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
16 const char git_attr__true
[] = "(builtin)true";
17 const char git_attr__false
[] = "\0(builtin)false";
18 static const char git_attr__unknown
[] = "(builtin)unknown";
19 #define ATTR__TRUE git_attr__true
20 #define ATTR__FALSE git_attr__false
21 #define ATTR__UNSET NULL
22 #define ATTR__UNKNOWN git_attr__unknown
24 /* This is a randomly chosen prime. */
32 struct git_attr
*next
;
35 char name
[FLEX_ARRAY
];
39 static struct git_attr_check
*check_all_attr
;
40 static struct git_attr
*(git_attr_hash
[HASHSIZE
]);
42 char *git_attr_name(struct git_attr
*attr
)
47 static unsigned hash_name(const char *name
, int namelen
)
53 val
= ((val
<< 7) | (val
>> 22)) ^ c
;
58 static int invalid_attr_name(const char *name
, int namelen
)
61 * Attribute name cannot begin with '-' and must consist of
62 * characters from [-A-Za-z0-9_.].
64 if (namelen
<= 0 || *name
== '-')
68 if (! (ch
== '-' || ch
== '.' || ch
== '_' ||
69 ('0' <= ch
&& ch
<= '9') ||
70 ('a' <= ch
&& ch
<= 'z') ||
71 ('A' <= ch
&& ch
<= 'Z')) )
77 static struct git_attr
*git_attr_internal(const char *name
, int len
)
79 unsigned hval
= hash_name(name
, len
);
80 unsigned pos
= hval
% HASHSIZE
;
83 for (a
= git_attr_hash
[pos
]; a
; a
= a
->next
) {
85 !memcmp(a
->name
, name
, len
) && !a
->name
[len
])
89 if (invalid_attr_name(name
, len
))
92 a
= xmalloc(sizeof(*a
) + len
+ 1);
93 memcpy(a
->name
, name
, len
);
96 a
->next
= git_attr_hash
[pos
];
97 a
->attr_nr
= attr_nr
++;
98 git_attr_hash
[pos
] = a
;
100 check_all_attr
= xrealloc(check_all_attr
,
101 sizeof(*check_all_attr
) * attr_nr
);
102 check_all_attr
[a
->attr_nr
].attr
= a
;
103 check_all_attr
[a
->attr_nr
].value
= ATTR__UNKNOWN
;
107 struct git_attr
*git_attr(const char *name
)
109 return git_attr_internal(name
, strlen(name
));
112 /* What does a matched pattern decide? */
114 struct git_attr
*attr
;
122 int flags
; /* EXC_FLAG_* */
126 * One rule, as from a .gitattributes file.
128 * If is_macro is true, then u.attr is a pointer to the git_attr being
131 * If is_macro is false, then u.pattern points at the filename pattern
132 * to which the rule applies. (The memory pointed to is part of the
133 * memory block allocated for the match_attr instance.)
135 * In either case, num_attr is the number of attributes affected by
136 * this rule, and state is an array listing them. The attributes are
137 * listed as they appear in the file (macros unexpanded).
142 struct git_attr
*attr
;
146 struct attr_state state
[FLEX_ARRAY
];
149 static const char blank
[] = " \t\r\n";
152 * Parse a whitespace-delimited attribute state (i.e., "attr",
153 * "-attr", "!attr", or "attr=value") from the string starting at src.
154 * If e is not NULL, write the results to *e. Return a pointer to the
155 * remainder of the string (with leading whitespace removed), or NULL
156 * if there was an error.
