2 * "git fast-export" builtin command
4 * Copyright (C) 2007 Johannes E. Schindelin
16 #include "string-list.h"
18 #include "parse-options.h"
22 static const char *fast_export_usage[] = {
23 N_("git fast-export [rev-list-opts]"),
28 static enum { ABORT, VERBATIM, WARN, WARN_STRIP, STRIP } signed_tag_mode = ABORT;
29 static enum { ERROR, DROP, REWRITE } tag_of_filtered_mode = ERROR;
30 static int fake_missing_tagger;
31 static int use_done_feature;
34 static struct string_list extra_refs = STRING_LIST_INIT_NODUP;
35 static struct refspec *refspecs;
36 static int refspecs_nr;
38 static int parse_opt_signed_tag_mode(const struct option *opt,
39 const char *arg, int unset)
41 if (unset || !strcmp(arg, "abort"))
42 signed_tag_mode = ABORT;
43 else if (!strcmp(arg, "verbatim") || !strcmp(arg, "ignore"))
44 signed_tag_mode = VERBATIM;
45 else if (!strcmp(arg, "warn"))
46 signed_tag_mode = WARN;
47 else if (!strcmp(arg, "warn-strip"))
48 signed_tag_mode = WARN_STRIP;
49 else if (!strcmp(arg, "strip"))
50 signed_tag_mode = STRIP;
52 return error("Unknown signed-tags mode: %s", arg);
56 static int parse_opt_tag_of_filtered_mode(const struct option *opt,
57 const char *arg, int unset)
59 if (unset || !strcmp(arg, "abort"))
60 tag_of_filtered_mode = ERROR;
61 else if (!strcmp(arg, "drop"))
62 tag_of_filtered_mode = DROP;
63 else if (!strcmp(arg, "rewrite"))
64 tag_of_filtered_mode = REWRITE;
66 return error("Unknown tag-of-filtered mode: %s", arg);
70 static struct decoration idnums;
71 static uint32_t last_idnum;
73 static int has_unshown_parent(struct commit *commit)
75 struct commit_list *parent;
77 for (parent = commit->parents; parent; parent = parent->next)
78 if (!(parent->item->object.flags & SHOWN) &&
79 !(parent->item->object.flags & UNINTERESTING))
84 /* Since intptr_t is C99, we do not use it here */
85 static inline uint32_t *mark_to_ptr(uint32_t mark)
87 return ((uint32_t *)NULL) + mark;
90 static inline uint32_t ptr_to_mark(void * mark)
92 return (uint32_t *)mark - (uint32_t *)NULL;
95 static inline void mark_object(struct object *object, uint32_t mark)
97 add_decoration(&idnums, object, mark_to_ptr(mark));
100 static inline void mark_next_object(struct object *object)
102 mark_object(object, ++last_idnum);
105 static int get_object_mark(struct object *object)
107 void *decoration = lookup_decoration(&idnums, object);
110 return ptr_to_mark(decoration);
113 static void show_progress(void)
115 static int counter = 0;
118 if ((++counter % progress) == 0)
119 printf("progress %d objects\n", counter);
122 static void export_blob(const unsigned char *sha1)
125 enum object_type type;
127 struct object *object;
133 if (is_null_sha1(sha1))
136 object = lookup_object(sha1);
137 if (object && object->flags & SHOWN)
140 buf = read_sha1_file(sha1, &type, &size);
142 die ("Could not read blob %s", sha1_to_hex(sha1));
143 if (check_sha1_signature(sha1, buf, size, typename(type)) < 0)
144 die("sha1 mismatch in blob %s", sha1_to_hex(sha1));
145 object = parse_object_buffer(sha1, type, size, buf, &eaten);
147 die("Could not read blob %s", sha1_to_hex(sha1));
149 mark_next_object(object);
151 printf("blob\nmark :%"PRIu32"\ndata %lu\n", last_idnum, size);
152 if (size && fwrite(buf, size, 1, stdout) != 1)
153 die_errno ("Could not write blob '%s'", sha1_to_hex(sha1));
158 object->flags |= SHOWN;
163 static int depth_first(const void *a_, const void *b_)
165 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
166 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
167 const char *name_a, *name_b;
168 int len_a, len_b, len;
171 name_a = a->one ? a->one->path : a->two->path;
172 name_b = b->one ? b->one->path : b->two->path;
174 len_a = strlen(name_a);
175 len_b = strlen(name_b);
176 len = (len_a < len_b) ? len_a : len_b;
178 /* strcmp will sort 'd' before 'd/e', we want 'd/e' before 'd' */
179 cmp = memcmp(name_a, name_b, len);
186 * Move 'R'ename entries last so that all references of the file
187 * appear in the output before it is renamed (e.g., when a file
188 * was copied and renamed in the same commit).
