This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Export Winsock error constants from POSIX.pm
[perl5.git] / ext / File-Glob / bsd_glob.c
CommitLineData
72b16652
GS
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
GS
17 * may be used to endorse or promote products derived from this software
18 * without specific prior written permission.
19 *
882ce583 20 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS" AND
72b16652
GS
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
GS
33#if defined(LIBC_SCCS) && !defined(lint)
34static char sccsid[] = "@(#)glob.c 8.3 (Berkeley) 10/13/93";
bac331f5
JH
35/* most changes between the version above and the one below have been ported:
36static char sscsid[]= "$OpenBSD: glob.c,v 1.8.10.1 2001/04/10 jason Exp $";
37 */
72b16652
GS
38#endif /* LIBC_SCCS and not lint */
39
40/*
41 * glob(3) -- a superset of the one defined in POSIX 1003.2.
42 *
43 * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
44 *
45 * Optional extra services, controlled by flags not defined by POSIX:
46 *
47 * GLOB_QUOTE:
48 * Escaping convention: \ inhibits any special meaning the following
49 * character might have (except \ at end of string is retained).
50 * GLOB_MAGCHAR:
51 * Set in gl_flags if pattern contained a globbing character.
52 * GLOB_NOMAGIC:
53 * Same as GLOB_NOCHECK, but it will only append pattern if it did
54 * not contain any magic characters. [Used in csh style globbing]
55 * GLOB_ALTDIRFUNC:
56 * Use alternately specified directory access functions.
57 * GLOB_TILDE:
58 * expand ~user/foo to the /home/dir/of/user/foo
59 * GLOB_BRACE:
60 * expand {1,2}{a,b} to 1a 1b 2a 2b
61 * gl_matchc:
62 * Number of matches in the current invocation of glob.
2d5e9e5d
JH
63 * GLOB_ALPHASORT:
64 * sort alphabetically like csh (case doesn't matter) instead of in ASCII
65 * order
72b16652
GS
66 */
67
68#include <EXTERN.h>
69#include <perl.h>
4f49e16e
GS
70#include <XSUB.h>
71
72b16652
GS
72#include "bsd_glob.h"
73#ifdef I_PWD
74# include <pwd.h>
75#else
fd8cd3a3 76#if defined(HAS_PASSWD) && !defined(VMS)
72b16652
GS
77 struct passwd *getpwnam(char *);
78 struct passwd *getpwuid(Uid_t);
79#endif
80#endif
81
82#ifndef MAXPATHLEN
83# ifdef PATH_MAX
84# define MAXPATHLEN PATH_MAX
f2100097 85# else
a9bada39 86# define MAXPATHLEN 1024
72b16652
GS
87# endif
88#endif
89
b8ef571c
JH
90#ifdef I_LIMITS
91#include <limits.h>
92#endif
93
94#ifndef ARG_MAX
e37778c2
NC
95# ifdef _SC_ARG_MAX
96# define ARG_MAX (sysconf(_SC_ARG_MAX))
b8ef571c 97# else
e37778c2
NC
98# ifdef _POSIX_ARG_MAX
99# define ARG_MAX _POSIX_ARG_MAX
b8ef571c 100# else
e37778c2
NC
101# ifdef WIN32
102# define ARG_MAX 14500 /* from VC's limits.h */
b8ef571c 103# else
e37778c2 104# define ARG_MAX 4096 /* from POSIX, be conservative */
b8ef571c
JH
105# endif
106# endif
107# endif
108#endif
109
3e5d0dec
GS
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 ']'
e37778c2 119#define BG_SEP '/'
220398a0
PM
120#ifdef DOSISH
121#define BG_SEP2 '\\'
122#endif
3e5d0dec
GS
123#define BG_STAR '*'
124#define BG_TILDE '~'
125#define BG_UNDERSCORE '_'
126#define BG_LBRACE '{'
127#define BG_RBRACE '}'
128#define BG_SLASH '/'
129#define BG_COMMA ','
72b16652
GS
130
131#ifndef GLOB_DEBUG
132
133#define M_QUOTE 0x8000
134#define M_PROTECT 0x4000
135#define M_MASK 0xffff
136#define M_ASCII 0x00ff
137
138typedef U16 Char;
139
140#else
141
142#define M_QUOTE 0x80
143#define M_PROTECT 0x40
144#define M_MASK 0xff
145#define M_ASCII 0x7f
146
147typedef U8 Char;
148
149#endif /* !GLOB_DEBUG */
150
151
152#define CHAR(c) ((Char)((c)&M_ASCII))
153#define META(c) ((Char)((c)|M_QUOTE))
154#define M_ALL META('*')
155#define M_END META(']')
156#define M_NOT META('!')
157#define M_ONE META('?')
