Commit | Line | Data |
72b16652 |
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. |
0e950d83 |
16 | * 3. Neither the name of the University nor the names of its contributors |
72b16652 |
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 | |
72b16652 |
33 | #if defined(LIBC_SCCS) && !defined(lint) |
34 | static char sccsid[] = "@(#)glob.c 8.3 (Berkeley) 10/13/93"; |
35 | #endif /* LIBC_SCCS and not lint */ |
36 | |
37 | /* |
38 | * glob(3) -- a superset of the one defined in POSIX 1003.2. |
39 | * |
40 | * The [!...] convention to negate a range is supported (SysV, Posix, ksh). |
41 | * |
42 | * Optional extra services, controlled by flags not defined by POSIX: |
43 | * |
44 | * GLOB_QUOTE: |
45 | * Escaping convention: \ inhibits any special meaning the following |
46 | * character might have (except \ at end of string is retained). |
47 | * GLOB_MAGCHAR: |
48 | * Set in gl_flags if pattern contained a globbing character. |
49 | * GLOB_NOMAGIC: |
50 | * Same as GLOB_NOCHECK, but it will only append pattern if it did |
51 | * not contain any magic characters. [Used in csh style globbing] |
52 | * GLOB_ALTDIRFUNC: |
53 | * Use alternately specified directory access functions. |
54 | * GLOB_TILDE: |
55 | * expand ~user/foo to the /home/dir/of/user/foo |
56 | * GLOB_BRACE: |
57 | * expand {1,2}{a,b} to 1a 1b 2a 2b |
58 | * gl_matchc: |
59 | * Number of matches in the current invocation of glob. |
2d5e9e5d |
60 | * GLOB_ALPHASORT: |
61 | * sort alphabetically like csh (case doesn't matter) instead of in ASCII |
62 | * order |
72b16652 |
63 | */ |
64 | |
65 | #include <EXTERN.h> |
66 | #include <perl.h> |
4f49e16e |
67 | #include <XSUB.h> |
68 | |
72b16652 |
69 | #include "bsd_glob.h" |
70 | #ifdef I_PWD |
71 | # include <pwd.h> |
72 | #else |
73 | #ifdef HAS_PASSWD |
74 | struct passwd *getpwnam(char *); |
75 | struct passwd *getpwuid(Uid_t); |
76 | #endif |
77 | #endif |
78 | |
79 | #ifndef MAXPATHLEN |
80 | # ifdef PATH_MAX |
81 | # define MAXPATHLEN PATH_MAX |
7369a524 |
82 | # ifdef MACOS_TRADITIONAL |
83 | # define MAXPATHLEN 255 |
84 | # else |
85 | # define MAXPATHLEN 1024 |
86 | # endif |
72b16652 |
87 | # endif |
88 | #endif |
89 | |
b8ef571c |
90 | #ifdef I_LIMITS |
91 | #include <limits.h> |
92 | #endif |
93 | |
94 | #ifndef ARG_MAX |
95 | # ifdef _SC_ARG_MAX |
96 | # define ARG_MAX (sysconf(_SC_ARG_MAX)) |
97 | # else |
98 | # ifdef _POSIX_ARG_MAX |
99 | # define ARG_MAX _POSIX_ARG_MAX |
100 | # else |
101 | # ifdef WIN32 |
102 | # define ARG_MAX 14500 /* from VC's limits.h */ |
103 | # else |
104 | # define ARG_MAX 4096 /* from POSIX, be conservative */ |
105 | # endif |
106 | # endif |
107 | # endif |
108 | #endif |
109 | |
3e5d0dec |
110 | #define BG_DOLLAR '$' |
111 | #define BG_DOT '.' |
112 | #define BG_EOS '\0' |
113 | #define BG_LBRACKET '[' |
114 | #define BG_NOT '!' |
115 | #define BG_QUESTION '?' |
116 | #define BG_QUOTE '\\' |
117 | #define BG_RANGE '-' |
118 | #define BG_RBRACKET ']' |
7369a524 |
119 | #ifdef MACOS_TRADITIONAL |
120 | # define BG_SEP ':' |
121 | #else |
122 | # define BG_SEP '/' |
123 | #endif |
220398a0 |
124 | #ifdef DOSISH |
125 | #define BG_SEP2 '\\' |
126 | #endif |
3e5d0dec |
127 | #define BG_STAR '*' |
128 | #define BG_TILDE '~' |
129 | #define BG_UNDERSCORE '_' |
130 | #define BG_LBRACE '{' |
131 | #define BG_RBRACE '}' |
132 | #define BG_SLASH '/' |
133 | #define BG_COMMA ',' |
72b16652 |
134 | |
135 | #ifndef GLOB_DEBUG |
136 | |
137 | #define M_QUOTE 0x8000 |
138 | #define M_PROTECT 0x4000 |
139 | #define M_MASK 0xffff |
140 | #define M_ASCII 0x00ff |
141 | |
142 | typedef U16 Char; |
143 | |
144 | #else |
145 | |
146 | #define M_QUOTE 0x80 |
147 | #define M_PROTECT 0x40 |
148 | #define M_MASK 0xff |
149 | #define M_ASCII 0x7f |
150 | |
151 | typedef U8 Char; |
152 | |
153 | #endif /* !GLOB_DEBUG */ |
154 | |
155 | |
156 | #define CHAR(c) ((Char)((c)&M_ASCII)) |
157 | #define META(c) ((Char)((c)|M_QUOTE)) |
158 | #define M_ALL META('*') |
159 | #define M_END META(']') |
160 | #define M_NOT META('!') |
