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
},
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
, ':');
116 if (!formatp
|| ep
< formatp
)
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 v
->ul
= num_parents(commit
);
275 sprintf(s
, "%lu", v
->ul
);
278 else if (!strcmp(name
, "parent")) {
279 int num
= num_parents(commit
);
281 struct commit_list
*parents
;
282 char *s
= xmalloc(41 * num
+ 1);
284 for (i
= 0, parents
= commit
->parents
;
286 parents
= parents
->next
, i
= i
+ 41) {
287 struct commit
*parent
= parents
->item
;
288 strcpy(s
+i
, sha1_to_hex(parent
->object
.sha1
));
298 static const char *find_wholine(const char *who
, int wholen
, const char *buf
, unsigned long sz
)
302 if (!strncmp(buf
, who
, wholen
) &&
304 return buf
+ wholen
+ 1;
305 eol
= strchr(buf
, '\n');
310 return ""; /* end of header */
316 static const char *copy_line(const char *buf
)
318 const char *eol
= strchr(buf
, '\n');
324 line
= xmalloc(len
+ 1);
325 memcpy(line
, buf
, len
);
330 static const char *copy_name(const char *buf
)
332 const char *eol
= strchr(buf
, '\n');
333 const char *eoname
= strstr(buf
, " <");
336 if (!(eoname
&& eol
&& eoname
< eol
))
339 line
= xmalloc(len
+ 1);
340 memcpy(line
, buf
, len
);
345 static const char *copy_email(const char *buf
)
347 const char *email
= strchr(buf
, '<');
348 const char *eoemail
= strchr(email
, '>');
351 if (!email
|| !eoemail
)
354 len
= eoemail
- email
;
355 line
= xmalloc(len
+ 1);
356 memcpy(line
, email
, len
);
361 static void grab_date(const char *buf
, struct atom_value
*v
, const char *atomname
)
363 const char *eoemail
= strstr(buf
, "> ");
365 unsigned long timestamp
;
367 enum date_mode date_mode
= DATE_NORMAL
;
371 * We got here because atomname ends in "date" or "date<something>";
372 * it's not possible that <something> is not ":<format>" because
373 * parse_atom() wouldn't have allowed it, so we can assume that no
374 * ":" means no format is specified, and use the default.
376 formatp
= strchr(atomname
, ':');
377 if (formatp
!= NULL
) {
379 date_mode
= parse_date_format(formatp
);
384 timestamp
= strtoul(eoemail
+ 2, &zone
, 10);
385 if (timestamp
== ULONG_MAX
)
387 tz
= strtol(zone
, NULL
, 10);
388 if ((tz
== LONG_MIN
|| tz
== LONG_MAX
) && errno
== ERANGE
)
390 v
->s
= xstrdup(show_date(timestamp
, tz
, date_mode
));
398 /* See grab_values */
399 static void grab_person(const char *who
, struct atom_value
*val
, int deref
, struct object
*obj
, void *buf
, unsigned long sz
)
402 int wholen
= strlen(who
);
403 const char *wholine
= NULL
;
405 for (i
= 0; i
< used_atom_cnt
; i
++) {
406 const char *name
= used_atom
[i
];
407 struct atom_value
*v
= &val
[i
];
408 if (!!deref
!= (*name
== '*'))
412 if (strncmp(who
, name
, wholen
))
414 if (name
[wholen
] != 0 &&
415 strcmp(name
+ wholen
, "name") &&
416 strcmp(name
+ wholen
, "email") &&
417 prefixcmp(name
+ wholen
, "date"))
420 wholine
= find_wholine(who
, wholen
, buf
, sz
);
422 return; /* no point looking for it */
423 if (name
[wholen
] == 0)
424 v
->s
= copy_line(wholine
);
425 else if (!strcmp(name
+ wholen
, "name"))
426 v
->s
= copy_name(wholine
);
427 else if (!strcmp(name
+ wholen
, "email"))
428 v
->s
= copy_email(wholine
);
429 else if (!prefixcmp(name
+ wholen
, "date"))
430 grab_date(wholine
, v
, name
);
433 /* For a tag or a commit object, if "creator" or "creatordate" is
434 * requested, do something special.
