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 static const char *attributes_file
;
25 /* This is a randomly chosen prime. */
33 struct git_attr
*next
;
36 char name
[FLEX_ARRAY
];
40 static struct git_attr_check
*check_all_attr
;
41 static struct git_attr
*(git_attr_hash
[HASHSIZE
]);
43 static unsigned hash_name(const char *name
, int namelen
)
49 val
= ((val
<< 7) | (val
>> 22)) ^ c
;
54 static int invalid_attr_name(const char *name
, int namelen
)
57 * Attribute name cannot begin with '-' and from
58 * [-A-Za-z0-9_.]. We'd specifically exclude '=' for now,
59 * as we might later want to allow non-binary value for
60 * attributes, e.g. "*.svg merge=special-merge-program-for-svg"
66 if (! (ch
== '-' || ch
== '.' || ch
== '_' ||
67 ('0' <= ch
&& ch
<= '9') ||
68 ('a' <= ch
&& ch
<= 'z') ||
69 ('A' <= ch
&& ch
<= 'Z')) )
75 static struct git_attr
*git_attr_internal(const char *name
, int len
)
77 unsigned hval
= hash_name(name
, len
);
78 unsigned pos
= hval
% HASHSIZE
;
81 for (a
= git_attr_hash
[pos
]; a
; a
= a
->next
) {
83 !memcmp(a
->name
, name
, len
) && !a
->name
[len
])
87 if (invalid_attr_name(name
, len
))
90 a
= xmalloc(sizeof(*a
) + len
+ 1);
91 memcpy(a
->name
, name
, len
);
94 a
->next
= git_attr_hash
[pos
];
95 a
->attr_nr
= attr_nr
++;
96 git_attr_hash
[pos
] = a
;
98 check_all_attr
= xrealloc(check_all_attr
,
99 sizeof(*check_all_attr
) * attr_nr
);
100 check_all_attr
[a
->attr_nr
].attr
= a
;
101 check_all_attr
[a
->attr_nr
].value
= ATTR__UNKNOWN
;
105 struct git_attr
*git_attr(const char *name
)
107 return git_attr_internal(name
, strlen(name
));
110 /* What does a matched pattern decide? */
112 struct git_attr
*attr
;
117 * One rule, as from a .gitattributes file.
119 * If is_macro is true, then u.attr is a pointer to the git_attr being
122 * If is_macro is false, then u.pattern points at the filename pattern
123 * to which the rule applies. (The memory pointed to is part of the
124 * memory block allocated for the match_attr instance.)
126 * In either case, num_attr is the number of attributes affected by
127 * this rule, and state is an array listing them. The attributes are
128 * listed as they appear in the file (macros unexpanded).
133 struct git_attr
*attr
;
137 struct attr_state state
[FLEX_ARRAY
];
140 static const char blank
[] = " \t\r\n";
143 * Parse a whitespace-delimited attribute state (i.e., "attr",
144 * "-attr", "!attr", or "attr=value") from the string starting at src.
145 * If e is not NULL, write the results to *e. Return a pointer to the
146 * remainder of the string (with leading whitespace removed), or NULL
147 * if there was an error.
149 static const char *parse_attr(const char *src
, int lineno
, const char *cp
,
150 struct attr_state
*e
)
152 const char *ep
, *equals
;
155 ep
= cp
+ strcspn(cp
, blank
);
156 equals
= strchr(cp
, '=');
157 if (equals
&& ep
< equals
)
164 if (*cp
== '-' || *cp
== '!') {
168 if (invalid_attr_name(cp
, len
)) {
170 "%.*s is not a valid attribute name: %s:%d\n",
171 len
, cp
, src
, lineno
);
175 if (*cp
== '-' || *cp
== '!') {
176 e
->setto
= (*cp
== '-') ? ATTR__FALSE
: ATTR__UNSET
;
181 e
->setto
= ATTR__TRUE
;
183 e
->setto
= xmemdupz(equals
+ 1, ep
- equals
- 1);
185 e
->attr
= git_attr_internal(cp
, len
);
187 return ep
+ strspn(ep
, blank
);
190 static struct match_attr
*parse_attr_line(const char *line
, const char *src
,
191 int lineno
, int macro_ok
)
195 const char *cp
, *name
, *states
;
196 struct match_attr
*res
= NULL
;
199 cp
= line
+ strspn(line
, blank
);
200 if (!*cp
|| *cp
== '#')
203 namelen
= strcspn(name
, blank
);
204 if (strlen(ATTRIBUTE_MACRO_PREFIX
) < namelen
&&
205 !prefixcmp(name
, ATTRIBUTE_MACRO_PREFIX
)) {
207 fprintf(stderr
, "%s not allowed: %s:%d\n",
212 name
+= strlen(ATTRIBUTE_MACRO_PREFIX
);
213 name
+= strspn(name
, blank
);
214 namelen
= strcspn(name
, blank
);
215 if (invalid_attr_name(name
, namelen
)) {
217 "%.*s is not a valid attribute name: %s:%d\n",
218 namelen
, name
, src
, lineno
);
225 states
= name
+ namelen
;
226 states
+= strspn(states
, blank
);
228 /* First pass to count the attr_states */
229 for (cp
= states
, num_attr
= 0; *cp
; num_attr
++) {
230 cp
= parse_attr(src
, lineno
, cp
, NULL
);
237 sizeof(struct attr_state
) * num_attr
+
238 (is_macro
? 0 : namelen
+ 1));
240 res
->u
.attr
= git_attr_internal(name
, namelen
);
242 res
->u
.pattern
= (char *)&(res
->state
[num_attr
]);
243 memcpy(res
->u
.pattern
, name
, namelen
);
244 res
->u
.pattern
[namelen
] = 0;
246 res
->is_macro
= is_macro
;
247 res
->num_attr
= num_attr
;
249 /* Second pass to fill the attr_states */
250 for (cp
= states
, i
= 0; *cp
; i
++) {
251 cp
= parse_attr(src
, lineno
, cp
, &(res
->state
[i
]));
258 * Like info/exclude and .gitignore, the attribute information can
259 * come from many places.