161 | #define M_ONE META('?') |
162 | #define M_RNG META('-') |
163 | #define M_SET META('[') |
164 | #define ismeta(c) (((c)&M_QUOTE) != 0) |
165 | |
166 | |
167 | static int compare(const void *, const void *); |
220398a0 |
168 | static int ci_compare(const void *, const void *); |
b8ef571c |
169 | static int g_Ctoc(const Char *, char *, STRLEN); |
72b16652 |
170 | static int g_lstat(Char *, Stat_t *, glob_t *); |
171 | static DIR *g_opendir(Char *, glob_t *); |
172 | static Char *g_strchr(Char *, int); |
72b16652 |
173 | static int g_stat(Char *, Stat_t *, glob_t *); |
174 | static int glob0(const Char *, glob_t *); |
b8ef571c |
175 | static int glob1(Char *, Char *, glob_t *, size_t *); |
176 | static int glob2(Char *, Char *, Char *, Char *, Char *, Char *, |
177 | glob_t *, size_t *); |
178 | static int glob3(Char *, Char *, Char *, Char *, Char *, Char *, |
179 | Char *, Char *, glob_t *, size_t *); |
180 | static int globextend(const Char *, glob_t *, size_t *); |
181 | static const Char * |
182 | globtilde(const Char *, Char *, size_t, glob_t *); |
72b16652 |
183 | static int globexp1(const Char *, glob_t *); |
184 | static int globexp2(const Char *, const Char *, glob_t *, int *); |
220398a0 |
185 | static int match(Char *, Char *, Char *, int); |
72b16652 |
186 | #ifdef GLOB_DEBUG |
187 | static void qprintf(const char *, Char *); |
188 | #endif /* GLOB_DEBUG */ |
189 | |
4f49e16e |
190 | #ifdef PERL_IMPLICIT_CONTEXT |
191 | static Direntry_t * my_readdir(DIR*); |
192 | |
193 | static Direntry_t * |
194 | my_readdir(DIR *d) |
195 | { |
196 | return PerlDir_read(d); |
197 | } |
198 | #else |
199 | #define my_readdir readdir |
200 | #endif |
201 | |
72b16652 |
202 | int |
203 | bsd_glob(const char *pattern, int flags, |
204 | int (*errfunc)(const char *, int), glob_t *pglob) |
205 | { |
206 | const U8 *patnext; |
207 | int c; |
b8ef571c |
208 | Char *bufnext, *bufend, patbuf[MAXPATHLEN]; |
72b16652 |
209 | |
210 | patnext = (U8 *) pattern; |
211 | if (!(flags & GLOB_APPEND)) { |
212 | pglob->gl_pathc = 0; |
213 | pglob->gl_pathv = NULL; |
214 | if (!(flags & GLOB_DOOFFS)) |
215 | pglob->gl_offs = 0; |
216 | } |
217 | pglob->gl_flags = flags & ~GLOB_MAGCHAR; |
218 | pglob->gl_errfunc = errfunc; |
219 | pglob->gl_matchc = 0; |
220 | |
221 | bufnext = patbuf; |
b8ef571c |
222 | bufend = bufnext + MAXPATHLEN - 1; |
220398a0 |
223 | #ifdef DOSISH |
224 | /* Nasty hack to treat patterns like "C:*" correctly. In this |
225 | * case, the * should match any file in the current directory |
226 | * on the C: drive. However, the glob code does not treat the |
227 | * colon specially, so it looks for files beginning "C:" in |
228 | * the current directory. To fix this, change the pattern to |
229 | * add an explicit "./" at the start (just after the drive |
230 | * letter and colon - ie change to "C:./*"). |
231 | */ |
232 | if (isalpha(pattern[0]) && pattern[1] == ':' && |
233 | pattern[2] != BG_SEP && pattern[2] != BG_SEP2 && |
234 | bufend - bufnext > 4) { |
235 | *bufnext++ = pattern[0]; |
236 | *bufnext++ = ':'; |
237 | *bufnext++ = '.'; |
238 | *bufnext++ = BG_SEP; |
239 | patnext += 2; |
240 | } |
241 | #endif |
72b16652 |
242 | if (flags & GLOB_QUOTE) { |
243 | /* Protect the quoted characters. */ |
3e5d0dec |
244 | while (bufnext < bufend && (c = *patnext++) != BG_EOS) |
245 | if (c == BG_QUOTE) { |
220398a0 |
246 | #ifdef DOSISH |
247 | /* To avoid backslashitis on Win32, |
248 | * we only treat \ as a quoting character |
249 | * if it precedes one of the |
250 | * metacharacters []-{}~\ |
251 | */ |
252 | if ((c = *patnext++) != '[' && c != ']' && |
253 | c != '-' && c != '{' && c != '}' && |
254 | c != '~' && c != '\\') { |
255 | #else |
3e5d0dec |
256 | if ((c = *patnext++) == BG_EOS) { |
220398a0 |
257 | #endif |
3e5d0dec |
258 | c = BG_QUOTE; |
72b16652 |
259 | --patnext; |
260 | } |
261 | *bufnext++ = c | M_PROTECT; |
b8ef571c |
262 | } else |
72b16652 |
263 | *bufnext++ = c; |
b8ef571c |
264 | } else |
