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 REALLOC_ARRAY(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
;
121 int flags
; /* EXC_FLAG_* */
125 * One rule, as from a .gitattributes file.
127 * If is_macro is true, then u.attr is a pointer to the git_attr being
130 * If is_macro is false, then u.pattern points at the filename pattern
131 * to which the rule applies. (The memory pointed to is part of the
132 * memory block allocated for the match_attr instance.)
134 * In either case, num_attr is the number of attributes affected by
135 * this rule, and state is an array listing them. The attributes are
136 * listed as they appear in the file (macros unexpanded).
141 struct git_attr
*attr
;
145 struct attr_state state
[FLEX_ARRAY
];
148 static const char blank
[] = " \t\r\n";
151 * Parse a whitespace-delimited attribute state (i.e., "attr",
152 * "-attr", "!attr", or "attr=value") from the string starting at src.
153 * If e is not NULL, write the results to *e. Return a pointer to the
154 * remainder of the string (with leading whitespace removed), or NULL
155 * if there was an error.
157 static const char *parse_attr(const char *src
, int lineno
, const char *cp
,
158 struct attr_state
*e
)
160 const char *ep
, *equals
;
163 ep
= cp
+ strcspn(cp
, blank
);
164 equals
= strchr(cp
, '=');
165 if (equals
&& ep
< equals
)
172 if (*cp
== '-' || *cp
== '!') {
176 if (invalid_attr_name(cp
, len
)) {
178 "%.*s is not a valid attribute name: %s:%d\n",
179 len
, cp
, src
, lineno
);
183 if (*cp
== '-' || *cp
== '!') {
184 e
->setto
= (*cp
== '-') ? ATTR__FALSE
: ATTR__UNSET
;
189 e
->setto
= ATTR__TRUE
;
191 e
->setto
= xmemdupz(equals
+ 1, ep
- equals
- 1);
193 e
->attr
= git_attr_internal(cp
, len
);
195 return ep
+ strspn(ep
, blank
);
198 static struct match_attr
*parse_attr_line(const char *line
, const char *src
,
199 int lineno
, int macro_ok
)
203 const char *cp
, *name
, *states
;
204 struct match_attr
*res
= NULL
;
207 cp
= line
+ strspn(line
, blank
);
208 if (!*cp
|| *cp
== '#')
211 namelen
= strcspn(name
, blank
);
212 if (strlen(ATTRIBUTE_MACRO_PREFIX
) < namelen
&&
213 starts_with(name
, ATTRIBUTE_MACRO_PREFIX
)) {
215 fprintf(stderr
, "%s not allowed: %s:%d\n",
220 name
+= strlen(ATTRIBUTE_MACRO_PREFIX
);
221 name
+= strspn(name
, blank
);
222 namelen
= strcspn(name
, blank
);
223 if (invalid_attr_name(name
, namelen
)) {
225 "%.*s is not a valid attribute name: %s:%d\n",
226 namelen
, name
, src
, lineno
);
233 states
= name
+ namelen
;
234 states
+= strspn(states
, blank
);
236 /* First pass to count the attr_states */
237 for (cp
= states
, num_attr
= 0; *cp
; num_attr
++) {
238 cp
= parse_attr(src
, lineno
, cp
, NULL
);
245 sizeof(struct attr_state
) * num_attr
+
246 (is_macro
? 0 : namelen
+ 1));
248 res
->u
.attr
= git_attr_internal(name
, namelen
);
250 char *p
= (char *)&(res
->state
[num_attr
]);
251 memcpy(p
, name
, namelen
);
252 res
->u
.pat
.pattern
= p
;
253 parse_exclude_pattern(&res
->u
.pat
.pattern
,
254 &res
->u
.pat
.patternlen
,
256 &res
->u
.pat
.nowildcardlen
);
257 if (res
->u
.pat
.flags
& EXC_FLAG_NEGATIVE
) {
258 warning(_("Negative patterns are ignored in git attributes\n"
259 "Use '\\!' for literal leading exclamation."));
263 res
->is_macro
= is_macro
;
264 res
->num_attr
= num_attr
;
266 /* Second pass to fill the attr_states */
267 for (cp
= states
, i
= 0; *cp
; i
++) {
268 cp
= parse_attr(src
, lineno
, cp
, &(res
->state
[i
]));
275 * Like info/exclude and .gitignore, the attribute information can
276 * come from many places.
278 * (1) .gitattribute file of the same directory;
279 * (2) .gitattribute file of the parent directory if (1) does not have
280 * any match; this goes recursively upwards, just like .gitignore.
281 * (3) $GIT_DIR/info/attributes, which overrides both of the above.
