1 /* modified from dos_glob.c to work with wchar_t */
2
3 /*
4 * Copyright (c) 1989, 1993
5 * The Regents of the University of California. All rights reserved.
6 *
7 * This code is derived from software contributed to Berkeley by
8 * Guido van Rossum.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 * notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 * notice, this list of conditions and the following disclaimer in the
17 * documentation and/or other materials provided with the distribution.
18 * 3. Neither the name of the University nor the names of its contributors
19 * may be used to endorse or promote products derived from this software
20 * without specific prior written permission.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * SUCH DAMAGE.
33 */
34
35 #if defined(LIBC_SCCS) && !defined(lint)
36 static char sccsid[] = "@(#)glob.c 8.3 (Berkeley) 10/13/93";
37 /* most changes between the version above and the one below have been ported:
38 static char sscsid[]= "$OpenBSD: glob.c,v 1.8.10.1 2001/04/10 jason Exp $";
39 */
40 #endif /* LIBC_SCCS and not lint */
41
42 /*
43 * glob(3) -- a superset of the one defined in POSIX 1003.2.
44 *
45 * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
46 *
47 * Optional extra services, controlled by flags not defined by POSIX:
48 *
49 * GLOB_QUOTE:
50 * Escaping convention: \ inhibits any special meaning the following
51 * character might have (except \ at end of string is retained).
52 * GLOB_MAGCHAR:
53 * Set in gl_flags if pattern contained a globbing character.
54 * GLOB_NOMAGIC:
55 * Same as GLOB_NOCHECK, but it will only append pattern if it did
56 * not contain any magic characters. [Used in csh style globbing]
57 * GLOB_ALTDIRFUNC:
58 * Use alternately specified directory access functions.
59 * GLOB_TILDE:
60 * expand ~user/foo to the /home/dir/of/user/foo
61 * GLOB_BRACE:
62 * expand {1,2}{a,b} to 1a 1b 2a 2b
63 * gl_matchc:
64 * Number of matches in the current invocation of glob.
65 * GLOB_ALPHASORT:
66 * sort alphabetically like csh (case doesn't matter) instead of in ASCII
67 * order
68 */
69
70 #include <wchar.h>
71 #include "dos_wglob.h"
72
73 //#define GLOB_DEBUG
74
75 #ifdef GLOB_DEBUG
76 void Rprintf(const char *, ...);
77 #endif
78
79 #define MAXPATHLEN 255
80 #define DOSISH
81 #define ARG_MAX 14500
82
83
84 #define BG_DOLLAR L'$'
85 #define BG_DOT L'.'
86 #define BG_EOS L'\0'
87 #define BG_LBRACKET L'['
88 #define BG_NOT L'!'
89 #define BG_QUESTION L'?'
90 #define BG_QUOTE L'\\'
91 #define BG_RANGE L'-'
92 #define BG_RBRACKET L']'
93 #define BG_SEP L'/'
94 #ifdef DOSISH /* true, cannot be set to false anymore */
95 #define BG_SEP2 L'\\'
96 #endif
97 #define BG_STAR L'*'
98 #define BG_TILDE L'~'
99 #define BG_UNDERSCORE L'_'
100 #define BG_LBRACE L'{'
101 #define BG_RBRACE L'}'
102 #define BG_SLASH L'/'
103 #define BG_COMMA L','
104
105
106 #include <stdlib.h>
107 #include <dirent.h>
108 #include <sys/types.h>
109 #include <sys/stat.h>
110 #include <string.h>
111 #include <wctype.h>
112
113
114 typedef size_t STRLEN;
115 typedef struct _stat Stat_t;
116 typedef struct _wdirent Direntry_t;
117
118
119 static int compare(const void *, const void *);
120 static int ci_compare(const void *, const void *);
121 static int g_Ctoc(const wchar_t *, wchar_t *, STRLEN);
122 static int g_lstat(wchar_t *, Stat_t *, wglob_t *);
123 static _WDIR *g_opendir(wchar_t *, wglob_t *);
124 static const wchar_t *
125 g_strchr(const wchar_t *, int);
126 static int glob0(const wchar_t *, wglob_t *);
127 static int glob1(wchar_t *, wchar_t *, wglob_t *, size_t *);
128 static int glob2(wchar_t *, wchar_t *, wchar_t *, wchar_t *, wchar_t *, wchar_t *,
129 wglob_t *, size_t *);
130 static int glob3(wchar_t *, wchar_t *, wchar_t *, wchar_t *, wchar_t *, wchar_t *,
131 wchar_t *, wchar_t *, wglob_t *, size_t *);
132 static int globextend(const wchar_t *, wglob_t *, size_t *);
133 static const wchar_t *
134 globtilde(const wchar_t *, wchar_t *, size_t, wglob_t *);
135 static int globexp1(const wchar_t *, wglob_t *);
136 static int globexp2(const wchar_t *, const wchar_t *, wglob_t *, int *);
137 static int match(wchar_t *, wchar_t *, wchar_t *, int);
138 #ifdef GLOB_DEBUG
139 static void qprintf(const char *, wchar_t *);
