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
)
93 if (*sp
== '*' && sp
< ep
)
96 die("malformed field name: %.*s", (int)(ep
-atom
), atom
);
98 /* Do we have the atom already used elsewhere? */
99 for (i
= 0; i
< used_atom_cnt
; i
++) {
100 int len
= strlen(used_atom
[i
]);
101 if (len
== ep
- atom
&& !memcmp(used_atom
[i
], atom
, len
))
105 /* Is the atom a valid one? */
106 for (i
= 0; i
< ARRAY_SIZE(valid_atom
); i
++) {
107 int len
= strlen(valid_atom
[i
].name
);
109 * If the atom name has a colon, strip it and everything after
110 * it off - it specifies the format for this entry, and
111 * shouldn't be used for checking against the valid_atom
114 const char *formatp
= strchr(sp
, ':');
115 if (!formatp
|| ep
< formatp
)
117 if (len
== formatp
- sp
&& !memcmp(valid_atom
[i
].name
, sp
, len
))
121 if (ARRAY_SIZE(valid_atom
) <= i
)
122 die("unknown field name: %.*s", (int)(ep
-atom
), atom
);
124 /* Add it in, including the deref prefix */
127 used_atom
= xrealloc(used_atom
,
128 (sizeof *used_atom
) * used_atom_cnt
);
129 used_atom_type
= xrealloc(used_atom_type
,
130 (sizeof(*used_atom_type
) * used_atom_cnt
));
131 used_atom
[at
] = xmemdupz(atom
, ep
- atom
);
132 used_atom_type
[at
] = valid_atom
[i
].cmp_type
;
137 * In a format string, find the next occurrence of %(atom).
139 static const char *find_next(const char *cp
)
143 /* %( is the start of an atom;
144 * %% is a quoted per-cent.
148 else if (cp
[1] == '%')
149 cp
++; /* skip over two % */
150 /* otherwise this is a singleton, literal % */
158 * Make sure the format string is well formed, and parse out
161 static void verify_format(const char *format
)
164 for (cp
= format
; *cp
&& (sp
= find_next(cp
)); ) {
165 const char *ep
= strchr(sp
, ')');
167 die("malformatted format string %s", sp
);
168 /* sp points at "%(" and ep points at the closing ")" */
169 parse_atom(sp
+ 2, ep
);
175 * Given an object name, read the object data and size, and return a
176 * "struct object". If the object data we are returning is also borrowed
177 * by the "struct object" representation, set *eaten as well---it is a
178 * signal from parse_object_buffer to us not to free the buffer.
180 static void *get_obj(const unsigned char *sha1
, struct object
**obj
, unsigned long *sz
, int *eaten
)
182 enum object_type type
;
183 void *buf
= read_sha1_file(sha1
, &type
, sz
);
186 *obj
= parse_object_buffer(sha1
, type
, *sz
, buf
, eaten
);
192 /* See grab_values */
193 static void grab_common_values(struct atom_value
*val
, int deref
, struct object
*obj
, void *buf
, unsigned long sz
)
197 for (i
= 0; i
< used_atom_cnt
; i
++) {
198 const char *name
= used_atom
[i
];
199 struct atom_value
*v
= &val
[i
];
200 if (!!deref
!= (*name
== '*'))
204 if (!strcmp(name
, "objecttype"))
205 v
->s
= typename(obj
->type
);
206 else if (!strcmp(name
, "objectsize")) {
207 char *s
= xmalloc(40);
208 sprintf(s
, "%lu", sz
);
212 else if (!strcmp(name
, "objectname")) {
213 char *s
= xmalloc(41);
214 strcpy(s
, sha1_to_hex(obj
->sha1
));
220 /* See grab_values */
221 static void grab_tag_values(struct atom_value
*val
, int deref
, struct object
*obj
, void *buf
, unsigned long sz
)
224 struct tag
*tag
= (struct tag
*) obj
;
226 for (i
= 0; i
< used_atom_cnt
; i
++) {
227 const char *name
= used_atom
[i
];
228 struct atom_value
*v
= &val
[i
];
229 if (!!deref
!= (*name
== '*'))
233 if (!strcmp(name
, "tag"))
238 static int num_parents(struct commit
*commit
)
240 struct commit_list
*parents
;
243 for (i
= 0, parents
= commit
->parents
;
245 parents
= parents
->next
)
250 /* See grab_values */