190 return (a->status == 'R') - (b->status == 'R');
193 static void print_path(const char *path)
195 int need_quote = quote_c_style(path, NULL, NULL, 0);
197 quote_c_style(path, NULL, stdout, 0);
198 else if (strchr(path, ' '))
199 printf("\"%s\"", path);
204 static void show_filemodify(struct diff_queue_struct *q,
205 struct diff_options *options, void *data)
210 * Handle files below a directory first, in case they are all deleted
211 * and the directory changes to a file or symlink.
213 qsort(q->queue, q->nr, sizeof(q->queue[0]), depth_first);
215 for (i = 0; i < q->nr; i++) {
216 struct diff_filespec *ospec = q->queue[i]->one;
217 struct diff_filespec *spec = q->queue[i]->two;
219 switch (q->queue[i]->status) {
220 case DIFF_STATUS_DELETED:
222 print_path(spec->path);
226 case DIFF_STATUS_COPIED:
227 case DIFF_STATUS_RENAMED:
228 printf("%c ", q->queue[i]->status);
229 print_path(ospec->path);
231 print_path(spec->path);
234 if (!hashcmp(ospec->sha1, spec->sha1) &&
235 ospec->mode == spec->mode)
239 case DIFF_STATUS_TYPE_CHANGED:
240 case DIFF_STATUS_MODIFIED:
241 case DIFF_STATUS_ADDED:
243 * Links refer to objects in another repositories;
244 * output the SHA-1 verbatim.
246 if (no_data || S_ISGITLINK(spec->mode))
247 printf("M %06o %s ", spec->mode,
248 sha1_to_hex(spec->sha1));
250 struct object *object = lookup_object(spec->sha1);
251 printf("M %06o :%d ", spec->mode,
252 get_object_mark(object));
254 print_path(spec->path);
259 die("Unexpected comparison status '%c' for %s, %s",
261 ospec->path ? ospec->path : "none",
262 spec->path ? spec->path : "none");
267 static const char *find_encoding(const char *begin, const char *end)
269 const char *needle = "\nencoding ";
272 bol = memmem(begin, end ? end - begin : strlen(begin),
273 needle, strlen(needle));
275 return git_commit_encoding;
276 bol += strlen(needle);
277 eol = strchrnul(bol, '\n');
282 static void handle_commit(struct commit *commit, struct rev_info *rev)
284 int saved_output_format = rev->diffopt.output_format;
285 const char *author, *author_end, *committer, *committer_end;
286 const char *encoding, *message;
287 char *reencoded = NULL;
288 struct commit_list *p;
291 rev->diffopt.output_format = DIFF_FORMAT_CALLBACK;
293 parse_commit_or_die(commit);
294 author = strstr(commit->buffer, "\nauthor ");
296 die ("Could not find author in commit %s",
297 sha1_to_hex(commit->object.sha1));
299 author_end = strchrnul(author, '\n');
300 committer = strstr(author_end, "\ncommitter ");
302 die ("Could not find committer in commit %s",
303 sha1_to_hex(commit->object.sha1));
305 committer_end = strchrnul(committer, '\n');
306 message = strstr(committer_end, "\n\n");
307 encoding = find_encoding(committer_end, message);
311 if (commit->parents &&
312 get_object_mark(&commit->parents->item->object) != 0 &&
314 parse_commit_or_die(commit->parents->item);
315 diff_tree_sha1(commit->parents->item->tree->object.sha1,
316 commit->tree->object.sha1, "", &rev->diffopt);
319 diff_root_tree_sha1(commit->tree->object.sha1,
322 /* Export the referenced blobs, and remember the marks. */
323 for (i = 0; i < diff_queued_diff.nr; i++)
324 if (!S_ISGITLINK(diff_queued_diff.queue[i]->two->mode))
325 export_blob(diff_queued_diff.queue[i]->two->sha1);
327 mark_next_object(&commit->object);
328 if (!is_encoding_utf8(encoding))
329 reencoded = reencode_string(message, "UTF-8", encoding);
330 if (!commit->parents)