436 if (strcmp(who
, "tagger") && strcmp(who
, "committer"))
437 return; /* "author" for commit object is not wanted */
439 wholine
= find_wholine(who
, wholen
, buf
, sz
);
442 for (i
= 0; i
< used_atom_cnt
; i
++) {
443 const char *name
= used_atom
[i
];
444 struct atom_value
*v
= &val
[i
];
445 if (!!deref
!= (*name
== '*'))
450 if (!prefixcmp(name
, "creatordate"))
451 grab_date(wholine
, v
, name
);
452 else if (!strcmp(name
, "creator"))
453 v
->s
= copy_line(wholine
);
457 static void find_subpos(const char *buf
, unsigned long sz
, const char **sub
, const char **body
)
460 const char *eol
= strchr(buf
, '\n');
463 if (eol
[1] == '\n') {
465 break; /* found end of header */
473 *sub
= buf
; /* first non-empty line */
474 buf
= strchr(buf
, '\n');
476 return; /* no body */
478 buf
++; /* skip blank between subject and body */
482 /* See grab_values */
483 static void grab_sub_body_contents(struct atom_value
*val
, int deref
, struct object
*obj
, void *buf
, unsigned long sz
)
486 const char *subpos
= NULL
, *bodypos
= NULL
;
488 for (i
= 0; i
< used_atom_cnt
; i
++) {
489 const char *name
= used_atom
[i
];
490 struct atom_value
*v
= &val
[i
];
491 if (!!deref
!= (*name
== '*'))
495 if (strcmp(name
, "subject") &&
496 strcmp(name
, "body") &&
497 strcmp(name
, "contents"))
500 find_subpos(buf
, sz
, &subpos
, &bodypos
);
504 if (!strcmp(name
, "subject"))
505 v
->s
= copy_line(subpos
);
506 else if (!strcmp(name
, "body"))
507 v
->s
= xstrdup(bodypos
);
508 else if (!strcmp(name
, "contents"))
509 v
->s
= xstrdup(subpos
);
513 /* We want to have empty print-string for field requests
514 * that do not apply (e.g. "authordate" for a tag object)
516 static void fill_missing_values(struct atom_value
*val
)
519 for (i
= 0; i
< used_atom_cnt
; i
++) {
520 struct atom_value
*v
= &val
[i
];
527 * val is a list of atom_value to hold returned values. Extract
528 * the values for atoms in used_atom array out of (obj, buf, sz).
529 * when deref is false, (obj, buf, sz) is the object that is
530 * pointed at by the ref itself; otherwise it is the object the
531 * ref (which is a tag) refers to.
533 static void grab_values(struct atom_value
*val
, int deref
, struct object
*obj
, void *buf
, unsigned long sz
)
535 grab_common_values(val
, deref
, obj
, buf
, sz
);
538 grab_tag_values(val
, deref
, obj
, buf
, sz
);
539 grab_sub_body_contents(val
, deref
, obj
, buf
, sz
);
540 grab_person("tagger", val
, deref
, obj
, buf
, sz
);
543 grab_commit_values(val
, deref
, obj
, buf
, sz
);
544 grab_sub_body_contents(val
, deref
, obj
, buf
, sz
);
545 grab_person("author", val
, deref
, obj
, buf
, sz
);
546 grab_person("committer", val
, deref
, obj
, buf
, sz
);
549 // grab_tree_values(val, deref, obj, buf, sz);
552 // grab_blob_values(val, deref, obj, buf, sz);
555 die("Eh? Object of type %d?", obj
->type
);
560 * Parse the object referred by ref, and grab needed value.
562 static void populate_value(struct refinfo
*ref
)
568 const unsigned char *tagged
;
570 ref
->value
= xcalloc(sizeof(struct atom_value
), used_atom_cnt
);
572 buf
= get_obj(ref
->objectname
, &obj
, &size
, &eaten
);
574 die("missing object %s for %s",
575 sha1_to_hex(ref
->objectname
), ref
->refname
);
577 die("parse_object_buffer failed on %s for %s",
578 sha1_to_hex(ref
->objectname
), ref
->refname
);
580 /* Fill in specials first */
581 for (i
= 0; i
< used_atom_cnt
; i
++) {
582 const char *name
= used_atom
[i
];
583 struct atom_value
*v
= &ref
->value
[i
];
584 if (!strcmp(name
, "refname"))
586 else if (!strcmp(name
, "*refname")) {
587 int len
= strlen(ref
->refname
);
588 char *s
= xmalloc(len
+ 4);
589 sprintf(s
, "%s^{}", ref
->refname
);
594 grab_values(ref
->value
, 0, obj
, buf
, size
);
598 /* If there is no atom that wants to know about tagged
599 * object, we are done.