158 static const char *parse_attr(const char *src
, int lineno
, const char *cp
,
159 struct attr_state
*e
)
161 const char *ep
, *equals
;
164 ep
= cp
+ strcspn(cp
, blank
);
165 equals
= strchr(cp
, '=');
166 if (equals
&& ep
< equals
)
173 if (*cp
== '-' || *cp
== '!') {
177 if (invalid_attr_name(cp
, len
)) {
179 "%.*s is not a valid attribute name: %s:%d\n",
180 len
, cp
, src
, lineno
);
184 if (*cp
== '-' || *cp
== '!') {
185 e
->setto
= (*cp
== '-') ? ATTR__FALSE
: ATTR__UNSET
;
190 e
->setto
= ATTR__TRUE
;
192 e
->setto
= xmemdupz(equals
+ 1, ep
- equals
- 1);
194 e
->attr
= git_attr_internal(cp
, len
);
196 return ep
+ strspn(ep
, blank
);
199 static struct match_attr
*parse_attr_line(const char *line
, const char *src
,
200 int lineno
, int macro_ok
)
204 const char *cp
, *name
, *states
;
205 struct match_attr
*res
= NULL
;
208 cp
= line
+ strspn(line
, blank
);
209 if (!*cp
|| *cp
== '#')
212 namelen
= strcspn(name
, blank
);
213 if (strlen(ATTRIBUTE_MACRO_PREFIX
) < namelen
&&
214 !prefixcmp(name
, ATTRIBUTE_MACRO_PREFIX
)) {
216 fprintf(stderr
, "%s not allowed: %s:%d\n",
221 name
+= strlen(ATTRIBUTE_MACRO_PREFIX
);
222 name
+= strspn(name
, blank
);
223 namelen
= strcspn(name
, blank
);
224 if (invalid_attr_name(name
, namelen
)) {
226 "%.*s is not a valid attribute name: %s:%d\n",
227 namelen
, name
, src
, lineno
);
234 states
= name
+ namelen
;
235 states
+= strspn(states
, blank
);
237 /* First pass to count the attr_states */
238 for (cp
= states
, num_attr
= 0; *cp
; num_attr
++) {
239 cp
= parse_attr(src
, lineno
, cp
, NULL
);
246 sizeof(struct attr_state
) * num_attr
+
247 (is_macro
? 0 : namelen
+ 1));
249 res
->u
.attr
= git_attr_internal(name
, namelen
);
251 char *p
= (char *)&(res
->state
[num_attr
]);
252 memcpy(p
, name
, namelen
);
253 res
->u
.pat
.pattern
= p
;
254 parse_exclude_pattern(&res
->u
.pat
.pattern
,
255 &res
->u
.pat
.patternlen
,
257 &res
->u
.pat
.nowildcardlen
);
258 if (res
->u
.pat
.flags
& EXC_FLAG_NEGATIVE
) {
259 warning(_("Negative patterns are ignored in git attributes\n"
260 "Use '\\!' for literal leading exclamation."));
264 res
->is_macro
= is_macro
;
265 res
->num_attr
= num_attr
;
267 /* Second pass to fill the attr_states */
268 for (cp
= states
, i
= 0; *cp
; i
++) {
269 cp
= parse_attr(src
, lineno
, cp
, &(res
->state
[i
]));
276 * Like info/exclude and .gitignore, the attribute information can
277 * come from many places.
279 * (1) .gitattribute file of the same directory;
280 * (2) .gitattribute file of the parent directory if (1) does not have
281 * any match; this goes recursively upwards, just like .gitignore.
282 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
284 * In the same file, later entries override the earlier match, so in the
285 * global list, we would have entries from info/attributes the earliest
286 * (reading the file from top to bottom), .gitattribute of the root
287 * directory (again, reading the file from top to bottom) down to the
288 * current directory, and then scan the list backwards to find the first match.
