15 #define QUOTE_PYTHON 3
18 typedef enum { FIELD_STR, FIELD_ULONG, FIELD_TIME } cmp_type;
22 unsigned long ul; /* used for sorting when not FIELD_STR */
26 struct ref_sort *next;
27 int atom; /* index into used_atom array */
33 unsigned char objectname[20];
34 struct atom_value *value;
43 { "objectsize", FIELD_ULONG },
46 { "parent" }, /* NEEDSWORK: how to address 2nd and later parents? */
47 { "numparent", FIELD_ULONG },
54 { "authordate", FIELD_TIME },
58 { "committerdate", FIELD_TIME },
62 { "taggerdate", FIELD_TIME },
64 { "creatordate", FIELD_TIME },
71 * An atom is a valid field atom listed above, possibly prefixed with
72 * a "*" to denote deref_tag().
74 * We parse given format string and sort specifiers, and make a list
75 * of properties that we need to extract out of objects. refinfo
76 * structure will hold an array of values extracted that can be
77 * indexed with the "atom number", which is an index into this
80 static const char **used_atom;
81 static cmp_type *used_atom_type;
82 static int used_atom_cnt, sort_atom_limit, need_tagged;
85 * Used to parse format string and sort specifiers
87 static int parse_atom(const char *atom, const char *ep)
94 if (*sp == '*' && sp < ep)
97 die("malformed field name: %.*s", (int)(ep-atom), atom);
99 /* Do we have the atom already used elsewhere? */
100 for (i = 0; i < used_atom_cnt; i++) {
101 int len = strlen(used_atom[i]);
102 if (len == ep - atom && !memcmp(used_atom[i], atom, len))
106 /* Is the atom a valid one? */
107 for (i = 0; i < ARRAY_SIZE(valid_atom); i++) {
108 int len = strlen(valid_atom[i].name);
110 * If the atom name has a colon, strip it and everything after
111 * it off - it specifies the format for this entry, and
112 * shouldn't be used for checking against the valid_atom
115 const char *formatp = strchr(sp, ':');
118 if (len == formatp - sp && !memcmp(valid_atom[i].name, sp, len))
122 if (ARRAY_SIZE(valid_atom) <= i)
123 die("unknown field name: %.*s", (int)(ep-atom), atom);
125 /* Add it in, including the deref prefix */
128 used_atom = xrealloc(used_atom,
129 (sizeof *used_atom) * used_atom_cnt);
130 used_atom_type = xrealloc(used_atom_type,
131 (sizeof(*used_atom_type) * used_atom_cnt));
132 n = xmalloc(ep - atom + 1);
133 memcpy(n, atom, ep - atom);
136 used_atom_type[at] = valid_atom[i].cmp_type;
141 * In a format string, find the next occurrence of %(atom).
143 static const char *find_next(const char *cp)
147 /* %( is the start of an atom;
148 * %% is a quoted per-cent.
152 else if (cp[1] == '%')
153 cp++; /* skip over two % */
154 /* otherwise this is a singleton, literal % */
162 * Make sure the format string is well formed, and parse out
165 static void verify_format(const char *format)
168 for (cp = format; *cp && (sp = find_next(cp)); ) {
169 const char *ep = strchr(sp, ')');
171 die("malformatted format string %s", sp);
172 /* sp points at "%(" and ep points at the closing ")" */
173 parse_atom(sp + 2, ep);
179 * Given an object name, read the object data and size, and return a
180 * "struct object". If the object data we are returning is also borrowed
181 * by the "struct object" representation, set *eaten as well---it is a
182 * signal from parse_object_buffer to us not to free the buffer.
