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>
21 #include <sys/socket.h>
22 #include <sys/syslimits.h>
38 #include "got_error.h"
39 #include "got_object.h"
40 #include "got_repository.h"
41 #include "got_opentemp.h"
44 #include "got_lib_sha1.h"
45 #include "got_lib_delta.h"
46 #include "got_lib_inflate.h"
47 #include "got_lib_object.h"
48 #include "got_lib_object_parse.h"
49 #include "got_lib_object_cache.h"
50 #include "got_lib_pack.h"
51 #include "got_lib_privsep.h"
52 #include "got_lib_repository.h"
55 #define nitems(_a) (sizeof(_a) / sizeof((_a)[0]))
59 got_object_id_cmp(const struct got_object_id *id1,
60 const struct got_object_id *id2)
62 return memcmp(id1->sha1, id2->sha1, SHA1_DIGEST_LENGTH);
65 const struct got_error *
66 got_object_qid_alloc_partial(struct got_object_qid **qid)
68 const struct got_error *err = NULL;
70 *qid = malloc(sizeof(**qid));
72 return got_error_from_errno("malloc");
74 (*qid)->id = malloc(sizeof(*((*qid)->id)));
75 if ((*qid)->id == NULL) {
76 err = got_error_from_errno("malloc");
77 got_object_qid_free(*qid);
85 const struct got_error *
86 got_object_id_str(char **outbuf, struct got_object_id *id)
88 static const size_t len = SHA1_DIGEST_STRING_LENGTH;
90 *outbuf = malloc(len);
92 return got_error_from_errno("malloc");
94 if (got_sha1_digest_to_str(id->sha1, *outbuf, len) == NULL) {
97 return got_error(GOT_ERR_BAD_OBJ_ID_STR);
104 got_object_close(struct got_object *obj)
106 if (obj->refcnt > 0) {
112 if (obj->flags & GOT_OBJ_FLAG_DELTIFIED) {
113 struct got_delta *delta;
114 while (!SIMPLEQ_EMPTY(&obj->deltas.entries)) {
115 delta = SIMPLEQ_FIRST(&obj->deltas.entries);
116 SIMPLEQ_REMOVE_HEAD(&obj->deltas.entries, entry);
120 if (obj->flags & GOT_OBJ_FLAG_PACKED)
121 free(obj->path_packfile);
126 got_object_qid_free(struct got_object_qid *qid)
133 got_object_id_queue_free(struct got_object_id_queue *ids)
135 struct got_object_qid *qid;
137 while (!SIMPLEQ_EMPTY(ids)) {
138 qid = SIMPLEQ_FIRST(ids);
139 SIMPLEQ_REMOVE_HEAD(ids, entry);
140 got_object_qid_free(qid);
144 const struct got_error *
145 got_object_parse_header(struct got_object **obj, char *buf, size_t len)
147 const char *obj_labels[] = {
148 GOT_OBJ_LABEL_COMMIT,
153 const int obj_types[] = {
160 size_t size = 0, hdrlen = 0;
165 hdrlen = strnlen(buf, len) + 1 /* '\0' */;
167 return got_error(GOT_ERR_BAD_OBJ_HDR);
169 for (i = 0; i < nitems(obj_labels); i++) {
170 const char *label = obj_labels[i];
171 size_t label_len = strlen(label);
174 if (strncmp(buf, label, label_len) != 0)
178 if (len <= label_len)
179 return got_error(GOT_ERR_BAD_OBJ_HDR);
180 size = strtonum(buf + label_len, 0, LONG_MAX, &errstr);
182 return got_error(GOT_ERR_BAD_OBJ_HDR);
187 return got_error(GOT_ERR_BAD_OBJ_HDR);
189 *obj = calloc(1, sizeof(**obj));
191 return got_error_from_errno("calloc");
193 (*obj)->hdrlen = hdrlen;
198 const struct got_error *
199 got_object_read_header(struct got_object **obj, int fd)
201 const struct got_error *err;
202 struct got_inflate_buf zb;
204 const size_t zbsize = 64;
205 size_t outlen, totlen;
210 buf = malloc(zbsize);
212 return got_error_from_errno("malloc");