251 static void grab_commit_values(struct atom_value
*val
, int deref
, struct object
*obj
, void *buf
, unsigned long sz
)
254 struct commit
*commit
= (struct commit
*) obj
;
256 for (i
= 0; i
< used_atom_cnt
; i
++) {
257 const char *name
= used_atom
[i
];
258 struct atom_value
*v
= &val
[i
];
259 if (!!deref
!= (*name
== '*'))
263 if (!strcmp(name
, "tree")) {
264 char *s
= xmalloc(41);
265 strcpy(s
, sha1_to_hex(commit
->tree
->object
.sha1
));
268 if (!strcmp(name
, "numparent")) {
269 char *s
= xmalloc(40);
270 v
->ul
= num_parents(commit
);
271 sprintf(s
, "%lu", v
->ul
);
274 else if (!strcmp(name
, "parent")) {
275 int num
= num_parents(commit
);
277 struct commit_list
*parents
;
278 char *s
= xmalloc(41 * num
+ 1);
280 for (i
= 0, parents
= commit
->parents
;
282 parents
= parents
->next
, i
= i
+ 41) {
283 struct commit
*parent
= parents
->item
;
284 strcpy(s
+i
, sha1_to_hex(parent
->object
.sha1
));
294 static const char *find_wholine(const char *who
, int wholen
, const char *buf
, unsigned long sz
)
298 if (!strncmp(buf
, who
, wholen
) &&
300 return buf
+ wholen
+ 1;
301 eol
= strchr(buf
, '\n');
306 return ""; /* end of header */
312 static const char *copy_line(const char *buf
)
314 const char *eol
= strchr(buf
, '\n');
317 return xmemdupz(buf
, eol
- buf
);
320 static const char *copy_name(const char *buf
)
323 for (cp
= buf
; *cp
&& *cp
!= '\n'; cp
++) {
324 if (!strncmp(cp
, " <", 2))
325 return xmemdupz(buf
, cp
- buf
);
330 static const char *copy_email(const char *buf
)
332 const char *email
= strchr(buf
, '<');
333 const char *eoemail
= strchr(email
, '>');
334 if (!email
|| !eoemail
)
336 return xmemdupz(email
, eoemail
+ 1 - email
);
339 static void grab_date(const char *buf
, struct atom_value
*v
, const char *atomname
)
341 const char *eoemail
= strstr(buf
, "> ");
343 unsigned long timestamp
;
345 enum date_mode date_mode
= DATE_NORMAL
;
349 * We got here because atomname ends in "date" or "date<something>";
350 * it's not possible that <something> is not ":<format>" because
351 * parse_atom() wouldn't have allowed it, so we can assume that no
352 * ":" means no format is specified, and use the default.
354 formatp
= strchr(atomname
, ':');
355 if (formatp
!= NULL
) {
357 date_mode
= parse_date_format(formatp
);
362 timestamp
= strtoul(eoemail
+ 2, &zone
, 10);
363 if (timestamp
== ULONG_MAX
)
365 tz
= strtol(zone
, NULL
, 10);
366 if ((tz
== LONG_MIN
|| tz
== LONG_MAX
) && errno
== ERANGE
)
368 v
->s
= xstrdup(show_date(timestamp
, tz
, date_mode
));
376 /* See grab_values */
377 static void grab_person(const char *who
, struct atom_value
*val
, int deref
, struct object
*obj
, void *buf
, unsigned long sz
)
380 int wholen
= strlen(who
);
381 const char *wholine
= NULL
;
383 for (i
= 0; i
< used_atom_cnt
; i
++) {
384 const char *name
= used_atom
[i
];
385 struct atom_value
*v
= &val
[i
];
386 if (!!deref
!= (*name
== '*'))
390 if (strncmp(who
, name
, wholen
))
392 if (name
[wholen
] != 0 &&
393 strcmp(name
+ wholen
, "name") &&
394 strcmp(name
+ wholen
, "email") &&
395 prefixcmp(name
+ wholen
, "date"))
398 wholine
= find_wholine(who
, wholen
, buf
, sz
);
400 return; /* no point looking for it */
401 if (name
[wholen
] == 0)
402 v
->s
= copy_line(wholine
);
403 else if (!strcmp(name
+ wholen
, "name"))
404 v
->s
= copy_name(wholine
);
405 else if (!strcmp(name
+ wholen
, "email"))
406 v
->s
= copy_email(wholine
);
407 else if (!prefixcmp(name
+ wholen
, "date"))
408 grab_date(wholine
, v
, name
);
411 /* For a tag or a commit object, if "creator" or "creatordate" is
412 * requested, do something special.