283 * In the same file, later entries override the earlier match, so in the
284 * global list, we would have entries from info/attributes the earliest
285 * (reading the file from top to bottom), .gitattribute of the root
286 * directory (again, reading the file from top to bottom) down to the
287 * current directory, and then scan the list backwards to find the first match.
288 * This is exactly the same as what is_excluded() does in dir.c to deal with
292 static struct attr_stack
{
293 struct attr_stack
*prev
;
296 unsigned num_matches
;
298 struct match_attr
**attrs
;
301 static void free_attr_elem(struct attr_stack
*e
)
305 for (i
= 0; i
< e
->num_matches
; i
++) {
306 struct match_attr
*a
= e
->attrs
[i
];
308 for (j
= 0; j
< a
->num_attr
; j
++) {
309 const char *setto
= a
->state
[j
].setto
;
310 if (setto
== ATTR__TRUE
||
311 setto
== ATTR__FALSE
||
312 setto
== ATTR__UNSET
||
313 setto
== ATTR__UNKNOWN
)
316 free((char *) setto
);
324 static const char *builtin_attr
[] = {
325 "[attr]binary -diff -merge -text",
329 static void handle_attr_line(struct attr_stack
*res
,
335 struct match_attr
*a
;
337 a
= parse_attr_line(line
, src
, lineno
, macro_ok
);
340 ALLOC_GROW(res
->attrs
, res
->num_matches
+ 1, res
->alloc
);
341 res
->attrs
[res
->num_matches
++] = a
;
344 static struct attr_stack
*read_attr_from_array(const char **list
)
346 struct attr_stack
*res
;
350 res
= xcalloc(1, sizeof(*res
));
351 while ((line
= *(list
++)) != NULL
)
352 handle_attr_line(res
, line
, "[builtin]", ++lineno
, 1);
356 static enum git_attr_direction direction
;
357 static struct index_state
*use_index
;
359 static struct attr_stack
*read_attr_from_file(const char *path
, int macro_ok
)
361 FILE *fp
= fopen(path
, "r");
362 struct attr_stack
*res
;
367 if (errno
!= ENOENT
&& errno
!= ENOTDIR
)
368 warn_on_inaccessible(path
);
371 res
= xcalloc(1, sizeof(*res
));
372 while (fgets(buf
, sizeof(buf
), fp
))
373 handle_attr_line(res
, buf
, path
, ++lineno
, macro_ok
);
378 static struct attr_stack
*read_attr_from_index(const char *path
, int macro_ok
)
380 struct attr_stack
*res
;
384 buf
= read_blob_data_from_index(use_index
? use_index
: &the_index
, path
, NULL
);
388 res
= xcalloc(1, sizeof(*res
));
389 for (sp
= buf
; *sp
; ) {
392 for (ep
= sp
; *ep
&& *ep
!= '\n'; ep
++)
394 more
= (*ep
== '\n');
396 handle_attr_line(res
, sp
, path
, ++lineno
, macro_ok
);
403 static struct attr_stack
*read_attr(const char *path
, int macro_ok
)
405 struct attr_stack
*res
;
407 if (direction
== GIT_ATTR_CHECKOUT
) {
408 res
= read_attr_from_index(path
, macro_ok
);
410 res
= read_attr_from_file(path
, macro_ok
);
412 else if (direction
== GIT_ATTR_CHECKIN
) {
413 res
= read_attr_from_file(path
, macro_ok
);
416 * There is no checked out .gitattributes file there, but
417 * we might have it in the index. We allow operation in a
418 * sparsely checked out work tree, so read from it.