601 if (!need_tagged
|| (obj
->type
!= OBJ_TAG
))
604 /* If it is a tag object, see if we use a value that derefs
605 * the object, and if we do grab the object it refers to.
607 tagged
= ((struct tag
*)obj
)->tagged
->sha1
;
609 /* NEEDSWORK: This derefs tag only once, which
610 * is good to deal with chains of trust, but
611 * is not consistent with what deref_tag() does
612 * which peels the onion to the core.
614 buf
= get_obj(tagged
, &obj
, &size
, &eaten
);
616 die("missing object %s for %s",
617 sha1_to_hex(tagged
), ref
->refname
);
619 die("parse_object_buffer failed on %s for %s",
620 sha1_to_hex(tagged
), ref
->refname
);
621 grab_values(ref
->value
, 1, obj
, buf
, size
);
627 * Given a ref, return the value for the atom. This lazily gets value
628 * out of the object by calling populate value.
630 static void get_value(struct refinfo
*ref
, int atom
, struct atom_value
**v
)
634 fill_missing_values(ref
->value
);
636 *v
= &ref
->value
[atom
];
639 struct grab_ref_cbdata
{
640 struct refinfo
**grab_array
;
641 const char **grab_pattern
;
646 * A call-back given to for_each_ref(). It is unfortunate that we
647 * need to use global variables to pass extra information to this
650 static int grab_single_ref(const char *refname
, const unsigned char *sha1
, int flag
, void *cb_data
)
652 struct grab_ref_cbdata
*cb
= cb_data
;
656 if (*cb
->grab_pattern
) {
657 const char **pattern
;
658 int namelen
= strlen(refname
);
659 for (pattern
= cb
->grab_pattern
; *pattern
; pattern
++) {
660 const char *p
= *pattern
;
661 int plen
= strlen(p
);
663 if ((plen
<= namelen
) &&
664 !strncmp(refname
, p
, plen
) &&
665 (refname
[plen
] == '\0' ||
666 refname
[plen
] == '/'))
668 if (!fnmatch(p
, refname
, FNM_PATHNAME
))
675 /* We do not open the object yet; sort may only need refname
676 * to do its job and the resulting list may yet to be pruned
679 ref
= xcalloc(1, sizeof(*ref
));
680 ref
->refname
= xstrdup(refname
);
681 hashcpy(ref
->objectname
, sha1
);
684 cb
->grab_array
= xrealloc(cb
->grab_array
,
685 sizeof(*cb
->grab_array
) * (cnt
+ 1));
686 cb
->grab_array
[cnt
++] = ref
;
691 static int cmp_ref_sort(struct ref_sort
*s
, struct refinfo
*a
, struct refinfo
*b
)
693 struct atom_value
*va
, *vb
;
695 cmp_type cmp_type
= used_atom_type
[s
->atom
];
697 get_value(a
, s
->atom
, &va
);
698 get_value(b
, s
->atom
, &vb
);
701 cmp
= strcmp(va
->s
, vb
->s
);
706 else if (va
->ul
== vb
->ul
)
712 return (s
->reverse
) ? -cmp
: cmp
;
715 static struct ref_sort
*ref_sort
;
716 static int compare_refs(const void *a_
, const void *b_
)
718 struct refinfo
*a
= *((struct refinfo
**)a_
);
719 struct refinfo
*b
= *((struct refinfo
**)b_
);
722 for (s
= ref_sort
; s
; s
= s
->next
) {
723 int cmp
= cmp_ref_sort(s
, a
, b
);
730 static void sort_refs(struct ref_sort
*sort
, struct refinfo
**refs
, int num_refs
)
733 qsort(refs
, num_refs
, sizeof(struct refinfo
*), compare_refs
);
736 static void print_value(struct refinfo
*ref
, int atom
, int quote_style
)
738 struct atom_value
*v
;
739 get_value(ref
, atom
, &v
);
