2 * Handle git attributes. See gitattributes(5) for a description of
3 * the file syntax, and Documentation/technical/api-gitattributes.txt
4 * for a description of the API.
6 * One basic design decision here is that we are not going to support
7 * an insanely large number of attributes.
10 #define NO_THE_INDEX_COMPATIBILITY_MACROS
15 const char git_attr__true
[] = "(builtin)true";
16 const char git_attr__false
[] = "\0(builtin)false";
17 static const char git_attr__unknown
[] = "(builtin)unknown";
18 #define ATTR__TRUE git_attr__true
19 #define ATTR__FALSE git_attr__false
20 #define ATTR__UNSET NULL
21 #define ATTR__UNKNOWN git_attr__unknown
23 /* This is a randomly chosen prime. */
31 struct git_attr
*next
;
34 char name
[FLEX_ARRAY
];
38 static struct git_attr_check
*check_all_attr
;
39 static struct git_attr
*(git_attr_hash
[HASHSIZE
]);
41 char *git_attr_name(struct git_attr
*attr
)
46 static unsigned hash_name(const char *name
, int namelen
)
52 val
= ((val
<< 7) | (val
>> 22)) ^ c
;
57 static int invalid_attr_name(const char *name
, int namelen
)
60 * Attribute name cannot begin with '-' and must consist of
61 * characters from [-A-Za-z0-9_.].
63 if (namelen
<= 0 || *name
== '-')
67 if (! (ch
== '-' || ch
== '.' || ch
== '_' ||
68 ('0' <= ch
&& ch
<= '9') ||
69 ('a' <= ch
&& ch
<= 'z') ||
70 ('A' <= ch
&& ch
<= 'Z')) )
76 static struct git_attr
*git_attr_internal(const char *name
, int len
)
78 unsigned hval
= hash_name(name
, len
);
79 unsigned pos
= hval
% HASHSIZE
;
82 for (a
= git_attr_hash
[pos
]; a
; a
= a
->next
) {
84 !memcmp(a
->name
, name
, len
) && !a
->name
[len
])
88 if (invalid_attr_name(name
, len
))
91 a
= xmalloc(sizeof(*a
) + len
+ 1);
92 memcpy(a
->name
, name
, len
);
95 a
->next
= git_attr_hash
[pos
];
96 a
->attr_nr
= attr_nr
++;
97 git_attr_hash
[pos
] = a
;
99 check_all_attr
= xrealloc(check_all_attr
,
100 sizeof(*check_all_attr
) * attr_nr
);
101 check_all_attr
[a
->attr_nr
].attr
= a
;
102 check_all_attr
[a
->attr_nr
].value
= ATTR__UNKNOWN
;
106 struct git_attr
*git_attr(const char *name
)
108 return git_attr_internal(name
, strlen(name
));
111 /* What does a matched pattern decide? */
113 struct git_attr
*attr
;
118 * One rule, as from a .gitattributes file.
120 * If is_macro is true, then u.attr is a pointer to the git_attr being
123 * If is_macro is false, then u.pattern points at the filename pattern
124 * to which the rule applies. (The memory pointed to is part of the
125 * memory block allocated for the match_attr instance.)
127 * In either case, num_attr is the number of attributes affected by
128 * this rule, and state is an array listing them. The attributes are
129 * listed as they appear in the file (macros unexpanded).
134 struct git_attr
*attr
;
138 struct attr_state state
[FLEX_ARRAY
];
141 static const char blank
[] = " \t\r\n";
144 * Parse a whitespace-delimited attribute state (i.e., "attr",
145 * "-attr", "!attr", or "attr=value") from the string starting at src.
146 * If e is not NULL, write the results to *e. Return a pointer to the
147 * remainder of the string (with leading whitespace removed), or NULL
148 * if there was an error.