420 res
= read_attr_from_index(path
, macro_ok
);
423 res
= read_attr_from_index(path
, macro_ok
);
425 res
= xcalloc(1, sizeof(*res
));
430 static void debug_info(const char *what
, struct attr_stack
*elem
)
432 fprintf(stderr
, "%s: %s\n", what
, elem
->origin
? elem
->origin
: "()");
434 static void debug_set(const char *what
, const char *match
, struct git_attr
*attr
, const void *v
)
436 const char *value
= v
;
438 if (ATTR_TRUE(value
))
440 else if (ATTR_FALSE(value
))
442 else if (ATTR_UNSET(value
))
443 value
= "unspecified";
445 fprintf(stderr
, "%s: %s => %s (%s)\n",
446 what
, attr
->name
, (char *) value
, match
);
448 #define debug_push(a) debug_info("push", (a))
449 #define debug_pop(a) debug_info("pop", (a))
451 #define debug_push(a) do { ; } while (0)
452 #define debug_pop(a) do { ; } while (0)
453 #define debug_set(a,b,c,d) do { ; } while (0)
456 static void drop_attr_stack(void)
459 struct attr_stack
*elem
= attr_stack
;
460 attr_stack
= elem
->prev
;
461 free_attr_elem(elem
);
465 static const char *git_etc_gitattributes(void)
467 static const char *system_wide
;
469 system_wide
= system_path(ETC_GITATTRIBUTES
);
473 static int git_attr_system(void)
475 return !git_env_bool("GIT_ATTR_NOSYSTEM", 0);
478 static void bootstrap_attr_stack(void)
480 struct attr_stack
*elem
;
485 elem
= read_attr_from_array(builtin_attr
);
487 elem
->prev
= attr_stack
;
490 if (git_attr_system()) {
491 elem
= read_attr_from_file(git_etc_gitattributes(), 1);
494 elem
->prev
= attr_stack
;
499 if (!git_attributes_file
)
500 git_attributes_file
= xdg_config_home("attributes");
501 if (git_attributes_file
) {
502 elem
= read_attr_from_file(git_attributes_file
, 1);
505 elem
->prev
= attr_stack
;
510 if (!is_bare_repository() || direction
== GIT_ATTR_INDEX
) {
511 elem
= read_attr(GITATTRIBUTES_FILE
, 1);
512 elem
->origin
= xstrdup("");
514 elem
->prev
= attr_stack
;
519 elem
= read_attr_from_file(git_path(INFOATTRIBUTES_FILE
), 1);
521 elem
= xcalloc(1, sizeof(*elem
));
523 elem
->prev
= attr_stack
;
527 static void prepare_attr_stack(const char *path
, int dirlen
)
529 struct attr_stack
*elem
, *info
;
534 * At the bottom of the attribute stack is the built-in
535 * set of attribute definitions, followed by the contents
536 * of $(prefix)/etc/gitattributes and a file specified by
537 * core.attributesfile. Then, contents from
538 * .gitattribute files from directories closer to the
539 * root to the ones in deeper directories are pushed
540 * to the stack. Finally, at the very top of the stack
541 * we always keep the contents of $GIT_DIR/info/attributes.
543 * When checking, we use entries from near the top of the
544 * stack, preferring $GIT_DIR/info/attributes, then
545 * .gitattributes in deeper directories to shallower ones,
546 * and finally use the built-in set as the default.
548 bootstrap_attr_stack();
551 * Pop the "info" one that is always at the top of the stack.
554 attr_stack
= info
->prev
;
557 * Pop the ones from directories that are not the prefix of
558 * the path we are checking. Break out of the loop when we see
559 * the root one (whose origin is an empty string "") or the builtin
560 * one (whose origin is NULL) without popping it.
562 while (attr_stack
->origin
) {
563 int namelen
= strlen(attr_stack
->origin
);
566 if (namelen
<= dirlen
&&
567 !strncmp(elem
->origin
, path
, namelen
) &&
568 (!namelen
|| path
[namelen
] == '/'))
572 attr_stack
= elem
->prev
;
573 free_attr_elem(elem
);
577 * Read from parent directories and push them down
579 if (!is_bare_repository() || direction
== GIT_ATTR_INDEX
) {
581 * bootstrap_attr_stack() should have added, and the
582 * above loop should have stopped before popping, the
583 * root element whose attr_stack->origin is set to an
586 struct strbuf pathbuf
= STRBUF_INIT
;
588 assert(attr_stack
->origin
);
590 len
= strlen(attr_stack
->origin
);
593 cp
= memchr(path
+ len
+ 1, '/', dirlen
- len
- 1);
596 strbuf_add(&pathbuf
, path
, cp
- path
);
597 strbuf_addch(&pathbuf
, '/');
598 strbuf_addstr(&pathbuf
, GITATTRIBUTES_FILE
);
599 elem
= read_attr(pathbuf
.buf
, 0);
600 strbuf_setlen(&pathbuf
, cp
- path
);
601 elem
->origin
= strbuf_detach(&pathbuf
, &elem
->originlen
);
602 elem
->prev
= attr_stack
;
607 strbuf_release(&pathbuf
);
611 * Finally push the "info" one at the top of the stack.