140 #endif /* GLOB_DEBUG */
141
142 /*
143 Protected and meta characters are from Unicode Private Use Area.
144 The DOS version set the upper bit of bytes for meta characters.
145 An earlier wchar_t version set the top bits of wchar_t.
146 */
147
148 #define M_MASK 0xffff /* pointless in current setup */
149
150 #define P_LBRACKET (wchar_t) 0xfdd0
151 #define P_RBRACKET (wchar_t) 0xfdd1
152 #define P_RANGE (wchar_t) 0xfdd2
153 #define P_LBRACE (wchar_t) 0xfdd3
154 #define P_RBRACE (wchar_t) 0xfdd4
155 #define P_TILDE (wchar_t) 0xfdd5
156 #define P_QUOTE (wchar_t) 0xfdd6
157 #define M_ALL (wchar_t) 0xfdd7
158 #define M_END (wchar_t) 0xfdd8
159 #define M_NOT (wchar_t) 0xfdd9
160 #define M_ONE (wchar_t) 0xfdda
161 #define M_RNG (wchar_t) 0xfddb
162 #define M_SET (wchar_t) 0xfddc
163
WC_PROTECT(wchar_t c)164 static wchar_t WC_PROTECT(wchar_t c)
165 {
166 /* []-{}~\ */
167
168 switch(c) {
169 case BG_LBRACKET: return P_LBRACKET;
170 case BG_RBRACKET: return P_RBRACKET;
171 case BG_RANGE: return P_RANGE;
172 case BG_LBRACE: return P_LBRACE;
173 case BG_RBRACE: return P_RBRACE;
174 case BG_TILDE: return P_TILDE;
175 case BG_QUOTE: return P_QUOTE;
176 default:
177 /* not reachable */
178 return c;
179 }
180 }
181
ismeta(wchar_t c)182 static int ismeta(wchar_t c)
183 {
184 switch(c) {
185 case M_ALL:
186 case M_END:
187 case M_NOT:
188 case M_ONE:
189 case M_RNG:
190 case M_SET:
191 return 1;
192 default:
193 return 0;
194 }
195 }
196
CHAR(wchar_t c)197 static wchar_t CHAR(wchar_t c)
198 {
199 switch(c) {
200 case P_LBRACKET: return BG_LBRACKET;
201 case P_RBRACKET: return BG_RBRACKET;
202 case P_RANGE: return BG_RANGE;
203 case P_LBRACE: return BG_LBRACE;
204 case P_RBRACE: return BG_RBRACE;
205 case P_TILDE: return BG_TILDE;
206 case P_QUOTE: return BG_QUOTE;
207 case M_ALL: return BG_STAR;
208 case M_END: return BG_RBRACKET;
209 case M_NOT: return BG_NOT;
210 case M_ONE: return BG_QUESTION;
211 case M_RNG: return BG_RANGE;
212 case M_SET: return BG_LBRACKET;
213 default:
214 return c;
215 }
216 }
217
218 int
dos_wglob(const wchar_t * pattern,int flags,int (* errfunc)(const wchar_t *,int),wglob_t * pglob)219 dos_wglob(const wchar_t *pattern, int flags,
220 int (*errfunc)(const wchar_t *, int), wglob_t *pglob)
221 {
222 const wchar_t *patnext;
223 int c;
224 wchar_t *bufnext, *bufend, patbuf[MAXPATHLEN];
225
226 patnext = pattern;
227 #if 1
228 if (!(flags & GLOB_APPEND)) {
229 pglob->gl_pathc = 0;
230 pglob->gl_pathv = NULL;
231 if (!(flags & GLOB_DOOFFS))
232 pglob->gl_offs = 0;
233 }
234 #else
235 pglob->gl_pathc = 0;
236 pglob->gl_pathv = NULL;
237 pglob->gl_offs = 0;
238 #endif
239 pglob->gl_flags = flags & ~GLOB_MAGCHAR;
240 pglob->gl_errfunc = errfunc;
241 pglob->gl_matchc = 0;
242
243 bufnext = patbuf;
244 bufend = bufnext + MAXPATHLEN - 1;
245 #ifdef DOSISH /* true */
246 /* Nasty hack to treat patterns like "C:*" correctly. In this
247 * case, the * should match any file in the current directory
248 * on the C: drive. However, the glob code does not treat the
249 * colon specially, so it looks for files beginning "C:" in
250 * the current directory. To fix this, change the pattern to
251 * add an explicit "./" at the start (just after the drive
252 * letter and colon - ie change to "C:./").
