]>
Commit | Line | Data |
---|---|---|
1fd5e000 CF |
1 | /* |
2 | * Copyright (c) 1989, 1993 | |
3 | * The Regents of the University of California. All rights reserved. | |
4 | * | |
5 | * This code is derived from software contributed to Berkeley by | |
6 | * Guido van Rossum. | |
7 | * | |
8 | * Redistribution and use in source and binary forms, with or without | |
9 | * modification, are permitted provided that the following conditions | |
10 | * are met: | |
11 | * 1. Redistributions of source code must retain the above copyright | |
12 | * notice, this list of conditions and the following disclaimer. | |
13 | * 2. Redistributions in binary form must reproduce the above copyright | |
14 | * notice, this list of conditions and the following disclaimer in the | |
15 | * documentation and/or other materials provided with the distribution. | |
1fd5e000 CF |
16 | * 4. Neither the name of the University nor the names of its contributors |
17 | * may be used to endorse or promote products derived from this software | |
18 | * without specific prior written permission. | |
19 | * | |
20 | * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND | |
21 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
22 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE | |
23 | * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE | |
24 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
25 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS | |
26 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) | |
27 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT | |
28 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY | |
29 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF | |
30 | * SUCH DAMAGE. | |
31 | */ | |
32 | ||
48f81ac6 CV |
33 | #ifdef __CYGWIN__ |
34 | #include "winsup.h" | |
35 | #endif | |
36 | ||
733c9ed3 CV |
37 | #if defined(LIBC_SCCS) && !defined(lint) |
38 | static char sccsid[] = "@(#)glob.c 8.3 (Berkeley) 10/13/93"; | |
39 | #endif /* LIBC_SCCS and not lint */ | |
d7bcd2a1 CV |
40 | #include <sys/cdefs.h> |
41 | __FBSDID("$FreeBSD: src/lib/libc/gen/glob.c,v 1.28 2010/05/12 17:44:00 gordon Exp $"); | |
1fd5e000 CF |
42 | |
43 | /* | |
44 | * glob(3) -- a superset of the one defined in POSIX 1003.2. | |
45 | * | |
46 | * The [!...] convention to negate a range is supported (SysV, Posix, ksh). | |
47 | * | |
48 | * Optional extra services, controlled by flags not defined by POSIX: | |
49 | * | |
50 | * GLOB_QUOTE: | |
51 | * Escaping convention: \ inhibits any special meaning the following | |
52 | * character might have (except \ at end of string is retained). | |
53 | * GLOB_MAGCHAR: | |
54 | * Set in gl_flags if pattern contained a globbing character. | |
55 | * GLOB_NOMAGIC: | |
56 | * Same as GLOB_NOCHECK, but it will only append pattern if it did | |
57 | * not contain any magic characters. [Used in csh style globbing] | |
58 | * GLOB_ALTDIRFUNC: | |
59 | * Use alternately specified directory access functions. | |
60 | * GLOB_TILDE: | |
61 | * expand ~user/foo to the /home/dir/of/user/foo | |
62 | * GLOB_BRACE: | |
75a57bf0 | 63 | * expand {1,2}{a,b} to 1a 1b 2a 2b |
1fd5e000 CF |
64 | * gl_matchc: |
65 | * Number of matches in the current invocation of glob. | |
66 | */ | |
67 | ||
733c9ed3 CV |
68 | /* |
69 | * Some notes on multibyte character support: | |
70 | * 1. Patterns with illegal byte sequences match nothing - even if | |
71 | * GLOB_NOCHECK is specified. | |
72 | * 2. Illegal byte sequences in filenames are handled by treating them as | |
73 | * single-byte characters with a value of the first byte of the sequence | |
74 | * cast to wchar_t. | |
75 | * 3. State-dependent encodings are not currently supported. | |
76 | */ | |
77 | ||
d7bcd2a1 CV |
78 | #include <sys/param.h> |
79 | #include <sys/stat.h> | |
1fd5e000 | 80 | |
6ccb6bcf | 81 | #include <ctype.h> |
1fd5e000 | 82 | #include <dirent.h> |
d7bcd2a1 | 83 | #include <errno.h> |
1fd5e000 | 84 | #include <glob.h> |
d7bcd2a1 | 85 | #include <limits.h> |
1fd5e000 | 86 | #include <pwd.h> |
d7bcd2a1 CV |
87 | #include <stdint.h> |
88 | #include <stdio.h> | |
1fd5e000 | 89 | #include <stdlib.h> |
d7bcd2a1 | 90 | #include <string.h> |
1fd5e000 | 91 | #include <unistd.h> |
733c9ed3 CV |