331 printf("reset %s\n", (const char*)commit->util);
332 printf("commit %s\nmark :%"PRIu32"\n%.*s\n%.*s\ndata %u\n%s",
333 (const char *)commit->util, last_idnum,
334 (int)(author_end - author), author,
335 (int)(committer_end - committer), committer,
337 ? strlen(reencoded) : message
338 ? strlen(message) : 0),
339 reencoded ? reencoded : message ? message : "");
342 for (i = 0, p = commit->parents; p; p = p->next) {
343 int mark = get_object_mark(&p->item->object);
347 printf("from :%d\n", mark);
349 printf("merge :%d\n", mark);
354 printf("deleteall\n");
355 log_tree_diff_flush(rev);
356 rev->diffopt.output_format = saved_output_format;
363 static void handle_tail(struct object_array *commits, struct rev_info *revs)
365 struct commit *commit;
366 while (commits->nr) {
367 commit = (struct commit *)commits->objects[commits->nr - 1].item;
368 if (has_unshown_parent(commit))
370 handle_commit(commit, revs);
375 static void handle_tag(const char *name, struct tag *tag)
378 enum object_type type;
380 const char *tagger, *tagger_end, *message;
381 size_t message_size = 0;
382 struct object *tagged;
386 /* Trees have no identifier in fast-export output, thus we have no way
387 * to output tags of trees, tags of tags of trees, etc. Simply omit
390 tagged = tag->tagged;
391 while (tagged->type == OBJ_TAG) {
392 tagged = ((struct tag *)tagged)->tagged;
394 if (tagged->type == OBJ_TREE) {
395 warning("Omitting tag %s,\nsince tags of trees (or tags of tags of trees, etc.) are not supported.",
396 sha1_to_hex(tag->object.sha1));
400 buf = read_sha1_file(tag->object.sha1, &type, &size);
402 die ("Could not read tag %s", sha1_to_hex(tag->object.sha1));
403 message = memmem(buf, size, "\n\n", 2);
406 message_size = strlen(message);
408 tagger = memmem(buf, message ? message - buf : size, "\ntagger ", 8);
410 if (fake_missing_tagger)
411 tagger = "tagger Unspecified Tagger "
412 "<unspecified-tagger> 0 +0000";
415 tagger_end = tagger + strlen(tagger);
418 tagger_end = strchrnul(tagger, '\n');
421 /* handle signed tags */
423 const char *signature = strstr(message,
424 "\n-----BEGIN PGP SIGNATURE-----\n");
426 switch(signed_tag_mode) {
428 die ("Encountered signed tag %s; use "
429 "--signed-tags=<mode> to handle it.",
430 sha1_to_hex(tag->object.sha1));
432 warning ("Exporting signed tag %s",
433 sha1_to_hex(tag->object.sha1));
438 warning ("Stripping signature from tag %s",
439 sha1_to_hex(tag->object.sha1));
442 message_size = signature + 1 - message;
447 /* handle tag->tagged having been filtered out due to paths specified */
448 tagged = tag->tagged;
449 tagged_mark = get_object_mark(tagged);
451 switch(tag_of_filtered_mode) {
453 die ("Tag %s tags unexported object; use "
454 "--tag-of-filtered-object=<mode> to handle it.",
455 sha1_to_hex(tag->object.sha1));
457 /* Ignore this tag altogether */
460 if (tagged->type != OBJ_COMMIT) {
461 die ("Tag %s tags unexported %s!",
462 sha1_to_hex(tag->object.sha1),
463 typename(tagged->type));
465 p = (struct commit *)tagged;
467 if (p->parents && p->parents->next)
469 if (p->object.flags & UNINTERESTING)
471 if (!(p->object.flags & TREESAME))
474 die ("Can't find replacement commit for tag %s\n",
475 sha1_to_hex(tag->object.sha1));
476 p = p->parents->item;
478 tagged_mark = get_object_mark(&p->object);
482 if (starts_with(name, "refs/tags/"))
484 printf("tag %s\nfrom :%d\n%.*s%sdata %d\n%.*s\n",
