This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Add plain ~ expansion for Windows system in File::Glob
[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 *
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
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
e37778c2 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 *);
174static int glob3(Char *, Char *, Char *, Char *, Char *, Char *,
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))
605 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
be708cc0 606 pglob->gl_pathc - oldpathc, sizeof(char *),
2d5e9e5d
JH
607 (pglob->gl_flags & (GLOB_ALPHASORT|GLOB_NOCASE))
608 ? ci_compare : compare);
72b16652
GS
609 pglob->gl_flags = oldflags;
610 return(0);
611}
612
613static int
220398a0
PM
614ci_compare(const void *p, const void *q)
615{
b8ef571c
JH
616 const char *pp = *(const char **)p;
617 const char *qq = *(const char **)q;
618 int ci;
619 while (*pp && *qq) {
3f939f22 620 if (toLOWER(*pp) != toLOWER(*qq))
b8ef571c
JH
621 break;
622 ++pp;
623 ++qq;
624 }
3f939f22 625 ci = toLOWER(*pp) - toLOWER(*qq);
b8ef571c
JH
626 if (ci == 0)
627 return compare(p, q);
628 return ci;
220398a0
PM
629}
630
631static int
72b16652
GS
632compare(const void *p, const void *q)
633{
634 return(strcmp(*(char **)p, *(char **)q));
635}
636
637static int
b8ef571c 638glob1(Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp)
72b16652 639{
b8ef571c 640 Char pathbuf[MAXPATHLEN];
72b16652
GS
641
642 /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
3e5d0dec 643 if (*pattern == BG_EOS)
72b16652 644 return(0);
b8ef571c
JH
645 return(glob2(pathbuf, pathbuf+MAXPATHLEN-1,
646 pathbuf, pathbuf+MAXPATHLEN-1,
647 pattern, pattern_last, pglob, limitp));
72b16652
GS
648}
649
650/*
651 * The functions glob2 and glob3 are mutually recursive; there is one level
652 * of recursion for each segment in the pattern that contains one or more
653 * meta characters.
654 */
655static int
b8ef571c
JH
656glob2(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
657 Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp)
72b16652
GS
658{
659 Stat_t sb;
660 Char *p, *q;
661 int anymeta;
662
663 /*
664 * Loop over pattern segments until end of pattern or until
665 * segment with meta character found.
666 */
667 for (anymeta = 0;;) {
3e5d0dec
GS
668 if (*pattern == BG_EOS) { /* End of pattern? */
669 *pathend = BG_EOS;
72b16652
GS
670 if (g_lstat(pathbuf, &sb, pglob))
671 return(0);
72b16652
GS
672
673 if (((pglob->gl_flags & GLOB_MARK) &&
220398a0
PM
674 pathend[-1] != BG_SEP
675#ifdef DOSISH
676 && pathend[-1] != BG_SEP2
677#endif
b8ef571c
JH
678 ) && (S_ISDIR(sb.st_mode) ||
679 (S_ISLNK(sb.st_mode) &&
72b16652
GS
680 (g_stat(pathbuf, &sb, pglob) == 0) &&
681 S_ISDIR(sb.st_mode)))) {
b8ef571c
JH
682 if (pathend+1 > pathend_last)
683 return (1);
3e5d0dec
GS
684 *pathend++ = BG_SEP;
685 *pathend = BG_EOS;
72b16652
GS
686 }
687 ++pglob->gl_matchc;
688#ifdef GLOB_DEBUG
689 printf("calling globextend from glob2\n");
690#endif /* GLOB_DEBUG */
b8ef571c 691 return(globextend(pathbuf, pglob, limitp));
72b16652
GS
692 }
693