289 * This is exactly the same as what is_excluded() does in dir.c to deal with
293 static struct attr_stack
{
294 struct attr_stack
*prev
;
297 unsigned num_matches
;
299 struct match_attr
**attrs
;
302 static void free_attr_elem(struct attr_stack
*e
)
306 for (i
= 0; i
< e
->num_matches
; i
++) {
307 struct match_attr
*a
= e
->attrs
[i
];
309 for (j
= 0; j
< a
->num_attr
; j
++) {
310 const char *setto
= a
->state
[j
].setto
;
311 if (setto
== ATTR__TRUE
||
312 setto
== ATTR__FALSE
||
313 setto
== ATTR__UNSET
||
314 setto
== ATTR__UNKNOWN
)
317 free((char *) setto
);
325 static const char *builtin_attr
[] = {
326 "[attr]binary -diff -merge -text",
330 static void handle_attr_line(struct attr_stack
*res
,
336 struct match_attr
*a
;
338 a
= parse_attr_line(line
, src
, lineno
, macro_ok
);
341 if (res
->alloc
<= res
->num_matches
) {
342 res
->alloc
= alloc_nr(res
->num_matches
);
343 res
->attrs
= xrealloc(res
->attrs
,
344 sizeof(struct match_attr
*) *
347 res
->attrs
[res
->num_matches
++] = a
;
350 static struct attr_stack
*read_attr_from_array(const char **list
)
352 struct attr_stack
*res
;
356 res
= xcalloc(1, sizeof(*res
));
357 while ((line
= *(list
++)) != NULL
)
358 handle_attr_line(res
, line
, "[builtin]", ++lineno
, 1);
362 static enum git_attr_direction direction
;
363 static struct index_state
*use_index
;
365 static struct attr_stack
*read_attr_from_file(const char *path
, int macro_ok
)
367 FILE *fp
= fopen(path
, "r");
368 struct attr_stack
*res
;
373 if (errno
!= ENOENT
&& errno
!= ENOTDIR
)
374 warn_on_inaccessible(path
);
377 res
= xcalloc(1, sizeof(*res
));
378 while (fgets(buf
, sizeof(buf
), fp
))
379 handle_attr_line(res
, buf
, path
, ++lineno
, macro_ok
);
384 static struct attr_stack
*read_attr_from_index(const char *path
, int macro_ok
)
386 struct attr_stack
*res
;
390 buf
= read_blob_data_from_index(use_index
? use_index
: &the_index
, path
, NULL
);
394 res
= xcalloc(1, sizeof(*res
));
395 for (sp
= buf
; *sp
; ) {
398 for (ep
= sp
; *ep
&& *ep
!= '\n'; ep
++)
400 more
= (*ep
== '\n');
402 handle_attr_line(res
, sp
, path
, ++lineno
, macro_ok
);
409 static struct attr_stack
*read_attr(const char *path
, int macro_ok
)
411 struct attr_stack
*res
;
413 if (direction
== GIT_ATTR_CHECKOUT
) {
414 res
= read_attr_from_index(path
, macro_ok
);
416 res
= read_attr_from_file(path
, macro_ok
);
418 else if (direction
== GIT_ATTR_CHECKIN
) {
419 res
= read_attr_from_file(path
, macro_ok
);
422 * There is no checked out .gitattributes file there, but
423 * we might have it in the index. We allow operation in a
424 * sparsely checked out work tree, so read from it.
426 res
= read_attr_from_index(path
, macro_ok
);
429 res
= read_attr_from_index(path
, macro_ok
);
431 res
= xcalloc(1, sizeof(*res
));
436 static void debug_info(const char *what
, struct attr_stack
*elem
)
438 fprintf(stderr
, "%s: %s\n", what
, elem
->origin
? elem
->origin
: "()");
440 static void debug_set(const char *what
, const char *match
, struct git_attr
*attr
, const void *v
)
442 const char *value
= v
;
444 if (ATTR_TRUE(value
))
446 else if (ATTR_FALSE(value
))
448 else if (ATTR_UNSET(value
))
449 value
= "unspecified";
451 fprintf(stderr
, "%s: %s => %s (%s)\n",
452 what
, attr
->name
, (char *) value
, match
);
454 #define debug_push(a) debug_info("push", (a))
455 #define debug_pop(a) debug_info("pop", (a))
457 #define debug_push(a) do { ; } while (0)
458 #define debug_pop(a) do { ; } while (0)
