2 * Copyright (c) 2018 Stefan Sperling <stsp@openbsd.org>
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.
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.
18 #include <sys/queue.h>
29 #include "got_error.h"
30 #include "got_object.h"
31 #include "got_repository.h"
39 #define MIN(_a,_b) ((_a) < (_b) ? (_a) : (_b))
43 #define nitems(_a) (sizeof(_a) / sizeof((_a)[0]))
46 #define GOT_OBJ_TAG_COMMIT "commit"
47 #define GOT_OBJ_TAG_TREE "tree"
48 #define GOT_OBJ_TAG_BLOB "blob"
50 #define GOT_COMMIT_TAG_TREE "tree "
51 #define GOT_COMMIT_TAG_PARENT "parent "
52 #define GOT_COMMIT_TAG_AUTHOR "author "
53 #define GOT_COMMIT_TAG_COMMITTER "committer "
56 got_object_id_str(struct got_object_id *id, char *buf, size_t size)
58 return got_sha1_digest_to_str(id->sha1, buf, size);
61 const struct got_error *
62 got_parse_object_id(struct got_object_id **id, const char *buf)
64 *id = calloc(1, sizeof(**id));
66 return got_error(GOT_ERR_NO_MEM);
67 if (!got_parse_sha1_digest((*id)->sha1, buf)) {
70 return got_error(GOT_ERR_BAD_OBJ_ID_STR);
77 got_object_id_cmp(struct got_object_id *id1, struct got_object_id *id2)
79 return memcmp(id1->sha1, id2->sha1, SHA1_DIGEST_LENGTH);
83 got_object_get_type(struct got_object *obj)
86 case GOT_OBJ_TYPE_COMMIT:
87 case GOT_OBJ_TYPE_TREE:
88 case GOT_OBJ_TYPE_BLOB:
89 case GOT_OBJ_TYPE_TAG:
100 static const struct got_error *
101 parse_object_header(struct got_object **obj, char *buf, size_t len)
103 const char *obj_tags[] = {
108 const int obj_types[] = {
114 size_t size = 0, hdrlen = 0;
116 char *p = strchr(buf, '\0');
119 return got_error(GOT_ERR_BAD_OBJ_HDR);
121 hdrlen = strlen(buf) + 1 /* '\0' */;
123 for (i = 0; i < nitems(obj_tags); i++) {
124 const char *tag = obj_tags[i];
125 size_t tlen = strlen(tag);
128 if (strncmp(buf, tag, tlen) != 0)
133 return got_error(GOT_ERR_BAD_OBJ_HDR);
134 size = strtonum(buf + tlen, 0, LONG_MAX, &errstr);
136 return got_error(GOT_ERR_BAD_OBJ_HDR);
141 return got_error(GOT_ERR_BAD_OBJ_HDR);
143 *obj = calloc(1, sizeof(**obj));
145 return got_error(GOT_ERR_NO_MEM);
147 (*obj)->hdrlen = hdrlen;
152 static const struct got_error *
153 read_object_header(struct got_object **obj, struct got_repository *repo,
156 const struct got_error *err;
157 struct got_zstream_buf zb;
160 const size_t zbsize = 64;
161 size_t outlen, totlen;
164 buf = calloc(zbsize, sizeof(char));
166 return got_error(GOT_ERR_NO_MEM);
168 err = got_inflate_init(&zb, zbsize);
175 err = got_inflate_read(&zb, f, &outlen);
178 if (strchr(zb.outbuf, '\0') == NULL) {
179 buf = recallocarray(buf, 1 + i, 2 + i, zbsize);
181 err = got_error(GOT_ERR_NO_MEM);
185 memcpy(buf + totlen, zb.outbuf, outlen);
188 } while (strchr(zb.outbuf, '\0') == NULL);
190 err = parse_object_header(obj, buf, totlen);
192 got_inflate_end(&zb);
196 static const struct got_error *
197 object_path(char **path, struct got_object_id *id, struct got_repository *repo)
199 const struct got_error *err = NULL;
200 char hex[SHA1_DIGEST_STRING_LENGTH];
201 char *path_objects = got_repo_get_path_objects(repo);
203 if (path_objects == NULL)
204 return got_error(GOT_ERR_NO_MEM);
206 got_object_id_str(id, hex, sizeof(hex));
208 if (asprintf(path, "%s/%.2x/%s", path_objects,
209 id->sha1[0], hex + 2) == -1)
210 err = got_error(GOT_ERR_NO_MEM);
216 static const struct got_error *