486 (int)(tagger_end - tagger), tagger,
487 tagger == tagger_end ? "" : "\n",
488 (int)message_size, (int)message_size, message ? message : "");
491 static struct commit *get_commit(struct rev_cmdline_entry *e, char *full_name)
493 switch (e->item->type) {
495 return (struct commit *)e->item;
497 struct tag *tag = (struct tag *)e->item;
499 /* handle nested tags */
500 while (tag && tag->object.type == OBJ_TAG) {
501 parse_object(tag->object.sha1);
502 string_list_append(&extra_refs, full_name)->util = tag;
503 tag = (struct tag *)tag->tagged;
506 die("Tag %s points nowhere?", e->name);
507 return (struct commit *)tag;
515 static void get_tags_and_duplicates(struct rev_cmdline_info *info)
519 for (i = 0; i < info->nr; i++) {
520 struct rev_cmdline_entry *e = info->rev + i;
521 unsigned char sha1[20];
522 struct commit *commit;
525 if (e->flags & UNINTERESTING)
528 if (dwim_ref(e->name, strlen(e->name), sha1, &full_name) != 1)
533 private = apply_refspecs(refspecs, refspecs_nr, full_name);
540 commit = get_commit(e, full_name);
542 warning("%s: Unexpected object of type %s, skipping.",
544 typename(e->item->type));
548 switch(commit->object.type) {
552 export_blob(commit->object.sha1);
554 default: /* OBJ_TAG (nested tags) is already handled */
555 warning("Tag points to object of unexpected type %s, skipping.",
556 typename(commit->object.type));
561 * This ref will not be updated through a commit, lets make
562 * sure it gets properly updated eventually.
564 if (commit->util || commit->object.flags & SHOWN)
565 string_list_append(&extra_refs, full_name)->util = commit;
567 commit->util = full_name;
571 static void handle_tags_and_duplicates(void)
573 struct commit *commit;
576 for (i = extra_refs.nr - 1; i >= 0; i--) {
577 const char *name = extra_refs.items[i].string;
578 struct object *object = extra_refs.items[i].util;
579 switch (object->type) {
581 handle_tag(name, (struct tag *)object);
584 /* create refs pointing to already seen commits */
585 commit = (struct commit *)object;
586 printf("reset %s\nfrom :%d\n\n", name,
587 get_object_mark(&commit->object));
594 static void export_marks(char *file)
598 struct object_decoration *deco = idnums.hash;
602 f = fopen(file, "w");
604 die_errno("Unable to open marks file %s for writing.", file);
606 for (i = 0; i < idnums.size; i++) {
607 if (deco->base && deco->base->type == 1) {
608 mark = ptr_to_mark(deco->decoration);
609 if (fprintf(f, ":%"PRIu32" %s\n", mark,
610 sha1_to_hex(deco->base->sha1)) < 0) {
621 error("Unable to write marks file %s.", file);
624 static void import_marks(char *input_file)
627 FILE *f = fopen(input_file, "r");
629 die_errno("cannot read '%s'", input_file);
631 while (fgets(line, sizeof(line), f)) {
633 char *line_end, *mark_end;
634 unsigned char sha1[20];
635 struct object *object;
636 struct commit *commit;
637 enum object_type type;
639 line_end = strchr(line, '\n');
640 if (line[0] != ':' || !line_end)
641 die("corrupt mark line: %s", line);
644 mark = strtoumax(line + 1, &mark_end, 10);
645 if (!mark || mark_end == line + 1
646 || *mark_end != ' ' || get_sha1_hex(mark_end + 1, sha1))
647 die("corrupt mark line: %s", line);
649 if (last_idnum < mark)
652 type = sha1_object_info(sha1, NULL);
654 die("object not found: %s", sha1_to_hex(sha1));
656 if (type != OBJ_COMMIT)
660 commit = lookup_commit(sha1);
662 die("not a commit? can't happen: %s", sha1_to_hex(sha1));
664 object = &commit->object;