414 if (strcmp(who
, "tagger") && strcmp(who
, "committer"))
415 return; /* "author" for commit object is not wanted */
417 wholine
= find_wholine(who
, wholen
, buf
, sz
);
420 for (i
= 0; i
< used_atom_cnt
; i
++) {
421 const char *name
= used_atom
[i
];
422 struct atom_value
*v
= &val
[i
];
423 if (!!deref
!= (*name
== '*'))
428 if (!prefixcmp(name
, "creatordate"))
429 grab_date(wholine
, v
, name
);
430 else if (!strcmp(name
, "creator"))
431 v
->s
= copy_line(wholine
);
435 static void find_subpos(const char *buf
, unsigned long sz
, const char **sub
, const char **body
)
438 const char *eol
= strchr(buf
, '\n');
441 if (eol
[1] == '\n') {
443 break; /* found end of header */
451 *sub
= buf
; /* first non-empty line */
452 buf
= strchr(buf
, '\n');
454 return; /* no body */
456 buf
++; /* skip blank between subject and body */
460 /* See grab_values */
461 static void grab_sub_body_contents(struct atom_value
*val
, int deref
, struct object
*obj
, void *buf
, unsigned long sz
)
464 const char *subpos
= NULL
, *bodypos
= NULL
;
466 for (i
= 0; i
< used_atom_cnt
; i
++) {
467 const char *name
= used_atom
[i
];
468 struct atom_value
*v
= &val
[i
];
469 if (!!deref
!= (*name
== '*'))
473 if (strcmp(name
, "subject") &&
474 strcmp(name
, "body") &&
475 strcmp(name
, "contents"))
478 find_subpos(buf
, sz
, &subpos
, &bodypos
);
482 if (!strcmp(name
, "subject"))
483 v
->s
= copy_line(subpos
);
484 else if (!strcmp(name
, "body"))
485 v
->s
= xstrdup(bodypos
);
486 else if (!strcmp(name
, "contents"))
487 v
->s
= xstrdup(subpos
);
491 /* We want to have empty print-string for field requests
492 * that do not apply (e.g. "authordate" for a tag object)
494 static void fill_missing_values(struct atom_value
*val
)
497 for (i
= 0; i
< used_atom_cnt
; i
++) {
498 struct atom_value
*v
= &val
[i
];
505 * val is a list of atom_value to hold returned values. Extract
506 * the values for atoms in used_atom array out of (obj, buf, sz).
507 * when deref is false, (obj, buf, sz) is the object that is
508 * pointed at by the ref itself; otherwise it is the object the
509 * ref (which is a tag) refers to.
511 static void grab_values(struct atom_value
*val
, int deref
, struct object
*obj
, void *buf
, unsigned long sz
)
513 grab_common_values(val
, deref
, obj
, buf
, sz
);
516 grab_tag_values(val
, deref
, obj
, buf
, sz
);
517 grab_sub_body_contents(val
, deref
, obj
, buf
, sz
);
518 grab_person("tagger", val
, deref
, obj
, buf
, sz
);
521 grab_commit_values(val
, deref
, obj
, buf
, sz
);
522 grab_sub_body_contents(val
, deref
, obj
, buf
, sz
);
523 grab_person("author", val
, deref
, obj
, buf
, sz
);
524 grab_person("committer", val
, deref
, obj
, buf
, sz
);
527 // grab_tree_values(val, deref, obj, buf, sz);
530 // grab_blob_values(val, deref, obj, buf, sz);
533 die("Eh? Object of type %d?", obj
->type
);
538 * Parse the object referred by ref, and grab needed value.
540 static void populate_value(struct refinfo
*ref
)
546 const unsigned char *tagged
;
548 ref
->value
= xcalloc(sizeof(struct atom_value
), used_atom_cnt
);
550 buf
= get_obj(ref
->objectname
, &obj
, &size
, &eaten
);
552 die("missing object %s for %s",
553 sha1_to_hex(ref
->objectname
), ref
->refname
);
555 die("parse_object_buffer failed on %s for %s",
556 sha1_to_hex(ref
->objectname
), ref
->refname
);
558 /* Fill in specials first */
559 for (i
= 0; i
< used_atom_cnt
; i
++) {
560 const char *name
= used_atom
[i
];
561 struct atom_value
*v
= &ref
->value
[i
];
562 if (!strcmp(name
, "refname"))
564 else if (!strcmp(name
, "*refname")) {
565 int len
= strlen(ref
->refname
);
566 char *s
= xmalloc(len
+ 4);
567 sprintf(s
, "%s^{}", ref
->refname
);
572 grab_values(ref
->value
, 0, obj
, buf
, size
);
576 /* If there is no atom that wants to know about tagged
577 * object, we are done.