184 static void *get_obj(const unsigned char *sha1, struct object **obj, unsigned long *sz, int *eaten)
186 enum object_type type;
187 void *buf = read_sha1_file(sha1, &type, sz);
190 *obj = parse_object_buffer(sha1, type, *sz, buf, eaten);
196 /* See grab_values */
197 static void grab_common_values(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
201 for (i = 0; i < used_atom_cnt; i++) {
202 const char *name = used_atom[i];
203 struct atom_value *v = &val[i];
204 if (!!deref != (*name == '*'))
208 if (!strcmp(name, "objecttype"))
209 v->s = typename(obj->type);
210 else if (!strcmp(name, "objectsize")) {
211 char *s = xmalloc(40);
212 sprintf(s, "%lu", sz);
216 else if (!strcmp(name, "objectname")) {
217 char *s = xmalloc(41);
218 strcpy(s, sha1_to_hex(obj->sha1));
224 /* See grab_values */
225 static void grab_tag_values(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
228 struct tag *tag = (struct tag *) obj;
230 for (i = 0; i < used_atom_cnt; i++) {
231 const char *name = used_atom[i];
232 struct atom_value *v = &val[i];
233 if (!!deref != (*name == '*'))
237 if (!strcmp(name, "tag"))
242 static int num_parents(struct commit *commit)
244 struct commit_list *parents;
247 for (i = 0, parents = commit->parents;
249 parents = parents->next)
254 /* See grab_values */
255 static void grab_commit_values(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
258 struct commit *commit = (struct commit *) obj;
260 for (i = 0; i < used_atom_cnt; i++) {
261 const char *name = used_atom[i];
262 struct atom_value *v = &val[i];
263 if (!!deref != (*name == '*'))
267 if (!strcmp(name, "tree")) {
268 char *s = xmalloc(41);
269 strcpy(s, sha1_to_hex(commit->tree->object.sha1));
272 if (!strcmp(name, "numparent")) {
273 char *s = xmalloc(40);
274 sprintf(s, "%lu", v->ul);
276 v->ul = num_parents(commit);
278 else if (!strcmp(name, "parent")) {
279 int num = num_parents(commit);
281 struct commit_list *parents;
282 char *s = xmalloc(42 * num);
284 for (i = 0, parents = commit->parents;
286 parents = parents->next, i = i + 42) {
287 struct commit *parent = parents->item;
288 strcpy(s+i, sha1_to_hex(parent->object.sha1));
296 static const char *find_wholine(const char *who, int wholen, const char *buf, unsigned long sz)
300 if (!strncmp(buf, who, wholen) &&
302 return buf + wholen + 1;
303 eol = strchr(buf, '\n');
308 return ""; /* end of header */
314 static const char *copy_line(const char *buf)
316 const char *eol = strchr(buf, '\n');
322 line = xmalloc(len + 1);
323 memcpy(line, buf, len);
328 static const char *copy_name(const char *buf)
330 const char *eol = strchr(buf, '\n');
331 const char *eoname = strstr(buf, " <");
334 if (!(eoname && eol && eoname < eol))
337 line = xmalloc(len + 1);
338 memcpy(line, buf, len);
343 static const char *copy_email(const char *buf)
345 const char *email = strchr(buf, '<');
346 const char *eoemail = strchr(email, '>');
349 if (!email || !eoemail)
352 len = eoemail - email;
353 line = xmalloc(len + 1);
354 memcpy(line, email, len);
359 static void grab_date(const char *buf, struct atom_value *v, const char *atomname)
361 const char *eoemail = strstr(buf, "> ");
363 unsigned long timestamp;
365 enum date_mode date_mode = DATE_NORMAL;
369 * We got here because atomname ends in "date" or "date<something>";
370 * it's not possible that <something> is not ":<format>" because
371 * parse_atom() wouldn't have allowed it, so we can assume that no
372 * ":" means no format is specified, and use the default.