158#define M_RNG META('-')
159#define M_SET META('[')
160#define ismeta(c) (((c)&M_QUOTE) != 0)
161
162
163static int compare(const void *, const void *);
220398a0 164static int ci_compare(const void *, const void *);
b8ef571c 165static int g_Ctoc(const Char *, char *, STRLEN);
72b16652
GS
166static int g_lstat(Char *, Stat_t *, glob_t *);
167static DIR *g_opendir(Char *, glob_t *);
168static Char *g_strchr(Char *, int);
72b16652
GS
169static int g_stat(Char *, Stat_t *, glob_t *);
170static int glob0(const Char *, glob_t *);
b8ef571c
JH
171static int glob1(Char *, Char *, glob_t *, size_t *);
172static int glob2(Char *, Char *, Char *, Char *, Char *, Char *,
173 glob_t *, size_t *);
46a847d3 174static int glob3(Char *, Char *, Char *, Char *, Char *,
b8ef571c
JH
175 Char *, Char *, glob_t *, size_t *);
176static int globextend(const Char *, glob_t *, size_t *);
177static const Char *
178 globtilde(const Char *, Char *, size_t, glob_t *);
72b16652
GS
179static int globexp1(const Char *, glob_t *);
180static int globexp2(const Char *, const Char *, glob_t *, int *);
220398a0 181static int match(Char *, Char *, Char *, int);
72b16652
GS
182#ifdef GLOB_DEBUG
183static void qprintf(const char *, Char *);
184#endif /* GLOB_DEBUG */
185
4f49e16e
GS
186#ifdef PERL_IMPLICIT_CONTEXT
187static Direntry_t * my_readdir(DIR*);
188
189static Direntry_t *
190my_readdir(DIR *d)
191{
011f1a1a 192#ifndef NETWARE
4f49e16e 193 return PerlDir_read(d);
011f1a1a
JH
194#else
195 return (DIR *)PerlDir_read(d);
196#endif
4f49e16e
GS
197}
198#else
31d9e88b
AB
199
200/* ReliantUNIX (OS formerly known as SINIX) defines readdir
201 * in LFS-mode to be a 64-bit version of readdir. */
202
203# ifdef sinix
204static Direntry_t * my_readdir(DIR*);
205
206static Direntry_t *
207my_readdir(DIR *d)
208{
209 return readdir(d);
210}
211# else
212
213# define my_readdir readdir
214
215# endif
216
4f49e16e
GS
217#endif
218
72b16652
GS
219int
220bsd_glob(const char *pattern, int flags,
221 int (*errfunc)(const char *, int), glob_t *pglob)
222{
223 const U8 *patnext;
224 int c;
b8ef571c 225 Char *bufnext, *bufend, patbuf[MAXPATHLEN];
72b16652 226 patnext = (U8 *) pattern;
e0e8a4dc
MHM
227 /* TODO: GLOB_APPEND / GLOB_DOOFFS aren't supported yet */
228#if 0
72b16652
GS
229 if (!(flags & GLOB_APPEND)) {
230 pglob->gl_pathc = 0;
231 pglob->gl_pathv = NULL;
232 if (!(flags & GLOB_DOOFFS))
233 pglob->gl_offs = 0;
234 }
e0e8a4dc
MHM
235#else
236 pglob->gl_pathc = 0;
237 pglob->gl_pathv = NULL;
238 pglob->gl_offs = 0;
239#endif
72b16652
GS
240 pglob->gl_flags = flags & ~GLOB_MAGCHAR;
241 pglob->gl_errfunc = errfunc;
242 pglob->gl_matchc = 0;
243
244 bufnext = patbuf;
b8ef571c 245 bufend = bufnext + MAXPATHLEN - 1;
220398a0
PM
246#ifdef DOSISH
247 /* Nasty hack to treat patterns like "C:*" correctly. In this
248 * case, the * should match any file in the current directory
249 * on the C: drive. However, the glob code does not treat the
250 * colon specially, so it looks for files beginning "C:" in
251 * the current directory. To fix this, change the pattern to
252 * add an explicit "./" at the start (just after the drive
c6c619a9 253 * letter and colon - ie change to "C:./").