265 | while (bufnext < bufend && (c = *patnext++) != BG_EOS) |
266 | *bufnext++ = c; |
3e5d0dec |
267 | *bufnext = BG_EOS; |
72b16652 |
268 | |
269 | if (flags & GLOB_BRACE) |
270 | return globexp1(patbuf, pglob); |
271 | else |
272 | return glob0(patbuf, pglob); |
273 | } |
274 | |
275 | /* |
276 | * Expand recursively a glob {} pattern. When there is no more expansion |
277 | * invoke the standard globbing routine to glob the rest of the magic |
278 | * characters |
279 | */ |
b8ef571c |
280 | static int |
281 | globexp1(const Char *pattern, glob_t *pglob) |
72b16652 |
282 | { |
283 | const Char* ptr = pattern; |
284 | int rv; |
285 | |
286 | /* Protect a single {}, for find(1), like csh */ |
3e5d0dec |
287 | if (pattern[0] == BG_LBRACE && pattern[1] == BG_RBRACE && pattern[2] == BG_EOS) |
72b16652 |
288 | return glob0(pattern, pglob); |
289 | |
3e5d0dec |
290 | while ((ptr = (const Char *) g_strchr((Char *) ptr, BG_LBRACE)) != NULL) |
72b16652 |
291 | if (!globexp2(ptr, pattern, pglob, &rv)) |
292 | return rv; |
293 | |
294 | return glob0(pattern, pglob); |
295 | } |
296 | |
297 | |
298 | /* |
299 | * Recursive brace globbing helper. Tries to expand a single brace. |
300 | * If it succeeds then it invokes globexp1 with the new pattern. |
301 | * If it fails then it tries to glob the rest of the pattern and returns. |
302 | */ |
b8ef571c |
303 | static int |
304 | globexp2(const Char *ptr, const Char *pattern, |
305 | glob_t *pglob, int *rv) |
72b16652 |
306 | { |
307 | int i; |
308 | Char *lm, *ls; |
309 | const Char *pe, *pm, *pl; |
b8ef571c |
310 | Char patbuf[MAXPATHLEN]; |
72b16652 |
311 | |
312 | /* copy part up to the brace */ |
313 | for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++) |
b8ef571c |
314 | ; |
315 | *lm = BG_EOS; |
72b16652 |
316 | ls = lm; |
317 | |
318 | /* Find the balanced brace */ |
319 | for (i = 0, pe = ++ptr; *pe; pe++) |
3e5d0dec |
320 | if (*pe == BG_LBRACKET) { |
72b16652 |
321 | /* Ignore everything between [] */ |
3e5d0dec |
322 | for (pm = pe++; *pe != BG_RBRACKET && *pe != BG_EOS; pe++) |
b8ef571c |
323 | ; |
3e5d0dec |
324 | if (*pe == BG_EOS) { |
72b16652 |
325 | /* |
3e5d0dec |
326 | * We could not find a matching BG_RBRACKET. |
327 | * Ignore and just look for BG_RBRACE |
72b16652 |
328 | */ |
329 | pe = pm; |
330 | } |
b8ef571c |
331 | } else if (*pe == BG_LBRACE) |
72b16652 |
332 | i++; |
3e5d0dec |
333 | else if (*pe == BG_RBRACE) { |
72b16652 |
334 | if (i == 0) |
335 | break; |
336 | i--; |
337 | } |
338 | |
339 | /* Non matching braces; just glob the pattern */ |
3e5d0dec |
340 | if (i != 0 || *pe == BG_EOS) { |
72b16652 |
341 | *rv = glob0(patbuf, pglob); |
342 | return 0; |
343 | } |
344 | |
b8ef571c |
345 | for (i = 0, pl = pm = ptr; pm <= pe; pm++) { |
72b16652 |
346 | switch (*pm) { |
3e5d0dec |
347 | case BG_LBRACKET: |
72b16652 |
348 | /* Ignore everything between [] */ |
3e5d0dec |
349 | for (pl = pm++; *pm != BG_RBRACKET && *pm != BG_EOS; pm++) |
b8ef571c |
350 | ; |
3e5d0dec |
351 | if (*pm == BG_EOS) { |
72b16652 |
352 | /* |
3e5d0dec |
353 | * We could not find a matching BG_RBRACKET. |
354 | * Ignore and just look for BG_RBRACE |
72b16652 |
355 | */ |
356 | pm = pl; |
357 | } |
358 | break; |
359 | |
3e5d0dec |
360 | case BG_LBRACE: |
72b16652 |
361 | i++; |
362 | break; |
363 | |
3e5d0dec |
364 | case BG_RBRACE: |
72b16652 |
365 | if (i) { |
b8ef571c |
366 | i--; |
367 | break; |
72b16652 |
368 | } |
369 | /* FALLTHROUGH */ |
3e5d0dec |
370 | case BG_COMMA: |
371 | if (i && *pm == BG_COMMA) |
72b16652 |
372 | break; |
373 | else { |
374 | /* Append the current string */ |
375 | for (lm = ls; (pl < pm); *lm++ = *pl++) |
b8ef571c |
376 | ; |
377 | |
72b16652 |
378 | /* |
379 | * Append the rest of the pattern after the |
380 | * closing brace |
381 | */ |
b8ef571c |
382 | for (pl = pe + 1; (*lm++ = *pl++) != BG_EOS; ) |
383 | ; |
72b16652 |
384 | |
385 | /* Expand the current pattern */ |
386 | #ifdef GLOB_DEBUG |
387 | qprintf("globexp2:", patbuf); |
388 | #endif /* GLOB_DEBUG */ |
389 | *rv = globexp1(patbuf, pglob); |