92 | #include <wchar.h> |
93 | ||
d7bcd2a1 | 94 | #include "collate.h" |
733c9ed3 | 95 | |
d7bcd2a1 CV |
96 | #ifdef __CYGWIN__ |
97 | #include <wctype.h> | |
733c9ed3 CV |
98 | #include "path.h" |
99 | #include "fhandler.h" | |
100 | #include "dtable.h" | |
733c9ed3 | 101 | #include "cygheap.h" |
c16d0946 | 102 | #include "cygwin/version.h" |
b1aae492 | 103 | |
d7bcd2a1 CV |
104 | #define getpwuid(uid) getpwuid32 (uid) |
105 | #define getuid() getuid32 () | |
106 | #define issetugid() (cygheap->user.issetuid ()) | |
107 | ||
d7bcd2a1 CV |
108 | #define CCHAR(c) (ignore_case_with_glob ? towlower (CHAR (c)) : CHAR (c)) |
109 | #define Cchar(c) (ignore_case_with_glob ? towlower (c) : (c)) | |
110 | #endif | |
733c9ed3 | 111 | |
1d928241 CV |
112 | #undef MAXPATHLEN |
113 | #define MAXPATHLEN 8192 | |
114 | ||
1fd5e000 CF |
115 | #define DOLLAR '$' |
116 | #define DOT '.' | |
117 | #define EOS '\0' | |
118 | #define LBRACKET '[' | |
119 | #define NOT '!' | |
120 | #define QUESTION '?' | |
121 | #define QUOTE '\\' | |
122 | #define RANGE '-' | |
123 | #define RBRACKET ']' | |
124 | #define SEP '/' | |
125 | #define STAR '*' | |
126 | #define TILDE '~' | |
127 | #define UNDERSCORE '_' | |
128 | #define LBRACE '{' | |
129 | #define RBRACE '}' | |
130 | #define SLASH '/' | |
131 | #define COMMA ',' | |
132 | ||
133 | #ifndef DEBUG | |
134 | ||
733c9ed3 CV |
135 | #define M_QUOTE 0x8000000000ULL |
136 | #define M_PROTECT 0x4000000000ULL | |
137 | #define M_MASK 0xffffffffffULL | |
138 | #define M_CHAR 0x00ffffffffULL | |
1fd5e000 | 139 | |
733c9ed3 | 140 | typedef uint_fast64_t Char; |
1fd5e000 CF |
141 | |
142 | #else | |
143 | ||
144 | #define M_QUOTE 0x80 | |
145 | #define M_PROTECT 0x40 | |
146 | #define M_MASK 0xff | |
733c9ed3 | 147 | #define M_CHAR 0x7f |
1fd5e000 CF |
148 | |
149 | typedef char Char; | |
150 | ||
151 | #endif | |
152 | ||
153 | ||
733c9ed3 | 154 | #define CHAR(c) ((Char)((c)&M_CHAR)) |
1fd5e000 CF |
155 | #define META(c) ((Char)((c)|M_QUOTE)) |
156 | #define M_ALL META('*') | |
157 | #define M_END META(']') | |
158 | #define M_NOT META('!') | |
159 | #define M_ONE META('?') | |
160 | #define M_RNG META('-') | |
161 | #define M_SET META('[') | |
162 | #define ismeta(c) (((c)&M_QUOTE) != 0) | |
163 | ||
164 | ||
733c9ed3 CV |
165 | static int compare(const void *, const void *); |
166 | static int g_Ctoc(const Char *, char *, size_t); | |
d7bcd2a1 | 167 | static int g_lstat(Char *, struct stat *, glob_t *); |
733c9ed3 | 168 | static DIR *g_opendir(Char *, glob_t *); |
d7bcd2a1 | 169 | static const Char *g_strchr(const Char *, wchar_t); |
1fd5e000 | 170 | #ifdef notdef |
733c9ed3 | 171 | static Char *g_strcat(Char *, const Char *); |
1fd5e000 | 172 | #endif |
d7bcd2a1 | 173 | static int g_stat(Char *, struct stat *, glob_t *); |
733c9ed3 CV |
174 | static int glob0(const Char *, glob_t *, size_t *); |
175 | static int glob1(Char *, glob_t *, size_t *); | |
176 | static int glob2(Char *, Char *, Char *, Char *, glob_t *, size_t *); | |
177 | static int glob3(Char *, Char *, Char *, Char *, Char *, glob_t *, size_t *); | |
178 | static int globextend(const Char *, glob_t *, size_t *); | |
46f5dd59 | 179 | static const Char * |
733c9ed3 CV |
180 | globtilde(const Char *, Char *, size_t, glob_t *); |
181 | static int globexp1(const Char *, glob_t *, size_t *); | |
182 | static int globexp2(const Char *, const Char *, glob_t *, int *, size_t *); | |
183 | static int match(Char *, Char *, Char *); | |
1fd5e000 | 184 | #ifdef DEBUG |
733c9ed3 | 185 | static void qprintf(const char *, Char *); |
1fd5e000 CF |
186 | #endif |
187 | ||
1fd5e000 | 188 | int |
733c9ed3 | 189 | glob(const char *pattern, int flags, int (*errfunc)(const char *, int), glob_t *pglob) |
1fd5e000 | 190 | { |
733c9ed3 CV |
191 | const char *patnext; |
192 | size_t limit; | |
193 | Char *bufnext, *bufend, patbuf[MAXPATHLEN], prot; | |
194 | mbstate_t mbs; | |
195 | wchar_t wc; | |
196 | size_t clen; | |
197 | ||
198 | patnext = pattern; | |
1fd5e000 CF |
199 | if (!(flags & GLOB_APPEND)) { |
200 | pglob->gl_pathc = 0; | |
201 | pglob->gl_pathv = NULL; | |
202 | if (!(flags & GLOB_DOOFFS)) | |
203 | pglob->gl_offs = 0; | |
204 | } | |
733c9ed3 CV |
205 | if (flags & GLOB_LIMIT) { |
206 | limit = pglob->gl_matchc; | |