220398a0
PM
254 */
255 if (isalpha(pattern[0]) && pattern[1] == ':' &&
256 pattern[2] != BG_SEP && pattern[2] != BG_SEP2 &&
257 bufend - bufnext > 4) {
258 *bufnext++ = pattern[0];
259 *bufnext++ = ':';
260 *bufnext++ = '.';
261 *bufnext++ = BG_SEP;
262 patnext += 2;
263 }
264#endif
be708cc0 265
72b16652
GS
266 if (flags & GLOB_QUOTE) {
267 /* Protect the quoted characters. */
3e5d0dec
GS
268 while (bufnext < bufend && (c = *patnext++) != BG_EOS)
269 if (c == BG_QUOTE) {
220398a0
PM
270#ifdef DOSISH
271 /* To avoid backslashitis on Win32,
272 * we only treat \ as a quoting character
273 * if it precedes one of the
274 * metacharacters []-{}~\
275 */
276 if ((c = *patnext++) != '[' && c != ']' &&
277 c != '-' && c != '{' && c != '}' &&
278 c != '~' && c != '\\') {
279#else
3e5d0dec 280 if ((c = *patnext++) == BG_EOS) {
220398a0 281#endif
3e5d0dec 282 c = BG_QUOTE;
72b16652
GS
283 --patnext;
284 }
285 *bufnext++ = c | M_PROTECT;
b8ef571c 286 } else
72b16652 287 *bufnext++ = c;
b8ef571c
JH
288 } else
289 while (bufnext < bufend && (c = *patnext++) != BG_EOS)
290 *bufnext++ = c;
3e5d0dec 291 *bufnext = BG_EOS;
72b16652
GS
292
293 if (flags & GLOB_BRACE)
294 return globexp1(patbuf, pglob);
295 else
296 return glob0(patbuf, pglob);
297}
298
299/*
300 * Expand recursively a glob {} pattern. When there is no more expansion
301 * invoke the standard globbing routine to glob the rest of the magic
302 * characters
303 */
b8ef571c
JH
304static int
305globexp1(const Char *pattern, glob_t *pglob)
72b16652
GS
306{
307 const Char* ptr = pattern;
308 int rv;
309
310 /* Protect a single {}, for find(1), like csh */
3e5d0dec 311 if (pattern[0] == BG_LBRACE && pattern[1] == BG_RBRACE && pattern[2] == BG_EOS)
72b16652
GS
312 return glob0(pattern, pglob);
313
3e5d0dec 314 while ((ptr = (const Char *) g_strchr((Char *) ptr, BG_LBRACE)) != NULL)
72b16652
GS
315 if (!globexp2(ptr, pattern, pglob, &rv))
316 return rv;
317
318 return glob0(pattern, pglob);
319}
320
321
322/*
323 * Recursive brace globbing helper. Tries to expand a single brace.
324 * If it succeeds then it invokes globexp1 with the new pattern.
325 * If it fails then it tries to glob the rest of the pattern and returns.
326 */
b8ef571c
JH
327static int
328globexp2(const Char *ptr, const Char *pattern,
329 glob_t *pglob, int *rv)
72b16652
GS
330{
331 int i;
332 Char *lm, *ls;
77348331 333 const Char *pe, *pm, *pm1, *pl;
b8ef571c 334 Char patbuf[MAXPATHLEN];
72b16652
GS
335
336 /* copy part up to the brace */
337 for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
b8ef571c
JH
338 ;
339 *lm = BG_EOS;
72b16652
GS
340 ls = lm;
341
342 /* Find the balanced brace */
343 for (i = 0, pe = ++ptr; *pe; pe++)
3e5d0dec 344 if (*pe == BG_LBRACKET) {
72b16652 345 /* Ignore everything between [] */
3e5d0dec 346 for (pm = pe++; *pe != BG_RBRACKET && *pe != BG_EOS; pe++)
b8ef571c 347 ;
3e5d0dec 348 if (*pe == BG_EOS) {
72b16652 349 /*
3e5d0dec
GS
350 * We could not find a matching BG_RBRACKET.
351 * Ignore and just look for BG_RBRACE
72b16652
GS
352 */
353 pe = pm;
354 }
b8ef571c 355 } else if (*pe == BG_LBRACE)
72b16652 356 i++;
3e5d0dec 357 else if (*pe == BG_RBRACE) {
72b16652
GS
358 if (i == 0)
359 break;
360 i--;
361 }
362
363 /* Non matching braces; just glob the pattern */
3e5d0dec 364 if (i != 0 || *pe == BG_EOS) {
72b16652
GS
365 *rv = glob0(patbuf, pglob);
366 return 0;
367 }
368
b8ef571c 369 for (i = 0, pl = pm = ptr; pm <= pe; pm++) {
72b16652 370 switch (*pm) {
3e5d0dec 371 case BG_LBRACKET:
72b16652 372 /* Ignore everything between [] */
77348331 373 for (pm1 = pm++; *pm != BG_RBRACKET && *pm != BG_EOS; pm++)
b8ef571c 374 ;
3e5d0dec 375 if (*pm == BG_EOS) {
72b16652 376 /*
3e5d0dec
GS
377 * We could not find a matching BG_RBRACKET.
378 * Ignore and just look for BG_RBRACE
72b16652 379 */
77348331 380 pm = pm1;
72b16652
GS
381 }
382 break;
383
3e5d0dec 384 case BG_LBRACE:
72b16652
GS
385 i++;
386 break;
387
3e5d0dec 388 case BG_RBRACE:
72b16652 389 if (i) {
b8ef571c
JH
390 i--;
391 break;
72b16652
GS
392 }
393 /* FALLTHROUGH */
3e5d0dec
GS
394 case BG_COMMA:
395 if (i && *pm == BG_COMMA)
72b16652
GS
396 break;
397 else {
398 /* Append the current string */
399 for (lm = ls; (pl < pm); *lm++ = *pl++)
b8ef571c
JH
400 ;
401
72b16652
GS
402 /*
403 * Append the rest of the pattern after the
404 * closing brace
405 */
b8ef571c
JH
406 for (pl = pe + 1; (*lm++ = *pl++) != BG_EOS; )
407 ;
72b16652
GS
408
409 /* Expand the current pattern */
410#ifdef GLOB_DEBUG
411 qprintf("globexp2:", patbuf);
412#endif /* GLOB_DEBUG */
413 *rv = globexp1(patbuf, pglob);
414
415 /* move after the comma, to the next string */
416 pl = pm + 1;
417 }
418 break;
419
420 default:
421 break;
422 }
b8ef571c 423 }
72b16652
GS
424 *rv = 0;
425 return 0;
426}
427
428
429
430/*
431 * expand tilde from the passwd file.