374 formatp = strchr(atomname, ':');
375 if (formatp != NULL) {
377 date_mode = parse_date_format(formatp);
382 timestamp = strtoul(eoemail + 2, &zone, 10);
383 if (timestamp == ULONG_MAX)
385 tz = strtol(zone, NULL, 10);
386 if ((tz == LONG_MIN || tz == LONG_MAX) && errno == ERANGE)
388 v->s = xstrdup(show_date(timestamp, tz, date_mode));
396 /* See grab_values */
397 static void grab_person(const char *who, struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
400 int wholen = strlen(who);
401 const char *wholine = NULL;
403 for (i = 0; i < used_atom_cnt; i++) {
404 const char *name = used_atom[i];
405 struct atom_value *v = &val[i];
406 if (!!deref != (*name == '*'))
410 if (strncmp(who, name, wholen))
412 if (name[wholen] != 0 &&
413 strcmp(name + wholen, "name") &&
414 strcmp(name + wholen, "email") &&
415 prefixcmp(name + wholen, "date"))
418 wholine = find_wholine(who, wholen, buf, sz);
420 return; /* no point looking for it */
421 if (name[wholen] == 0)
422 v->s = copy_line(wholine);
423 else if (!strcmp(name + wholen, "name"))
424 v->s = copy_name(wholine);
425 else if (!strcmp(name + wholen, "email"))
426 v->s = copy_email(wholine);
427 else if (!prefixcmp(name + wholen, "date"))
428 grab_date(wholine, v, name);
431 /* For a tag or a commit object, if "creator" or "creatordate" is
432 * requested, do something special.
434 if (strcmp(who, "tagger") && strcmp(who, "committer"))
435 return; /* "author" for commit object is not wanted */
437 wholine = find_wholine(who, wholen, buf, sz);
440 for (i = 0; i < used_atom_cnt; i++) {
441 const char *name = used_atom[i];
442 struct atom_value *v = &val[i];
443 if (!!deref != (*name == '*'))
448 if (!prefixcmp(name, "creatordate"))
449 grab_date(wholine, v, name);
450 else if (!strcmp(name, "creator"))
451 v->s = copy_line(wholine);
455 static void find_subpos(const char *buf, unsigned long sz, const char **sub, const char **body)
458 const char *eol = strchr(buf, '\n');
461 if (eol[1] == '\n') {
463 break; /* found end of header */
471 *sub = buf; /* first non-empty line */
472 buf = strchr(buf, '\n');
474 return; /* no body */
476 buf++; /* skip blank between subject and body */
480 /* See grab_values */
481 static void grab_sub_body_contents(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
484 const char *subpos = NULL, *bodypos = NULL;
486 for (i = 0; i < used_atom_cnt; i++) {
487 const char *name = used_atom[i];
488 struct atom_value *v = &val[i];
489 if (!!deref != (*name == '*'))
493 if (strcmp(name, "subject") &&
494 strcmp(name, "body") &&
495 strcmp(name, "contents"))
498 find_subpos(buf, sz, &subpos, &bodypos);
502 if (!strcmp(name, "subject"))
503 v->s = copy_line(subpos);
504 else if (!strcmp(name, "body"))
505 v->s = xstrdup(bodypos);
506 else if (!strcmp(name, "contents"))
507 v->s = xstrdup(subpos);
511 /* We want to have empty print-string for field requests
512 * that do not apply (e.g. "authordate" for a tag object)
514 static void fill_missing_values(struct atom_value *val)
517 for (i = 0; i < used_atom_cnt; i++) {
518 struct atom_value *v = &val[i];
525 * val is a list of atom_value to hold returned values. Extract
526 * the values for atoms in used_atom array out of (obj, buf, sz).
527 * when deref is false, (obj, buf, sz) is the object that is
528 * pointed at by the ref itself; otherwise it is the object the
529 * ref (which is a tag) refers to.
531 static void grab_values(struct atom_value *val, int deref, struct object *obj, void *buf, unsigned long sz)
533 grab_common_values(val, deref, obj, buf, sz);
536 grab_tag_values(val, deref, obj, buf, sz);
537 grab_sub_body_contents(val, deref, obj, buf, sz);
538 grab_person("tagger", val, deref, obj, buf, sz);
541 grab_commit_values(val, deref, obj, buf, sz);
542 grab_sub_body_contents(val, deref, obj, buf, sz);
543 grab_person("author", val, deref, obj, buf, sz);
544 grab_person("committer", val, deref, obj, buf, sz);
547 // grab_tree_values(val, deref, obj, buf, sz);
550 // grab_blob_values(val, deref, obj, buf, sz);
553 die("Eh? Object of type %d?", obj->type);
558 * Parse the object referred by ref, and grab needed value.