207 | if (limit == 0) | |
208 | limit = ARG_MAX; | |
209 | } else | |
210 | limit = 0; | |
1fd5e000 CF |
211 | pglob->gl_flags = flags & ~GLOB_MAGCHAR; |
212 | pglob->gl_errfunc = errfunc; | |
213 | pglob->gl_matchc = 0; | |
214 | ||
215 | bufnext = patbuf; | |
733c9ed3 CV |
216 | bufend = bufnext + MAXPATHLEN - 1; |
217 | if (flags & GLOB_NOESCAPE) { | |
218 | memset(&mbs, 0, sizeof(mbs)); | |
219 | while (bufend - bufnext >= MB_CUR_MAX) { | |
220 | clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs); | |
221 | if (clen == (size_t)-1 || clen == (size_t)-2) | |
222 | return (GLOB_NOMATCH); | |
223 | else if (clen == 0) | |
224 | break; | |
225 | *bufnext++ = wc; | |
226 | patnext += clen; | |
227 | } | |
228 | } else { | |
1fd5e000 | 229 | /* Protect the quoted characters. */ |
733c9ed3 CV |
230 | memset(&mbs, 0, sizeof(mbs)); |
231 | while (bufend - bufnext >= MB_CUR_MAX) { | |
232 | if (*patnext == QUOTE) { | |
233 | if (*++patnext == EOS) { | |
234 | *bufnext++ = QUOTE | M_PROTECT; | |
235 | continue; | |
1fd5e000 | 236 | } |
733c9ed3 CV |
237 | prot = M_PROTECT; |
238 | } else | |
239 | prot = 0; | |
240 | clen = mbrtowc(&wc, patnext, MB_LEN_MAX, &mbs); | |
241 | if (clen == (size_t)-1 || clen == (size_t)-2) | |
242 | return (GLOB_NOMATCH); | |
243 | else if (clen == 0) | |
244 | break; | |
245 | *bufnext++ = wc | prot; | |
246 | patnext += clen; | |
247 | } | |
1fd5e000 | 248 | } |
1fd5e000 CF |
249 | *bufnext = EOS; |
250 | ||
251 | if (flags & GLOB_BRACE) | |
733c9ed3 | 252 | return globexp1(patbuf, pglob, &limit); |
1fd5e000 | 253 | else |
733c9ed3 | 254 | return glob0(patbuf, pglob, &limit); |
1fd5e000 CF |
255 | } |
256 | ||
257 | /* | |
258 | * Expand recursively a glob {} pattern. When there is no more expansion | |
259 | * invoke the standard globbing routine to glob the rest of the magic | |
260 | * characters | |
261 | */ | |
733c9ed3 CV |
262 | static int |
263 | globexp1(const Char *pattern, glob_t *pglob, size_t *limit) | |
1fd5e000 CF |
264 | { |
265 | const Char* ptr = pattern; | |
266 | int rv; | |
267 | ||
268 | /* Protect a single {}, for find(1), like csh */ | |
269 | if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS) | |
733c9ed3 | 270 | return glob0(pattern, pglob, limit); |
1fd5e000 | 271 | |
d7bcd2a1 | 272 | while ((ptr = g_strchr(ptr, LBRACE)) != NULL) |
733c9ed3 | 273 | if (!globexp2(ptr, pattern, pglob, &rv, limit)) |
1fd5e000 CF |
274 | return rv; |
275 | ||
733c9ed3 | 276 | return glob0(pattern, pglob, limit); |
1fd5e000 CF |
277 | } |
278 | ||
279 | ||
280 | /* | |
281 | * Recursive brace globbing helper. Tries to expand a single brace. | |
282 | * If it succeeds then it invokes globexp1 with the new pattern. | |
283 | * If it fails then it tries to glob the rest of the pattern and returns. | |
284 | */ | |
733c9ed3 CV |
285 | static int |
286 | globexp2(const Char *ptr, const Char *pattern, glob_t *pglob, int *rv, size_t *limit) | |
1fd5e000 CF |
287 | { |
288 | int i; | |
289 | Char *lm, *ls; | |
733c9ed3 CV |
290 | const Char *pe, *pm, *pm1, *pl; |
291 | Char patbuf[MAXPATHLEN]; | |
1fd5e000 CF |
292 | |
293 | /* copy part up to the brace */ | |
294 | for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++) | |
295 | continue; | |
733c9ed3 | 296 | *lm = EOS; |
1fd5e000 CF |
297 | ls = lm; |
298 | ||
299 | /* Find the balanced brace */ | |
300 | for (i = 0, pe = ++ptr; *pe; pe++) | |
301 | if (*pe == LBRACKET) { | |
302 | /* Ignore everything between [] */ | |
303 | for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++) | |
304 | continue; | |
305 | if (*pe == EOS) { | |
75a57bf0 | 306 | /* |
1fd5e000 CF |
307 | * We could not find a matching RBRACKET. |
308 | * Ignore and just look for RBRACE | |
309 | */ | |
310 | pe = pm; | |
311 | } | |
312 | } | |
313 | else if (*pe == LBRACE) | |
314 | i++; | |
315 | else if (*pe == RBRACE) { | |
316 | if (i == 0) | |
317 | break; | |
318 | i--; | |
319 | } | |
320 | ||
321 | /* Non matching braces; just glob the pattern */ | |
322 | if (i != 0 || *pe == EOS) { | |
733c9ed3 | 323 | *rv = glob0(patbuf, pglob, limit); |
1fd5e000 CF |
324 | return 0; |
325 | } | |
326 | ||
327 | for (i = 0, pl = pm = ptr; pm <= pe; pm++) | |
328 | switch (*pm) { | |
329 | case LBRACKET: | |