150 static const char *parse_attr(const char *src
, int lineno
, const char *cp
,
151 struct attr_state
*e
)
153 const char *ep
, *equals
;
156 ep
= cp
+ strcspn(cp
, blank
);
157 equals
= strchr(cp
, '=');
158 if (equals
&& ep
< equals
)
165 if (*cp
== '-' || *cp
== '!') {
169 if (invalid_attr_name(cp
, len
)) {
171 "%.*s is not a valid attribute name: %s:%d\n",
172 len
, cp
, src
, lineno
);
176 if (*cp
== '-' || *cp
== '!') {
177 e
->setto
= (*cp
== '-') ? ATTR__FALSE
: ATTR__UNSET
;
182 e
->setto
= ATTR__TRUE
;
184 e
->setto
= xmemdupz(equals
+ 1, ep
- equals
- 1);
186 e
->attr
= git_attr_internal(cp
, len
);
188 return ep
+ strspn(ep
, blank
);
191 static struct match_attr
*parse_attr_line(const char *line
, const char *src
,
192 int lineno
, int macro_ok
)
196 const char *cp
, *name
, *states
;
197 struct match_attr
*res
= NULL
;
200 cp
= line
+ strspn(line
, blank
);
201 if (!*cp
|| *cp
== '#')
204 namelen
= strcspn(name
, blank
);
205 if (strlen(ATTRIBUTE_MACRO_PREFIX
) < namelen
&&
206 !prefixcmp(name
, ATTRIBUTE_MACRO_PREFIX
)) {
208 fprintf(stderr
, "%s not allowed: %s:%d\n",
213 name
+= strlen(ATTRIBUTE_MACRO_PREFIX
);
214 name
+= strspn(name
, blank
);
215 namelen
= strcspn(name
, blank
);
216 if (invalid_attr_name(name
, namelen
)) {
218 "%.*s is not a valid attribute name: %s:%d\n",
219 namelen
, name
, src
, lineno
);
226 states
= name
+ namelen
;
227 states
+= strspn(states
, blank
);
229 /* First pass to count the attr_states */
230 for (cp
= states
, num_attr
= 0; *cp
; num_attr
++) {
231 cp
= parse_attr(src
, lineno
, cp
, NULL
);
238 sizeof(struct attr_state
) * num_attr
+
239 (is_macro
? 0 : namelen
+ 1));
241 res
->u
.attr
= git_attr_internal(name
, namelen
);
243 res
->u
.pattern
= (char *)&(res
->state
[num_attr
]);
244 memcpy(res
->u
.pattern
, name
, namelen
);
245 res
->u
.pattern
[namelen
] = 0;
247 res
->is_macro
= is_macro
;
248 res
->num_attr
= num_attr
;
250 /* Second pass to fill the attr_states */
251 for (cp
= states
, i
= 0; *cp
; i
++) {
252 cp
= parse_attr(src
, lineno
, cp
, &(res
->state
[i
]));
259 * Like info/exclude and .gitignore, the attribute information can
260 * come from many places.
262 * (1) .gitattribute file of the same directory;
263 * (2) .gitattribute file of the parent directory if (1) does not have
264 * any match; this goes recursively upwards, just like .gitignore.
265 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
267 * In the same file, later entries override the earlier match, so in the
268 * global list, we would have entries from info/attributes the earliest
269 * (reading the file from top to bottom), .gitattribute of the root
270 * directory (again, reading the file from top to bottom) down to the
271 * current directory, and then scan the list backwards to find the first match.