613 info
->prev
= attr_stack
;
617 static int path_matches(const char *pathname
, int pathlen
,
619 const struct pattern
*pat
,
620 const char *base
, int baselen
)
622 const char *pattern
= pat
->pattern
;
623 int prefix
= pat
->nowildcardlen
;
624 int isdir
= (pathlen
&& pathname
[pathlen
- 1] == '/');
626 if ((pat
->flags
& EXC_FLAG_MUSTBEDIR
) && !isdir
)
629 if (pat
->flags
& EXC_FLAG_NODIR
) {
630 return match_basename(pathname
+ basename_offset
,
631 pathlen
- basename_offset
- isdir
,
633 pat
->patternlen
, pat
->flags
);
635 return match_pathname(pathname
, pathlen
- isdir
,
637 pattern
, prefix
, pat
->patternlen
, pat
->flags
);
640 static int macroexpand_one(int attr_nr
, int rem
);
642 static int fill_one(const char *what
, struct match_attr
*a
, int rem
)
644 struct git_attr_check
*check
= check_all_attr
;
647 for (i
= a
->num_attr
- 1; 0 < rem
&& 0 <= i
; i
--) {
648 struct git_attr
*attr
= a
->state
[i
].attr
;
649 const char **n
= &(check
[attr
->attr_nr
].value
);
650 const char *v
= a
->state
[i
].setto
;
652 if (*n
== ATTR__UNKNOWN
) {
654 a
->is_macro
? a
->u
.attr
->name
: a
->u
.pat
.pattern
,
658 rem
= macroexpand_one(attr
->attr_nr
, rem
);
664 static int fill(const char *path
, int pathlen
, int basename_offset
,
665 struct attr_stack
*stk
, int rem
)
668 const char *base
= stk
->origin
? stk
->origin
: "";
670 for (i
= stk
->num_matches
- 1; 0 < rem
&& 0 <= i
; i
--) {
671 struct match_attr
*a
= stk
->attrs
[i
];
674 if (path_matches(path
, pathlen
, basename_offset
,
675 &a
->u
.pat
, base
, stk
->originlen
))
676 rem
= fill_one("fill", a
, rem
);
681 static int macroexpand_one(int attr_nr
, int rem
)
683 struct attr_stack
*stk
;
684 struct match_attr
*a
= NULL
;
687 if (check_all_attr
[attr_nr
].value
!= ATTR__TRUE
)
690 for (stk
= attr_stack
; !a
&& stk
; stk
= stk
->prev
)
691 for (i
= stk
->num_matches
- 1; !a
&& 0 <= i
; i
--) {
692 struct match_attr
*ma
= stk
->attrs
[i
];
695 if (ma
->u
.attr
->attr_nr
== attr_nr
)
700 rem
= fill_one("expand", a
, rem
);
706 * Collect all attributes for path into the array pointed to by
709 static void collect_all_attrs(const char *path
)
711 struct attr_stack
*stk
;
712 int i
, pathlen
, rem
, dirlen
;
713 const char *cp
, *last_slash
= NULL
;
716 for (cp
= path
; *cp
; cp
++) {
717 if (*cp
== '/' && cp
[1])
722 basename_offset
= last_slash
+ 1 - path
;
723 dirlen
= last_slash
- path
;
729 prepare_attr_stack(path
, dirlen
);
730 for (i
= 0; i
< attr_nr
; i
++)
731 check_all_attr
[i
].value
= ATTR__UNKNOWN
;
734 for (stk
= attr_stack
; 0 < rem
&& stk
; stk
= stk
->prev
)
735 rem
= fill(path
, pathlen
, basename_offset
, stk
, rem
);
738 int git_check_attr(const char *path
, int num
, struct git_attr_check
*check
)
742 collect_all_attrs(path
);
744 for (i
= 0; i
< num
; i
++) {
745 const char *value
= check_all_attr
[check
[i
].attr
->attr_nr
].value
;
746 if (value
== ATTR__UNKNOWN
)
748 check
[i
].value
= value
;
754 int git_all_attrs(const char *path
, int *num
, struct git_attr_check
**check
)
758 collect_all_attrs(path
);
760 /* Count the number of attributes that are set. */
762 for (i
= 0; i
< attr_nr
; i
++) {
763 const char *value
= check_all_attr
[i
].value
;
764 if (value
!= ATTR__UNSET
&& value
!= ATTR__UNKNOWN
)
768 *check
= xmalloc(sizeof(**check
) * count
);
770 for (i
= 0; i
< attr_nr
; i
++) {
771 const char *value
= check_all_attr
[i
].value
;
772 if (value
!= ATTR__UNSET
&& value
!= ATTR__UNKNOWN
) {
773 (*check
)[j
].attr
= check_all_attr
[i
].attr
;
774 (*check
)[j
].value
= value
;
782 void git_attr_set_direction(enum git_attr_direction
new, struct index_state
*istate
)
784 enum git_attr_direction old
= direction
;
786 if (is_bare_repository() && new != GIT_ATTR_INDEX
)
787 die("BUG: non-INDEX attr direction in a bare repo");