330 | /* Ignore everything between [] */ | |
733c9ed3 | 331 | for (pm1 = pm++; *pm != RBRACKET && *pm != EOS; pm++) |
1fd5e000 CF |
332 | continue; |
333 | if (*pm == EOS) { | |
75a57bf0 | 334 | /* |
1fd5e000 CF |
335 | * We could not find a matching RBRACKET. |
336 | * Ignore and just look for RBRACE | |
337 | */ | |
733c9ed3 | 338 | pm = pm1; |
1fd5e000 CF |
339 | } |
340 | break; | |
341 | ||
342 | case LBRACE: | |
343 | i++; | |
344 | break; | |
345 | ||
346 | case RBRACE: | |
347 | if (i) { | |
348 | i--; | |
349 | break; | |
350 | } | |
351 | /* FALLTHROUGH */ | |
352 | case COMMA: | |
353 | if (i && *pm == COMMA) | |
354 | break; | |
355 | else { | |
356 | /* Append the current string */ | |
357 | for (lm = ls; (pl < pm); *lm++ = *pl++) | |
358 | continue; | |
75a57bf0 | 359 | /* |
1fd5e000 CF |
360 | * Append the rest of the pattern after the |
361 | * closing brace | |
362 | */ | |
363 | for (pl = pe + 1; (*lm++ = *pl++) != EOS;) | |
364 | continue; | |
365 | ||
366 | /* Expand the current pattern */ | |
367 | #ifdef DEBUG | |
368 | qprintf("globexp2:", patbuf); | |
369 | #endif | |
733c9ed3 | 370 | *rv = globexp1(patbuf, pglob, limit); |
1fd5e000 CF |
371 | |
372 | /* move after the comma, to the next string */ | |
373 | pl = pm + 1; | |
374 | } | |
375 | break; | |
376 | ||
377 | default: | |
378 | break; | |
379 | } | |
380 | *rv = 0; | |
381 | return 0; | |
382 | } | |
383 | ||
384 | ||
385 | ||
386 | /* | |
387 | * expand tilde from the passwd file. | |
388 | */ | |
389 | static const Char * | |
733c9ed3 | 390 | globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob) |
1fd5e000 CF |
391 | { |
392 | struct passwd *pwd; | |
393 | char *h; | |
394 | const Char *p; | |
733c9ed3 | 395 | Char *b, *eb; |
1fd5e000 CF |
396 | |
397 | if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE)) | |
398 | return pattern; | |
399 | ||
46f5dd59 CF |
400 | /* |
401 | * Copy up to the end of the string or / | |
733c9ed3 CV |
402 | */ |
403 | eb = &patbuf[patbuf_len - 1]; | |
404 | for (p = pattern + 1, h = (char *) patbuf; | |
405 | h < (char *)eb && *p && *p != SLASH; *h++ = *p++) | |
1fd5e000 CF |
406 | continue; |
407 | ||
408 | *h = EOS; | |
409 | ||
410 | if (((char *) patbuf)[0] == EOS) { | |
75a57bf0 | 411 | /* |
733c9ed3 CV |
412 | * handle a plain ~ or ~/ by expanding $HOME first (iff |
413 | * we're not running setuid or setgid) and then trying | |
414 | * the password file | |
1fd5e000 | 415 | */ |
d7bcd2a1 | 416 | if (issetugid() != 0 || |
733c9ed3 CV |
417 | (h = getenv("HOME")) == NULL) { |
418 | if (((h = getlogin()) != NULL && | |
419 | (pwd = getpwnam(h)) != NULL) || | |
d7bcd2a1 | 420 | (pwd = getpwuid(getuid())) != NULL) |
1fd5e000 | 421 | h = pwd->pw_dir; |
733c9ed3 CV |
422 | else |
423 | return pattern; | |
1fd5e000 CF |
424 | } |
425 | } | |
426 | else { | |
427 | /* | |
428 | * Expand a ~user | |
429 | */ | |
430 | if ((pwd = getpwnam((char*) patbuf)) == NULL) | |
431 | return pattern; | |
432 | else | |
433 | h = pwd->pw_dir; | |
434 | } | |
435 | ||
436 | /* Copy the home directory */ | |
733c9ed3 | 437 | for (b = patbuf; b < eb && *h; *b++ = *h++) |
1fd5e000 | 438 | continue; |
75a57bf0 | 439 | |
1fd5e000 | 440 | /* Append the rest of the pattern */ |
733c9ed3 | 441 | while (b < eb && (*b++ = *p++) != EOS) |
1fd5e000 | 442 | continue; |
733c9ed3 | 443 | *b = EOS; |
1fd5e000 CF |
444 | |
445 | return patbuf; | |
446 | } | |
75a57bf0 | 447 | |
1fd5e000 CF |
448 | |
449 | /* | |
450 | * The main glob() routine: compiles the pattern (optionally processing | |
451 | * quotes), calls glob1() to do the real pattern matching, and finally | |
452 | * sorts the list (unless unsorted operation is requested). Returns 0 | |
733c9ed3 | 453 | * if things went well, nonzero if errors occurred. |
1fd5e000 CF |
454 | */ |
455 | static int | |
733c9ed3 | 456 | glob0(const Char *pattern, glob_t *pglob, size_t *limit) |
1fd5e000 CF |
457 | { |
458 | const Char *qpatnext; | |
b3f40a5f | 459 | int err; |
733c9ed3 | 460 | size_t oldpathc; |
d7bcd2a1 | 461 | Char *bufnext, c, patbuf[MAXPATHLEN]; |
1fd5e000 | 462 | |
733c9ed3 | 463 | qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob); |
1fd5e000 CF |
464 | oldpathc = pglob->gl_pathc; |
465 | bufnext = patbuf; | |
466 | ||