432 */
433static const Char *
bac331f5 434globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob)
72b16652 435{
72b16652
GS
436 char *h;
437 const Char *p;
bac331f5 438 Char *b, *eb;
72b16652 439
3e5d0dec 440 if (*pattern != BG_TILDE || !(pglob->gl_flags & GLOB_TILDE))
72b16652
GS
441 return pattern;
442
443 /* Copy up to the end of the string or / */
bac331f5
JH
444 eb = &patbuf[patbuf_len - 1];
445 for (p = pattern + 1, h = (char *) patbuf;
7c436af3 446 h < (char*)eb && *p && *p != BG_SLASH; *h++ = (char)*p++)
b8ef571c 447 ;
72b16652 448
3e5d0dec 449 *h = BG_EOS;
72b16652 450
b8ef571c
JH
451#if 0
452 if (h == (char *)eb)
453 return what;
454#endif
455
3e5d0dec 456 if (((char *) patbuf)[0] == BG_EOS) {
72b16652
GS
457 /*
458 * handle a plain ~ or ~/ by expanding $HOME
459 * first and then trying the password file
528bd3ce 460 * or $USERPROFILE on DOSISH systems
72b16652
GS
461 */
462 if ((h = getenv("HOME")) == NULL) {
463#ifdef HAS_PASSWD
91f3b821 464 struct passwd *pwd;
72b16652
GS
465 if ((pwd = getpwuid(getuid())) == NULL)
466 return pattern;
467 else
468 h = pwd->pw_dir;
528bd3ce
DCW
469#elif DOSISH
470 /*
471 * When no passwd file, fallback to the USERPROFILE
472 * environment variable on DOSish systems.
473 */
474 if ((h = getenv("USERPROFILE")) == NULL) {
475 return pattern;
476 }
72b16652
GS
477#else
478 return pattern;
479#endif
480 }
b8ef571c 481 } else {
72b16652
GS
482 /*
483 * Expand a ~user
484 */
485#ifdef HAS_PASSWD
91f3b821 486 struct passwd *pwd;
72b16652
GS
487 if ((pwd = getpwnam((char*) patbuf)) == NULL)
488 return pattern;
489 else
490 h = pwd->pw_dir;
491#else
492 return pattern;
493#endif
494 }
495
496 /* Copy the home directory */
bac331f5 497 for (b = patbuf; b < eb && *h; *b++ = *h++)
b8ef571c 498 ;
72b16652
GS
499
500 /* Append the rest of the pattern */
bac331f5 501 while (b < eb && (*b++ = *p++) != BG_EOS)
b8ef571c 502 ;
bac331f5 503 *b = BG_EOS;
72b16652
GS
504
505 return patbuf;
506}
507
508
509/*
510 * The main glob() routine: compiles the pattern (optionally processing
511 * quotes), calls glob1() to do the real pattern matching, and finally
512 * sorts the list (unless unsorted operation is requested). Returns 0
513 * if things went well, nonzero if errors occurred. It is not an error
514 * to find no matches.