272 * This is exactly the same as what excluded() does in dir.c to deal with
276 static struct attr_stack
{
277 struct attr_stack
*prev
;
279 unsigned num_matches
;
281 struct match_attr
**attrs
;
284 static void free_attr_elem(struct attr_stack
*e
)
288 for (i
= 0; i
< e
->num_matches
; i
++) {
289 struct match_attr
*a
= e
->attrs
[i
];
291 for (j
= 0; j
< a
->num_attr
; j
++) {
292 const char *setto
= a
->state
[j
].setto
;
293 if (setto
== ATTR__TRUE
||
294 setto
== ATTR__FALSE
||
295 setto
== ATTR__UNSET
||
296 setto
== ATTR__UNKNOWN
)
299 free((char *) setto
);
306 static const char *builtin_attr
[] = {
307 "[attr]binary -diff -text",
311 static void handle_attr_line(struct attr_stack
*res
,
317 struct match_attr
*a
;
319 a
= parse_attr_line(line
, src
, lineno
, macro_ok
);
322 if (res
->alloc
<= res
->num_matches
) {
323 res
->alloc
= alloc_nr(res
->num_matches
);
324 res
->attrs
= xrealloc(res
->attrs
,
325 sizeof(struct match_attr
*) *
328 res
->attrs
[res
->num_matches
++] = a
;
331 static struct attr_stack
*read_attr_from_array(const char **list
)
333 struct attr_stack
*res
;
337 res
= xcalloc(1, sizeof(*res
));
338 while ((line
= *(list
++)) != NULL
)
339 handle_attr_line(res
, line
, "[builtin]", ++lineno
, 1);
343 static enum git_attr_direction direction
;
344 static struct index_state
*use_index
;
346 static struct attr_stack
*read_attr_from_file(const char *path
, int macro_ok
)
348 FILE *fp
= fopen(path
, "r");
349 struct attr_stack
*res
;
355 res
= xcalloc(1, sizeof(*res
));
356 while (fgets(buf
, sizeof(buf
), fp
))
357 handle_attr_line(res
, buf
, path
, ++lineno
, macro_ok
);
362 static void *read_index_data(const char *path
)
366 enum object_type type
;
368 struct index_state
*istate
= use_index
? use_index
: &the_index
;
371 pos
= index_name_pos(istate
, path
, len
);
374 * We might be in the middle of a merge, in which
375 * case we would read stage #2 (ours).
379 (pos
< 0 && i
< istate
->cache_nr
&&
380 !strcmp(istate
->cache
[i
]->name
, path
));
382 if (ce_stage(istate
->cache
[i
]) == 2)
387 data
= read_sha1_file(istate
->cache
[pos
]->sha1
, &type
, &sz
);
388 if (!data
|| type
!= OBJ_BLOB
) {
395 static struct attr_stack
*read_attr_from_index(const char *path
, int macro_ok
)
397 struct attr_stack
*res
;
401 buf
= read_index_data(path
);
405 res
= xcalloc(1, sizeof(*res
));
406 for (sp
= buf
; *sp
; ) {
409 for (ep
= sp
; *ep
&& *ep
!= '\n'; ep
++)
411 more
= (*ep
== '\n');
413 handle_attr_line(res
, sp
, path
, ++lineno
, macro_ok
);
420 static struct attr_stack
*read_attr(const char *path
, int macro_ok
)
422 struct attr_stack
*res
;
424 if (direction
== GIT_ATTR_CHECKOUT
) {
425 res
= read_attr_from_index(path
, macro_ok
);
427 res
= read_attr_from_file(path
, macro_ok
);
429 else if (direction
== GIT_ATTR_CHECKIN
) {
430 res
= read_attr_from_file(path
, macro_ok
);
433 * There is no checked out .gitattributes file there, but
434 * we might have it in the index. We allow operation in a
435 * sparsely checked out work tree, so read from it.