467 | /* We don't need to check for buffer overflow any more. */ | |
468 | while ((c = *qpatnext++) != EOS) { | |
469 | switch (c) { | |
470 | case LBRACKET: | |
471 | c = *qpatnext; | |
472 | if (c == NOT) | |
473 | ++qpatnext; | |
474 | if (*qpatnext == EOS || | |
d7bcd2a1 | 475 | g_strchr(qpatnext+1, RBRACKET) == NULL) { |
1fd5e000 CF |
476 | *bufnext++ = LBRACKET; |
477 | if (c == NOT) | |
478 | --qpatnext; | |
479 | break; | |
480 | } | |
481 | *bufnext++ = M_SET; | |
482 | if (c == NOT) | |
483 | *bufnext++ = M_NOT; | |
484 | c = *qpatnext++; | |
485 | do { | |
486 | *bufnext++ = CHAR(c); | |
487 | if (*qpatnext == RANGE && | |
488 | (c = qpatnext[1]) != RBRACKET) { | |
489 | *bufnext++ = M_RNG; | |
490 | *bufnext++ = CHAR(c); | |
491 | qpatnext += 2; | |
492 | } | |
493 | } while ((c = *qpatnext++) != RBRACKET); | |
494 | pglob->gl_flags |= GLOB_MAGCHAR; | |
495 | *bufnext++ = M_END; | |
496 | break; | |
497 | case QUESTION: | |
498 | pglob->gl_flags |= GLOB_MAGCHAR; | |
499 | *bufnext++ = M_ONE; | |
500 | break; | |
501 | case STAR: | |
502 | pglob->gl_flags |= GLOB_MAGCHAR; | |
75a57bf0 | 503 | /* collapse adjacent stars to one, |
1fd5e000 CF |
504 | * to avoid exponential behavior |
505 | */ | |
506 | if (bufnext == patbuf || bufnext[-1] != M_ALL) | |
507 | *bufnext++ = M_ALL; | |
508 | break; | |
509 | default: | |
510 | *bufnext++ = CHAR(c); | |
511 | break; | |
512 | } | |
513 | } | |
514 | *bufnext = EOS; | |
515 | #ifdef DEBUG | |
516 | qprintf("glob0:", patbuf); | |
517 | #endif | |
518 | ||
733c9ed3 | 519 | if ((err = glob1(patbuf, pglob, limit)) != 0) |
1fd5e000 CF |
520 | return(err); |
521 | ||
522 | /* | |
75a57bf0 | 523 | * If there was no match we are going to append the pattern |
1fd5e000 CF |
524 | * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified |
525 | * and the pattern did not contain any magic characters | |
526 | * GLOB_NOMAGIC is there just for compatibility with csh. | |
527 | */ | |
733c9ed3 CV |
528 | if (pglob->gl_pathc == oldpathc) { |
529 | if (((pglob->gl_flags & GLOB_NOCHECK) || | |
530 | ((pglob->gl_flags & GLOB_NOMAGIC) && | |
531 | !(pglob->gl_flags & GLOB_MAGCHAR)))) | |
532 | return(globextend(pattern, pglob, limit)); | |
533 | else | |
534 | return(GLOB_NOMATCH); | |
535 | } | |
536 | if (!(pglob->gl_flags & GLOB_NOSORT)) | |
1fd5e000 CF |
537 | qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc, |
538 | pglob->gl_pathc - oldpathc, sizeof(char *), compare); | |
539 | return(0); | |
540 | } | |
541 | ||
542 | static int | |
733c9ed3 | 543 | compare(const void *p, const void *q) |
1fd5e000 CF |
544 | { |
545 | return(strcmp(*(char **)p, *(char **)q)); | |
546 | } | |
547 | ||
548 | static int | |
733c9ed3 | 549 | glob1(Char *pattern, glob_t *pglob, size_t *limit) |
1fd5e000 | 550 | { |
733c9ed3 | 551 | Char pathbuf[MAXPATHLEN]; |
1fd5e000 CF |
552 | |
553 | /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */ | |
554 | if (*pattern == EOS) | |
555 | return(0); | |
733c9ed3 CV |
556 | return(glob2(pathbuf, pathbuf, pathbuf + MAXPATHLEN - 1, |
557 | pattern, pglob, limit)); | |
1fd5e000 CF |
558 | } |
559 | ||
560 | /* | |
561 | * The functions glob2 and glob3 are mutually recursive; there is one level | |
562 | * of recursion for each segment in the pattern that contains one or more | |
563 | * meta characters. | |
564 | */ | |
565 | static int | |
733c9ed3 CV |
566 | glob2(Char *pathbuf, Char *pathend, Char *pathend_last, Char *pattern, |
567 | glob_t *pglob, size_t *limit) | |
1fd5e000 | 568 | { |
d7bcd2a1 | 569 | struct stat sb; |
1fd5e000 CF |
570 | Char *p, *q; |
571 | int anymeta; | |
572 | ||
573 | /* | |
574 | * Loop over pattern segments until end of pattern or until | |
575 | * segment with meta character found. | |
576 | */ | |
577 | for (anymeta = 0;;) { | |
578 | if (*pattern == EOS) { /* End of pattern? */ | |
579 | *pathend = EOS; | |
580 | if (g_lstat(pathbuf, &sb, pglob)) | |
581 | return(0); | |
75a57bf0 | 582 | |
1fd5e000 CF |
583 | if (((pglob->gl_flags & GLOB_MARK) && |
584 | pathend[-1] != SEP) && (S_ISDIR(sb.st_mode) | |
585 | || (S_ISLNK(sb.st_mode) && | |
586 | (g_stat(pathbuf, &sb, pglob) == 0) && | |
587 | S_ISDIR(sb.st_mode)))) { | |
733c9ed3 CV |
588 | if (pathend + 1 > pathend_last) |
589 | return (GLOB_ABORTED); | |