390 | |
391 | /* move after the comma, to the next string */ |
392 | pl = pm + 1; |
393 | } |
394 | break; |
395 | |
396 | default: |
397 | break; |
398 | } |
b8ef571c |
399 | } |
72b16652 |
400 | *rv = 0; |
401 | return 0; |
402 | } |
403 | |
404 | |
405 | |
406 | /* |
407 | * expand tilde from the passwd file. |
408 | */ |
409 | static const Char * |
b8ef571c |
410 | globtilde(const Char *pattern, Char *patbuf, size_t limit, glob_t *pglob) |
72b16652 |
411 | { |
412 | struct passwd *pwd; |
413 | char *h; |
414 | const Char *p; |
415 | Char *b; |
416 | |
3e5d0dec |
417 | if (*pattern != BG_TILDE || !(pglob->gl_flags & GLOB_TILDE)) |
72b16652 |
418 | return pattern; |
419 | |
420 | /* Copy up to the end of the string or / */ |
3e5d0dec |
421 | for (p = pattern + 1, h = (char *) patbuf; *p && *p != BG_SLASH; |
72b16652 |
422 | *h++ = *p++) |
b8ef571c |
423 | ; |
72b16652 |
424 | |
3e5d0dec |
425 | *h = BG_EOS; |
72b16652 |
426 | |
b8ef571c |
427 | #if 0 |
428 | if (h == (char *)eb) |
429 | return what; |
430 | #endif |
431 | |
3e5d0dec |
432 | if (((char *) patbuf)[0] == BG_EOS) { |
72b16652 |
433 | /* |
434 | * handle a plain ~ or ~/ by expanding $HOME |
435 | * first and then trying the password file |
436 | */ |
437 | if ((h = getenv("HOME")) == NULL) { |
438 | #ifdef HAS_PASSWD |
439 | if ((pwd = getpwuid(getuid())) == NULL) |
440 | return pattern; |
441 | else |
442 | h = pwd->pw_dir; |
443 | #else |
444 | return pattern; |
445 | #endif |
446 | } |
b8ef571c |
447 | } else { |
72b16652 |
448 | /* |
449 | * Expand a ~user |
450 | */ |
451 | #ifdef HAS_PASSWD |
452 | if ((pwd = getpwnam((char*) patbuf)) == NULL) |
453 | return pattern; |
454 | else |
455 | h = pwd->pw_dir; |
456 | #else |
457 | return pattern; |
458 | #endif |
459 | } |
460 | |
461 | /* Copy the home directory */ |
462 | for (b = patbuf; *h; *b++ = *h++) |
b8ef571c |
463 | ; |
72b16652 |
464 | |
465 | /* Append the rest of the pattern */ |
3e5d0dec |
466 | while ((*b++ = *p++) != BG_EOS) |
b8ef571c |
467 | ; |
72b16652 |
468 | |
469 | return patbuf; |
470 | } |
471 | |
472 | |
473 | /* |
474 | * The main glob() routine: compiles the pattern (optionally processing |
475 | * quotes), calls glob1() to do the real pattern matching, and finally |
476 | * sorts the list (unless unsorted operation is requested). Returns 0 |
477 | * if things went well, nonzero if errors occurred. It is not an error |
478 | * to find no matches. |
479 | */ |
480 | static int |
481 | glob0(const Char *pattern, glob_t *pglob) |
482 | { |
483 | const Char *qpat, *qpatnext; |
484 | int c, err, oldflags, oldpathc; |
b8ef571c |
485 | Char *bufnext, patbuf[MAXPATHLEN]; |
486 | size_t limit = 0; |
72b16652 |
487 | |
7369a524 |
488 | #ifdef MACOS_TRADITIONAL |
489 | if ( (*pattern == BG_TILDE) && (pglob->gl_flags & GLOB_TILDE) ) { |
490 | return(globextend(pattern, pglob)); |
491 | } |
492 | #endif |
493 | |
b8ef571c |
494 | qpat = globtilde(pattern, patbuf, MAXPATHLEN, pglob); |
72b16652 |
495 | qpatnext = qpat; |
496 | oldflags = pglob->gl_flags; |
497 | oldpathc = pglob->gl_pathc; |
498 | bufnext = patbuf; |
499 | |
500 | /* We don't need to check for buffer overflow any more. */ |
3e5d0dec |
501 | while ((c = *qpatnext++) != BG_EOS) { |
72b16652 |
502 | switch (c) { |
3e5d0dec |
503 | case BG_LBRACKET: |
72b16652 |
504 | c = *qpatnext; |
3e5d0dec |
505 | if (c == BG_NOT) |
72b16652 |
506 | ++qpatnext; |
3e5d0dec |
507 | if (*qpatnext == BG_EOS || |
508 | g_strchr((Char *) qpatnext+1, BG_RBRACKET) == NULL) { |
509 | *bufnext++ = BG_LBRACKET; |
510 | if (c == BG_NOT) |
72b16652 |
511 | --qpatnext; |
512 | break; |
513 | } |
514 | *bufnext++ = M_SET; |
3e5d0dec |
515 | if (c == BG_NOT) |
72b16652 |
516 | *bufnext++ = M_NOT; |
517 | c = *qpatnext++; |
518 | do { |
519 | *bufnext++ = CHAR(c); |
3e5d0dec |
520 | if (*qpatnext == BG_RANGE && |
521 | (c = qpatnext[1]) != BG_RBRACKET) { |
72b16652 |
522 | *bufnext++ = M_RNG; |
523 | *bufnext++ = CHAR(c); |
524 | qpatnext += 2; |
525 | } |
3e5d0dec |