261 * (1) .gitattribute file of the same directory;
262 * (2) .gitattribute file of the parent directory if (1) does not have
263 * any match; this goes recursively upwards, just like .gitignore.
264 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
266 * In the same file, later entries override the earlier match, so in the
267 * global list, we would have entries from info/attributes the earliest
268 * (reading the file from top to bottom), .gitattribute of the root
269 * directory (again, reading the file from top to bottom) down to the
270 * current directory, and then scan the list backwards to find the first match.
271 * This is exactly the same as what excluded() does in dir.c to deal with
275 static struct attr_stack
{
276 struct attr_stack
*prev
;
278 unsigned num_matches
;
280 struct match_attr
**attrs
;
283 static void free_attr_elem(struct attr_stack
*e
)
287 for (i
= 0; i
< e
->num_matches
; i
++) {
288 struct match_attr
*a
= e
->attrs
[i
];
290 for (j
= 0; j
< a
->num_attr
; j
++) {
291 const char *setto
= a
->state
[j
].setto
;
292 if (setto
== ATTR__TRUE
||
293 setto
== ATTR__FALSE
||
294 setto
== ATTR__UNSET
||
295 setto
== ATTR__UNKNOWN
)
298 free((char *) setto
);
305 static const char *builtin_attr
[] = {
306 "[attr]binary -diff -text",
310 static void handle_attr_line(struct attr_stack
*res
,
316 struct match_attr
*a
;
318 a
= parse_attr_line(line
, src
, lineno
, macro_ok
);
321 if (res
->alloc
<= res
->num_matches
) {
322 res
->alloc
= alloc_nr(res
->num_matches
);
323 res
->attrs
= xrealloc(res
->attrs
,
324 sizeof(struct match_attr
*) *
327 res
->attrs
[res
->num_matches
++] = a
;
330 static struct attr_stack
*read_attr_from_array(const char **list
)
332 struct attr_stack
*res
;
336 res
= xcalloc(1, sizeof(*res
));
337 while ((line
= *(list
++)) != NULL
)
338 handle_attr_line(res
, line
, "[builtin]", ++lineno
, 1);
342 static enum git_attr_direction direction
;
343 static struct index_state
*use_index
;
345 static struct attr_stack
*read_attr_from_file(const char *path
, int macro_ok
)
347 FILE *fp
= fopen(path
, "r");
348 struct attr_stack
*res
;
354 res
= xcalloc(1, sizeof(*res
));
355 while (fgets(buf
, sizeof(buf
), fp
))
356 handle_attr_line(res
, buf
, path
, ++lineno
, macro_ok
);
361 static void *read_index_data(const char *path
)
365 enum object_type type
;
367 struct index_state
*istate
= use_index
? use_index
: &the_index
;
370 pos
= index_name_pos(istate
, path
, len
);
373 * We might be in the middle of a merge, in which
374 * case we would read stage #2 (ours).