459 #define debug_set(a,b,c,d) do { ; } while (0)
462 static void drop_attr_stack(void)
465 struct attr_stack
*elem
= attr_stack
;
466 attr_stack
= elem
->prev
;
467 free_attr_elem(elem
);
471 static const char *git_etc_gitattributes(void)
473 static const char *system_wide
;
475 system_wide
= system_path(ETC_GITATTRIBUTES
);
479 static int git_attr_system(void)
481 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
484 static void bootstrap_attr_stack(void)
486 struct attr_stack
*elem
;
487 char *xdg_attributes_file
;
492 elem
= read_attr_from_array(builtin_attr
);
494 elem
->prev
= attr_stack
;
497 if (git_attr_system()) {
498 elem
= read_attr_from_file(git_etc_gitattributes(), 1);
501 elem
->prev
= attr_stack
;
506 if (!git_attributes_file
) {
507 home_config_paths(NULL
, &xdg_attributes_file
, "attributes");
508 git_attributes_file
= xdg_attributes_file
;
510 if (git_attributes_file
) {
511 elem
= read_attr_from_file(git_attributes_file
, 1);
514 elem
->prev
= attr_stack
;
519 if (!is_bare_repository() || direction
== GIT_ATTR_INDEX
) {
520 elem
= read_attr(GITATTRIBUTES_FILE
, 1);
521 elem
->origin
= xstrdup("");
523 elem
->prev
= attr_stack
;
528 elem
= read_attr_from_file(git_path(INFOATTRIBUTES_FILE
), 1);
530 elem
= xcalloc(1, sizeof(*elem
));
532 elem
->prev
= attr_stack
;
536 static void prepare_attr_stack(const char *path
, int dirlen
)
538 struct attr_stack
*elem
, *info
;
543 * At the bottom of the attribute stack is the built-in
544 * set of attribute definitions, followed by the contents
545 * of $(prefix)/etc/gitattributes and a file specified by
546 * core.attributesfile. Then, contents from
547 * .gitattribute files from directories closer to the
548 * root to the ones in deeper directories are pushed
549 * to the stack. Finally, at the very top of the stack
550 * we always keep the contents of $GIT_DIR/info/attributes.
552 * When checking, we use entries from near the top of the
553 * stack, preferring $GIT_DIR/info/attributes, then
554 * .gitattributes in deeper directories to shallower ones,
555 * and finally use the built-in set as the default.
557 bootstrap_attr_stack();
560 * Pop the "info" one that is always at the top of the stack.
563 attr_stack
= info
->prev
;
566 * Pop the ones from directories that are not the prefix of
567 * the path we are checking. Break out of the loop when we see
568 * the root one (whose origin is an empty string "") or the builtin
569 * one (whose origin is NULL) without popping it.
571 while (attr_stack
->origin
) {
572 int namelen
= strlen(attr_stack
->origin
);
575 if (namelen
<= dirlen
&&
576 !strncmp(elem
->origin
, path
, namelen
) &&
577 (!namelen
|| path
[namelen
] == '/'))
581 attr_stack
= elem
->prev
;
582 free_attr_elem(elem
);
586 * Read from parent directories and push them down
588 if (!is_bare_repository() || direction
== GIT_ATTR_INDEX
) {
590 * bootstrap_attr_stack() should have added, and the
591 * above loop should have stopped before popping, the
592 * root element whose attr_stack->origin is set to an
595 struct strbuf pathbuf
= STRBUF_INIT
;
597 assert(attr_stack
->origin
);
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
, &elem
->originlen
);
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 struct pattern
*pat
,
629 const char *base
, int baselen
)
631 const char *pattern
= pat
->pattern
;
632 int prefix
= pat
->nowildcardlen
;
633 int isdir
= (pathlen
&& pathname
[pathlen
- 1] == '/');
635 if ((pat
->flags