253 */
254 if (iswalpha(pattern[0]) && pattern[1] == L':' &&
255 pattern[2] != BG_SEP && pattern[2] != BG_SEP2 &&
256 bufend - bufnext > 4) {
257 *bufnext++ = pattern[0];
258 *bufnext++ = L':';
259 *bufnext++ = L'.';
260 *bufnext++ = BG_SEP;
261 patnext += 2;
262 }
263
264 /* Hack from Tony Plate to allow UNC network drive specification:
265 * Without this code, '\\' (i.e., literally two backslashes inpattern)
266 * at the beginning of a path is not recognized as a network drive,
267 * because the GLOB_QUOTE loop below changes the two backslashes to one.
268 * So, in the case where there are two but not three backslashes at
269 * the beginning of the path, transfer these to the output.
270 */
271 if (patnext == pattern && bufend - bufnext > 2 &&
272 pattern[0] == BG_SEP2 && pattern[1] == BG_SEP2 &&
273 pattern[2] != BG_SEP2) {
274 *bufnext++ = pattern[0];
275 *bufnext++ = pattern[1];
276 patnext += 2;
277 }
278 #endif
279
280 if (flags & GLOB_QUOTE) {
281 /* Protect the quoted characters. */
282 while (bufnext < bufend && (c = *patnext++) != BG_EOS)
283 if (c == BG_QUOTE) {
284 #ifdef DOSISH /* true */
285 /* To avoid backslashitis on Win32,
286 * we only treat \ as a quoting character
287 * if it precedes one of the
288 * metacharacters []-{}~\
289 */
290 if ((c = *patnext++) != L'[' && c != L']' &&
291 c != L'-' && c != L'{' && c != L'}' &&
292 c != L'~' && c != L'\\') {
293 /* WC_PROTECT has to support all characters above */
294 #else
295 # error DOSISH must be true
296 if ((c = *patnext++) == BG_EOS) {
297 #endif
298 c = BG_QUOTE;
299 --patnext;
300 }
301 *bufnext++ = WC_PROTECT(c);
302 } else
303 *bufnext++ = c;
304 } else
305 while (bufnext < bufend && (c = *patnext++) != BG_EOS)
306 *bufnext++ = c;
307 *bufnext = BG_EOS;
308
309 if (flags & GLOB_BRACE)
310 return globexp1(patbuf, pglob);
311 else
312 return glob0(patbuf, pglob);
313 }
314
315 /*
316 * Expand recursively a glob {} pattern. When there is no more expansion
317 * invoke the standard globbing routine to glob the rest of the magic
318 * characters
319 */
320 static int
321 globexp1(const wchar_t *pattern, wglob_t *pglob)
322 {
323 const wchar_t* ptr = pattern;
324 int rv;
325
326 /* Protect a single {}, for find(1), like csh */
327 if (pattern[0] == BG_LBRACE && pattern[1] == BG_RBRACE && pattern[2] == BG_EOS)
328 return glob0(pattern, pglob);
329
330 while ((ptr = (const wchar_t *) g_strchr(ptr, BG_LBRACE)) != NULL)
331 if (!globexp2(ptr, pattern, pglob, &rv))
332 return rv;
333
334 return glob0(pattern, pglob);
335 }
336
337
338 /*
339 * Recursive brace globbing helper. Tries to expand a single brace.