579 if (!need_tagged
|| (obj
->type
!= OBJ_TAG
))
582 /* If it is a tag object, see if we use a value that derefs
583 * the object, and if we do grab the object it refers to.
585 tagged
= ((struct tag
*)obj
)->tagged
->sha1
;
587 /* NEEDSWORK: This derefs tag only once, which
588 * is good to deal with chains of trust, but
589 * is not consistent with what deref_tag() does
590 * which peels the onion to the core.
592 buf
= get_obj(tagged
, &obj
, &size
, &eaten
);
594 die("missing object %s for %s",
595 sha1_to_hex(tagged
), ref
->refname
);
597 die("parse_object_buffer failed on %s for %s",
598 sha1_to_hex(tagged
), ref
->refname
);
599 grab_values(ref
->value
, 1, obj
, buf
, size
);
605 * Given a ref, return the value for the atom. This lazily gets value
606 * out of the object by calling populate value.
608 static void get_value(struct refinfo
*ref
, int atom
, struct atom_value
**v
)
612 fill_missing_values(ref
->value
);
614 *v
= &ref
->value
[atom
];
617 struct grab_ref_cbdata
{
618 struct refinfo
**grab_array
;
619 const char **grab_pattern
;
624 * A call-back given to for_each_ref(). It is unfortunate that we
625 * need to use global variables to pass extra information to this
628 static int grab_single_ref(const char *refname
, const unsigned char *sha1
, int flag
, void *cb_data
)
630 struct grab_ref_cbdata
*cb
= cb_data
;
634 if (*cb
->grab_pattern
) {
635 const char **pattern
;
636 int namelen
= strlen(refname
);
637 for (pattern
= cb
->grab_pattern
; *pattern
; pattern
++) {
638 const char *p
= *pattern
;
639 int plen
= strlen(p
);
641 if ((plen
<= namelen
) &&
642 !strncmp(refname
, p
, plen
) &&
643 (refname
[plen
] == '\0' ||
644 refname
[plen
] == '/'))
646 if (!fnmatch(p
, refname
, FNM_PATHNAME
))
653 /* We do not open the object yet; sort may only need refname
654 * to do its job and the resulting list may yet to be pruned
657 ref
= xcalloc(1, sizeof(*ref
));
658 ref
->refname
= xstrdup(refname
);
659 hashcpy(ref
->objectname
, sha1
);
662 cb
->grab_array
= xrealloc(cb
->grab_array
,
663 sizeof(*cb
->grab_array
) * (cnt
+ 1));
664 cb
->grab_array
[cnt
++] = ref
;
669 static int cmp_ref_sort(struct ref_sort
*s
, struct refinfo
*a
, struct refinfo
*b
)
671 struct atom_value
*va
, *vb
;
673 cmp_type cmp_type
= used_atom_type
[s
->atom
];
675 get_value(a
, s
->atom
, &va
);
676 get_value(b
, s
->atom
, &vb
);
679 cmp
= strcmp(va
->s
, vb
->s
);
684 else if (va
->ul
== vb
->ul
)
690 return (s
->reverse
) ? -cmp
: cmp
;
693 static struct ref_sort
*ref_sort
;
694 static int compare_refs(const void *a_
, const void *b_
)
696 struct refinfo
*a
= *((struct refinfo
**)a_
);
697 struct refinfo
*b
= *((struct refinfo
**)b_
);
700 for (s
= ref_sort
; s
; s
= s
->next
) {
701 int cmp
= cmp_ref_sort(s
, a
, b
);
708 static void sort_refs(struct ref_sort
*sort
, struct refinfo
**refs
, int num_refs
)
711 qsort(refs
, num_refs
, sizeof(struct refinfo
*), compare_refs
);
714 static void print_value(struct refinfo
*ref
, int atom
, int quote_style
)
716 struct atom_value
*v
;
717 get_value(ref
, atom
, &v
);