437 res
= read_attr_from_index(path
, macro_ok
);
440 res
= read_attr_from_index(path
, macro_ok
);
442 res
= xcalloc(1, sizeof(*res
));
447 static void debug_info(const char *what
, struct attr_stack
*elem
)
449 fprintf(stderr
, "%s: %s\n", what
, elem
->origin
? elem
->origin
: "()");
451 static void debug_set(const char *what
, const char *match
, struct git_attr
*attr
, const void *v
)
453 const char *value
= v
;
455 if (ATTR_TRUE(value
))
457 else if (ATTR_FALSE(value
))
459 else if (ATTR_UNSET(value
))
460 value
= "unspecified";
462 fprintf(stderr
, "%s: %s => %s (%s)\n",
463 what
, attr
->name
, (char *) value
, match
);
465 #define debug_push(a) debug_info("push", (a))
466 #define debug_pop(a) debug_info("pop", (a))
468 #define debug_push(a) do { ; } while (0)
469 #define debug_pop(a) do { ; } while (0)
470 #define debug_set(a,b,c,d) do { ; } while (0)
473 static void drop_attr_stack(void)
476 struct attr_stack
*elem
= attr_stack
;
477 attr_stack
= elem
->prev
;
478 free_attr_elem(elem
);
482 static const char *git_etc_gitattributes(void)
484 static const char *system_wide
;
486 system_wide
= system_path(ETC_GITATTRIBUTES
);
490 static int git_attr_system(void)
492 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
495 static void bootstrap_attr_stack(void)
498 struct attr_stack
*elem
;
500 elem
= read_attr_from_array(builtin_attr
);
502 elem
->prev
= attr_stack
;
505 if (git_attr_system()) {
506 elem
= read_attr_from_file(git_etc_gitattributes(), 1);
509 elem
->prev
= attr_stack
;
514 if (git_attributes_file
) {
515 elem
= read_attr_from_file(git_attributes_file
, 1);
518 elem
->prev
= attr_stack
;
523 if (!is_bare_repository() || direction
== GIT_ATTR_INDEX
) {
524 elem
= read_attr(GITATTRIBUTES_FILE
, 1);
525 elem
->origin
= xstrdup("");
526 elem
->prev
= attr_stack
;
531 elem
= read_attr_from_file(git_path(INFOATTRIBUTES_FILE
), 1);
533 elem
= xcalloc(1, sizeof(*elem
));
535 elem
->prev
= attr_stack
;
540 static void prepare_attr_stack(const char *path
)
542 struct attr_stack
*elem
, *info
;
546 cp
= strrchr(path
, '/');
553 * At the bottom of the attribute stack is the built-in
554 * set of attribute definitions, followed by the contents
555 * of $(prefix)/etc/gitattributes and a file specified by
556 * core.attributesfile. Then, contents from
557 * .gitattribute files from directories closer to the
558 * root to the ones in deeper directories are pushed
559 * to the stack. Finally, at the very top of the stack
560 * we always keep the contents of $GIT_DIR/info/attributes.
562 * When checking, we use entries from near the top of the
563 * stack, preferring $GIT_DIR/info/attributes, then
564 * .gitattributes in deeper directories to shallower ones,
565 * and finally use the built-in set as the default.
567 bootstrap_attr_stack();
570 * Pop the "info" one that is always at the top of the stack.
573 attr_stack
= info
->prev
;
576 * Pop the ones from directories that are not the prefix of
577 * the path we are checking.
579 while (attr_stack
&& attr_stack
->origin
) {
580 int namelen
= strlen(attr_stack
->origin
);
583 if (namelen
<= dirlen
&&
584 !strncmp(elem
->origin
, path
, namelen
))
588 attr_stack
= elem
->prev
;
589 free_attr_elem(elem
);
593 * Read from parent directories and push them down
595 if (!is_bare_repository() || direction
== GIT_ATTR_INDEX
) {
596 struct strbuf pathbuf
= STRBUF_INIT
;
599 len
= strlen(attr_stack
->origin
);
602 cp
= memchr(path
+ len
+ 1, '/', dirlen
- len
- 1);
605 strbuf_add(&pathbuf
, path
, cp
- path
);
606 strbuf_addch(&pathbuf
, '/');
607 strbuf_addstr(&pathbuf
, GITATTRIBUTES_FILE
);
608 elem
= read_attr(pathbuf
.buf
, 0);
609 strbuf_setlen(&pathbuf
, cp
- path
);
610 elem
->origin
= strbuf_detach(&pathbuf
, NULL
);
611 elem
->prev
= attr_stack
;
616 strbuf_release(&pathbuf
);
620 * Finally push the "info" one at the top of the stack.