340 * If it succeeds then it invokes globexp1 with the new pattern.
341 * If it fails then it tries to glob the rest of the pattern and returns.
342 */
343 static int
344 globexp2(const wchar_t *ptr, const wchar_t *pattern,
345 wglob_t *pglob, int *rv)
346 {
347 int i;
348 wchar_t *lm, *ls;
349 const wchar_t *pe, *pm, *pl;
350 wchar_t patbuf[MAXPATHLEN];
351
352 /* copy part up to the brace */
353 for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
354 ;
355 *lm = BG_EOS;
356 ls = lm;
357
358 /* Find the balanced brace */
359 for (i = 0, pe = ++ptr; *pe; pe++)
360 if (*pe == BG_LBRACKET) {
361 /* Ignore everything between [] */
362 for (pm = pe++; *pe != BG_RBRACKET && *pe != BG_EOS; pe++)
363 ;
364 if (*pe == BG_EOS) {
365 /*
366 * We could not find a matching BG_RBRACKET.
367 * Ignore and just look for BG_RBRACE
368 */
369 pe = pm;
370 }
371 } else if (*pe == BG_LBRACE)
372 i++;
373 else if (*pe == BG_RBRACE) {
374 if (i == 0)
375 break;
376 i--;
377 }
378
379 /* Non matching braces; just glob the pattern */
380 if (i != 0 || *pe == BG_EOS) {
381 *rv = glob0(patbuf, pglob);
382 return 0;
383 }
384
385 for (i = 0, pl = pm = ptr; pm <= pe; pm++) {
386 switch (*pm) {
387 case BG_LBRACKET:
388 /* Ignore everything between [] */
389 for (pl = pm++; *pm != BG_RBRACKET && *pm != BG_EOS; pm++)
390 ;
391 if (*pm == BG_EOS) {
392 /*
393 * We could not find a matching BG_RBRACKET.
394 * Ignore and just look for BG_RBRACE
395 */
396 pm = pl;
397 }
398 break;
399
400 case BG_LBRACE:
401 i++;
402 break;
403
404 case BG_RBRACE:
405 if (i) {
406 i--;
407 break;
408 }
409 /* FALLTHROUGH */
410 case BG_COMMA:
411 if (i && *pm == BG_COMMA)
412 break;
413 else {
414 /* Append the current string */
415 for (lm = ls; (pl < pm); *lm++ = *pl++)
416 ;
417
418 /*
419 * Append the rest of the pattern after the
420 * closing brace
421 */
422 for (pl = pe + 1; (*lm++ = *pl++) != BG_EOS; )
423 ;
424
425 /* Expand the current pattern */
426 #ifdef GLOB_DEBUG
427 qprintf("globexp2:", patbuf);
428 #endif /* GLOB_DEBUG */
429 *rv = globexp1(patbuf, pglob);
430
431 /* move after the comma, to the next string */
432 pl = pm + 1;
433 }
434 break;
435
436 default:
437 break;
438 }
439 }
440 *rv = 0;
441 return 0;
442 }
443
444
445
446 /*
447 * expand tilde from the passwd file: not supported.
448 */
449 static const wchar_t *
450 globtilde(const wchar_t *pattern, wchar_t *patbuf, size_t patbuf_len, wglob_t *pglob)
451 {
452 wchar_t *h;
453 const wchar_t *p;
454 wchar_t *b, *eb;
455
456 if (*pattern != BG_TILDE || !(pglob->gl_flags & GLOB_TILDE))
457 return pattern;
458
459 /* Copy up to the end of the string or / */
460 eb = &patbuf[patbuf_len - 1];
461 for (p = pattern + 1, h = (wchar_t *) patbuf;
462 h < eb && *p && *p != BG_SLASH; *h++ = *p++)
463 ;
464
465 *h = BG_EOS;
466
467 if (((wchar_t *) patbuf)[0] == BG_EOS) {
468 /*
469 * handle a plain ~ or ~/ by expanding $HOME
470 * first and then trying the password file
471 */
472 if ((h = _wgetenv(L"R_USER")) == NULL) {
473 return pattern;
474 }
475 } else {
476 /*
477 * Expand a ~user
478 */
479 return pattern;
480 }
481
482 /* Copy the home directory */
483 for (b = patbuf; b < eb && *h; *b++ = *h++)
484 ;
485
486 /* Append the rest of the pattern */
487 while (b < eb && (*b++ = *p++) != BG_EOS)
488 ;
489 *b = BG_EOS;
490
491 return patbuf;
492 }
493
494
495 /*
496 * The main glob() routine: compiles the pattern (optionally processing
497 * quotes), calls glob1() to do the real pattern matching, and finally
498 * sorts the list (unless unsorted operation is requested). Returns 0
499 * if things went well, nonzero if errors occurred. It is not an error
500 * to find no matches.