214 err = got_inflate_init(&zb, buf, zbsize);
220 err = got_inflate_read_fd(&zb, fd, &outlen);
226 if (memchr(zb.outbuf, '\0', outlen) == NULL) {
229 newbuf = recallocarray(buf, nbuf - 1, nbuf, zbsize);
230 if (newbuf == NULL) {
231 err = got_error_from_errno("recallocarray");
235 zb.outbuf = newbuf + totlen;
236 zb.outlen = (nbuf * zbsize) - totlen;
238 } while (memchr(zb.outbuf, '\0', outlen) == NULL);
240 err = got_object_parse_header(obj, buf, totlen);
243 got_inflate_end(&zb);
247 struct got_commit_object *
248 got_object_commit_alloc_partial(void)
250 struct got_commit_object *commit;
252 commit = calloc(1, sizeof(*commit));
255 commit->tree_id = malloc(sizeof(*commit->tree_id));
256 if (commit->tree_id == NULL) {
261 SIMPLEQ_INIT(&commit->parent_ids);
266 const struct got_error *
267 got_object_commit_add_parent(struct got_commit_object *commit,
270 const struct got_error *err = NULL;
271 struct got_object_qid *qid;
273 err = got_object_qid_alloc_partial(&qid);
277 if (!got_parse_sha1_digest(qid->id->sha1, id_str)) {
278 err = got_error(GOT_ERR_BAD_OBJ_DATA);
279 got_object_qid_free(qid);
283 SIMPLEQ_INSERT_TAIL(&commit->parent_ids, qid, entry);
289 static const struct got_error *
290 parse_gmtoff(time_t *gmtoff, const char *tzstr)
293 const char *p = tzstr;
301 return got_error(GOT_ERR_BAD_OBJ_DATA);
303 if (!isdigit(*p) && !isdigit(*(p + 1)))
304 return got_error(GOT_ERR_BAD_OBJ_DATA);
305 h = (((*p - '0') * 10) + (*(p + 1) - '0'));
308 if (!isdigit(*p) && !isdigit(*(p + 1)))
309 return got_error(GOT_ERR_BAD_OBJ_DATA);
310 m = ((*p - '0') * 10) + (*(p + 1) - '0');
312 *gmtoff = (h * 60 * 60 + m * 60) * sign;
316 static const struct got_error *
317 parse_commit_time(time_t *time, time_t *gmtoff, char *committer)
319 const struct got_error *err = NULL;
323 /* Parse and strip off trailing timezone indicator string. */
324 space = strrchr(committer, ' ');
326 return got_error(GOT_ERR_BAD_OBJ_DATA);
327 tzstr = strdup(space + 1);
329 return got_error_from_errno("strdup");
330 err = parse_gmtoff(gmtoff, tzstr);
336 /* Timestamp is separated from committer name + email by space. */
337 space = strrchr(committer, ' ');
339 return got_error(GOT_ERR_BAD_OBJ_DATA);
341 /* Timestamp parsed here is expressed as UNIX timestamp (UTC). */
342 *time = strtonum(space + 1, 0, INT64_MAX, &errstr);
344 return got_error(GOT_ERR_BAD_OBJ_DATA);
346 /* Strip off parsed time information, leaving just author and email. */
353 got_object_commit_close(struct got_commit_object *commit)
355 if (commit->refcnt > 0) {
357 if (commit->refcnt > 0)
361 got_object_id_queue_free(&commit->parent_ids);
362 free(commit->tree_id);
363 free(commit->author);
364 free(commit->committer);
365 free(commit->logmsg);
369 struct got_object_id *
370 got_object_commit_get_tree_id(struct got_commit_object *commit)
372 return commit->tree_id;
376 got_object_commit_get_nparents(struct got_commit_object *commit)
378 return commit->nparents;
381 const struct got_object_id_queue *
382 got_object_commit_get_parent_ids(struct got_commit_object *commit)
384 return &commit->parent_ids;
388 got_object_commit_get_author(struct got_commit_object *commit)