378 (pos
< 0 && i
< istate
->cache_nr
&&
379 !strcmp(istate
->cache
[i
]->name
, path
));
381 if (ce_stage(istate
->cache
[i
]) == 2)
386 data
= read_sha1_file(istate
->cache
[pos
]->sha1
, &type
, &sz
);
387 if (!data
|| type
!= OBJ_BLOB
) {
394 static struct attr_stack
*read_attr_from_index(const char *path
, int macro_ok
)
396 struct attr_stack
*res
;
400 buf
= read_index_data(path
);
404 res
= xcalloc(1, sizeof(*res
));
405 for (sp
= buf
; *sp
; ) {
408 for (ep
= sp
; *ep
&& *ep
!= '\n'; ep
++)
410 more
= (*ep
== '\n');
412 handle_attr_line(res
, sp
, path
, ++lineno
, macro_ok
);
419 static struct attr_stack
*read_attr(const char *path
, int macro_ok
)
421 struct attr_stack
*res
;
423 if (direction
== GIT_ATTR_CHECKOUT
) {
424 res
= read_attr_from_index(path
, macro_ok
);
426 res
= read_attr_from_file(path
, macro_ok
);
428 else if (direction
== GIT_ATTR_CHECKIN
) {
429 res
= read_attr_from_file(path
, macro_ok
);
432 * There is no checked out .gitattributes file there, but
433 * we might have it in the index. We allow operation in a
434 * sparsely checked out work tree, so read from it.
436 res
= read_attr_from_index(path
, macro_ok
);
439 res
= read_attr_from_index(path
, macro_ok
);
441 res
= xcalloc(1, sizeof(*res
));
446 static void debug_info(const char *what
, struct attr_stack
*elem
)
448 fprintf(stderr
, "%s: %s\n", what
, elem
->origin
? elem
->origin
: "()");
450 static void debug_set(const char *what
, const char *match
, struct git_attr
*attr
, const void *v
)
452 const char *value
= v
;
454 if (ATTR_TRUE(value
))
456 else if (ATTR_FALSE(value
))
458 else if (ATTR_UNSET(value
))
459 value
= "unspecified";
461 fprintf(stderr
, "%s: %s => %s (%s)\n",
462 what
, attr
->name
, (char *) value
, match
);
464 #define debug_push(a) debug_info("push", (a))
465 #define debug_pop(a) debug_info("pop", (a))
467 #define debug_push(a) do { ; } while (0)
468 #define debug_pop(a) do { ; } while (0)
469 #define debug_set(a,b,c,d) do { ; } while (0)
472 static void drop_attr_stack(void)
475 struct attr_stack
*elem
= attr_stack
;
476 attr_stack
= elem
->prev
;
477 free_attr_elem(elem
);
481 static const char *git_etc_gitattributes(void)
483 static const char *system_wide
;
485 system_wide
= system_path(ETC_GITATTRIBUTES
);
489 static int git_attr_system(void)
491 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
494 static int git_attr_config(const char *var
, const char *value
, void *dummy
)
496 if (!strcmp(var
, "core.attributesfile"))
497 return git_config_pathname(&attributes_file
, var
, value
);
502 static void bootstrap_attr_stack(void)
505 struct attr_stack
*elem
;
507 elem
= read_attr_from_array(builtin_attr
);
509 elem
->prev
= attr_stack
;
512 if (git_attr_system()) {
513 elem
= read_attr_from_file(git_etc_gitattributes(), 1);
516 elem
->prev
= attr_stack
;
521 git_config(git_attr_config
, NULL
);
522 if (attributes_file
) {
523 elem
= read_attr_from_file(attributes_file
, 1);
526 elem
->prev
= attr_stack
;
531 if (!is_bare_repository() || direction
== GIT_ATTR_INDEX
) {
532 elem
= read_attr(GITATTRIBUTES_FILE
, 1);
533 elem
->origin
= strdup("");
534 elem
->prev
= attr_stack
;
539 elem
= read_attr_from_file(git_path(INFOATTRIBUTES_FILE
), 1);
541 elem
= xcalloc(1, sizeof(*elem
));
543 elem
->prev
= attr_stack
;
548 static void prepare_attr_stack(const char *path
, int dirlen
)
550 struct attr_stack
*elem
, *info
;
552 struct strbuf pathbuf
;
554 strbuf_init(&pathbuf
, dirlen
+2+strlen(GITATTRIBUTES_FILE
));
557 * At the bottom of the attribute stack is the built-in
558 * set of attribute definitions, followed by the contents
559 * of $(prefix)/etc/gitattributes and a file specified by
560 * core.attributesfile. Then, contents from
561 * .gitattribute files from directories closer to the
562 * root to the ones in deeper directories are pushed
563 * to the stack. Finally, at the very top of the stack
564 * we always keep the contents of $GIT_DIR/info/attributes.
566 * When checking, we use entries from near the top of the
567 * stack, preferring $GIT_DIR/info/attributes, then
568 * .gitattributes in deeper directories to shallower ones,
569 * and finally use the built-in set as the default.
572 bootstrap_attr_stack();
575 * Pop the "info" one that is always at the top of the stack.