501 */
502 static int
503 glob0(const wchar_t *pattern, wglob_t *pglob)
504 {
505 const wchar_t *qpat, *qpatnext;
506 int c, err, oldflags, oldpathc;
507 wchar_t *bufnext, patbuf[MAXPATHLEN];
508 size_t limit = 0;
509
510 qpat = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
511 qpatnext = qpat;
512 oldflags = pglob->gl_flags;
513 oldpathc = pglob->gl_pathc;
514 bufnext = patbuf;
515
516 /* We don't need to check for buffer overflow any more. */
517 while ((c = *qpatnext++) != BG_EOS) {
518 switch (c) {
519 case BG_LBRACKET:
520 c = *qpatnext;
521 if (c == BG_NOT)
522 ++qpatnext;
523 if (*qpatnext == BG_EOS ||
524 g_strchr((wchar_t *) qpatnext+1, BG_RBRACKET) == NULL) {
525 *bufnext++ = BG_LBRACKET;
526 if (c == BG_NOT)
527 --qpatnext;
528 break;
529 }
530 *bufnext++ = M_SET;
531 if (c == BG_NOT)
532 *bufnext++ = M_NOT;
533 c = *qpatnext++;
534 do {
535 *bufnext++ = CHAR(c);
536 if (*qpatnext == BG_RANGE &&
537 (c = qpatnext[1]) != BG_RBRACKET) {
538 *bufnext++ = M_RNG;
539 *bufnext++ = CHAR(c);
540 qpatnext += 2;
541 }
542 } while ((c = *qpatnext++) != BG_RBRACKET);
543 pglob->gl_flags |= GLOB_MAGCHAR;
544 *bufnext++ = M_END;
545 break;
546 case BG_QUESTION:
547 pglob->gl_flags |= GLOB_MAGCHAR;
548 *bufnext++ = M_ONE;
549 break;
550 case BG_STAR:
551 pglob->gl_flags |= GLOB_MAGCHAR;
552 /* collapse adjacent stars to one,
553 * to avoid exponential behavior
554 */
555 if (bufnext == patbuf || bufnext[-1] != M_ALL)
556 *bufnext++ = M_ALL;
557 break;
558 default:
559 *bufnext++ = CHAR(c);
560 break;
561 }
562 }
563 *bufnext = BG_EOS;
564 #ifdef GLOB_DEBUG
565 qprintf("glob0:", patbuf);
566 #endif /* GLOB_DEBUG */
567
568 if ((err = glob1(patbuf, patbuf+MAXPATHLEN-1, pglob, &limit)) != 0) {
569 pglob->gl_flags = oldflags;
570 return(err);
571 }
572
573 /*
574 * If there was no match we are going to append the pattern
575 * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
576 * and the pattern did not contain any magic characters
577 * GLOB_NOMAGIC is there just for compatibility with csh.