217 open_loose_object(FILE **f, struct got_object *obj, struct got_repository *repo)
219 const struct got_error *err = NULL;
222 err = object_path(&path, &obj->id, repo);
225 *f = fopen(path, "rb");
227 err = got_error_from_errno();
235 const struct got_error *
236 got_object_open(struct got_object **obj, struct got_repository *repo,
237 struct got_object_id *id)
239 const struct got_error *err = NULL;
243 err = object_path(&path, id, repo);
247 f = fopen(path, "rb");
249 if (errno != ENOENT) {
250 err = got_error_from_errno();
253 err = got_packfile_open_object(obj, id, repo);
257 err = got_error(GOT_ERR_NO_OBJ);
259 err = read_object_header(obj, repo, f);
262 memcpy((*obj)->id.sha1, id->sha1, SHA1_DIGEST_LENGTH);
273 got_object_close(struct got_object *obj)
275 if (obj->flags & GOT_OBJ_FLAG_DELTIFIED) {
276 struct got_delta *delta;
277 while (!SIMPLEQ_EMPTY(&obj->deltas.entries)) {
278 delta = SIMPLEQ_FIRST(&obj->deltas.entries);
279 SIMPLEQ_REMOVE_HEAD(&obj->deltas.entries, entry);
280 got_delta_close(delta);
283 if (obj->flags & GOT_OBJ_FLAG_PACKED)
284 free(obj->path_packfile);
289 commit_object_valid(struct got_commit_object *commit)
298 for (i = 0; i < SHA1_DIGEST_LENGTH; i++) {
299 if (commit->tree_id->sha1[i] == 0)
302 if (n == SHA1_DIGEST_LENGTH)
308 static const struct got_error *
309 parse_commit_object(struct got_commit_object **commit, char *buf, size_t len)
311 const struct got_error *err = NULL;
314 ssize_t remain = (ssize_t)len;
316 *commit = calloc(1, sizeof(**commit));
318 return got_error(GOT_ERR_NO_MEM);
319 (*commit)->tree_id = calloc(1, sizeof(*(*commit)->tree_id));
320 if ((*commit)->tree_id == NULL) {
323 return got_error(GOT_ERR_NO_MEM);
326 SIMPLEQ_INIT(&(*commit)->parent_ids);
328 tlen = strlen(GOT_COMMIT_TAG_TREE);
329 if (strncmp(s, GOT_COMMIT_TAG_TREE, tlen) == 0) {
331 if (remain < SHA1_DIGEST_STRING_LENGTH) {
332 err = got_error(GOT_ERR_BAD_OBJ_DATA);
336 if (!got_parse_sha1_digest((*commit)->tree_id->sha1, s)) {
337 err = got_error(GOT_ERR_BAD_OBJ_DATA);
340 remain -= SHA1_DIGEST_STRING_LENGTH;
341 s += SHA1_DIGEST_STRING_LENGTH;
343 err = got_error(GOT_ERR_BAD_OBJ_DATA);
347 tlen = strlen(GOT_COMMIT_TAG_PARENT);
348 while (strncmp(s, GOT_COMMIT_TAG_PARENT, tlen) == 0) {
349 struct got_parent_id *pid;
352 if (remain < SHA1_DIGEST_STRING_LENGTH) {
353 err = got_error(GOT_ERR_BAD_OBJ_DATA);
357 pid = calloc(1, sizeof(*pid));
359 err = got_error(GOT_ERR_NO_MEM);
362 pid->id = calloc(1, sizeof(*pid->id));
363 if (pid->id == NULL) {
365 err = got_error(GOT_ERR_NO_MEM);
369 if (!got_parse_sha1_digest(pid->id->sha1, s)) {
370 err = got_error(GOT_ERR_BAD_OBJ_DATA);
375 SIMPLEQ_INSERT_TAIL(&(*commit)->parent_ids, pid, entry);
376 (*commit)->nparents++;
378 remain -= SHA1_DIGEST_STRING_LENGTH;
379 s += SHA1_DIGEST_STRING_LENGTH;
382 tlen = strlen(GOT_COMMIT_TAG_AUTHOR);
383 if (strncmp(s, GOT_COMMIT_TAG_AUTHOR, tlen) == 0) {
388 err = got_error(GOT_ERR_BAD_OBJ_DATA);
394 err = got_error(GOT_ERR_BAD_OBJ_DATA);
398 (*commit)->author = strdup(s);
399 if ((*commit)->author == NULL) {
400 err = got_error(GOT_ERR_NO_MEM);
403 s += strlen((*commit)->author) + 1;
404 remain -= strlen((*commit)->author) + 1;
407 tlen = strlen(GOT_COMMIT_TAG_COMMITTER);
408 if (strncmp(s, GOT_COMMIT_TAG_COMMITTER, tlen) == 0) {
413 err = got_error(GOT_ERR_BAD_OBJ_DATA);