390 return commit->author;
394 got_object_commit_get_author_time(struct got_commit_object *commit)
396 return commit->author_time;
399 time_t got_object_commit_get_author_gmtoff(struct got_commit_object *commit)
401 return commit->author_gmtoff;
405 got_object_commit_get_committer(struct got_commit_object *commit)
407 return commit->committer;
411 got_object_commit_get_committer_time(struct got_commit_object *commit)
413 return commit->committer_time;
417 got_object_commit_get_committer_gmtoff(struct got_commit_object *commit)
419 return commit->committer_gmtoff;
422 const struct got_error *
423 got_object_commit_get_logmsg(char **logmsg, struct got_commit_object *commit)
425 const struct got_error *err = NULL;
426 char *msg0, *msg, *line, *s;
432 msg0 = strdup(commit->logmsg);
434 return got_error_from_errno("strdup");
436 /* Copy log message line by line to strip out unusual headers... */
439 if ((line = strsep(&msg, "\n")) == NULL)
443 if (line[0] != '\0' &&
444 strncmp(line, GOT_COMMIT_LABEL_TREE,
445 strlen(GOT_COMMIT_LABEL_TREE)) != 0 &&
446 strncmp(line, GOT_COMMIT_LABEL_AUTHOR,
447 strlen(GOT_COMMIT_LABEL_AUTHOR)) != 0 &&
448 strncmp(line, GOT_COMMIT_LABEL_PARENT,
449 strlen(GOT_COMMIT_LABEL_PARENT)) != 0 &&
450 strncmp(line, GOT_COMMIT_LABEL_COMMITTER,
451 strlen(GOT_COMMIT_LABEL_COMMITTER)) != 0)
458 if (asprintf(&s, "%s%s\n",
459 *logmsg ? *logmsg : "", line) == -1) {
460 err = got_error_from_errno("asprintf");
468 /* Trim redundant trailing whitespace. */
469 len = strlen(*logmsg);
470 while (len > 1 && isspace((unsigned char)(*logmsg)[len - 2]) &&
471 isspace((unsigned char)(*logmsg)[len - 1])) {
472 (*logmsg)[len - 1] = '\0';
485 got_object_commit_get_logmsg_raw(struct got_commit_object *commit)
487 return commit->logmsg;
490 const struct got_error *
491 got_object_parse_commit(struct got_commit_object **commit, char *buf,
494 const struct got_error *err = NULL;
497 ssize_t remain = (ssize_t)len;
500 return got_error(GOT_ERR_BAD_OBJ_DATA);
502 *commit = got_object_commit_alloc_partial();
504 return got_error_from_errno("got_object_commit_alloc_partial");
506 label_len = strlen(GOT_COMMIT_LABEL_TREE);
507 if (strncmp(s, GOT_COMMIT_LABEL_TREE, label_len) == 0) {
509 if (remain < SHA1_DIGEST_STRING_LENGTH) {
510 err = got_error(GOT_ERR_BAD_OBJ_DATA);
514 if (!got_parse_sha1_digest((*commit)->tree_id->sha1, s)) {
515 err = got_error(GOT_ERR_BAD_OBJ_DATA);
518 remain -= SHA1_DIGEST_STRING_LENGTH;
519 s += SHA1_DIGEST_STRING_LENGTH;
521 err = got_error(GOT_ERR_BAD_OBJ_DATA);
525 label_len = strlen(GOT_COMMIT_LABEL_PARENT);
526 while (strncmp(s, GOT_COMMIT_LABEL_PARENT, label_len) == 0) {
528 if (remain < SHA1_DIGEST_STRING_LENGTH) {
529 err = got_error(GOT_ERR_BAD_OBJ_DATA);
533 err = got_object_commit_add_parent(*commit, s);
537 remain -= SHA1_DIGEST_STRING_LENGTH;
538 s += SHA1_DIGEST_STRING_LENGTH;
541 label_len = strlen(GOT_COMMIT_LABEL_AUTHOR);
542 if (strncmp(s, GOT_COMMIT_LABEL_AUTHOR, label_len) == 0) {
548 err = got_error(GOT_ERR_BAD_OBJ_DATA);
552 p = memchr(s, '\n', remain);
554 err = got_error(GOT_ERR_BAD_OBJ_DATA);