1fd5e000 CF |
590 | *pathend++ = SEP; |
591 | *pathend = EOS; | |
592 | } | |
593 | ++pglob->gl_matchc; | |
fa2d9fc5 | 594 | return(globextend(pathbuf, pglob, limit)); |
1fd5e000 CF |
595 | } |
596 | ||
597 | /* Find end of next segment, copy tentatively to pathend. */ | |
598 | q = pathend; | |
599 | p = pattern; | |
600 | while (*p != EOS && *p != SEP) { | |
601 | if (ismeta(*p)) | |
602 | anymeta = 1; | |
733c9ed3 CV |
603 | if (q + 1 > pathend_last) |
604 | return (GLOB_ABORTED); | |
1fd5e000 CF |
605 | *q++ = *p++; |
606 | } | |
607 | ||
608 | if (!anymeta) { /* No expansion, do next segment. */ | |
609 | pathend = q; | |
610 | pattern = p; | |
733c9ed3 CV |
611 | while (*pattern == SEP) { |
612 | if (pathend + 1 > pathend_last) | |
613 | return (GLOB_ABORTED); | |
1fd5e000 | 614 | *pathend++ = *pattern++; |
733c9ed3 | 615 | } |
1fd5e000 | 616 | } else /* Need expansion, recurse. */ |
733c9ed3 CV |
617 | return(glob3(pathbuf, pathend, pathend_last, pattern, p, |
618 | pglob, limit)); | |
1fd5e000 CF |
619 | } |
620 | /* NOTREACHED */ | |
621 | } | |
622 | ||
623 | static int | |
733c9ed3 CV |
624 | glob3(Char *pathbuf, Char *pathend, Char *pathend_last, |
625 | Char *pattern, Char *restpattern, | |
626 | glob_t *pglob, size_t *limit) | |
1fd5e000 | 627 | { |
733c9ed3 | 628 | struct dirent *dp; |
1fd5e000 CF |
629 | DIR *dirp; |
630 | int err; | |
631 | char buf[MAXPATHLEN]; | |
632 | ||
633 | /* | |
634 | * The readdirfunc declaration can't be prototyped, because it is | |
635 | * assigned, below, to two functions which are prototyped in glob.h | |
636 | * and dirent.h as taking pointers to differently typed opaque | |
637 | * structures. | |
d7bcd2a1 CV |
638 | * CYGWIN: Needs prototype and subsequently wild casting to avoid |
639 | * compiler error. | |
1fd5e000 | 640 | */ |
733c9ed3 | 641 | struct dirent *(*readdirfunc)(void *); |
1fd5e000 | 642 | |
733c9ed3 CV |
643 | if (pathend > pathend_last) |
644 | return (GLOB_ABORTED); | |
1fd5e000 CF |
645 | *pathend = EOS; |
646 | errno = 0; | |
75a57bf0 | 647 | |
1fd5e000 CF |
648 | if ((dirp = g_opendir(pathbuf, pglob)) == NULL) { |
649 | /* TODO: don't call for ENOENT or ENOTDIR? */ | |
650 | if (pglob->gl_errfunc) { | |
733c9ed3 CV |
651 | if (g_Ctoc(pathbuf, buf, sizeof(buf))) |
652 | return (GLOB_ABORTED); | |
1fd5e000 CF |
653 | if (pglob->gl_errfunc(buf, errno) || |
654 | pglob->gl_flags & GLOB_ERR) | |
733c9ed3 | 655 | return (GLOB_ABORTED); |
1fd5e000 | 656 | } |
d7bcd2a1 | 657 | return(0); |
1fd5e000 CF |
658 | } |
659 | ||
660 | err = 0; | |
661 | ||
662 | /* Search directory for matching names. */ | |
663 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) | |
664 | readdirfunc = pglob->gl_readdir; | |
665 | else | |
733c9ed3 | 666 | readdirfunc = (dirent*(*)(void*)) readdir; |
1fd5e000 | 667 | while ((dp = (*readdirfunc)(dirp))) { |
733c9ed3 CV |
668 | char *sc; |
669 | Char *dc; | |
670 | wchar_t wc; | |
671 | size_t clen; | |
672 | mbstate_t mbs; | |
1fd5e000 CF |
673 | |
674 | /* Initial DOT must be matched literally. */ | |
675 | if (dp->d_name[0] == DOT && *pattern != DOT) | |
676 | continue; | |
733c9ed3 CV |
677 | memset(&mbs, 0, sizeof(mbs)); |
678 | dc = pathend; | |
679 | sc = dp->d_name; | |
680 | while (dc < pathend_last) { | |
681 | clen = mbrtowc(&wc, sc, MB_LEN_MAX, &mbs); | |
682 | if (clen == (size_t)-1 || clen == (size_t)-2) { | |
683 | wc = *sc; | |
684 | clen = 1; | |
685 | memset(&mbs, 0, sizeof(mbs)); | |
686 | } | |
687 | if ((*dc++ = wc) == EOS) | |
688 | break; | |
689 | sc += clen; | |
690 | } | |
1fd5e000 CF |
691 | if (!match(pathend, pattern, restpattern)) { |
692 | *pathend = EOS; | |
693 | continue; | |
694 | } | |
733c9ed3 CV |
695 | err = glob2(pathbuf, --dc, pathend_last, restpattern, |
696 | pglob, limit); | |
1fd5e000 CF |
697 | if (err) |
698 | break; | |
699 | } | |
700 | ||
701 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) | |
702 | (*pglob->gl_closedir)(dirp); | |
703 | else | |
704 | closedir(dirp); | |
705 | return(err); | |
706 | } | |
707 | ||
708 | ||
709 | /* | |
733c9ed3 | 710 | * Extend the gl_pathv member of a glob_t structure to accomodate a new item, |
1fd5e000 CF |
711 | * add the new item, and update gl_pathc. |
712 | * | |