560 static void populate_value(struct refinfo *ref)
566 const unsigned char *tagged;
568 ref->value = xcalloc(sizeof(struct atom_value), used_atom_cnt);
570 buf = get_obj(ref->objectname, &obj, &size, &eaten);
572 die("missing object %s for %s",
573 sha1_to_hex(ref->objectname), ref->refname);
575 die("parse_object_buffer failed on %s for %s",
576 sha1_to_hex(ref->objectname), ref->refname);
578 /* Fill in specials first */
579 for (i = 0; i < used_atom_cnt; i++) {
580 const char *name = used_atom[i];
581 struct atom_value *v = &ref->value[i];
582 if (!strcmp(name, "refname"))
584 else if (!strcmp(name, "*refname")) {
585 int len = strlen(ref->refname);
586 char *s = xmalloc(len + 4);
587 sprintf(s, "%s^{}", ref->refname);
592 grab_values(ref->value, 0, obj, buf, size);
596 /* If there is no atom that wants to know about tagged
597 * object, we are done.
599 if (!need_tagged || (obj->type != OBJ_TAG))
602 /* If it is a tag object, see if we use a value that derefs
603 * the object, and if we do grab the object it refers to.
605 tagged = ((struct tag *)obj)->tagged->sha1;
607 /* NEEDSWORK: This derefs tag only once, which
608 * is good to deal with chains of trust, but
609 * is not consistent with what deref_tag() does
610 * which peels the onion to the core.
612 buf = get_obj(tagged, &obj, &size, &eaten);
614 die("missing object %s for %s",
615 sha1_to_hex(tagged), ref->refname);
617 die("parse_object_buffer failed on %s for %s",
618 sha1_to_hex(tagged), ref->refname);
619 grab_values(ref->value, 1, obj, buf, size);
625 * Given a ref, return the value for the atom. This lazily gets value
626 * out of the object by calling populate value.
628 static void get_value(struct refinfo *ref, int atom, struct atom_value **v)
632 fill_missing_values(ref->value);
634 *v = &ref->value[atom];
637 struct grab_ref_cbdata {
638 struct refinfo **grab_array;
639 const char **grab_pattern;
644 * A call-back given to for_each_ref(). It is unfortunate that we
645 * need to use global variables to pass extra information to this
648 static int grab_single_ref(const char *refname, const unsigned char *sha1, int flag, void *cb_data)
650 struct grab_ref_cbdata *cb = cb_data;
654 if (*cb->grab_pattern) {
655 const char **pattern;
656 int namelen = strlen(refname);
657 for (pattern = cb->grab_pattern; *pattern; pattern++) {
658 const char *p = *pattern;
659 int plen = strlen(p);
661 if ((plen <= namelen) &&
662 !strncmp(refname, p, plen) &&
663 (refname[plen] == '\0' ||
664 refname[plen] == '/'))
666 if (!fnmatch(p, refname, FNM_PATHNAME))
673 /* We do not open the object yet; sort may only need refname
674 * to do its job and the resulting list may yet to be pruned
677 ref = xcalloc(1, sizeof(*ref));
678 ref->refname = xstrdup(refname);
679 hashcpy(ref->objectname, sha1);
682 cb->grab_array = xrealloc(cb->grab_array,
683 sizeof(*cb->grab_array) * (cnt + 1));
684 cb->grab_array[cnt++] = ref;
689 static int cmp_ref_sort(struct ref_sort *s, struct refinfo *a, struct refinfo *b)
691 struct atom_value *va, *vb;
693 cmp_type cmp_type = used_atom_type[s->atom];
695 get_value(a, s->atom, &va);
696 get_value(b, s->atom, &vb);
699 cmp = strcmp(va->s, vb->s);
704 else if (va->ul == vb->ul)
710 return (s->reverse) ? -cmp : cmp;
713 static struct ref_sort *ref_sort;
714 static int compare_refs(const void *a_, const void *b_)
716 struct refinfo *a = *((struct refinfo **)a_);