578 */
579 if (pglob->gl_pathc == oldpathc &&
580 ((pglob->gl_flags & GLOB_NOCHECK) ||
581 ((pglob->gl_flags & GLOB_NOMAGIC) &&
582 !(pglob->gl_flags & GLOB_MAGCHAR))))
583 {
584 #ifdef GLOB_DEBUG
585 Rprintf("calling globextend from glob0\n");
586 #endif /* GLOB_DEBUG */
587 pglob->gl_flags = oldflags;
588 return(globextend(qpat, pglob, &limit));
589 }
590 else if (!(pglob->gl_flags & GLOB_NOSORT))
591 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
592 pglob->gl_pathc - oldpathc, sizeof(wchar_t *),
593 (pglob->gl_flags & (GLOB_ALPHASORT|GLOB_NOCASE))
594 ? ci_compare : compare);
595 pglob->gl_flags = oldflags;
596 return(0);
597 }
598
599 static int
600 ci_compare(const void *p, const void *q)
601 {
602 const wchar_t *pp = *(const wchar_t **)p;
603 const wchar_t *qq = *(const wchar_t **)q;
604 int ci;
605 while (*pp && *qq) {
606 if (towlower(*pp) != towlower(*qq))
607 break;
608 ++pp;
609 ++qq;
610 }
611 ci = towlower(*pp) - towlower(*qq);
612 if (ci == 0)
613 return compare(p, q);
614 return ci;
615 }
616
617 static int
618 compare(const void *p, const void *q)
619 {
620 return(wcscmp(*(wchar_t **)p, *(wchar_t **)q));
621 }
622
623 static int
624 glob1(wchar_t *pattern, wchar_t *pattern_last, wglob_t *pglob, size_t *limitp)
625 {
626 wchar_t pathbuf[MAXPATHLEN];
627
628 /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
629 if (*pattern == BG_EOS) return(0);
630 return(glob2(pathbuf, pathbuf+MAXPATHLEN-1,
631 pathbuf, pathbuf+MAXPATHLEN-1,
632 pattern, pattern_last, pglob, limitp));
633 }
634
635 /*
636 * The functions glob2 and glob3 are mutually recursive; there is one level
637 * of recursion for each segment in the pattern that contains one or more
638 * meta characters.
639 */
640 static int
641 glob2(wchar_t *pathbuf, wchar_t *pathbuf_last, wchar_t *pathend, wchar_t *pathend_last,
642 wchar_t *pattern, wchar_t *pattern_last, wglob_t *pglob, size_t *limitp)
643 {
644 Stat_t sb;
645 wchar_t *p, *q;
646 int anymeta;
647
648 /*
649 * Loop over pattern segments until end of pattern or until
650 * segment with meta character found.
651 */
652 for (anymeta = 0;;) {
653 if (*pattern == BG_EOS) { /* End of pattern? */
654 *pathend = BG_EOS;
655 if (g_lstat(pathbuf, &sb, pglob)) return(0);
656
657 if (((pglob->gl_flags & GLOB_MARK) &&
658 pathend[-1] != BG_SEP
659 #ifdef DOSISH /* true */
660 && pathend[-1] != BG_SEP2
661 #endif
662 ) && S_ISDIR(sb.st_mode) ) {
663 if (pathend+1 > pathend_last)
664 return (1);
665 *pathend++ = BG_SEP;
666 *pathend = BG_EOS;
667 }
668 ++pglob->gl_matchc;
669 #ifdef GLOB_DEBUG
670 Rprintf("calling globextend from glob2\n");
671 #endif /* GLOB_DEBUG */
672 return(globextend(pathbuf, pglob, limitp));
673 }
674
675 /* Find end of next segment, copy tentatively to pathend. */
676 q = pathend;
677 p = pattern;
678 while (*p != BG_EOS && *p != BG_SEP
679 #ifdef DOSISH /* true */
680 && *p != BG_SEP2
681 #endif
682 ) {
683 if (ismeta(*p)) anymeta = 1;
684 if (q+1 > pathend_last) return (1);
685 *q++ = *p++;
686 }
687
688 if (!anymeta) { /* No expansion, do next segment. */
689 pathend = q;
690 pattern = p;
691 while (*pattern == BG_SEP
692 #ifdef DOSISH /* true */
693 || *pattern == BG_SEP2
694 #endif
695 ) {
696 if (pathend+1 > pathend_last) return (1);
697 *pathend++ = *pattern++;
698 }
699 } else
700 /* Need expansion, recurse. */
701 return(glob3(pathbuf, pathbuf_last, pathend,
702 pathend_last, pattern, pattern_last,
703 p, pattern_last, pglob, limitp));
704 }
705 /* NOTREACHED */
706 }
707
708 static int
709 glob3(wchar_t *pathbuf, wchar_t *pathbuf_last, wchar_t *pathend, wchar_t *pathend_last,