526 | } while ((c = *qpatnext++) != BG_RBRACKET); |
72b16652 |
527 | pglob->gl_flags |= GLOB_MAGCHAR; |
528 | *bufnext++ = M_END; |
529 | break; |
3e5d0dec |
530 | case BG_QUESTION: |
72b16652 |
531 | pglob->gl_flags |= GLOB_MAGCHAR; |
532 | *bufnext++ = M_ONE; |
533 | break; |
3e5d0dec |
534 | case BG_STAR: |
72b16652 |
535 | pglob->gl_flags |= GLOB_MAGCHAR; |
536 | /* collapse adjacent stars to one, |
537 | * to avoid exponential behavior |
538 | */ |
539 | if (bufnext == patbuf || bufnext[-1] != M_ALL) |
b8ef571c |
540 | *bufnext++ = M_ALL; |
72b16652 |
541 | break; |
542 | default: |
543 | *bufnext++ = CHAR(c); |
544 | break; |
545 | } |
546 | } |
3e5d0dec |
547 | *bufnext = BG_EOS; |
72b16652 |
548 | #ifdef GLOB_DEBUG |
549 | qprintf("glob0:", patbuf); |
550 | #endif /* GLOB_DEBUG */ |
551 | |
b8ef571c |
552 | if ((err = glob1(patbuf, patbuf+MAXPATHLEN-1, pglob, &limit)) != 0) { |
72b16652 |
553 | pglob->gl_flags = oldflags; |
554 | return(err); |
555 | } |
556 | |
557 | /* |
558 | * If there was no match we are going to append the pattern |
559 | * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified |
560 | * and the pattern did not contain any magic characters |
561 | * GLOB_NOMAGIC is there just for compatibility with csh. |
562 | */ |
563 | if (pglob->gl_pathc == oldpathc && |
564 | ((pglob->gl_flags & GLOB_NOCHECK) || |
565 | ((pglob->gl_flags & GLOB_NOMAGIC) && |
566 | !(pglob->gl_flags & GLOB_MAGCHAR)))) |
567 | { |
568 | #ifdef GLOB_DEBUG |
569 | printf("calling globextend from glob0\n"); |
570 | #endif /* GLOB_DEBUG */ |
571 | pglob->gl_flags = oldflags; |
b8ef571c |
572 | return(globextend(qpat, pglob, &limit)); |
72b16652 |
573 | } |
574 | else if (!(pglob->gl_flags & GLOB_NOSORT)) |
575 | qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc, |
220398a0 |
576 | pglob->gl_pathc - oldpathc, sizeof(char *), |
2d5e9e5d |
577 | (pglob->gl_flags & (GLOB_ALPHASORT|GLOB_NOCASE)) |
578 | ? ci_compare : compare); |
72b16652 |
579 | pglob->gl_flags = oldflags; |
580 | return(0); |
581 | } |
582 | |
583 | static int |
220398a0 |
584 | ci_compare(const void *p, const void *q) |
585 | { |
b8ef571c |
586 | const char *pp = *(const char **)p; |
587 | const char *qq = *(const char **)q; |
588 | int ci; |
589 | while (*pp && *qq) { |
590 | if (tolower(*pp) != tolower(*qq)) |
591 | break; |
592 | ++pp; |
593 | ++qq; |
594 | } |
595 | ci = tolower(*pp) - tolower(*qq); |
596 | if (ci == 0) |
597 | return compare(p, q); |
598 | return ci; |
220398a0 |
599 | } |
600 | |
601 | static int |
72b16652 |
602 | compare(const void *p, const void *q) |
603 | { |
604 | return(strcmp(*(char **)p, *(char **)q)); |
605 | } |
606 | |
607 | static int |
b8ef571c |
608 | glob1(Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp) |
72b16652 |
609 | { |
b8ef571c |
610 | Char pathbuf[MAXPATHLEN]; |
72b16652 |
611 | |
612 | /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */ |
3e5d0dec |
613 | if (*pattern == BG_EOS) |
72b16652 |
614 | return(0); |
b8ef571c |
615 | return(glob2(pathbuf, pathbuf+MAXPATHLEN-1, |
616 | pathbuf, pathbuf+MAXPATHLEN-1, |
617 | pattern, pattern_last, pglob, limitp)); |
72b16652 |
618 | } |
619 | |
620 | /* |
621 | * The functions glob2 and glob3 are mutually recursive; there is one level |
622 | * of recursion for each segment in the pattern that contains one or more |
623 | * meta characters. |
624 | */ |
625 | static int |
b8ef571c |
626 | glob2(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last, |
627 | Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp) |
72b16652 |
628 | { |
629 | Stat_t sb; |
630 | Char *p, *q; |
631 | int anymeta; |
632 | |
633 | /* |
634 | * Loop over pattern segments until end of pattern or until |
635 | * segment with meta character found. |
636 | */ |
637 | for (anymeta = 0;;) { |
3e5d0dec |
638 | if (*pattern == BG_EOS) { /* End of pattern? */ |
639 | *pathend = BG_EOS; |
72b16652 |
640 | |
72b16652 |
641 | if (g_lstat(pathbuf, &sb, pglob)) |
642 | return(0); |
72b16652 |
643 | |
644 | if (((pglob->gl_flags & GLOB_MARK) && |