740 switch (quote_style
) {
745 sq_quote_print(stdout
, v
->s
);
748 perl_quote_print(stdout
, v
->s
);
751 python_quote_print(stdout
, v
->s
);
754 tcl_quote_print(stdout
, v
->s
);
759 static int hex1(char ch
)
761 if ('0' <= ch
&& ch
<= '9')
763 else if ('a' <= ch
&& ch
<= 'f')
764 return ch
- 'a' + 10;
765 else if ('A' <= ch
&& ch
<= 'F')
766 return ch
- 'A' + 10;
769 static int hex2(const char *cp
)
772 return (hex1(cp
[0]) << 4) | hex1(cp
[1]);
777 static void emit(const char *cp
, const char *ep
)
779 while (*cp
&& (!ep
|| cp
< ep
)) {
784 int ch
= hex2(cp
+ 1);
797 static void show_ref(struct refinfo
*info
, const char *format
, int quote_style
)
799 const char *cp
, *sp
, *ep
;
801 for (cp
= format
; *cp
&& (sp
= find_next(cp
)); cp
= ep
+ 1) {
802 ep
= strchr(sp
, ')');
805 print_value(info
, parse_atom(sp
+ 2, ep
), quote_style
);
808 sp
= cp
+ strlen(cp
);
814 static struct ref_sort
*default_sort(void)
816 static const char cstr_name
[] = "refname";
818 struct ref_sort
*sort
= xcalloc(1, sizeof(*sort
));
821 sort
->atom
= parse_atom(cstr_name
, cstr_name
+ strlen(cstr_name
));
825 int cmd_for_each_ref(int ac
, const char **av
, const char *prefix
)
828 const char *format
= NULL
;
829 struct ref_sort
*sort
= NULL
, **sort_tail
= &sort
;
831 int quote_style
= -1; /* unspecified yet */
832 struct refinfo
**refs
;
833 struct grab_ref_cbdata cbdata
;
835 for (i
= 1; i
< ac
; i
++) {
836 const char *arg
= av
[i
];
839 if (!strcmp(arg
, "--")) {
843 if (!prefixcmp(arg
, "--format=")) {
845 die("more than one --format?");
849 if (!strcmp(arg
, "-s") || !strcmp(arg
, "--shell") ) {
850 if (0 <= quote_style
)
851 die("more than one quoting style?");
852 quote_style
= QUOTE_SHELL
;
855 if (!strcmp(arg
, "-p") || !strcmp(arg
, "--perl") ) {
856 if (0 <= quote_style
)
857 die("more than one quoting style?");
858 quote_style
= QUOTE_PERL
;
861 if (!strcmp(arg
, "--python") ) {
862 if (0 <= quote_style
)
863 die("more than one quoting style?");
864 quote_style
= QUOTE_PYTHON
;
867 if (!strcmp(arg
, "--tcl") ) {
868 if (0 <= quote_style
)
869 die("more than one quoting style?");
870 quote_style
= QUOTE_TCL
;
873 if (!prefixcmp(arg
, "--count=")) {
875 die("more than one --count?");
876 maxcount
= atoi(arg
+ 8);
878 die("The number %s did not parse", arg
);
881 if (!prefixcmp(arg
, "--sort=")) {
882 struct ref_sort
*s
= xcalloc(1, sizeof(*s
));
887 sort_tail
= &s
->next
;
895 sort
->atom
= parse_atom(arg
, arg
+len
);
901 quote_style
= QUOTE_NONE
;
904 sort
= default_sort();
905 sort_atom_limit
= used_atom_cnt
;
907 format
= "%(objectname) %(objecttype)\t%(refname)";
909 verify_format(format
);
911 memset(&cbdata
, 0, sizeof(cbdata
));
912 cbdata
.grab_pattern
= av
+ i
;
913 for_each_ref(grab_single_ref
, &cbdata
);
914 refs
= cbdata
.grab_array
;
915 num_refs
= cbdata
.grab_cnt
;
917 for (i
= 0; i
< used_atom_cnt
; i
++) {
918 if (used_atom
[i
][0] == '*') {
924 sort_refs(sort
, refs
, num_refs
);
926 if (!maxcount
|| num_refs
< maxcount
)
928 for (i
= 0; i
< maxcount
; i
++)
929 show_ref(refs
[i
], format
, quote_style
);