666 if (object->flags & SHOWN)
667 error("Object %s already has a mark", sha1_to_hex(sha1));
669 mark_object(object, mark);
671 object->flags |= SHOWN;
676 static void handle_deletes(void)
679 for (i = 0; i < refspecs_nr; i++) {
680 struct refspec *refspec = &refspecs[i];
684 printf("reset %s\nfrom %s\n\n",
685 refspec->dst, sha1_to_hex(null_sha1));
689 int cmd_fast_export(int argc, const char **argv, const char *prefix)
691 struct rev_info revs;
692 struct object_array commits = OBJECT_ARRAY_INIT;
693 struct commit *commit;
694 char *export_filename = NULL, *import_filename = NULL;
695 uint32_t lastimportid;
696 struct string_list refspecs_list = STRING_LIST_INIT_NODUP;
697 struct option options[] = {
698 OPT_INTEGER(0, "progress", &progress,
699 N_("show progress after <n> objects")),
700 OPT_CALLBACK(0, "signed-tags", &signed_tag_mode, N_("mode"),
701 N_("select handling of signed tags"),
702 parse_opt_signed_tag_mode),
703 OPT_CALLBACK(0, "tag-of-filtered-object", &tag_of_filtered_mode, N_("mode"),
704 N_("select handling of tags that tag filtered objects"),
705 parse_opt_tag_of_filtered_mode),
706 OPT_STRING(0, "export-marks", &export_filename, N_("file"),
707 N_("Dump marks to this file")),
708 OPT_STRING(0, "import-marks", &import_filename, N_("file"),
709 N_("Import marks from this file")),
710 OPT_BOOL(0, "fake-missing-tagger", &fake_missing_tagger,
711 N_("Fake a tagger when tags lack one")),
712 OPT_BOOL(0, "full-tree", &full_tree,
713 N_("Output full tree for each commit")),
714 OPT_BOOL(0, "use-done-feature", &use_done_feature,
715 N_("Use the done feature to terminate the stream")),
716 OPT_BOOL(0, "no-data", &no_data, N_("Skip output of blob data")),
717 OPT_STRING_LIST(0, "refspec", &refspecs_list, N_("refspec"),
718 N_("Apply refspec to exported refs")),
723 usage_with_options (fast_export_usage, options);
725 /* we handle encodings */
726 git_config(git_default_config, NULL);
728 init_revisions(&revs, prefix);
730 revs.show_source = 1;
731 revs.rewrite_parents = 1;
732 argc = parse_options(argc, argv, prefix, options, fast_export_usage,
733 PARSE_OPT_KEEP_ARGV0 | PARSE_OPT_KEEP_UNKNOWN);
734 argc = setup_revisions(argc, argv, &revs, NULL);
736 usage_with_options (fast_export_usage, options);
738 if (refspecs_list.nr) {
739 const char **refspecs_str;
742 refspecs_str = xmalloc(sizeof(*refspecs_str) * refspecs_list.nr);
743 for (i = 0; i < refspecs_list.nr; i++)
744 refspecs_str[i] = refspecs_list.items[i].string;
746 refspecs_nr = refspecs_list.nr;
747 refspecs = parse_fetch_refspec(refspecs_nr, refspecs_str);
749 string_list_clear(&refspecs_list, 1);
753 if (use_done_feature)
754 printf("feature done\n");
757 import_marks(import_filename);
758 lastimportid = last_idnum;
760 if (import_filename && revs.prune_data.nr)
763 get_tags_and_duplicates(&revs.cmdline);
765 if (prepare_revision_walk(&revs))
766 die("revision walk setup failed");
767 revs.diffopt.format_callback = show_filemodify;
768 DIFF_OPT_SET(&revs.diffopt, RECURSIVE);
769 while ((commit = get_revision(&revs))) {
770 if (has_unshown_parent(commit)) {
771 add_object_array(&commit->object, NULL, &commits);
774 handle_commit(commit, &revs);
775 handle_tail(&commits, &revs);
779 handle_tags_and_duplicates();
782 if (export_filename && lastimportid != last_idnum)
783 export_marks(export_filename);
785 if (use_done_feature)
788 free_refspec(refspecs_nr, refspecs);