220398a0 |
645 | pathend[-1] != BG_SEP |
646 | #ifdef DOSISH |
647 | && pathend[-1] != BG_SEP2 |
648 | #endif |
b8ef571c |
649 | ) && (S_ISDIR(sb.st_mode) || |
650 | (S_ISLNK(sb.st_mode) && |
72b16652 |
651 | (g_stat(pathbuf, &sb, pglob) == 0) && |
652 | S_ISDIR(sb.st_mode)))) { |
b8ef571c |
653 | if (pathend+1 > pathend_last) |
654 | return (1); |
3e5d0dec |
655 | *pathend++ = BG_SEP; |
656 | *pathend = BG_EOS; |
72b16652 |
657 | } |
658 | ++pglob->gl_matchc; |
659 | #ifdef GLOB_DEBUG |
660 | printf("calling globextend from glob2\n"); |
661 | #endif /* GLOB_DEBUG */ |
b8ef571c |
662 | return(globextend(pathbuf, pglob, limitp)); |
72b16652 |
663 | } |
664 | |
665 | /* Find end of next segment, copy tentatively to pathend. */ |
666 | q = pathend; |
667 | p = pattern; |
220398a0 |
668 | while (*p != BG_EOS && *p != BG_SEP |
669 | #ifdef DOSISH |
670 | && *p != BG_SEP2 |
671 | #endif |
672 | ) { |
72b16652 |
673 | if (ismeta(*p)) |
674 | anymeta = 1; |
b8ef571c |
675 | if (q+1 > pathend_last) |
676 | return (1); |
72b16652 |
677 | *q++ = *p++; |
678 | } |
679 | |
680 | if (!anymeta) { /* No expansion, do next segment. */ |
681 | pathend = q; |
682 | pattern = p; |
220398a0 |
683 | while (*pattern == BG_SEP |
684 | #ifdef DOSISH |
685 | || *pattern == BG_SEP2 |
686 | #endif |
b8ef571c |
687 | ) { |
688 | if (pathend+1 > pathend_last) |
689 | return (1); |
72b16652 |
690 | *pathend++ = *pattern++; |
b8ef571c |
691 | } |
692 | } else |
693 | /* Need expansion, recurse. */ |
694 | return(glob3(pathbuf, pathbuf_last, pathend, |
695 | pathend_last, pattern, pattern_last, |
696 | p, pattern_last, pglob, limitp)); |
72b16652 |
697 | } |
698 | /* NOTREACHED */ |
699 | } |
700 | |
701 | static int |
b8ef571c |
702 | glob3(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last, |
703 | Char *pattern, Char *pattern_last, |
704 | Char *restpattern, Char *restpattern_last, glob_t *pglob, size_t *limitp) |
72b16652 |
705 | { |
706 | register Direntry_t *dp; |
707 | DIR *dirp; |
708 | int err; |
220398a0 |
709 | int nocase; |
72b16652 |
710 | char buf[MAXPATHLEN]; |
711 | |
712 | /* |
713 | * The readdirfunc declaration can't be prototyped, because it is |
714 | * assigned, below, to two functions which are prototyped in glob.h |
715 | * and dirent.h as taking pointers to differently typed opaque |
716 | * structures. |
717 | */ |
cb359b41 |
718 | Direntry_t *(*readdirfunc)(DIR*); |
72b16652 |
719 | |
b8ef571c |
720 | if (pathend > pathend_last) |
721 | return (1); |
3e5d0dec |
722 | *pathend = BG_EOS; |
72b16652 |
723 | errno = 0; |
724 | |
f0963acb |
725 | #ifdef VMS |
726 | { |
727 | Char *q = pathend; |
728 | if (q - pathbuf > 5) { |
729 | q -= 5; |
730 | if (q[0] == '.' && tolower(q[1]) == 'd' && tolower(q[2]) == 'i' |
731 | && tolower(q[3]) == 'r' && q[4] == '/') |
732 | { |
733 | q[0] = '/'; |
734 | q[1] = BG_EOS; |
735 | pathend = q+1; |
736 | } |
737 | } |
738 | } |
739 | #endif |
72b16652 |
740 | if ((dirp = g_opendir(pathbuf, pglob)) == NULL) { |
741 | /* TODO: don't call for ENOENT or ENOTDIR? */ |
742 | if (pglob->gl_errfunc) { |
b8ef571c |
743 | if (g_Ctoc(pathbuf, buf, sizeof(buf))) |
744 | return (GLOB_ABEND); |
72b16652 |
745 | if (pglob->gl_errfunc(buf, errno) || |
746 | (pglob->gl_flags & GLOB_ERR)) |
747 | return (GLOB_ABEND); |
748 | } |
749 | return(0); |
750 | } |
751 | |
752 | err = 0; |
220398a0 |
753 | nocase = ((pglob->gl_flags & GLOB_NOCASE) != 0); |
72b16652 |
754 | |
755 | /* Search directory for matching names. */ |
756 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) |
b8ef571c |
757 | readdirfunc = (Direntry_t *(*)(DIR *))pglob->gl_readdir; |
72b16652 |
758 | else |
4f49e16e |
759 | readdirfunc = my_readdir; |
72b16652 |
760 | while ((dp = (*readdirfunc)(dirp))) { |
761 | register U8 *sc; |
762 | register Char *dc; |
763 | |
3e5d0dec |
764 | /* Initial BG_DOT must be matched literally. */ |
765 | if (dp->d_name[0] == BG_DOT && *pattern != BG_DOT) |
72b16652 |
766 | continue; |
b8ef571c |
767 | dc = pathend; |