515 */
516static int
517glob0(const Char *pattern, glob_t *pglob)
518{
519 const Char *qpat, *qpatnext;
520 int c, err, oldflags, oldpathc;
b8ef571c
JH
521 Char *bufnext, patbuf[MAXPATHLEN];
522 size_t limit = 0;
72b16652 523
b8ef571c 524 qpat = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
72b16652
GS
525 qpatnext = qpat;
526 oldflags = pglob->gl_flags;
527 oldpathc = pglob->gl_pathc;
528 bufnext = patbuf;
529
530 /* We don't need to check for buffer overflow any more. */
3e5d0dec 531 while ((c = *qpatnext++) != BG_EOS) {
72b16652 532 switch (c) {
3e5d0dec 533 case BG_LBRACKET:
72b16652 534 c = *qpatnext;
3e5d0dec 535 if (c == BG_NOT)
72b16652 536 ++qpatnext;
3e5d0dec
GS
537 if (*qpatnext == BG_EOS ||
538 g_strchr((Char *) qpatnext+1, BG_RBRACKET) == NULL) {
539 *bufnext++ = BG_LBRACKET;
540 if (c == BG_NOT)
72b16652
GS
541 --qpatnext;
542 break;
543 }
544 *bufnext++ = M_SET;
3e5d0dec 545 if (c == BG_NOT)
72b16652
GS
546 *bufnext++ = M_NOT;
547 c = *qpatnext++;
548 do {
549 *bufnext++ = CHAR(c);
3e5d0dec
GS
550 if (*qpatnext == BG_RANGE &&
551 (c = qpatnext[1]) != BG_RBRACKET) {
72b16652
GS
552 *bufnext++ = M_RNG;
553 *bufnext++ = CHAR(c);
554 qpatnext += 2;
555 }
3e5d0dec 556 } while ((c = *qpatnext++) != BG_RBRACKET);
72b16652
GS
557 pglob->gl_flags |= GLOB_MAGCHAR;
558 *bufnext++ = M_END;
559 break;
3e5d0dec 560 case BG_QUESTION:
72b16652
GS
561 pglob->gl_flags |= GLOB_MAGCHAR;
562 *bufnext++ = M_ONE;
563 break;
3e5d0dec 564 case BG_STAR:
72b16652
GS
565 pglob->gl_flags |= GLOB_MAGCHAR;
566 /* collapse adjacent stars to one,
567 * to avoid exponential behavior
568 */
569 if (bufnext == patbuf || bufnext[-1] != M_ALL)
b8ef571c 570 *bufnext++ = M_ALL;
72b16652
GS
571 break;
572 default:
573 *bufnext++ = CHAR(c);
574 break;
575 }
576 }
3e5d0dec 577 *bufnext = BG_EOS;
72b16652
GS
578#ifdef GLOB_DEBUG
579 qprintf("glob0:", patbuf);
580#endif /* GLOB_DEBUG */
581
b8ef571c 582 if ((err = glob1(patbuf, patbuf+MAXPATHLEN-1, pglob, &limit)) != 0) {
72b16652
GS
583 pglob->gl_flags = oldflags;
584 return(err);
585 }
586
587 /*
588 * If there was no match we are going to append the pattern
589 * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
590 * and the pattern did not contain any magic characters
591 * GLOB_NOMAGIC is there just for compatibility with csh.
592 */
593 if (pglob->gl_pathc == oldpathc &&
594 ((pglob->gl_flags & GLOB_NOCHECK) ||
595 ((pglob->gl_flags & GLOB_NOMAGIC) &&
596 !(pglob->gl_flags & GLOB_MAGCHAR))))
597 {
598#ifdef GLOB_DEBUG
599 printf("calling globextend from glob0\n");
600#endif /* GLOB_DEBUG */
601 pglob->gl_flags = oldflags;
b8ef571c 602 return(globextend(qpat, pglob, &limit));
72b16652
GS
603 }
604 else if (!(pglob->gl_flags & GLOB_NOSORT))
a08678bc 605 if (pglob->gl_pathv)
72b16652 606 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
be708cc0 607 pglob->gl_pathc - oldpathc, sizeof(char *),
2d5e9e5d
JH
608 (pglob->gl_flags & (GLOB_ALPHASORT|GLOB_NOCASE))
609 ? ci_compare : compare);
72b16652
GS
610 pglob->gl_flags = oldflags;
611 return(0);
612}
613
614static int
220398a0
PM
615ci_compare(const void *p, const void *q)
616{
b8ef571c
JH
617 const char *pp = *(const char **)p;
618 const char *qq = *(const char **)q;
619 int ci;
620 while (*pp && *qq) {
d22b930b 621 if (toFOLD(*pp) != toFOLD(*qq))
b8ef571c
JH
622 break;
623 ++pp;
624 ++qq;
625 }
d22b930b 626 ci = toFOLD(*pp) - toFOLD(*qq);
b8ef571c
JH
627 if (ci == 0)
628 return compare(p, q);
629 return ci;
220398a0
PM
630}
631
632static int
72b16652
GS
633compare(const void *p, const void *q)
634{
635 return(strcmp(*(char **)p, *(char **)q));
636}
637
638static int
b8ef571c 639glob1(Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp)
72b16652 640{
b8ef571c 641 Char pathbuf[MAXPATHLEN];
72b16652 642
46a847d3
DM
643 assert(pattern < pattern_last);
644
72b16652 645 /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
3e5d0dec 646 if (*pattern == BG_EOS)
72b16652 647 return(0);
b8ef571c
JH
648 return(glob2(pathbuf, pathbuf+MAXPATHLEN-1,
649 pathbuf, pathbuf+MAXPATHLEN-1,
650 pattern, pattern_last, pglob, limitp));
72b16652
GS
651}
652
653/*
654 * The functions glob2 and glob3 are mutually recursive; there is one level
655 * of recursion for each segment in the pattern that contains one or more
656 * meta characters.
657 */
658static int
b8ef571c
JH
659glob2(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
660 Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp)
72b16652
GS
661{
662 Stat_t sb;
663 Char *p, *q;
664 int anymeta;
665
46a847d3
DM
666 assert(pattern < pattern_last);
667
72b16652
GS
668 /*
669 * Loop over pattern segments until end of pattern or until
670 * segment with meta character found.