710 wchar_t *pattern, wchar_t *pattern_last,
711 wchar_t *restpattern, wchar_t *restpattern_last, wglob_t *pglob, size_t *limitp)
712 {
713 Direntry_t *dp;
714 _WDIR *dirp;
715 int err;
716 int nocase;
717 wchar_t buf[MAXPATHLEN];
718
719 if (pathend > pathend_last)
720 return (1);
721 *pathend = BG_EOS;
722 errno = 0;
723
724 if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
725 /* TODO: don't call for ENOENT or ENOTDIR? */
726 if (pglob->gl_errfunc) {
727 if (g_Ctoc(pathbuf, buf, sizeof(buf)))
728 return (GLOB_ABEND);
729 if (pglob->gl_errfunc(buf, errno) ||
730 (pglob->gl_flags & GLOB_ERR))
731 return (GLOB_ABEND);
732 }
733 return(0);
734 }
735
736 err = 0;
737 nocase = ((pglob->gl_flags & GLOB_NOCASE) != 0);
738
739 /* Search directory for matching names. */
740 while ((dp = _wreaddir(dirp))) {
741 wchar_t *sc, *dc;
742
743 /* Initial BG_DOT must be matched literally. */
744 if (dp->d_name[0] == BG_DOT && *pattern != BG_DOT)
745 continue;
746 dc = pathend;
747 sc = dp->d_name;
748 while (dc < pathend_last && (*dc++ = *sc++) != BG_EOS)
749 ;
750 if (dc >= pathend_last) {
751 *dc = BG_EOS;
752 err = 1;
753 break;
754 }
755
756 if (!match(pathend, pattern, restpattern, nocase)) {
757 *pathend = BG_EOS;
758 continue;
759 }
760 err = glob2(pathbuf, pathbuf_last, --dc, pathend_last,
761 restpattern, restpattern_last, pglob, limitp);
762 if (err)
763 break;
764 }
765
766 _wclosedir(dirp);
767 return(err);
768 }
769
770
771 #include <R_ext/RS.h> /* for Calloc, Realloc, Free */
772
773 /*
774 * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
775 * add the new item, and update gl_pathc.
776 *
777 * This assumes the BSD realloc, which only copies the block when its size
778 * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
779 * behavior.
780 *
781 * Return 0 if new item added, error code if memory couldn't be allocated.
782 *
783 * Invariant of the glob_t structure:
784 * Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
785 * gl_pathv points to (gl_offs + gl_pathc + 1) items.
786 */
787 static int
788 globextend(const wchar_t *path, wglob_t *pglob, size_t *limitp)
789 {
790 wchar_t **pathv;
791 int i;
792 STRLEN newsize, len;
793 wchar_t *copy;
794 const wchar_t *p;
795
796 #ifdef GLOB_DEBUG
797 Rprintf("Adding ");
798 for (p = path; *p; p++)
799 (void)Rprintf("%c", CHAR(*p));
800 Rprintf("\n");
801 #endif /* GLOB_DEBUG */
802
803 newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
804 if (pglob->gl_pathv)
805 pathv = Realloc(pglob->gl_pathv, newsize, wchar_t *);
806 else
807 pathv = Calloc(newsize, wchar_t *);
808 if (pathv == NULL) {
809 if (pglob->gl_pathv) {
810 Free(pglob->gl_pathv);
811 pglob->gl_pathv = NULL;
812 }
813 return(GLOB_NOSPACE);
814 }
815
816 if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
817 /* first time around -- clear initial gl_offs items */
818 pathv += pglob->gl_offs;
819 for (i = pglob->gl_offs; --i >= 0; )
820 *--pathv = NULL;
821 }
822 pglob->gl_pathv = pathv;
823
824 for (p = path; *p++;)
825 ;
826 len = (STRLEN)(p - path);
827 *limitp += len;
828 copy = Calloc(p-path, wchar_t);
829 if (copy != NULL) {
830 if (g_Ctoc(path, copy, len)) {
831 Free(copy);
832 return(GLOB_NOSPACE);
833 }
834 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
835 }
836 pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
837
838 if ((pglob->gl_flags & GLOB_LIMIT) &&
839 newsize + *limitp >= ARG_MAX) {
840 errno = 0;
841 return(GLOB_NOSPACE);
842 }
843
844 return(copy == NULL ? GLOB_NOSPACE : 0);
845 }
846
847
848 /*
849 * pattern matching function for filenames. Each occurrence of the *
850 * pattern causes a recursion level.
