2 * Copyright (c) 2018, 2019 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.
17 #include <sys/types.h>
19 #include <sys/queue.h>
36 #include "got_error.h"
37 #include "got_object.h"
38 #include "got_repository.h"
39 #include "got_opentemp.h"
42 #include "got_lib_sha1.h"
43 #include "got_lib_delta.h"
44 #include "got_lib_inflate.h"
45 #include "got_lib_object.h"
46 #include "got_lib_object_idcache.h"
47 #include "got_lib_object_cache.h"
48 #include "got_lib_object_parse.h"
49 #include "got_lib_pack.h"
50 #include "got_lib_repository.h"
53 #define MIN(_a,_b) ((_a) < (_b) ? (_a) : (_b))
57 #define nitems(_a) (sizeof((_a)) / sizeof((_a)[0]))
60 struct got_object_id *
61 got_object_get_id(struct got_object *obj)
66 const struct got_error *
67 got_object_get_id_str(char **outbuf, struct got_object *obj)
69 return got_object_id_str(outbuf, &obj->id);
72 const struct got_error *
73 got_object_get_type(int *type, struct got_repository *repo,
74 struct got_object_id *id)
76 const struct got_error *err = NULL;
77 struct got_object *obj;
79 err = got_object_open(&obj, repo, id);
84 case GOT_OBJ_TYPE_COMMIT:
85 case GOT_OBJ_TYPE_TREE:
86 case GOT_OBJ_TYPE_BLOB:
87 case GOT_OBJ_TYPE_TAG:
91 err = got_error(GOT_ERR_OBJ_TYPE);
95 got_object_close(obj);
99 const struct got_error *
100 got_object_get_path(char **path, struct got_object_id *id,
101 struct got_repository *repo)
103 const struct got_error *err = NULL;
109 path_objects = got_repo_get_path_objects(repo);
110 if (path_objects == NULL)
111 return got_error_from_errno("got_repo_get_path_objects");
113 err = got_object_id_str(&hex, id);
117 if (asprintf(path, "%s/%.2x/%s", path_objects,
118 id->sha1[0], hex + 2) == -1)
119 err = got_error_from_errno("asprintf");
127 const struct got_error *
128 got_object_open_loose_fd(int *fd, struct got_object_id *id,
129 struct got_repository *repo)
131 const struct got_error *err = NULL;
134 err = got_object_get_path(&path, id, repo);
137 *fd = open(path, O_RDONLY | O_NOFOLLOW | O_CLOEXEC);
139 err = got_error_from_errno2("open", path);
147 const struct got_error *
148 got_object_open_by_id_str(struct got_object **obj, struct got_repository *repo,
151 struct got_object_id id;
153 if (!got_parse_sha1_digest(id.sha1, id_str))
154 return got_error_path(id_str, GOT_ERR_BAD_OBJ_ID_STR);
156 return got_object_open(obj, repo, &id);
159 const struct got_error *
160 got_object_resolve_id_str(struct got_object_id **id,
161 struct got_repository *repo, const char *id_str)
163 const struct got_error *err = NULL;
164 struct got_object *obj;
166 err = got_object_open_by_id_str(&obj, repo, id_str);
170 *id = got_object_id_dup(got_object_get_id(obj));
171 got_object_close(obj);
173 return got_error_from_errno("got_object_id_dup");
178 const struct got_error *
179 got_object_qid_alloc(struct got_object_qid **qid, struct got_object_id *id)
181 *qid = calloc(1, sizeof(**qid));
183 return got_error_from_errno("calloc");
185 memcpy(&(*qid)->id, id, sizeof((*qid)->id));
189 const struct got_error *
190 got_object_id_queue_copy(const struct got_object_id_queue *src,
191 struct got_object_id_queue *dest)
193 const struct got_error *err;
194 struct got_object_qid *qid;
196 STAILQ_FOREACH(qid, src, entry) {
197 struct got_object_qid *new;
199 * Deep-copy the object ID only. Let the caller deal
200 * with setting up the new->data pointer if needed.