671 */
672 for (anymeta = 0;;) {
3e5d0dec
GS
673 if (*pattern == BG_EOS) { /* End of pattern? */
674 *pathend = BG_EOS;
72b16652
GS
675 if (g_lstat(pathbuf, &sb, pglob))
676 return(0);
72b16652
GS
677
678 if (((pglob->gl_flags & GLOB_MARK) &&
220398a0
PM
679 pathend[-1] != BG_SEP
680#ifdef DOSISH
681 && pathend[-1] != BG_SEP2
682#endif
b8ef571c
JH
683 ) && (S_ISDIR(sb.st_mode) ||
684 (S_ISLNK(sb.st_mode) &&
72b16652
GS
685 (g_stat(pathbuf, &sb, pglob) == 0) &&
686 S_ISDIR(sb.st_mode)))) {
b8ef571c
JH
687 if (pathend+1 > pathend_last)
688 return (1);
3e5d0dec
GS
689 *pathend++ = BG_SEP;
690 *pathend = BG_EOS;
72b16652
GS
691 }
692 ++pglob->gl_matchc;
693#ifdef GLOB_DEBUG
694 printf("calling globextend from glob2\n");
695#endif /* GLOB_DEBUG */
b8ef571c 696 return(globextend(pathbuf, pglob, limitp));
72b16652
GS
697 }
698
699 /* Find end of next segment, copy tentatively to pathend. */
700 q = pathend;
701 p = pattern;
220398a0
PM
702 while (*p != BG_EOS && *p != BG_SEP
703#ifdef DOSISH
704 && *p != BG_SEP2
705#endif
706 ) {
46a847d3 707 assert(p < pattern_last);
72b16652
GS
708 if (ismeta(*p))
709 anymeta = 1;
b8ef571c
JH
710 if (q+1 > pathend_last)
711 return (1);
72b16652
GS
712 *q++ = *p++;
713 }
714
715 if (!anymeta) { /* No expansion, do next segment. */
716 pathend = q;
717 pattern = p;
220398a0
PM
718 while (*pattern == BG_SEP
719#ifdef DOSISH
720 || *pattern == BG_SEP2
721#endif
b8ef571c 722 ) {
46a847d3 723 assert(p < pattern_last);
b8ef571c
JH
724 if (pathend+1 > pathend_last)
725 return (1);
72b16652 726 *pathend++ = *pattern++;
b8ef571c
JH
727 }
728 } else
729 /* Need expansion, recurse. */
730 return(glob3(pathbuf, pathbuf_last, pathend,
46a847d3 731 pathend_last, pattern,
b8ef571c 732 p, pattern_last, pglob, limitp));
72b16652
GS
733 }
734 /* NOTREACHED */
735}
736
737static int
b8ef571c 738glob3(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
46a847d3 739 Char *pattern,
b8ef571c 740 Char *restpattern, Char *restpattern_last, glob_t *pglob, size_t *limitp)
72b16652 741{
5aaab254 742 Direntry_t *dp;
72b16652
GS
743 DIR *dirp;
744 int err;
220398a0 745 int nocase;
72b16652
GS
746 char buf[MAXPATHLEN];
747
748 /*
749 * The readdirfunc declaration can't be prototyped, because it is
750 * assigned, below, to two functions which are prototyped in glob.h
751 * and dirent.h as taking pointers to differently typed opaque
752 * structures.
753 */
cb359b41 754 Direntry_t *(*readdirfunc)(DIR*);
72b16652 755
300feaf1
MB
756 assert(pattern < restpattern_last);
757 assert(restpattern < restpattern_last);
758
b8ef571c
JH
759 if (pathend > pathend_last)
760 return (1);
3e5d0dec 761 *pathend = BG_EOS;
72b16652
GS
762 errno = 0;
763
f0963acb
GS
764#ifdef VMS
765 {
bac331f5
JH
766 Char *q = pathend;
767 if (q - pathbuf > 5) {
768 q -= 5;
769 if (q[0] == '.' &&
770 tolower(q[1]) == 'd' && tolower(q[2]) == 'i' &&
771 tolower(q[3]) == 'r' && q[4] == '/')
772 {
773 q[0] = '/';
774 q[1] = BG_EOS;
775 pathend = q+1;
776 }
777 }
f0963acb
GS
778 }
779#endif
be708cc0 780
72b16652
GS
781 if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
782 /* TODO: don't call for ENOENT or ENOTDIR? */
783 if (pglob->gl_errfunc) {
b8ef571c
JH
784 if (g_Ctoc(pathbuf, buf, sizeof(buf)))
785 return (GLOB_ABEND);
72b16652
GS
786 if (pglob->gl_errfunc(buf, errno) ||
787 (pglob->gl_flags & GLOB_ERR))
788 return (GLOB_ABEND);
789 }
790 return(0);
791 }
792
793 err = 0;
220398a0 794 nocase = ((pglob->gl_flags & GLOB_NOCASE) != 0);
72b16652
GS
795
796 /* Search directory for matching names. */
797 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
b8ef571c 798 readdirfunc = (Direntry_t *(*)(DIR *))pglob->gl_readdir;
72b16652 799 else
7ec40f37 800 readdirfunc = (Direntry_t *(*)(DIR *))my_readdir;
72b16652 801 while ((dp = (*readdirfunc)(dirp))) {
5aaab254
KW
802 U8 *sc;
803 Char *dc;
72b16652 804
3e5d0dec
GS
805 /* Initial BG_DOT must be matched literally. */
806 if (dp->d_name[0] == BG_DOT && *pattern != BG_DOT)
72b16652 807 continue;
b8ef571c
JH
808 dc = pathend;
809 sc = (U8 *) dp->d_name;
810 while (dc < pathend_last && (*dc++ = *sc++) != BG_EOS)
811 ;
812 if (dc >= pathend_last) {
813 *dc = BG_EOS;
814 err = 1;
815 break;
816 }
817
220398a0 818 if (!match(pathend, pattern, restpattern, nocase)) {
3e5d0dec 819 *pathend = BG_EOS;
72b16652
GS
820 continue;
821 }
b8ef571c
JH
822 err = glob2(pathbuf, pathbuf_last, --dc, pathend_last,
823 restpattern, restpattern_last, pglob, limitp);
72b16652
GS
824 if (err)
825 break;
826 }
827
828 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
829 (*pglob->gl_closedir)(dirp);
830 else
4f49e16e 831 PerlDir_close(dirp);
72b16652
GS
832 return(err);
833}
834
835
836/*
b7b1e41b 837 * Extend the gl_pathv member of a glob_t structure to accommodate a new item,
72b16652
GS
838 * add the new item, and update gl_pathc.
