1 /* exclude.c -- exclude file names
3 Copyright (C) 1992, 1993, 1994, 1997, 1999, 2000, 2001, 2002, 2003,
4 2004, 2005, 2006, 2007, 2009 Free Software Foundation, Inc.
6 This program is free software: you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 3 of the License, or
9 (at your option) any later version.
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program. If not, see <http://www.gnu.org/licenses/>. */
19 /* Written by Paul Eggert <eggert@twinsun.com>
20 and Sergey Poznyakoff <gray@gnu.org>.
21 Thanks to Phil Proudman <phil@proudman51.freeserve.co.uk>
22 for improvement suggestions. */
43 # include "unlocked-io.h"
46 /* Non-GNU systems lack these options, so we don't need to check them. */
48 # define FNM_CASEFOLD 0
51 # define FNM_EXTMATCH 0
53 #ifndef FNM_LEADING_DIR
54 # define FNM_LEADING_DIR 0
57 verify (((EXCLUDE_ANCHORED | EXCLUDE_INCLUDE | EXCLUDE_WILDCARDS)
58 & (FNM_PATHNAME | FNM_NOESCAPE | FNM_PERIOD | FNM_LEADING_DIR
59 | FNM_CASEFOLD | FNM_EXTMATCH))
63 /* Exclusion patterns are grouped into a singly-linked list of
64 "exclusion segments". Each segment represents a set of patterns
65 that can be matches using the same algorithm. Non-wildcard
66 patterns are kept in hash tables, to speed up searches. Wildcard
67 patterns are stored as arrays of patterns. */
70 /* An exclude pattern-options pair. The options are fnmatch options
71 ORed with EXCLUDE_* options. */
79 /* An array of pattern-options pairs. */
81 struct exclude_pattern
83 struct patopts *exclude;
90 exclude_hash, /* a hash table of excluded names */
91 exclude_pattern /* an array of exclude patterns */
94 struct exclude_segment
96 struct exclude_segment *next; /* next segment in list */
97 enum exclude_type type; /* type of this segment */
98 int options; /* common options for this segment */
101 Hash_table *table; /* for type == exclude_hash */
102 struct exclude_pattern pat; /* for type == exclude_pattern */
106 /* The exclude structure keeps a singly-linked list of exclude segments */
109 struct exclude_segment *head, *tail;
112 /* Return true if str has wildcard characters */
114 fnmatch_pattern_has_wildcards (const char *str, int options)
116 const char *cset = "\\?*[]";
117 if (options & FNM_NOESCAPE)
121 size_t n = strcspn (str, cset);
124 else if (str[n] == '\\')
136 /* Return a newly allocated and empty exclude list. */
141 return xzalloc (sizeof *new_exclude ());
144 /* Calculate the hash of string. */
146 string_hasher (void const *data, size_t n_buckets)
148 char const *p = data;
149 return hash_string (p, n_buckets);
152 /* Ditto, for case-insensitive hashes */
154 string_hasher_ci (void const *data, size_t n_buckets)
156 char const *p = data;
157 mbui_iterator_t iter;
160 for (mbui_init (iter, p); mbui_avail (iter); mbui_advance (iter))
162 mbchar_t m = mbui_cur (iter);
166 wc = towlower (m.wc);
170 value = (value * 31 + wc) % n_buckets;
176 /* compare two strings for equality */
178 string_compare (void const *data1, void const *data2)
180 char const *p1 = data1;
181 char const *p2 = data2;
182 return strcmp (p1, p2) == 0;
185 /* compare two strings for equality, case-insensitive */
187 string_compare_ci (void const *data1, void const *data2)
189 char const *p1 = data1;
190 char const *p2 = data2;
191 return mbscasecmp (p1, p2) == 0;
195 string_free (void *data)
200 /* Create new exclude segment of given TYPE and OPTIONS, and attach it
201 to the tail of list in EX */
202 static struct exclude_segment *
203 new_exclude_segment (struct exclude *ex, enum exclude_type type, int options)
205 struct exclude_segment *sp = xzalloc (sizeof (struct exclude_segment));
207 sp->options = options;
210 case exclude_pattern:
214 sp->v.table = hash_initialize (0, NULL,
215 (options & FNM_CASEFOLD) ?
218 (options & FNM_CASEFOLD) ?
232 /* Free a single exclude segment */
234 free_exclude_segment (struct exclude_segment *seg)
238 case exclude_pattern:
239 free (seg->v.pat.exclude);
243 hash_free (seg->v.table);
249 /* Free the storage associated with an exclude list. */
251 free_exclude (struct exclude *ex)
253 struct exclude_segment *seg;
254 for (seg = ex->head; seg; )
256 struct exclude_segment *next = seg->next;
257 free_exclude_segment (seg);
263 /* Return zero if PATTERN matches F, obeying OPTIONS, except that
264 (unlike fnmatch) wildcards are disabled in PATTERN. */
267 fnmatch_no_wildcards (char const *pattern, char const *f, int options)
269 if (! (options & FNM_LEADING_DIR))
270 return ((options & FNM_CASEFOLD)
271 ? mbscasecmp (pattern, f)
272 : strcmp (pattern, f));
273 else if (! (options & FNM_CASEFOLD))
275 size_t patlen = strlen (pattern);
276 int r = strncmp (pattern, f, patlen);