622 info
->prev
= attr_stack
;
626 static int path_matches(const char *pathname
, int pathlen
,
628 const char *base
, int baselen
)
630 if (!strchr(pattern
, '/')) {
632 const char *basename
= strrchr(pathname
, '/');
633 basename
= basename
? basename
+ 1 : pathname
;
634 return (fnmatch(pattern
, basename
, 0) == 0);
637 * match with FNM_PATHNAME; the pattern has base implicitly
642 if (pathlen
< baselen
||
643 (baselen
&& pathname
[baselen
] != '/') ||
644 strncmp(pathname
, base
, baselen
))
648 return fnmatch(pattern
, pathname
+ baselen
, FNM_PATHNAME
) == 0;
651 static int macroexpand_one(int attr_nr
, int rem
);
653 static int fill_one(const char *what
, struct match_attr
*a
, int rem
)
655 struct git_attr_check
*check
= check_all_attr
;
658 for (i
= a
->num_attr
- 1; 0 < rem
&& 0 <= i
; i
--) {
659 struct git_attr
*attr
= a
->state
[i
].attr
;
660 const char **n
= &(check
[attr
->attr_nr
].value
);
661 const char *v
= a
->state
[i
].setto
;
663 if (*n
== ATTR__UNKNOWN
) {
665 a
->is_macro
? a
->u
.attr
->name
: a
->u
.pattern
,
669 rem
= macroexpand_one(attr
->attr_nr
, rem
);
675 static int fill(const char *path
, int pathlen
, struct attr_stack
*stk
, int rem
)
678 const char *base
= stk
->origin
? stk
->origin
: "";
680 for (i
= stk
->num_matches
- 1; 0 < rem
&& 0 <= i
; i
--) {
681 struct match_attr
*a
= stk
->attrs
[i
];
684 if (path_matches(path
, pathlen
,
685 a
->u
.pattern
, base
, strlen(base
)))
686 rem
= fill_one("fill", a
, rem
);
691 static int macroexpand_one(int attr_nr
, int rem
)
693 struct attr_stack
*stk
;
694 struct match_attr
*a
= NULL
;
697 if (check_all_attr
[attr_nr
].value
!= ATTR__TRUE
)
700 for (stk
= attr_stack
; !a
&& stk
; stk
= stk
->prev
)
701 for (i
= stk
->num_matches
- 1; !a
&& 0 <= i
; i
--) {
702 struct match_attr
*ma
= stk
->attrs
[i
];
705 if (ma
->u
.attr
->attr_nr
== attr_nr
)
710 rem
= fill_one("expand", a
, rem
);
716 * Collect all attributes for path into the array pointed to by
719 static void collect_all_attrs(const char *path
)
721 struct attr_stack
*stk
;
724 prepare_attr_stack(path
);
725 for (i
= 0; i
< attr_nr
; i
++)
726 check_all_attr
[i
].value
= ATTR__UNKNOWN
;
728 pathlen
= strlen(path
);
730 for (stk
= attr_stack
; 0 < rem
&& stk
; stk
= stk
->prev
)
731 rem
= fill(path
, pathlen
, stk
, rem
);
734 int git_check_attr(const char *path
, int num
, struct git_attr_check
*check
)
738 collect_all_attrs(path
);
740 for (i
= 0; i
< num
; i
++) {
741 const char *value
= check_all_attr
[check
[i
].attr
->attr_nr
].value
;
742 if (value
== ATTR__UNKNOWN
)
744 check
[i
].value
= value
;
750 int git_all_attrs(const char *path
, int *num
, struct git_attr_check
**check
)
754 collect_all_attrs(path
);
756 /* Count the number of attributes that are set. */
758 for (i
= 0; i
< attr_nr
; i
++) {
759 const char *value
= check_all_attr
[i
].value
;
760 if (value
!= ATTR__UNSET
&& value
!= ATTR__UNKNOWN
)
764 *check
= xmalloc(sizeof(**check
) * count
);
766 for (i
= 0; i
< attr_nr
; i
++) {
767 const char *value
= check_all_attr
[i
].value
;
768 if (value
!= ATTR__UNSET
&& value
!= ATTR__UNKNOWN
) {
769 (*check
)[j
].attr
= check_all_attr
[i
].attr
;
770 (*check
)[j
].value
= value
;
778 void git_attr_set_direction(enum git_attr_direction
new, struct index_state
*istate
)
780 enum git_attr_direction old
= direction
;
782 if (is_bare_repository() && new != GIT_ATTR_INDEX
)
783 die("BUG: non-INDEX attr direction in a bare repo");