& EXC_FLAG_MUSTBEDIR
) && !isdir
)
638 if (pat
->flags
& EXC_FLAG_NODIR
) {
639 return match_basename(pathname
+ basename_offset
,
640 pathlen
- basename_offset
- isdir
,
642 pat
->patternlen
, pat
->flags
);
644 return match_pathname(pathname
, pathlen
- isdir
,
646 pattern
, prefix
, pat
->patternlen
, pat
->flags
);
649 static int macroexpand_one(int attr_nr
, int rem
);
651 static int fill_one(const char *what
, struct match_attr
*a
, int rem
)
653 struct git_attr_check
*check
= check_all_attr
;
656 for (i
= a
->num_attr
- 1; 0 < rem
&& 0 <= i
; i
--) {
657 struct git_attr
*attr
= a
->state
[i
].attr
;
658 const char **n
= &(check
[attr
->attr_nr
].value
);
659 const char *v
= a
->state
[i
].setto
;
661 if (*n
== ATTR__UNKNOWN
) {
663 a
->is_macro
? a
->u
.attr
->name
: a
->u
.pat
.pattern
,
667 rem
= macroexpand_one(attr
->attr_nr
, rem
);
673 static int fill(const char *path
, int pathlen
, int basename_offset
,
674 struct attr_stack
*stk
, int rem
)
677 const char *base
= stk
->origin
? stk
->origin
: "";
679 for (i
= stk
->num_matches
- 1; 0 < rem
&& 0 <= i
; i
--) {
680 struct match_attr
*a
= stk
->attrs
[i
];
683 if (path_matches(path
, pathlen
, basename_offset
,
684 &a
->u
.pat
, base
, stk
->originlen
))
685 rem
= fill_one("fill", a
, rem
);
690 static int macroexpand_one(int attr_nr
, int rem
)
692 struct attr_stack
*stk
;
693 struct match_attr
*a
= NULL
;
696 if (check_all_attr
[attr_nr
].value
!= ATTR__TRUE
)
699 for (stk
= attr_stack
; !a
&& stk
; stk
= stk
->prev
)
700 for (i
= stk
->num_matches
- 1; !a
&& 0 <= i
; i
--) {
701 struct match_attr
*ma
= stk
->attrs
[i
];
704 if (ma
->u
.attr
->attr_nr
== attr_nr
)
709 rem
= fill_one("expand", a
, rem
);
715 * Collect all attributes for path into the array pointed to by
718 static void collect_all_attrs(const char *path
)
720 struct attr_stack
*stk
;
721 int i
, pathlen
, rem
, dirlen
;
722 const char *cp
, *last_slash
= NULL
;
725 for (cp
= path
; *cp
; cp
++) {
726 if (*cp
== '/' && cp
[1])
731 basename_offset
= last_slash
+ 1 - path
;
732 dirlen
= last_slash
- path
;
738 prepare_attr_stack(path
, dirlen
);
739 for (i
= 0; i
< attr_nr
; i
++)
740 check_all_attr
[i
].value
= ATTR__UNKNOWN
;
743 for (stk
= attr_stack
; 0 < rem
&& stk
; stk
= stk
->prev
)
744 rem
= fill(path
, pathlen
, basename_offset
, stk
, rem
);
747 int git_check_attr(const char *path
, int num
, struct git_attr_check
*check
)
751 collect_all_attrs(path
);
753 for (i
= 0; i
< num
; i
++) {
754 const char *value
= check_all_attr
[check
[i
].attr
->attr_nr
].value
;
755 if (value
== ATTR__UNKNOWN
)
757 check
[i
].value
= value
;
763 int git_all_attrs(const char *path
, int *num
, struct git_attr_check
**check
)
767 collect_all_attrs(path
);
769 /* Count the number of attributes that are set. */
771 for (i
= 0; i
< attr_nr
; i
++) {
772 const char *value
= check_all_attr
[i
].value
;
773 if (value
!= ATTR__UNSET
&& value
!= ATTR__UNKNOWN
)
777 *check
= xmalloc(sizeof(**check
) * count
);
779 for (i
= 0; i
< attr_nr
; i
++) {
780 const char *value
= check_all_attr
[i
].value
;
781 if (value
!= ATTR__UNSET
&& value
!= ATTR__UNKNOWN
) {
782 (*check
)[j
].attr
= check_all_attr
[i
].attr
;
783 (*check
)[j
].value
= value
;
791 void git_attr_set_direction(enum git_attr_direction
new, struct index_state
*istate
)
793 enum git_attr_direction old
= direction
;
795 if (is_bare_repository() && new != GIT_ATTR_INDEX
)
796 die("BUG: non-INDEX attr direction in a bare repo");