202 err = got_object_qid_alloc(&new, &qid->id);
204 got_object_id_queue_free(dest);
207 STAILQ_INSERT_TAIL(dest, new, entry);
214 got_object_tree_get_nentries(struct got_tree_object *tree)
216 return tree->nentries;
219 struct got_tree_entry *
220 got_object_tree_get_first_entry(struct got_tree_object *tree)
222 return got_object_tree_get_entry(tree, 0);
225 struct got_tree_entry *
226 got_object_tree_get_last_entry(struct got_tree_object *tree)
228 return got_object_tree_get_entry(tree, tree->nentries - 1);
231 struct got_tree_entry *
232 got_object_tree_get_entry(struct got_tree_object *tree, int i)
234 if (i < 0 || i >= tree->nentries)
236 return &tree->entries[i];
240 got_tree_entry_get_mode(struct got_tree_entry *te)
246 got_tree_entry_get_name(struct got_tree_entry *te)
251 struct got_object_id *
252 got_tree_entry_get_id(struct got_tree_entry *te)
257 const struct got_error *
258 got_object_blob_read_to_str(char **s, struct got_blob_object *blob)
260 const struct got_error *err = NULL;
261 size_t len, totlen, hdrlen, offset;
265 hdrlen = got_object_blob_get_hdrlen(blob);
271 err = got_object_blob_read_block(&len, blob);
278 totlen += len - hdrlen;
279 p = realloc(*s, totlen + 1);
281 err = got_error_from_errno("realloc");
287 /* Skip blob object header first time around. */
289 got_object_blob_get_read_buf(blob) + hdrlen, len - hdrlen);
298 const struct got_error *
299 got_tree_entry_get_symlink_target(char **link_target, struct got_tree_entry *te,
300 struct got_repository *repo)
302 const struct got_error *err = NULL;
303 struct got_blob_object *blob = NULL;
308 if (!got_object_tree_entry_is_symlink(te))
309 return got_error(GOT_ERR_TREE_ENTRY_TYPE);
311 fd = got_opentempfd();
313 err = got_error_from_errno("got_opentempfd");
317 err = got_object_open_as_blob(&blob, repo,
318 got_tree_entry_get_id(te), PATH_MAX, fd);
322 err = got_object_blob_read_to_str(link_target, blob);
324 if (fd != -1 && close(fd) == -1 && err == NULL)
325 err = got_error_from_errno("close");
327 got_object_blob_close(blob);
336 got_tree_entry_get_index(struct got_tree_entry *te)
341 struct got_tree_entry *
342 got_tree_entry_get_next(struct got_tree_object *tree,
343 struct got_tree_entry *te)
345 return got_object_tree_get_entry(tree, te->idx + 1);
348 struct got_tree_entry *
349 got_tree_entry_get_prev(struct got_tree_object *tree,
350 struct got_tree_entry *te)
352 return got_object_tree_get_entry(tree, te->idx - 1);
355 const struct got_error *
356 got_object_blob_close(struct got_blob_object *blob)
358 const struct got_error *err = NULL;
359 free(blob->read_buf);
360 if (blob->f && fclose(blob->f) == EOF)
361 err = got_error_from_errno("fclose");
368 got_object_blob_rewind(struct got_blob_object *blob)
375 got_object_blob_id_str(struct got_blob_object *blob, char *buf, size_t size)
377 return got_sha1_digest_to_str(blob->id.sha1, buf, size);
381 got_object_blob_get_hdrlen(struct got_blob_object *blob)
387 got_object_blob_get_read_buf(struct got_blob_object *blob)
389 return blob->read_buf;
392 const struct got_error *
393 got_object_blob_read_block(size_t *outlenp, struct got_blob_object *blob)
397 n = fread(blob->read_buf, 1, blob->blocksize, blob->f);
398 if (n == 0 && ferror(blob->f))
399 return got_ferror(blob->f, GOT_ERR_IO);
404 const struct got_error *
405 got_object_blob_dump_to_file(off_t *filesize, int *nlines,
406 off_t **line_offsets, FILE *outfile, struct got_blob_object *blob)
408 const struct got_error *err = NULL;
409 size_t n, len, hdrlen;
412 const int alloc_chunksz = 512;
414 off_t off = 0, total_len = 0;
417 *line_offsets = NULL;
423 hdrlen = got_object_blob_get_hdrlen(blob);
425 err = got_object_blob_read_block(&len, blob);
430 buf = got_object_blob_get_read_buf(blob);
433 if (line_offsets && *line_offsets == NULL) {
434 /* Have some data but perhaps no '\n'. */