768 | sc = (U8 *) dp->d_name; |
769 | while (dc < pathend_last && (*dc++ = *sc++) != BG_EOS) |
770 | ; |
771 | if (dc >= pathend_last) { |
772 | *dc = BG_EOS; |
773 | err = 1; |
774 | break; |
775 | } |
776 | |
220398a0 |
777 | if (!match(pathend, pattern, restpattern, nocase)) { |
3e5d0dec |
778 | *pathend = BG_EOS; |
72b16652 |
779 | continue; |
780 | } |
b8ef571c |
781 | err = glob2(pathbuf, pathbuf_last, --dc, pathend_last, |
782 | restpattern, restpattern_last, pglob, limitp); |
72b16652 |
783 | if (err) |
784 | break; |
785 | } |
786 | |
787 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) |
788 | (*pglob->gl_closedir)(dirp); |
789 | else |
4f49e16e |
790 | PerlDir_close(dirp); |
72b16652 |
791 | return(err); |
792 | } |
793 | |
794 | |
795 | /* |
796 | * Extend the gl_pathv member of a glob_t structure to accomodate a new item, |
797 | * add the new item, and update gl_pathc. |
798 | * |
799 | * This assumes the BSD realloc, which only copies the block when its size |
800 | * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic |
801 | * behavior. |
802 | * |
803 | * Return 0 if new item added, error code if memory couldn't be allocated. |
804 | * |
805 | * Invariant of the glob_t structure: |
806 | * Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and |
807 | * gl_pathv points to (gl_offs + gl_pathc + 1) items. |
808 | */ |
809 | static int |
b8ef571c |
810 | globextend(const Char *path, glob_t *pglob, size_t *limitp) |
72b16652 |
811 | { |
812 | register char **pathv; |
813 | register int i; |
b8ef571c |
814 | STRLEN newsize, len; |
72b16652 |
815 | char *copy; |
816 | const Char *p; |
817 | |
818 | #ifdef GLOB_DEBUG |
819 | printf("Adding "); |
820 | for (p = path; *p; p++) |
821 | (void)printf("%c", CHAR(*p)); |
822 | printf("\n"); |
3e5d0dec |
823 | #endif /* GLOB_DEBUG */ |
72b16652 |
824 | |
b8ef571c |
825 | newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs); |
4f49e16e |
826 | if (pglob->gl_pathv) |
b8ef571c |
827 | pathv = Renew(pglob->gl_pathv,newsize,char*); |
4f49e16e |
828 | else |
b8ef571c |
829 | New(0,pathv,newsize,char*); |
830 | if (pathv == NULL) { |
831 | if (pglob->gl_pathv) { |
832 | Safefree(pglob->gl_pathv); |
833 | pglob->gl_pathv = NULL; |
834 | } |
72b16652 |
835 | return(GLOB_NOSPACE); |
b8ef571c |
836 | } |
72b16652 |
837 | |
838 | if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) { |
839 | /* first time around -- clear initial gl_offs items */ |
840 | pathv += pglob->gl_offs; |
841 | for (i = pglob->gl_offs; --i >= 0; ) |
842 | *--pathv = NULL; |
843 | } |
844 | pglob->gl_pathv = pathv; |
845 | |
846 | for (p = path; *p++;) |
b8ef571c |
847 | ; |
848 | len = (STRLEN)(p - path); |
4f49e16e |
849 | New(0, copy, p-path, char); |
850 | if (copy != NULL) { |
b8ef571c |
851 | if (g_Ctoc(path, copy, len)) { |
852 | Safefree(copy); |
853 | return(GLOB_NOSPACE); |
854 | } |
72b16652 |
855 | pathv[pglob->gl_offs + pglob->gl_pathc++] = copy; |
856 | } |
857 | pathv[pglob->gl_offs + pglob->gl_pathc] = NULL; |
b8ef571c |
858 | |
859 | if ((pglob->gl_flags & GLOB_LIMIT) && |
860 | newsize + *limitp >= ARG_MAX) { |
861 | errno = 0; |
862 | return(GLOB_NOSPACE); |
863 | } |
72b16652 |
864 | return(copy == NULL ? GLOB_NOSPACE : 0); |
865 | } |
866 | |
867 | |
868 | /* |
869 | * pattern matching function for filenames. Each occurrence of the * |
870 | * pattern causes a recursion level. |
871 | */ |
872 | static int |
220398a0 |
873 | match(register Char *name, register Char *pat, register Char *patend, int nocase) |
72b16652 |
874 | { |
875 | int ok, negate_range; |
876 | Char c, k; |
877 | |
878 | while (pat < patend) { |
879 | c = *pat++; |
880 | switch (c & M_MASK) { |
881 | case M_ALL: |
882 | if (pat == patend) |
883 | return(1); |
884 | do |
220398a0 |
885 | if (match(name, pat, patend, nocase)) |
72b16652 |
886 | return(1); |
b8ef571c |
887 | while (*name++ != BG_EOS) |
888 | ; |
72b16652 |
889 | return(0); |
890 | case M_ONE: |
3e5d0dec |
891 | if (*name++ == BG_EOS) |
72b16652 |
892 | return(0); |