559 err = parse_commit_time(&(*commit)->author_time,
560 &(*commit)->author_gmtoff, s);
563 (*commit)->author = strdup(s);
564 if ((*commit)->author == NULL) {
565 err = got_error_from_errno("strdup");
572 label_len = strlen(GOT_COMMIT_LABEL_COMMITTER);
573 if (strncmp(s, GOT_COMMIT_LABEL_COMMITTER, label_len) == 0) {
579 err = got_error(GOT_ERR_BAD_OBJ_DATA);
583 p = memchr(s, '\n', remain);
585 err = got_error(GOT_ERR_BAD_OBJ_DATA);
590 err = parse_commit_time(&(*commit)->committer_time,
591 &(*commit)->committer_gmtoff, s);
594 (*commit)->committer = strdup(s);
595 if ((*commit)->committer == NULL) {
596 err = got_error_from_errno("strdup");
603 (*commit)->logmsg = strndup(s, remain);
604 if ((*commit)->logmsg == NULL) {
605 err = got_error_from_errno("strndup");
610 got_object_commit_close(*commit);
617 got_object_tree_entry_close(struct got_tree_entry *te)
625 got_object_tree_entries_close(struct got_tree_entries *entries)
627 struct got_tree_entry *te;
629 while (!SIMPLEQ_EMPTY(&entries->head)) {
630 te = SIMPLEQ_FIRST(&entries->head);
631 SIMPLEQ_REMOVE_HEAD(&entries->head, entry);
632 got_object_tree_entry_close(te);
637 got_object_tree_close(struct got_tree_object *tree)
639 if (tree->refcnt > 0) {
641 if (tree->refcnt > 0)
645 got_object_tree_entries_close(&tree->entries);
649 struct got_tree_entry *
650 got_alloc_tree_entry_partial(void)
652 struct got_tree_entry *te;
654 te = malloc(sizeof(*te));
658 te->id = malloc(sizeof(*te->id));
659 if (te->id == NULL) {
666 static const struct got_error *
667 parse_tree_entry(struct got_parsed_tree_entry **pte, const char **name,
668 size_t *elen, char *buf,
672 const struct got_error *err = NULL;
677 *pte = malloc(sizeof(**pte));
679 return got_error_from_errno("malloc");
681 *elen = strnlen(buf, maxlen) + 1;
682 if (*elen > maxlen) {
685 return got_error(GOT_ERR_BAD_OBJ_DATA);
688 space = memchr(buf, ' ', *elen);
689 if (space == NULL || space <= buf) {
690 err = got_error(GOT_ERR_BAD_OBJ_DATA);
698 if (*p < '0' && *p > '7') {
699 err = got_error(GOT_ERR_BAD_OBJ_DATA);
703 (*pte)->mode |= *p - '0';
707 if (*elen > maxlen || maxlen - *elen < SHA1_DIGEST_LENGTH) {
708 err = got_error(GOT_ERR_BAD_OBJ_DATA);
714 *elen += SHA1_DIGEST_LENGTH;
723 const struct got_error *
724 got_object_parse_tree(struct got_pathlist_head *entries, int *nentries,
725 uint8_t *buf, size_t len)
727 const struct got_error *err = NULL;
732 return NULL; /* tree is empty */
735 struct got_parsed_tree_entry *pte;
736 struct got_pathlist_entry *new = NULL;
740 err = parse_tree_entry(&pte, &name, &elen, buf, remain);
743 err = got_pathlist_insert(&new, entries, name, pte);
747 err = got_error(GOT_ERR_TREE_DUP_ENTRY);
756 err = got_error(GOT_ERR_BAD_OBJ_DATA);
761 got_pathlist_free(entries);
768 got_object_tag_close(struct got_tag_object *tag)
770 if (tag->refcnt > 0) {
782 const struct got_error *
783 got_object_parse_tag(struct got_tag_object **tag, uint8_t *buf, size_t len)
785 const struct got_error *err = NULL;
791 return got_error(GOT_ERR_BAD_OBJ_DATA);
793 *tag = calloc(1, sizeof(**tag));
795 return got_error_from_errno("calloc");
797 label_len = strlen(GOT_TAG_LABEL_OBJECT);