839 *
840 * This assumes the BSD realloc, which only copies the block when its size
841 * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
842 * behavior.
843 *
844 * Return 0 if new item added, error code if memory couldn't be allocated.
845 *
846 * Invariant of the glob_t structure:
847 * Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
848 * gl_pathv points to (gl_offs + gl_pathc + 1) items.
849 */
850static int
b8ef571c 851globextend(const Char *path, glob_t *pglob, size_t *limitp)
72b16652 852{
5aaab254
KW
853 char **pathv;
854 int i;
b8ef571c 855 STRLEN newsize, len;
72b16652
GS
856 char *copy;
857 const Char *p;
858
859#ifdef GLOB_DEBUG
860 printf("Adding ");
861 for (p = path; *p; p++)
862 (void)printf("%c", CHAR(*p));
863 printf("\n");
3e5d0dec 864#endif /* GLOB_DEBUG */
72b16652 865
b8ef571c 866 newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
4f49e16e 867 if (pglob->gl_pathv)
b8ef571c 868 pathv = Renew(pglob->gl_pathv,newsize,char*);
4f49e16e 869 else
a02a5408 870 Newx(pathv,newsize,char*);
b8ef571c
JH
871 if (pathv == NULL) {
872 if (pglob->gl_pathv) {
873 Safefree(pglob->gl_pathv);
874 pglob->gl_pathv = NULL;
875 }
72b16652 876 return(GLOB_NOSPACE);
b8ef571c 877 }
72b16652
GS
878
879 if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
880 /* first time around -- clear initial gl_offs items */
881 pathv += pglob->gl_offs;
882 for (i = pglob->gl_offs; --i >= 0; )
883 *--pathv = NULL;
884 }
885 pglob->gl_pathv = pathv;
886
887 for (p = path; *p++;)
b8ef571c
JH
888 ;
889 len = (STRLEN)(p - path);
bac331f5 890 *limitp += len;
a02a5408 891 Newx(copy, p-path, char);
4f49e16e 892 if (copy != NULL) {
b8ef571c
JH
893 if (g_Ctoc(path, copy, len)) {
894 Safefree(copy);
895 return(GLOB_NOSPACE);
896 }
72b16652
GS
897 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
898 }
899 pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
b8ef571c
JH
900
901 if ((pglob->gl_flags & GLOB_LIMIT) &&
46a847d3 902 newsize + *limitp >= (unsigned long)ARG_MAX) {
b8ef571c
JH
903 errno = 0;
904 return(GLOB_NOSPACE);
905 }
bac331f5 906
72b16652
GS
907 return(copy == NULL ? GLOB_NOSPACE : 0);
908}
909
910
911/*
912 * pattern matching function for filenames. Each occurrence of the *
913 * pattern causes a recursion level.
