {
344 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
345 			goto done;
346 		}	
347 
348 		pid = calloc(1, sizeof(*pid));
349 		if (pid == NULL) {
350 			err = got_error(GOT_ERR_NO_MEM);
351 			goto done;
352 		}
353 		s += tlen;
354 		if (!got_parse_sha1_digest(pid->id.sha1, s)) {
355 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
356 			goto done;
357 		}
358 		SIMPLEQ_INSERT_TAIL(&(*commit)->parent_ids, pid, entry);
359 		(*commit)->nparents++;
360 
361 		s += SHA1_DIGEST_STRING_LENGTH;
362 	}
363 
364 	tlen = strlen(GOT_COMMIT_TAG_AUTHOR);
365 	if (strncmp(s, GOT_COMMIT_TAG_AUTHOR, tlen) == 0) {
366 		char *p;
367 
368 		remain -= tlen;
369 		if (remain <= 0) {
370 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
371 			goto done;
372 		}
373 		s += tlen;
374 		p = strchr(s, '\n');
375 		if (p == NULL) {
376 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
377 			goto done;
378 		}
379 		*p = '\0';
380 		(*commit)->author = strdup(s);
381 		if ((*commit)->author == NULL) {
382 			err = got_error(GOT_ERR_NO_MEM);
383 			goto done;
384 		}
385 		s += strlen((*commit)->author) + 1;
386 	}
387 
388 	tlen = strlen(GOT_COMMIT_TAG_COMMITTER);
389 	if (strncmp(s, GOT_COMMIT_TAG_COMMITTER, tlen) == 0) {
390 		char *p;
391 
392 		remain -= tlen;
393 		if (remain <= 0) {
394 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
395 			goto done;
396 		}
397 		s += tlen;
398 		p = strchr(s, '\n');
399 		if (p == NULL) {
400 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
401 			goto done;
402 		}
403 		*p = '\0';
404 		(*commit)->committer = strdup(s);
405 		if ((*commit)->committer == NULL) {
406 			err = got_error(GOT_ERR_NO_MEM);
407 			goto done;
408 		}
409 		s += strlen((*commit)->committer) + 1;
410 	}
411 
412 	(*commit)->logmsg = strdup(s);
413 done:
414 	if (err)
415 		got_object_commit_close(*commit);
416 	return err;
417 }
418 
419 static const struct got_error *
420 read_commit_object(struct got_commit_object **commit,
421     struct got_repository *repo, struct got_object *obj, const char *path)
422 {
423 	const struct got_error *err = NULL;
424 	FILE *f;
425 	struct got_zstream_buf zb;
426 	size_t len;
427 	char *p;
428 	int i, ret;
429 
430 	f = fopen(path, "rb");
431 	if (f == NULL)
432 		return got_error(GOT_ERR_BAD_PATH);
433 
434 	err = inflate_init(&zb, 8192);
435 	if (err) {
436 		fclose(f);
437 		return err;
438 	}
439 
440 	do {
441 		err = inflate_read(&zb, f, &len);
442 		if (err || len == 0)
443 			break;
444 	} while (len < obj->hdrlen + obj->size);
445 
446 	if (len < obj->hdrlen + obj->size) {
447 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
448 		goto done;
449 	}
450 
451 	/* Skip object header. */
452 	len -= obj->hdrlen;
453 	err = parse_commit_object(commit, zb.outbuf + obj->hdrlen, len);
454 done:
455 	inflate_end(&zb);
456 	fclose(f);
457 	return err;
458 }
459 
460 const struct got_error *
461 got_object_commit_open(struct got_commit_object **commit,
462     struct got_repository *repo, struct got_object *obj)
463 {
464 	const struct got_error *err = NULL;
465 	char *path = NULL;
466 
467 	if (obj->type != GOT_OBJ_TYPE_COMMIT)
468 		return got_error(GOT_ERR_OBJ_TYPE);
469 
470 	err = object_path(&path, &obj->id, repo);
471 	if (err)
472 		return err;
473 
474 	err = read_commit_object(commit, repo, obj, path);
475 	free(path);
476 	return err;
477 }
478 
479 void
480 got_object_commit_close(struct got_commit_object *commit)
481 {
482 	struct got_parent_id *pid;
483 
484 	while (!SIMPLEQ_EMPTY(&commit->parent_ids)) {
485 		pid = SIMPLEQ_FIRST(&commit->parent_ids);
486 		SIMPLEQ_REMOVE_HEAD(&commit->parent_ids, entry);
487 		free(pid);
488 	}
489 
490 	free(commit->author);
491 	free(commit->committer);
492 	free(commit->logmsg);
493 	free(commit);
494 }