713 | * This assumes the BSD realloc, which only copies the block when its size | |
714 | * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic | |
715 | * behavior. | |
716 | * | |
717 | * Return 0 if new item added, error code if memory couldn't be allocated. | |
718 | * | |
719 | * Invariant of the glob_t structure: | |
720 | * Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and | |
721 | * gl_pathv points to (gl_offs + gl_pathc + 1) items. | |
722 | */ | |
723 | static int | |
733c9ed3 | 724 | globextend(const Char *path, glob_t *pglob, size_t *limit) |
1fd5e000 | 725 | { |
733c9ed3 CV |
726 | char **pathv; |
727 | size_t i, newsize, len; | |
1fd5e000 CF |
728 | char *copy; |
729 | const Char *p; | |
730 | ||
733c9ed3 CV |
731 | if (*limit && pglob->gl_pathc > *limit) { |
732 | errno = 0; | |
733 | return (GLOB_NOSPACE); | |
734 | } | |
735 | ||
1fd5e000 | 736 | newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs); |
75a57bf0 | 737 | pathv = pglob->gl_pathv ? |
733c9ed3 CV |
738 | (char **) realloc((char *)pglob->gl_pathv, newsize) : |
739 | (char **) malloc(newsize); | |
740 | if (pathv == NULL) { | |
741 | if (pglob->gl_pathv) { | |
742 | free(pglob->gl_pathv); | |
743 | pglob->gl_pathv = NULL; | |
744 | } | |
1fd5e000 | 745 | return(GLOB_NOSPACE); |
733c9ed3 | 746 | } |
1fd5e000 CF |
747 | |
748 | if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) { | |
749 | /* first time around -- clear initial gl_offs items */ | |
750 | pathv += pglob->gl_offs; | |
733c9ed3 | 751 | for (i = pglob->gl_offs + 1; --i > 0; ) |
1fd5e000 CF |
752 | *--pathv = NULL; |
753 | } | |
754 | pglob->gl_pathv = pathv; | |
755 | ||
756 | for (p = path; *p++;) | |
757 | continue; | |
733c9ed3 CV |
758 | len = MB_CUR_MAX * (size_t)(p - path); /* XXX overallocation */ |
759 | if ((copy = (char *) malloc(len)) != NULL) { | |
760 | if (g_Ctoc(path, copy, len)) { | |
761 | free(copy); | |
762 | return (GLOB_NOSPACE); | |
763 | } | |
1fd5e000 CF |
764 | pathv[pglob->gl_offs + pglob->gl_pathc++] = copy; |
765 | } | |
766 | pathv[pglob->gl_offs + pglob->gl_pathc] = NULL; | |
767 | return(copy == NULL ? GLOB_NOSPACE : 0); | |
768 | } | |
769 | ||
1fd5e000 CF |
770 | /* |
771 | * pattern matching function for filenames. Each occurrence of the * | |
772 | * pattern causes a recursion level. | |
773 | */ | |
774 | static int | |
733c9ed3 | 775 | match(Char *name, Char *pat, Char *patend) |
1fd5e000 CF |
776 | { |
777 | int ok, negate_range; | |
778 | Char c, k; | |
779 | ||
780 | while (pat < patend) { | |
781 | c = *pat++; | |
782 | switch (c & M_MASK) { | |
783 | case M_ALL: | |
784 | if (pat == patend) | |
785 | return(1); | |
75a57bf0 | 786 | do |
1fd5e000 CF |
787 | if (match(name, pat, patend)) |
788 | return(1); | |
789 | while (*name++ != EOS); | |
790 | return(0); | |
791 | case M_ONE: | |
792 | if (*name++ == EOS) | |
793 | return(0); | |
794 | break; | |
795 | case M_SET: | |
796 | ok = 0; | |
797 | if ((k = *name++) == EOS) | |
798 | return(0); | |
799 | if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS) | |
800 | ++pat; | |
d7bcd2a1 | 801 | while (((c = *pat++) & M_MASK) != M_END) |
733c9ed3 | 802 | if ((*pat & M_MASK) == M_RNG) { |
d7bcd2a1 CV |
803 | if (__collate_load_error ? |
804 | CCHAR(c) <= CCHAR(k) && CCHAR(k) <= CCHAR(pat[1]) : | |
805 | __collate_range_cmp(CCHAR(c), CCHAR(k)) <= 0 | |
806 | && __collate_range_cmp(CCHAR(k), CCHAR(pat[1])) <= 0 | |
807 | ) | |
808 | ok = 1; | |
809 | pat += 2; | |
733c9ed3 | 810 | } else if (c == k) |
d7bcd2a1 | 811 | ok = 1; |
1fd5e000 CF |
812 | if (ok == negate_range) |
813 | return(0); | |
814 | break; | |
815 | default: | |
d7bcd2a1 | 816 | if (Cchar(*name++) != Cchar(c)) |
733c9ed3 | 817 | return(0); |
1fd5e000 CF |
818 | break; |
819 | } | |
820 | } | |
821 | return(*name == EOS); | |
822 | } | |
823 | ||
824 | /* Free allocated data belonging to a glob_t structure. */ | |
825 | void | |
733c9ed3 | 826 | globfree(glob_t *pglob) |
1fd5e000 | 827 | { |
733c9ed3 CV |
828 | size_t i; |
829 | char **pp; | |
1fd5e000 CF |
830 | |
831 | if (pglob->gl_pathv != NULL) { | |
832 | pp = pglob->gl_pathv + pglob->gl_offs; | |
833 | for (i = pglob->gl_pathc; i--; ++pp) | |
834 | if (*pp) | |
835 | free(*pp); | |
836 | free(pglob->gl_pathv); | |
733c9ed3 | 837 | pglob->gl_pathv = NULL; |