717 struct refinfo *b = *((struct refinfo **)b_);
720 for (s = ref_sort; s; s = s->next) {
721 int cmp = cmp_ref_sort(s, a, b);
728 static void sort_refs(struct ref_sort *sort, struct refinfo **refs, int num_refs)
731 qsort(refs, num_refs, sizeof(struct refinfo *), compare_refs);
734 static void print_value(struct refinfo *ref, int atom, int quote_style)
736 struct atom_value *v;
737 get_value(ref, atom, &v);
738 switch (quote_style) {
743 sq_quote_print(stdout, v->s);
746 perl_quote_print(stdout, v->s);
749 python_quote_print(stdout, v->s);
752 tcl_quote_print(stdout, v->s);
757 static int hex1(char ch)
759 if ('0' <= ch && ch <= '9')
761 else if ('a' <= ch && ch <= 'f')
762 return ch - 'a' + 10;
763 else if ('A' <= ch && ch <= 'F')
764 return ch - 'A' + 10;
767 static int hex2(const char *cp)
770 return (hex1(cp[0]) << 4) | hex1(cp[1]);
775 static void emit(const char *cp, const char *ep)
777 while (*cp && (!ep || cp < ep)) {
782 int ch = hex2(cp + 1);
795 static void show_ref(struct refinfo *info, const char *format, int quote_style)
797 const char *cp, *sp, *ep;
799 for (cp = format; *cp && (sp = find_next(cp)); cp = ep + 1) {
800 ep = strchr(sp, ')');
803 print_value(info, parse_atom(sp + 2, ep), quote_style);
806 sp = cp + strlen(cp);
812 static struct ref_sort *default_sort(void)
814 static const char cstr_name[] = "refname";
816 struct ref_sort *sort = xcalloc(1, sizeof(*sort));
819 sort->atom = parse_atom(cstr_name, cstr_name + strlen(cstr_name));
823 int cmd_for_each_ref(int ac, const char **av, const char *prefix)
826 const char *format = NULL;
827 struct ref_sort *sort = NULL, **sort_tail = &sort;
829 int quote_style = -1; /* unspecified yet */
830 struct refinfo **refs;
831 struct grab_ref_cbdata cbdata;
833 for (i = 1; i < ac; i++) {
834 const char *arg = av[i];
837 if (!strcmp(arg, "--")) {
841 if (!prefixcmp(arg, "--format=")) {
843 die("more than one --format?");
847 if (!strcmp(arg, "-s") || !strcmp(arg, "--shell") ) {
848 if (0 <= quote_style)
849 die("more than one quoting style?");
850 quote_style = QUOTE_SHELL;
853 if (!strcmp(arg, "-p") || !strcmp(arg, "--perl") ) {
854 if (0 <= quote_style)
855 die("more than one quoting style?");
856 quote_style = QUOTE_PERL;
859 if (!strcmp(arg, "--python") ) {
860 if (0 <= quote_style)
861 die("more than one quoting style?");
862 quote_style = QUOTE_PYTHON;
865 if (!strcmp(arg, "--tcl") ) {
866 if (0 <= quote_style)
867 die("more than one quoting style?");
868 quote_style = QUOTE_TCL;
871 if (!prefixcmp(arg, "--count=")) {
873 die("more than one --count?");
874 maxcount = atoi(arg + 8);
876 die("The number %s did not parse", arg);
879 if (!prefixcmp(arg, "--sort=")) {
880 struct ref_sort *s = xcalloc(1, sizeof(*s));
885 sort_tail = &s->next;
893 sort->atom = parse_atom(arg, arg+len);
899 quote_style = QUOTE_NONE;
902 sort = default_sort();
903 sort_atom_limit = used_atom_cnt;
905 format = "%(objectname) %(objecttype)\t%(refname)";
907 verify_format(format);
909 memset(&cbdata, 0, sizeof(cbdata));
910 cbdata.grab_pattern = av + i;
911 for_each_ref(grab_single_ref, &cbdata);
912 refs = cbdata.grab_array;
913 num_refs = cbdata.grab_cnt;
915 for (i = 0; i < used_atom_cnt; i++) {
916 if (used_atom[i][0] == '*') {
922 sort_refs(sort, refs, num_refs);
924 if (!maxcount || num_refs < maxcount)
926 for (i = 0; i < maxcount; i++)
927 show_ref(refs[i], format, quote_style);