578 attr_stack
= info
->prev
;
581 * Pop the ones from directories that are not the prefix of
582 * the path we are checking.
584 while (attr_stack
&& attr_stack
->origin
) {
585 int namelen
= strlen(attr_stack
->origin
);
588 if (namelen
<= dirlen
&&
589 !strncmp(elem
->origin
, path
, namelen
))
593 attr_stack
= elem
->prev
;
594 free_attr_elem(elem
);
598 * Read from parent directories and push them down
600 if (!is_bare_repository() || direction
== GIT_ATTR_INDEX
) {
604 len
= strlen(attr_stack
->origin
);
607 strbuf_reset(&pathbuf
);
608 strbuf_add(&pathbuf
, path
, dirlen
);
609 strbuf_addch(&pathbuf
, '/');
610 cp
= strchr(pathbuf
.buf
+ len
+ 1, '/');
611 strcpy(cp
+ 1, GITATTRIBUTES_FILE
);
612 elem
= read_attr(pathbuf
.buf
, 0);
614 elem
->origin
= strdup(pathbuf
.buf
);
615 elem
->prev
= attr_stack
;
621 strbuf_release(&pathbuf
);
624 * Finally push the "info" one at the top of the stack.
626 info
->prev
= attr_stack
;
630 static int path_matches(const char *pathname
, int pathlen
,
632 const char *base
, int baselen
)
634 if (!strchr(pattern
, '/')) {
636 const char *basename
= strrchr(pathname
, '/');
637 basename
= basename
? basename
+ 1 : pathname
;
638 return (fnmatch(pattern
, basename
, 0) == 0);
641 * match with FNM_PATHNAME; the pattern has base implicitly
646 if (pathlen
< baselen
||
647 (baselen
&& pathname
[baselen
] != '/') ||
648 strncmp(pathname
, base
, baselen
))
652 return fnmatch(pattern
, pathname
+ baselen
, FNM_PATHNAME
) == 0;
655 static int macroexpand_one(int attr_nr
, int rem
);
657 static int fill_one(const char *what
, struct match_attr
*a
, int rem
)
659 struct git_attr_check
*check
= check_all_attr
;
662 for (i
= a
->num_attr
- 1; 0 < rem
&& 0 <= i
; i
--) {
663 struct git_attr
*attr
= a
->state
[i
].attr
;
664 const char **n
= &(check
[attr
->attr_nr
].value
);
665 const char *v
= a
->state
[i
].setto
;
667 if (*n
== ATTR__UNKNOWN
) {
669 a
->is_macro
? a
->u
.attr
->name
: a
->u
.pattern
,
673 rem
= macroexpand_one(attr
->attr_nr
, rem
);
679 static int fill(const char *path
, int pathlen
, struct attr_stack
*stk
, int rem
)
682 const char *base
= stk
->origin
? stk
->origin
: "";
684 for (i
= stk
->num_matches
- 1; 0 < rem
&& 0 <= i
; i
--) {
685 struct match_attr
*a
= stk
->attrs
[i
];
688 if (path_matches(path
, pathlen
,
689 a
->u
.pattern
, base
, strlen(base
)))
690 rem
= fill_one("fill", a
, rem
);
695 static int macroexpand_one(int attr_nr
, int rem
)
697 struct attr_stack
*stk
;
698 struct match_attr
*a
= NULL
;
701 if (check_all_attr
[attr_nr
].value
!= ATTR__TRUE
)
704 for (stk
= attr_stack
; !a
&& stk
; stk
= stk
->prev
)
705 for (i
= stk
->num_matches
- 1; !a
&& 0 <= i
; i
--) {
706 struct match_attr
*ma
= stk
->attrs
[i
];
709 if (ma
->u
.attr
->attr_nr
== attr_nr
)
714 rem
= fill_one("expand", a
, rem
);
719 int git_checkattr(const char *path
, int num
, struct git_attr_check
*check
)
721 struct attr_stack
*stk
;
723 int dirlen
, pathlen
, i
, rem
;
725 bootstrap_attr_stack();
726 for (i
= 0; i
< attr_nr
; i
++)
727 check_all_attr
[i
].value
= ATTR__UNKNOWN
;
729 pathlen
= strlen(path
);
730 cp
= strrchr(path
, '/');
735 prepare_attr_stack(path
, dirlen
);
737 for (stk
= attr_stack
; 0 < rem
&& stk
; stk
= stk
->prev
)
738 rem
= fill(path
, pathlen
, stk
, rem
);
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 void git_attr_set_direction(enum git_attr_direction
new, struct index_state
*istate
)
752 enum git_attr_direction old
= direction
;
754 if (is_bare_repository() && new != GIT_ATTR_INDEX
)
755 die("BUG: non-INDEX attr direction in a bare repo");