419 err = got_error(GOT_ERR_BAD_OBJ_DATA);
423 (*commit)->committer = strdup(s);
424 if ((*commit)->committer == NULL) {
425 err = got_error(GOT_ERR_NO_MEM);
428 s += strlen((*commit)->committer) + 1;
429 remain -= strlen((*commit)->committer) + 1;
432 (*commit)->logmsg = strndup(s, remain);
433 if ((*commit)->logmsg == NULL) {
434 err = got_error(GOT_ERR_NO_MEM);
439 got_object_commit_close(*commit);
446 tree_entry_close(struct got_tree_entry *te)
453 static const struct got_error *
454 parse_tree_entry(struct got_tree_entry **te, size_t *elen, char *buf,
457 char *p = buf, *space;
458 const struct got_error *err = NULL;
459 char hex[SHA1_DIGEST_STRING_LENGTH];
461 *te = calloc(1, sizeof(**te));
463 return got_error(GOT_ERR_NO_MEM);
465 (*te)->id = calloc(1, sizeof(*(*te)->id));
466 if ((*te)->id == NULL) {
469 return got_error(GOT_ERR_NO_MEM);
472 *elen = strlen(buf) + 1;
473 if (*elen > maxlen) {
476 return got_error(GOT_ERR_BAD_OBJ_DATA);
479 space = strchr(buf, ' ');
483 return got_error(GOT_ERR_BAD_OBJ_DATA);
486 if (*p < '0' && *p > '7') {
487 err = got_error(GOT_ERR_BAD_OBJ_DATA);
491 (*te)->mode |= *p - '0';
495 (*te)->name = strdup(space + 1);
496 if (*elen > maxlen || maxlen - *elen < SHA1_DIGEST_LENGTH) {
497 err = got_error(GOT_ERR_BAD_OBJ_DATA);
500 buf += strlen(buf) + 1;
501 memcpy((*te)->id->sha1, buf, SHA1_DIGEST_LENGTH);
502 *elen += SHA1_DIGEST_LENGTH;
505 tree_entry_close(*te);
511 static const struct got_error *
512 parse_tree_object(struct got_tree_object **tree, struct got_repository *repo,
513 char *buf, size_t len)
515 const struct got_error *err;
519 *tree = calloc(1, sizeof(**tree));
521 return got_error(GOT_ERR_NO_MEM);
523 SIMPLEQ_INIT(&(*tree)->entries);
526 struct got_tree_entry *te;
529 err = parse_tree_entry(&te, &elen, buf, remain);
533 SIMPLEQ_INSERT_TAIL(&(*tree)->entries, te, entry);
539 got_object_tree_close(*tree);
540 return got_error(GOT_ERR_BAD_OBJ_DATA);
546 static const struct got_error *
547 read_to_mem(uint8_t **outbuf, size_t *outlen, FILE *f)
549 const struct got_error *err = NULL;
550 static const size_t blocksize = 512;
551 size_t n, total, remain;
557 buf = calloc(1, blocksize);
559 return got_error(GOT_ERR_NO_MEM);
566 newbuf = reallocarray(buf, 1, total + blocksize);
567 if (newbuf == NULL) {
568 err = got_error(GOT_ERR_NO_MEM);
574 n = fread(buf, 1, remain, f);
577 err = got_ferror(f, GOT_ERR_IO);
595 static const struct got_error *
596 read_commit_object(struct got_commit_object **commit,
597 struct got_repository *repo, struct got_object *obj, FILE *f)
599 const struct got_error *err = NULL;
604 if (obj->flags & GOT_OBJ_FLAG_PACKED)
605 err = read_to_mem(&p, &len, f);
607 err = got_inflate_to_mem(&p, &len, f);
611 if (len < obj->hdrlen + obj->size) {
612 err = got_error(GOT_ERR_BAD_OBJ_DATA);
616 /* Skip object header. */
618 err = parse_commit_object(commit, p + obj->hdrlen, len);
624 const struct got_error *
625 got_object_commit_open(struct got_commit_object **commit,
626 struct got_repository *repo, struct got_object *obj)
628 const struct got_error *err = NULL;
631 if (obj->type != GOT_OBJ_TYPE_COMMIT)
632 return got_error(GOT_ERR_OBJ_TYPE);
634 if (obj->flags & GOT_OBJ_FLAG_PACKED)
635 err = got_packfile_extract_object(&f, obj, repo);
637 err = open_loose_object(&f, obj, repo);
641 err = read_commit_object(commit, repo, obj, f);