718 switch (quote_style
) {
723 sq_quote_print(stdout
, v
->s
);
726 perl_quote_print(stdout
, v
->s
);
729 python_quote_print(stdout
, v
->s
);
732 tcl_quote_print(stdout
, v
->s
);
737 static int hex1(char ch
)
739 if ('0' <= ch
&& ch
<= '9')
741 else if ('a' <= ch
&& ch
<= 'f')
742 return ch
- 'a' + 10;
743 else if ('A' <= ch
&& ch
<= 'F')
744 return ch
- 'A' + 10;
747 static int hex2(const char *cp
)
750 return (hex1(cp
[0]) << 4) | hex1(cp
[1]);
755 static void emit(const char *cp
, const char *ep
)
757 while (*cp
&& (!ep
|| cp
< ep
)) {
762 int ch
= hex2(cp
+ 1);
775 static void show_ref(struct refinfo
*info
, const char *format
, int quote_style
)
777 const char *cp
, *sp
, *ep
;
779 for (cp
= format
; *cp
&& (sp
= find_next(cp
)); cp
= ep
+ 1) {
780 ep
= strchr(sp
, ')');
783 print_value(info
, parse_atom(sp
+ 2, ep
), quote_style
);
786 sp
= cp
+ strlen(cp
);
792 static struct ref_sort
*default_sort(void)
794 static const char cstr_name
[] = "refname";
796 struct ref_sort
*sort
= xcalloc(1, sizeof(*sort
));
799 sort
->atom
= parse_atom(cstr_name
, cstr_name
+ strlen(cstr_name
));
803 int cmd_for_each_ref(int ac
, const char **av
, const char *prefix
)
806 const char *format
= NULL
;
807 struct ref_sort
*sort
= NULL
, **sort_tail
= &sort
;
809 int quote_style
= -1; /* unspecified yet */
810 struct refinfo
**refs
;
811 struct grab_ref_cbdata cbdata
;
813 for (i
= 1; i
< ac
; i
++) {
814 const char *arg
= av
[i
];
817 if (!strcmp(arg
, "--")) {
821 if (!prefixcmp(arg
, "--format=")) {
823 die("more than one --format?");
827 if (!strcmp(arg
, "-s") || !strcmp(arg
, "--shell") ) {
828 if (0 <= quote_style
)
829 die("more than one quoting style?");
830 quote_style
= QUOTE_SHELL
;
833 if (!strcmp(arg
, "-p") || !strcmp(arg
, "--perl") ) {
834 if (0 <= quote_style
)
835 die("more than one quoting style?");
836 quote_style
= QUOTE_PERL
;
839 if (!strcmp(arg
, "--python") ) {
840 if (0 <= quote_style
)
841 die("more than one quoting style?");
842 quote_style
= QUOTE_PYTHON
;
845 if (!strcmp(arg
, "--tcl") ) {
846 if (0 <= quote_style
)
847 die("more than one quoting style?");
848 quote_style
= QUOTE_TCL
;
851 if (!prefixcmp(arg
, "--count=")) {
853 die("more than one --count?");
854 maxcount
= atoi(arg
+ 8);
856 die("The number %s did not parse", arg
);
859 if (!prefixcmp(arg
, "--sort=")) {
860 struct ref_sort
*s
= xcalloc(1, sizeof(*s
));
865 sort_tail
= &s
->next
;
873 sort
->atom
= parse_atom(arg
, arg
+len
);
879 quote_style
= QUOTE_NONE
;
882 sort
= default_sort();
883 sort_atom_limit
= used_atom_cnt
;
885 format
= "%(objectname) %(objecttype)\t%(refname)";
887 verify_format(format
);
889 memset(&cbdata
, 0, sizeof(cbdata
));
890 cbdata
.grab_pattern
= av
+ i
;
891 for_each_ref(grab_single_ref
, &cbdata
);
892 refs
= cbdata
.grab_array
;
893 num_refs
= cbdata
.grab_cnt
;
895 for (i
= 0; i
< used_atom_cnt
; i
++) {
896 if (used_atom
[i
][0] == '*') {
902 sort_refs(sort
, refs
, num_refs
);
904 if (!maxcount
|| num_refs
< maxcount
)
906 for (i
= 0; i
< maxcount
; i
++)
907 show_ref(refs
[i
], format
, quote_style
);