1fd5e000 CF |
838 | } |
839 | } | |
840 | ||
841 | static DIR * | |
733c9ed3 | 842 | g_opendir(Char *str, glob_t *pglob) |
1fd5e000 CF |
843 | { |
844 | char buf[MAXPATHLEN]; | |
845 | ||
846 | if (!*str) | |
847 | strcpy(buf, "."); | |
733c9ed3 CV |
848 | else { |
849 | if (g_Ctoc(str, buf, sizeof(buf))) | |
850 | return (NULL); | |
851 | } | |
1fd5e000 CF |
852 | |
853 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) | |
733c9ed3 | 854 | return (DIR *) ((*pglob->gl_opendir)((const char *) buf)); |
1fd5e000 CF |
855 | |
856 | return(opendir(buf)); | |
857 | } | |
858 | ||
61522196 CV |
859 | #ifdef __x86_64__ |
860 | #define CYGWIN_gl_stat(sfptr) ((*pglob->sfptr) (buf, sb)) | |
861 | #else | |
2f263187 | 862 | static void |
61522196 | 863 | stat32_to_stat64 (struct __stat32 *src, struct stat *dst) |
2f263187 CV |
864 | { |
865 | dst->st_dev = src->st_dev; | |
866 | dst->st_ino = src->st_ino; | |
867 | dst->st_mode = src->st_mode; | |
868 | dst->st_nlink = src->st_nlink; | |
869 | dst->st_uid = src->st_uid; | |
870 | dst->st_gid = src->st_gid; | |
871 | dst->st_rdev = src->st_rdev; | |
872 | dst->st_size = src->st_size; | |
c4e6ff48 CV |
873 | dst->st_atim = src->st_atim; |
874 | dst->st_mtim = src->st_mtim; | |
875 | dst->st_ctim = src->st_ctim; | |
7113f5da | 876 | dst->st_birthtim = src->st_mtim; |
2f263187 CV |
877 | dst->st_blksize = src->st_blksize; |
878 | dst->st_blocks = src->st_blocks; | |
879 | } | |
880 | ||
d7bcd2a1 CV |
881 | #define CYGWIN_gl_stat(sfptr) \ |
882 | ({ int ret; \ | |
883 | struct __stat32 lsb; \ | |
884 | if (CYGWIN_VERSION_CHECK_FOR_USING_BIG_TYPES) \ | |
885 | ret = (*pglob->sfptr) (buf, sb); \ | |
61522196 | 886 | else if (!(ret = (*pglob->sfptr) (buf, (struct stat *) &lsb))) \ |
d7bcd2a1 CV |
887 | stat32_to_stat64 (&lsb, sb); \ |
888 | ret; \ | |
889 | }) | |
61522196 | 890 | #endif |
d7bcd2a1 | 891 | |
1fd5e000 | 892 | static int |
d7bcd2a1 | 893 | g_lstat(Char *fn, struct stat *sb, glob_t *pglob) |
1fd5e000 CF |
894 | { |
895 | char buf[MAXPATHLEN]; | |
896 | ||
733c9ed3 CV |
897 | if (g_Ctoc(fn, buf, sizeof(buf))) { |
898 | errno = ENAMETOOLONG; | |
899 | return (-1); | |
900 | } | |
d7bcd2a1 CV |
901 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) |
902 | return CYGWIN_gl_stat (gl_lstat); | |
acb56175 | 903 | return(lstat64(buf, sb)); |
1fd5e000 CF |
904 | } |
905 | ||
906 | static int | |
d7bcd2a1 | 907 | g_stat(Char *fn, struct stat *sb, glob_t *pglob) |
1fd5e000 CF |
908 | { |
909 | char buf[MAXPATHLEN]; | |
910 | ||
733c9ed3 CV |
911 | if (g_Ctoc(fn, buf, sizeof(buf))) { |
912 | errno = ENAMETOOLONG; | |
913 | return (-1); | |
914 | } | |
d7bcd2a1 CV |
915 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) |
916 | return CYGWIN_gl_stat (gl_stat); | |
acb56175 | 917 | return(stat64(buf, sb)); |
1fd5e000 CF |
918 | } |
919 | ||
d7bcd2a1 CV |
920 | static const Char * |
921 | g_strchr(const Char *str, wchar_t ch) | |
1fd5e000 | 922 | { |
733c9ed3 | 923 | |
1fd5e000 CF |
924 | do { |
925 | if (*str == ch) | |
926 | return (str); | |
927 | } while (*str++); | |
928 | return (NULL); | |
929 | } | |
930 | ||
733c9ed3 CV |
931 | static int |
932 | g_Ctoc(const Char *str, char *buf, size_t len) | |
1fd5e000 | 933 | { |
733c9ed3 CV |
934 | mbstate_t mbs; |
935 | size_t clen; | |
936 | ||
937 | memset(&mbs, 0, sizeof(mbs)); | |
938 | while (len >= (size_t) MB_CUR_MAX) { | |
939 | clen = wcrtomb(buf, *str, &mbs); | |
940 | if (clen == (size_t)-1) | |
941 | return (1); | |
942 | if (*str == L'\0') | |
943 | return (0); | |
944 | str++; | |
945 | buf += clen; | |
946 | len -= clen; | |
947 | } | |
948 | return (1); | |
1fd5e000 CF |
949 | } |
950 | ||
951 | #ifdef DEBUG | |
75a57bf0 | 952 | static void |
733c9ed3 | 953 | qprintf(const char *str, Char *s) |
1fd5e000 | 954 | { |
733c9ed3 | 955 | Char *p; |
1fd5e000 CF |
956 | |
957 | (void)printf("%s:\n", str); | |
958 | for (p = s; *p; p++) | |
959 | (void)printf("%c", CHAR(*p)); | |
960 | (void)printf("\n"); | |
961 | for (p = s; *p; p++) | |
962 | (void)printf("%c", *p & M_PROTECT ? '"' : ' '); | |
963 | (void)printf("\n"); | |
964 | for (p = s; *p; p++) | |
965 | (void)printf("%c", ismeta(*p) ? '_' : ' '); | |
966 | (void)printf("\n"); | |
967 | } | |
968 | #endif |