436 nalloc = alloc_chunksz;
437 *line_offsets = calloc(nalloc,
438 sizeof(**line_offsets));
439 if (*line_offsets == NULL)
440 return got_error_from_errno("calloc");
442 /* Skip forward over end of first line. */
449 /* Scan '\n' offsets in remaining chunk of data. */
451 if (buf[i] != '\n') {
456 if (line_offsets && nalloc < *nlines) {
457 size_t n = *nlines + alloc_chunksz;
458 off_t *o = recallocarray(*line_offsets,
459 nalloc, n, sizeof(**line_offsets));
462 *line_offsets = NULL;
463 return got_error_from_errno(
470 off = total_len + i - hdrlen + 1;
471 (*line_offsets)[*nlines - 1] = off;
476 /* Skip blob object header first time around. */
477 n = fwrite(buf + hdrlen, 1, len - hdrlen, outfile);
478 if (n != len - hdrlen)
479 return got_ferror(outfile, GOT_ERR_IO);
480 total_len += len - hdrlen;
484 if (fflush(outfile) != 0)
485 return got_error_from_errno("fflush");
489 *filesize = total_len;
495 got_object_tag_get_name(struct got_tag_object *tag)
501 got_object_tag_get_object_type(struct got_tag_object *tag)
503 return tag->obj_type;
506 struct got_object_id *
507 got_object_tag_get_object_id(struct got_tag_object *tag)
513 got_object_tag_get_tagger_time(struct got_tag_object *tag)
515 return tag->tagger_time;
519 got_object_tag_get_tagger_gmtoff(struct got_tag_object *tag)
521 return tag->tagger_gmtoff;
525 got_object_tag_get_tagger(struct got_tag_object *tag)
531 got_object_tag_get_message(struct got_tag_object *tag)
536 static struct got_tree_entry *
537 find_entry_by_name(struct got_tree_object *tree, const char *name, size_t len)
541 /* Note that tree entries are sorted in strncmp() order. */
542 for (i = 0; i < tree->nentries; i++) {
543 struct got_tree_entry *te = &tree->entries[i];
544 int cmp = strncmp(te->name, name, len);
549 if (te->name[len] == '\0')
555 struct got_tree_entry *
556 got_object_tree_find_entry(struct got_tree_object *tree, const char *name)
558 return find_entry_by_name(tree, name, strlen(name));
561 const struct got_error *
562 got_object_tree_find_path(struct got_object_id **id, mode_t *mode,
563 struct got_repository *repo, struct got_tree_object *tree,
566 const struct got_error *err = NULL;
567 struct got_tree_object *subtree = NULL;
568 struct got_tree_entry *te = NULL;
581 struct got_tree_object *next_tree;
590 te = find_entry_by_name(subtree, seg, seglen);
592 err = got_error_path(path, GOT_ERR_NO_TREE_ENTRY);
603 err = got_object_open_as_tree(&next_tree, repo,
609 got_object_tree_close(subtree);
615 *id = got_object_id_dup(&te->id);
617 return got_error_from_errno("got_object_id_dup");
621 err = got_error_path(path, GOT_ERR_NO_TREE_ENTRY);
623 if (subtree && subtree != tree)
624 got_object_tree_close(subtree);
628 const struct got_error *
629 got_object_id_by_path(struct got_object_id **id, struct got_repository *repo,
630 struct got_commit_object *commit, const char *path)
632 const struct got_error *err = NULL;
633 struct got_tree_object *tree = NULL;
637 /* Handle opening of root of commit's tree. */
638 if (got_path_is_root_dir(path)) {
639 *id = got_object_id_dup(commit->tree_id);
641 err = got_error_from_errno("got_object_id_dup");
643 err = got_object_open_as_tree(&tree, repo, commit->tree_id);
646 err = got_object_tree_find_path(id, NULL, repo, tree, path);
650 got_object_tree_close(tree);
655 * Normalize file mode bits to avoid false positive tree entry differences
656 * in case tree entries have unexpected mode bits set.
659 normalize_mode_for_comparison(mode_t mode)
662 * For directories, the only relevant bit is the IFDIR bit.
663 * This allows us to detect paths changing from a directory
664 * to a file and vice versa.
667 return mode & S_IFDIR;
670 * For symlinks, the only relevant bit is the IFLNK bit.
671 * This allows us to detect paths changing from a symlinks
672 * to a file or directory and vice versa.