914 */
915static int
5aaab254 916match(Char *name, Char *pat, Char *patend, int nocase)
72b16652
GS
917{
918 int ok, negate_range;
919 Char c, k;
920
921 while (pat < patend) {
922 c = *pat++;
923 switch (c & M_MASK) {
924 case M_ALL:
925 if (pat == patend)
926 return(1);
927 do
220398a0 928 if (match(name, pat, patend, nocase))
72b16652 929 return(1);
b8ef571c
JH
930 while (*name++ != BG_EOS)
931 ;
72b16652
GS
932 return(0);
933 case M_ONE:
3e5d0dec 934 if (*name++ == BG_EOS)
72b16652
GS
935 return(0);
936 break;
937 case M_SET:
938 ok = 0;
3e5d0dec 939 if ((k = *name++) == BG_EOS)
72b16652 940 return(0);
3e5d0dec 941 if ((negate_range = ((*pat & M_MASK) == M_NOT)) != BG_EOS)
72b16652
GS
942 ++pat;
943 while (((c = *pat++) & M_MASK) != M_END)
944 if ((*pat & M_MASK) == M_RNG) {
220398a0
PM
945 if (nocase) {
946 if (tolower(c) <= tolower(k) && tolower(k) <= tolower(pat[1]))
947 ok = 1;
948 } else {
949 if (c <= k && k <= pat[1])
950 ok = 1;
951 }
72b16652 952 pat += 2;
220398a0 953 } else if (nocase ? (tolower(c) == tolower(k)) : (c == k))
72b16652
GS
954 ok = 1;
955 if (ok == negate_range)
956 return(0);
957 break;
958 default:
220398a0
PM
959 k = *name++;
960 if (nocase ? (tolower(k) != tolower(c)) : (k != c))
72b16652
GS
961 return(0);
962 break;
963 }
964 }
3e5d0dec 965 return(*name == BG_EOS);
72b16652
GS
966}
967
968/* Free allocated data belonging to a glob_t structure. */
969void
970bsd_globfree(glob_t *pglob)
971{
5aaab254
KW
972 int i;
973 char **pp;
72b16652
GS
974
975 if (pglob->gl_pathv != NULL) {
976 pp = pglob->gl_pathv + pglob->gl_offs;
977 for (i = pglob->gl_pathc; i--; ++pp)
978 if (*pp)
4f49e16e
GS
979 Safefree(*pp);
980 Safefree(pglob->gl_pathv);
b8ef571c 981 pglob->gl_pathv = NULL;
72b16652
GS
982 }
983}
984
985static DIR *
5aaab254 986g_opendir(Char *str, glob_t *pglob)
72b16652
GS
987{
988 char buf[MAXPATHLEN];
989
7369a524 990 if (!*str) {
28f0d0ec 991 my_strlcpy(buf, ".", sizeof(buf));
7369a524 992 } else {
b8ef571c
JH
993 if (g_Ctoc(str, buf, sizeof(buf)))
994 return(NULL);
7369a524 995 }
72b16652
GS
996
997 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
8f77bfdb 998 return((DIR*)(*pglob->gl_opendir)(buf));
bac331f5
JH
999
1000 return(PerlDir_open(buf));
72b16652
GS
1001}
1002
72b16652 1003static int
5aaab254 1004g_lstat(Char *fn, Stat_t *sb, glob_t *pglob)
72b16652
GS
1005{
1006 char buf[MAXPATHLEN];
1007
b8ef571c
JH
1008 if (g_Ctoc(fn, buf, sizeof(buf)))
1009 return(-1);
72b16652
GS
1010 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1011 return((*pglob->gl_lstat)(buf, sb));
4f49e16e
GS
1012#ifdef HAS_LSTAT
1013 return(PerlLIO_lstat(buf, sb));
1014#else
1015 return(PerlLIO_stat(buf, sb));
72b16652 1016#endif /* HAS_LSTAT */
4f49e16e 1017}
72b16652
GS
1018
1019static int
5aaab254 1020g_stat(Char *fn, Stat_t *sb, glob_t *pglob)
72b16652
GS
1021{
1022 char buf[MAXPATHLEN];
1023
b8ef571c
JH
1024 if (g_Ctoc(fn, buf, sizeof(buf)))
1025 return(-1);
72b16652
GS
1026 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1027 return((*pglob->gl_stat)(buf, sb));
4f49e16e 1028 return(PerlLIO_stat(buf, sb));
72b16652
GS
1029}
1030
1031static Char *
1032g_strchr(Char *str, int ch)
1033{
1034 do {
1035 if (*str == ch)
1036 return (str);
1037 } while (*str++);
1038 return (NULL);
1039}
1040
b8ef571c 1041static int
5aaab254 1042g_Ctoc(const Char *str, char *buf, STRLEN len)
72b16652 1043{
b8ef571c 1044 while (len--) {
7c436af3 1045 if ((*buf++ = (char)*str++) == BG_EOS)
b8ef571c
JH
1046 return (0);
1047 }
1048 return (1);
72b16652
GS
1049}
1050
1051#ifdef GLOB_DEBUG
1052static void
5aaab254 1053qprintf(const char *str, Char *s)
72b16652 1054{
5aaab254 1055 Char *p;
72b16652
GS
1056
1057 (void)printf("%s:\n", str);
1058 for (p = s; *p; p++)
1059 (void)printf("%c", CHAR(*p));
1060 (void)printf("\n");
1061 for (p = s; *p; p++)
1062 (void)printf("%c", *p & M_PROTECT ? '"' : ' ');
1063 (void)printf("\n");
1064 for (p = s; *p; p++)
1065 (void)printf("%c", ismeta(*p) ? '_' : ' ');
1066 (void)printf("\n");
1067}
1068#endif /* GLOB_DEBUG */