798 if (strncmp(s, GOT_TAG_LABEL_OBJECT, label_len) == 0) {
800 if (remain < SHA1_DIGEST_STRING_LENGTH) {
801 err = got_error(GOT_ERR_BAD_OBJ_DATA);
805 if (!got_parse_sha1_digest((*tag)->id.sha1, s)) {
806 err = got_error(GOT_ERR_BAD_OBJ_DATA);
809 remain -= SHA1_DIGEST_STRING_LENGTH;
810 s += SHA1_DIGEST_STRING_LENGTH;
812 err = got_error(GOT_ERR_BAD_OBJ_DATA);
817 err = got_error(GOT_ERR_BAD_OBJ_DATA);
821 label_len = strlen(GOT_TAG_LABEL_TYPE);
822 if (strncmp(s, GOT_TAG_LABEL_TYPE, label_len) == 0) {
825 err = got_error(GOT_ERR_BAD_OBJ_DATA);
829 if (strncmp(s, GOT_OBJ_LABEL_COMMIT,
830 strlen(GOT_OBJ_LABEL_COMMIT)) == 0) {
831 (*tag)->obj_type = GOT_OBJ_TYPE_COMMIT;
832 label_len = strlen(GOT_OBJ_LABEL_COMMIT);
835 } else if (strncmp(s, GOT_OBJ_LABEL_TREE,
836 strlen(GOT_OBJ_LABEL_TREE)) == 0) {
837 (*tag)->obj_type = GOT_OBJ_TYPE_TREE;
838 label_len = strlen(GOT_OBJ_LABEL_TREE);
841 } else if (strncmp(s, GOT_OBJ_LABEL_BLOB,
842 strlen(GOT_OBJ_LABEL_BLOB)) == 0) {
843 (*tag)->obj_type = GOT_OBJ_TYPE_BLOB;
844 label_len = strlen(GOT_OBJ_LABEL_BLOB);
847 } else if (strncmp(s, GOT_OBJ_LABEL_TAG,
848 strlen(GOT_OBJ_LABEL_TAG)) == 0) {
849 (*tag)->obj_type = GOT_OBJ_TYPE_TAG;
850 label_len = strlen(GOT_OBJ_LABEL_TAG);
854 err = got_error(GOT_ERR_BAD_OBJ_DATA);
858 if (remain <= 0 || *s != '\n') {
859 err = got_error(GOT_ERR_BAD_OBJ_DATA);
865 err = got_error(GOT_ERR_BAD_OBJ_DATA);
869 err = got_error(GOT_ERR_BAD_OBJ_DATA);
873 label_len = strlen(GOT_TAG_LABEL_TAG);
874 if (strncmp(s, GOT_TAG_LABEL_TAG, label_len) == 0) {
879 err = got_error(GOT_ERR_BAD_OBJ_DATA);
883 p = memchr(s, '\n', remain);
885 err = got_error(GOT_ERR_BAD_OBJ_DATA);
890 (*tag)->tag = strndup(s, slen);
891 if ((*tag)->tag == NULL) {
892 err = got_error_from_errno("strndup");
898 err = got_error(GOT_ERR_BAD_OBJ_DATA);
902 err = got_error(GOT_ERR_BAD_OBJ_DATA);
906 label_len = strlen(GOT_TAG_LABEL_TAGGER);
907 if (strncmp(s, GOT_TAG_LABEL_TAGGER, label_len) == 0) {
913 err = got_error(GOT_ERR_BAD_OBJ_DATA);
917 p = memchr(s, '\n', remain);
919 err = got_error(GOT_ERR_BAD_OBJ_DATA);
924 err = parse_commit_time(&(*tag)->tagger_time,
925 &(*tag)->tagger_gmtoff, s);
928 (*tag)->tagger = strdup(s);
929 if ((*tag)->tagger == NULL) {
930 err = got_error_from_errno("strdup");
936 err = got_error(GOT_ERR_BAD_OBJ_DATA);
940 /* Some old tags in the Linux git repo have no tagger. */
941 (*tag)->tagger = strdup("");
942 if ((*tag)->tagger == NULL) {
943 err = got_error_from_errno("strdup");
948 (*tag)->tagmsg = strndup(s, remain);
949 if ((*tag)->tagmsg == NULL) {
950 err = got_error_from_errno("strndup");
955 got_object_tag_close(*tag);
961 const struct got_error *
962 got_read_file_to_mem(uint8_t **outbuf, size_t *outlen, FILE *f)
964 const struct got_error *err = NULL;
965 static const size_t blocksize = 512;
966 size_t n, total, remain;
972 buf = malloc(blocksize);
974 return got_error_from_errno("malloc");
981 newbuf = reallocarray(buf, 1, total + blocksize);
982 if (newbuf == NULL) {
983 err = got_error_from_errno("reallocarray");
989 n = fread(buf + total, 1, remain, f);
992 err = got_ferror(f, GOT_ERR_IO);