647 got_object_commit_close(struct got_commit_object *commit)
649 struct got_parent_id *pid;
651 while (!SIMPLEQ_EMPTY(&commit->parent_ids)) {
652 pid = SIMPLEQ_FIRST(&commit->parent_ids);
653 SIMPLEQ_REMOVE_HEAD(&commit->parent_ids, entry);
658 free(commit->tree_id);
659 free(commit->author);
660 free(commit->committer);
661 free(commit->logmsg);
665 static const struct got_error *
666 read_tree_object(struct got_tree_object **tree,
667 struct got_repository *repo, struct got_object *obj, FILE *f)
669 const struct got_error *err = NULL;
674 if (obj->flags & GOT_OBJ_FLAG_PACKED)
675 err = read_to_mem(&p, &len, f);
677 err = got_inflate_to_mem(&p, &len, f);
681 if (len < obj->hdrlen + obj->size) {
682 err = got_error(GOT_ERR_BAD_OBJ_DATA);
686 /* Skip object header. */
688 err = parse_tree_object(tree, repo, p + obj->hdrlen, len);
694 const struct got_error *
695 got_object_tree_open(struct got_tree_object **tree,
696 struct got_repository *repo, struct got_object *obj)
698 const struct got_error *err = NULL;
701 if (obj->type != GOT_OBJ_TYPE_TREE)
702 return got_error(GOT_ERR_OBJ_TYPE);
704 if (obj->flags & GOT_OBJ_FLAG_PACKED)
705 err = got_packfile_extract_object(&f, obj, repo);
707 err = open_loose_object(&f, obj, repo);
711 err = read_tree_object(tree, repo, obj, f);
717 got_object_tree_close(struct got_tree_object *tree)
719 struct got_tree_entry *te;
721 while (!SIMPLEQ_EMPTY(&tree->entries)) {
722 te = SIMPLEQ_FIRST(&tree->entries);
723 SIMPLEQ_REMOVE_HEAD(&tree->entries, entry);
724 tree_entry_close(te);
730 const struct got_error *
731 got_object_blob_open(struct got_blob_object **blob,
732 struct got_repository *repo, struct got_object *obj, size_t blocksize)
734 const struct got_error *err = NULL;
736 if (obj->type != GOT_OBJ_TYPE_BLOB)
737 return got_error(GOT_ERR_OBJ_TYPE);
739 if (blocksize < obj->hdrlen)
740 return got_error(GOT_ERR_NO_SPACE);
742 *blob = calloc(1, sizeof(**blob));
744 return got_error(GOT_ERR_NO_MEM);
746 if (obj->flags & GOT_OBJ_FLAG_PACKED) {
747 (*blob)->read_buf = calloc(1, blocksize);
748 if ((*blob)->read_buf == NULL)
749 return got_error(GOT_ERR_NO_MEM);
750 err = got_packfile_extract_object(&((*blob)->f), obj, repo);
754 err = open_loose_object(&((*blob)->f), obj, repo);
760 err = got_inflate_init(&(*blob)->zb, blocksize);
767 (*blob)->read_buf = (*blob)->zb.outbuf;
768 (*blob)->flags |= GOT_BLOB_F_COMPRESSED;
771 (*blob)->hdrlen = obj->hdrlen;
772 (*blob)->blocksize = blocksize;
773 memcpy(&(*blob)->id.sha1, obj->id.sha1, SHA1_DIGEST_LENGTH);
779 got_object_blob_close(struct got_blob_object *blob)
781 if (blob->flags & GOT_BLOB_F_COMPRESSED)
782 got_inflate_end(&blob->zb);
784 free(blob->read_buf);
790 got_object_blob_id_str(struct got_blob_object *blob, char *buf, size_t size)
792 return got_sha1_digest_to_str(blob->id.sha1, buf, size);
796 got_object_blob_get_hdrlen(struct got_blob_object *blob)
802 got_object_blob_get_read_buf(struct got_blob_object *blob)
804 return blob->read_buf;
807 const struct got_error *
808 got_object_blob_read_block(size_t *outlenp, struct got_blob_object *blob)
812 if (blob->flags & GOT_BLOB_F_COMPRESSED)
813 return got_inflate_read(&blob->zb, blob->f, outlenp);
815 n = fread(blob->read_buf, 1, blob->blocksize, blob->f);
816 if (n == 0 && ferror(blob->f))
817 return got_ferror(blob->f, GOT_ERR_IO);