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