675 return mode & S_IFLNK;
677 /* For files, the only change we care about is the executable bit. */
678 return mode & S_IXUSR;
681 const struct got_error *
682 got_object_tree_path_changed(int *changed,
683 struct got_tree_object *tree01, struct got_tree_object *tree02,
684 const char *path, struct got_repository *repo)
686 const struct got_error *err = NULL;
687 struct got_tree_object *tree1 = NULL, *tree2 = NULL;
688 struct got_tree_entry *te1 = NULL, *te2 = NULL;
694 /* We not do support comparing the root path. */
695 if (got_path_is_root_dir(path))
696 return got_error_path(path, GOT_ERR_BAD_PATH);
706 struct got_tree_object *next_tree1, *next_tree2;
716 te1 = find_entry_by_name(tree1, seg, seglen);
718 err = got_error(GOT_ERR_NO_OBJ);
723 te2 = find_entry_by_name(tree2, seg, seglen);
726 mode1 = normalize_mode_for_comparison(te1->mode);
727 mode2 = normalize_mode_for_comparison(te2->mode);
728 if (mode1 != mode2) {
733 if (got_object_id_cmp(&te1->id, &te2->id) == 0) {
739 if (*s == '\0') { /* final path element */
748 err = got_object_open_as_tree(&next_tree1, repo,
754 got_object_tree_close(tree1);
758 err = got_object_open_as_tree(&next_tree2, repo,
764 got_object_tree_close(tree2);
768 got_object_tree_close(tree2);
774 if (tree1 && tree1 != tree01)
775 got_object_tree_close(tree1);
776 if (tree2 && tree2 != tree02)
777 got_object_tree_close(tree2);
781 const struct got_error *
782 got_object_tree_entry_dup(struct got_tree_entry **new_te,
783 struct got_tree_entry *te)
785 const struct got_error *err = NULL;
787 *new_te = calloc(1, sizeof(**new_te));
789 return got_error_from_errno("calloc");
791 (*new_te)->mode = te->mode;
792 memcpy((*new_te)->name, te->name, sizeof((*new_te)->name));
793 memcpy(&(*new_te)->id, &te->id, sizeof((*new_te)->id));
798 got_object_tree_entry_is_submodule(struct got_tree_entry *te)
800 return (te->mode & S_IFMT) == (S_IFDIR | S_IFLNK);
804 got_object_tree_entry_is_symlink(struct got_tree_entry *te)
806 /* S_IFDIR check avoids confusing symlinks with submodules. */
807 return ((te->mode & (S_IFDIR | S_IFLNK)) == S_IFLNK);
810 static const struct got_error *
811 resolve_symlink(char **link_target, const char *path,
812 struct got_commit_object *commit, struct got_repository *repo)
814 const struct got_error *err = NULL;
816 char *name, *parent_path = NULL;
817 struct got_object_id *tree_obj_id = NULL;
818 struct got_tree_object *tree = NULL;
819 struct got_tree_entry *te = NULL;
823 if (strlcpy(buf, path, sizeof(buf)) >= sizeof(buf))
824 return got_error(GOT_ERR_NO_SPACE);
826 name = basename(buf);
828 return got_error_from_errno2("basename", path);
830 err = got_path_dirname(&parent_path, path);
834 err = got_object_id_by_path(&tree_obj_id, repo, commit,
837 if (err->code == GOT_ERR_NO_TREE_ENTRY) {
838 /* Display the complete path in error message. */
839 err = got_error_path(path, err->code);
844 err = got_object_open_as_tree(&tree, repo, tree_obj_id);
848 te = got_object_tree_find_entry(tree, name);
850 err = got_error_path(path, GOT_ERR_NO_TREE_ENTRY);
854 if (got_object_tree_entry_is_symlink(te)) {
855 err = got_tree_entry_get_symlink_target(link_target, te, repo);
858 if (!got_path_is_absolute(*link_target)) {
860 if (asprintf(&abspath, "%s/%s", parent_path,
861 *link_target) == -1) {
862 err = got_error_from_errno("asprintf");
866 *link_target = malloc(PATH_MAX);
867 if (*link_target == NULL) {
868 err = got_error_from_errno("malloc");
871 err = got_canonpath(abspath, *link_target, PATH_MAX);
881 got_object_tree_close(tree);
889 const struct got_error *
890 got_object_resolve_symlinks(char **link_target, const char *path,
891 struct got_commit_object *commit, struct got_repository *repo)
893 const struct got_error *err = NULL;
894 char *next_target = NULL;
895 int max_recursion = 40; /* matches Git */
900 err = resolve_symlink(&next_target,
901 *link_target ? *link_target : path, commit, repo);
906 if (--max_recursion == 0) {
907 err = got_error_path(path, GOT_ERR_RECURSION);
911 *link_target = next_target;
913 } while (next_target);
919 got_object_commit_retain(struct got_commit_object *commit)
924 const struct got_error *
925 got_object_raw_alloc(struct got_raw_object **obj, uint8_t *outbuf, int *outfd,
926 size_t hdrlen, off_t size)
928 const struct got_error *err = NULL;
933 *obj = calloc(1, sizeof(**obj));
935 err = got_error_from_errno("calloc");
939 (*obj)->tempfile_idx = -1;
942 (*obj)->data = outbuf;
945 if (fstat(*outfd, &sb) == -1) {
946 err = got_error_from_errno("fstat");
950 if (sb.st_size != tot) {
951 err = got_error(GOT_ERR_PRIVSEP_LEN);
954 #ifndef GOT_PACK_NO_MMAP
955 if (tot > 0 && tot <= SIZE_MAX) {
956 (*obj)->data = mmap(NULL, tot, PROT_READ,
957 MAP_PRIVATE, *outfd, 0);
958 if ((*obj)->data == MAP_FAILED) {
959 if (errno != ENOMEM) {
960 err = got_error_from_errno("mmap");
971 (*obj)->f = fdopen(*outfd, "r");
972 if ((*obj)->f == NULL) {
973 err = got_error_from_errno("fdopen");
979 (*obj)->hdrlen = hdrlen;
984 got_object_raw_close(*obj);