893 | break; |
894 | case M_SET: |
895 | ok = 0; |
3e5d0dec |
896 | if ((k = *name++) == BG_EOS) |
72b16652 |
897 | return(0); |
3e5d0dec |
898 | if ((negate_range = ((*pat & M_MASK) == M_NOT)) != BG_EOS) |
72b16652 |
899 | ++pat; |
900 | while (((c = *pat++) & M_MASK) != M_END) |
901 | if ((*pat & M_MASK) == M_RNG) { |
220398a0 |
902 | if (nocase) { |
903 | if (tolower(c) <= tolower(k) && tolower(k) <= tolower(pat[1])) |
904 | ok = 1; |
905 | } else { |
906 | if (c <= k && k <= pat[1]) |
907 | ok = 1; |
908 | } |
72b16652 |
909 | pat += 2; |
220398a0 |
910 | } else if (nocase ? (tolower(c) == tolower(k)) : (c == k)) |
72b16652 |
911 | ok = 1; |
912 | if (ok == negate_range) |
913 | return(0); |
914 | break; |
915 | default: |
220398a0 |
916 | k = *name++; |
917 | if (nocase ? (tolower(k) != tolower(c)) : (k != c)) |
72b16652 |
918 | return(0); |
919 | break; |
920 | } |
921 | } |
3e5d0dec |
922 | return(*name == BG_EOS); |
72b16652 |
923 | } |
924 | |
925 | /* Free allocated data belonging to a glob_t structure. */ |
926 | void |
927 | bsd_globfree(glob_t *pglob) |
928 | { |
929 | register int i; |
930 | register char **pp; |
931 | |
932 | if (pglob->gl_pathv != NULL) { |
933 | pp = pglob->gl_pathv + pglob->gl_offs; |
934 | for (i = pglob->gl_pathc; i--; ++pp) |
935 | if (*pp) |
4f49e16e |
936 | Safefree(*pp); |
937 | Safefree(pglob->gl_pathv); |
b8ef571c |
938 | pglob->gl_pathv = NULL; |
72b16652 |
939 | } |
940 | } |
941 | |
942 | static DIR * |
943 | g_opendir(register Char *str, glob_t *pglob) |
944 | { |
945 | char buf[MAXPATHLEN]; |
946 | |
7369a524 |
947 | if (!*str) { |
948 | #ifdef MACOS_TRADITIONAL |
949 | strcpy(buf, ":"); |
950 | #else |
72b16652 |
951 | strcpy(buf, "."); |
7369a524 |
952 | #endif |
953 | } else { |
b8ef571c |
954 | if (g_Ctoc(str, buf, sizeof(buf))) |
955 | return(NULL); |
7369a524 |
956 | } |
72b16652 |
957 | |
958 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) |
959 | return((*pglob->gl_opendir)(buf)); |
4f49e16e |
960 | else |
961 | return(PerlDir_open(buf)); |
72b16652 |
962 | } |
963 | |
72b16652 |
964 | static int |
965 | g_lstat(register Char *fn, Stat_t *sb, glob_t *pglob) |
966 | { |
967 | char buf[MAXPATHLEN]; |
968 | |
b8ef571c |
969 | if (g_Ctoc(fn, buf, sizeof(buf))) |
970 | return(-1); |
72b16652 |
971 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) |
972 | return((*pglob->gl_lstat)(buf, sb)); |
4f49e16e |
973 | #ifdef HAS_LSTAT |
974 | return(PerlLIO_lstat(buf, sb)); |
975 | #else |
976 | return(PerlLIO_stat(buf, sb)); |
72b16652 |
977 | #endif /* HAS_LSTAT */ |
4f49e16e |
978 | } |
72b16652 |
979 | |
980 | static int |
981 | g_stat(register Char *fn, Stat_t *sb, glob_t *pglob) |
982 | { |
983 | char buf[MAXPATHLEN]; |
984 | |
b8ef571c |
985 | if (g_Ctoc(fn, buf, sizeof(buf))) |
986 | return(-1); |
72b16652 |
987 | if (pglob->gl_flags & GLOB_ALTDIRFUNC) |
988 | return((*pglob->gl_stat)(buf, sb)); |
4f49e16e |
989 | return(PerlLIO_stat(buf, sb)); |
72b16652 |
990 | } |
991 | |
992 | static Char * |
993 | g_strchr(Char *str, int ch) |
994 | { |
995 | do { |
996 | if (*str == ch) |
997 | return (str); |
998 | } while (*str++); |
999 | return (NULL); |
1000 | } |
1001 | |
b8ef571c |
1002 | static int |
1003 | g_Ctoc(register const Char *str, char *buf, STRLEN len) |
72b16652 |
1004 | { |
b8ef571c |
1005 | while (len--) { |
1006 | if ((*buf++ = *str++) == BG_EOS) |
1007 | return (0); |
1008 | } |
1009 | return (1); |
72b16652 |
1010 | } |
1011 | |
1012 | #ifdef GLOB_DEBUG |
1013 | static void |
1014 | qprintf(const char *str, register Char *s) |
1015 | { |
1016 | register Char *p; |
1017 | |
1018 | (void)printf("%s:\n", str); |
1019 | for (p = s; *p; p++) |
1020 | (void)printf("%c", CHAR(*p)); |
1021 | (void)printf("\n"); |
1022 | for (p = s; *p; p++) |
1023 | (void)printf("%c", *p & M_PROTECT ? '"' : ' '); |
1024 | (void)printf("\n"); |
1025 | for (p = s; *p; p++) |
1026 | (void)printf("%c", ismeta(*p) ? '_' : ' '); |
1027 | (void)printf("\n"); |
1028 | } |
1029 | #endif /* GLOB_DEBUG */ |