287 /* Walk through a copy of F, seeing whether P matches any prefix
290 FIXME: This is an O(N**2) algorithm; it should be O(N).
291 Also, the copy should not be necessary. However, fixing this
292 will probably involve a change to the mbs* API. */
294 char *fcopy = xstrdup (f);
297 for (p = fcopy; ; *p++ = '/')
302 r = mbscasecmp (pattern, fcopy);
312 exclude_fnmatch (char const *pattern, char const *f, int options)
314 int (*matcher) (char const *, char const *, int) =
315 (options & EXCLUDE_WILDCARDS
317 : fnmatch_no_wildcards);
318 bool matched = ((*matcher) (pattern, f, options) == 0);
321 if (! (options & EXCLUDE_ANCHORED))
322 for (p = f; *p && ! matched; p++)
323 if (*p == '/' && p[1] != '/')
324 matched = ((*matcher) (pattern, p + 1, options) == 0);
329 /* Return true if the exclude_pattern segment SEG excludes F. */
332 excluded_file_pattern_p (struct exclude_segment const *seg, char const *f)
334 size_t exclude_count = seg->v.pat.exclude_count;
335 struct patopts const *exclude = seg->v.pat.exclude;
337 bool excluded = !! (exclude[0].options & EXCLUDE_INCLUDE);
339 /* Scan through the options, until they change excluded */
340 for (i = 0; i < exclude_count; i++)
342 char const *pattern = exclude[i].pattern;
343 int options = exclude[i].options;
344 if (excluded != exclude_fnmatch (pattern, f, options))
350 /* Return true if the exclude_hash segment SEG excludes F.
351 BUFFER is an auxiliary storage of the same length as F (with nul
352 terminator included) */
354 excluded_file_name_p (struct exclude_segment const *seg, char const *f,
357 int options = seg->options;
358 bool excluded = !! (options & EXCLUDE_INCLUDE);
359 Hash_table *table = seg->v.table;
363 /* initialize the pattern */
368 if (hash_lookup (table, buffer))
370 if (options & FNM_LEADING_DIR)
372 char *p = strrchr (buffer, '/');
382 if (!(options & EXCLUDE_ANCHORED))
395 /* Return true if EX excludes F. */
398 excluded_file_name (struct exclude const *ex, char const *f)
400 struct exclude_segment *seg;
402 char *filename = NULL;
404 /* If no patterns are given, the default is to include. */
408 /* Otherwise, the default is the opposite of the first option. */
409 excluded = !! (ex->head->options & EXCLUDE_INCLUDE);
410 /* Scan through the segments, seeing whether they change status from
411 excluded to included or vice versa. */
412 for (seg = ex->head; seg; seg = seg->next)
418 case exclude_pattern:
419 rc = excluded_file_pattern_p (seg, f);
424 filename = xmalloc (strlen (f) + 1);
425 rc = excluded_file_name_p (seg, f, filename);
438 /* Append to EX the exclusion PATTERN with OPTIONS. */
441 add_exclude (struct exclude *ex, char const *pattern, int options)
443 struct exclude_segment *seg;
445 if ((options & EXCLUDE_WILDCARDS)
446 && fnmatch_pattern_has_wildcards (pattern, options))
448 struct exclude_pattern *pat;
449 struct patopts *patopts;
451 if (ex->tail && ex->tail->type == exclude_pattern
452 && ((ex->tail->options & EXCLUDE_INCLUDE) ==
453 (options & EXCLUDE_INCLUDE)))
456 seg = new_exclude_segment (ex, exclude_pattern, options);
459 if (pat->exclude_count == pat->exclude_alloc)
460 pat->exclude = x2nrealloc (pat->exclude, &pat->exclude_alloc,
461 sizeof *pat->exclude);
462 patopts = &pat->exclude[pat->exclude_count++];
463 patopts->pattern = pattern;
464 patopts->options = options;
469 #define EXCLUDE_HASH_FLAGS (EXCLUDE_INCLUDE|EXCLUDE_ANCHORED|\
470 FNM_LEADING_DIR|FNM_CASEFOLD)
471 if (ex->tail && ex->tail->type == exclude_hash
472 && ((ex->tail->options & EXCLUDE_HASH_FLAGS) ==
473 (options & EXCLUDE_HASH_FLAGS)))
476 seg = new_exclude_segment (ex, exclude_hash, options);
478 str = xstrdup (pattern);
479 p = hash_insert (seg->v.table, str);
485 /* Use ADD_FUNC to append to EX the patterns in FILE_NAME, each with
486 OPTIONS. LINE_END terminates each pattern in the file. If
487 LINE_END is a space character, ignore trailing spaces and empty
488 lines in FILE. Return -1 on failure, 0 on success. */
491 add_exclude_file (void (*add_func) (struct exclude *, char const *, int),
492 struct exclude *ex, char const *file_name, int options,
495 bool use_stdin = file_name[0] == '-' && !file_name[1];
501 size_t buf_alloc = 0;
502 size_t buf_count = 0;
508 else if (! (in = fopen (file_name, "r")))
511 while ((c = getc (in)) != EOF)
513 if (buf_count == buf_alloc)
514 buf = x2realloc (buf, &buf_alloc);
515 buf[buf_count++] = c;
521 if (!use_stdin && fclose (in) != 0)
524 buf = xrealloc (buf, buf_count + 1);
525 buf[buf_count] = line_end;
526 lim = buf + buf_count + ! (buf_count == 0 || buf[buf_count - 1] == line_end);
529 for (p = buf; p < lim; p++)
532 char *pattern_end = p;
534 if (isspace ((unsigned char) line_end))
536 for (; ; pattern_end--)
537 if (pattern_end == pattern)
539 else if (! isspace ((unsigned char) pattern_end[-1]))
544 (*add_func) (ex, pattern, options);