851 */
852 static int
853 match(wchar_t *name, wchar_t *pat, wchar_t *patend, int nocase)
854 {
855 int ok, negate_range;
856 wchar_t c, k;
857
858 while (pat < patend) {
859 c = *pat++;
860 switch (c & M_MASK) {
861 case M_ALL:
862 if (pat == patend)
863 return(1);
864 do
865 if (match(name, pat, patend, nocase))
866 return(1);
867 while (*name++ != BG_EOS)
868 ;
869 return(0);
870 case M_ONE:
871 if (*name++ == BG_EOS)
872 return(0);
873 break;
874 case M_SET:
875 ok = 0;
876 if ((k = *name++) == BG_EOS)
877 return(0);
878 if ((negate_range = ((*pat & M_MASK) == M_NOT)) != BG_EOS)
879 ++pat;
880 while (((c = *pat++) & M_MASK) != M_END)
881 if ((*pat & M_MASK) == M_RNG) {
882 if (nocase) {
883 if (towlower(c) <= towlower(k) && towlower(k) <= towlower(pat[1]))
884 ok = 1;
885 } else {
886 if (c <= k && k <= pat[1])
887 ok = 1;
888 }
889 pat += 2;
890 } else if (nocase ? (towlower(c) == towlower(k)) : (c == k))
891 ok = 1;
892 if (ok == negate_range)
893 return(0);
894 break;
895 default:
896 k = *name++;
897 if (nocase ? (towlower(k) != towlower(c)) : (k != c))
898 return(0);
899 break;
900 }
901 }
902 return(*name == BG_EOS);
903 }
904
905 /* Free allocated data belonging to a wglob_t structure. */
906 void
907 dos_wglobfree(wglob_t *pglob)
908 {
909 int i;
910 wchar_t **pp;
911
912 if (pglob->gl_pathv != NULL) {
913 pp = pglob->gl_pathv + pglob->gl_offs;
914 for (i = pglob->gl_pathc; i--; ++pp)
915 if (*pp)
916 Free(*pp);
917 Free(pglob->gl_pathv);
918 pglob->gl_pathv = NULL;
919 }
920 }
921
922 static _WDIR *
923 g_opendir(wchar_t *str, wglob_t *pglob)
924 {
925 wchar_t buf[MAXPATHLEN];
926
927 if (!*str) wcscpy(buf, L".");
928 else
929 if (g_Ctoc(str, buf, sizeof(buf))) return(NULL);
930 return _wopendir(buf);
931 }
932
933 static int
934 g_lstat(wchar_t *fn, Stat_t *sb, wglob_t *pglob)
935 {
936 wchar_t buf[MAXPATHLEN];
937
938 if (g_Ctoc(fn, buf, sizeof(buf)))
939 return(-1);
940 return(_wstat(buf, sb));
941 }
942
943 static const wchar_t *
944 g_strchr(const wchar_t *str, int ch)
945 {
946 do {
947 if (*str == ch)
948 return (str);
949 } while (*str++);
950 return (NULL);
951 }
952
953 static int
954 g_Ctoc(const wchar_t *str, wchar_t *buf, STRLEN len)
955 {
956 while (len--)
957 if ((*buf++ = *str++) == BG_EOS) return 0;
958 return 1;
959 }
960
961 #ifdef GLOB_DEBUG
962 static void
963 qprintf(const char *str, wchar_t *s)
964 {
965 wchar_t *p;
966
967 (void)Rprintf("%s:\n", str);
968 for (p = s; *p; p++)
969 (void)Rprintf("%lc", CHAR(*p));
970 (void)Rprintf("\n");
971 #if 0
972 for (p = s; *p; p++)
973 (void)Rprintf("%lc", *p & M_PROTECT ? L'"' : L' ');
974 (void)Rprintf("\n");
975 #endif
976 for (p = s; *p; p++)
977 (void)Rprintf("%lc", ismeta(*p) ? L'_' : L' ');
978 (void)Rprintf("\n");
979 }
980 #endif /* GLOB_DEBUG */
981