694 /* Find end of next segment, copy tentatively to pathend. */
695 q = pathend;
696 p = pattern;
220398a0
PM
697 while (*p != BG_EOS && *p != BG_SEP
698#ifdef DOSISH
699 && *p != BG_SEP2
700#endif
701 ) {
72b16652
GS
702 if (ismeta(*p))
703 anymeta = 1;
b8ef571c
JH
704 if (q+1 > pathend_last)
705 return (1);
72b16652
GS
706 *q++ = *p++;
707 }
708
709 if (!anymeta) { /* No expansion, do next segment. */
710 pathend = q;
711 pattern = p;
220398a0
PM
712 while (*pattern == BG_SEP
713#ifdef DOSISH
714 || *pattern == BG_SEP2
715#endif
b8ef571c
JH
716 ) {
717 if (pathend+1 > pathend_last)
718 return (1);
72b16652 719 *pathend++ = *pattern++;
b8ef571c
JH
720 }
721 } else
722 /* Need expansion, recurse. */
723 return(glob3(pathbuf, pathbuf_last, pathend,
724 pathend_last, pattern, pattern_last,
725 p, pattern_last, pglob, limitp));
72b16652
GS
726 }
727 /* NOTREACHED */
728}
729
730static int
b8ef571c
JH
731glob3(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
732 Char *pattern, Char *pattern_last,
733 Char *restpattern, Char *restpattern_last, glob_t *pglob, size_t *limitp)
72b16652
GS
734{
735 register Direntry_t *dp;
736 DIR *dirp;
737 int err;
220398a0 738 int nocase;
72b16652
GS
739 char buf[MAXPATHLEN];
740
741 /*
742 * The readdirfunc declaration can't be prototyped, because it is
743 * assigned, below, to two functions which are prototyped in glob.h
744 * and dirent.h as taking pointers to differently typed opaque
745 * structures.
746 */
cb359b41 747 Direntry_t *(*readdirfunc)(DIR*);
72b16652 748
b8ef571c
JH
749 if (pathend > pathend_last)
750 return (1);
3e5d0dec 751 *pathend = BG_EOS;
72b16652
GS
752 errno = 0;
753
f0963acb
GS
754#ifdef VMS
755 {
bac331f5
JH
756 Char *q = pathend;
757 if (q - pathbuf > 5) {
758 q -= 5;
759 if (q[0] == '.' &&
760 tolower(q[1]) == 'd' && tolower(q[2]) == 'i' &&
761 tolower(q[3]) == 'r' && q[4] == '/')
762 {
763 q[0] = '/';
764 q[1] = BG_EOS;
765 pathend = q+1;
766 }
767 }
f0963acb
GS
768 }
769#endif
be708cc0 770
72b16652
GS
771 if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
772 /* TODO: don't call for ENOENT or ENOTDIR? */
773 if (pglob->gl_errfunc) {
b8ef571c
JH
774 if (g_Ctoc(pathbuf, buf, sizeof(buf)))
775 return (GLOB_ABEND);
72b16652
GS
776 if (pglob->gl_errfunc(buf, errno) ||
777 (pglob->gl_flags & GLOB_ERR))
778 return (GLOB_ABEND);
779 }
780 return(0);
781 }
782
783 err = 0;
220398a0 784 nocase = ((pglob->gl_flags & GLOB_NOCASE) != 0);
72b16652
GS
785
786 /* Search directory for matching names. */
787 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
b8ef571c 788 readdirfunc = (Direntry_t *(*)(DIR *))pglob->gl_readdir;
72b16652 789 else
7ec40f37 790 readdirfunc = (Direntry_t *(*)(DIR *))my_readdir;
72b16652
GS
791 while ((dp = (*readdirfunc)(dirp))) {
792 register U8 *sc;
793 register Char *dc;
794
3e5d0dec
GS
795 /* Initial BG_DOT must be matched literally. */
796 if (dp->d_name[0] == BG_DOT && *pattern != BG_DOT)
72b16652 797 continue;
b8ef571c
JH
798 dc = pathend;
799 sc = (U8 *) dp->d_name;
800 while (dc < pathend_last && (*dc++ = *sc++) != BG_EOS)
801 ;
802 if (dc >= pathend_last) {
803 *dc = BG_EOS;
804 err = 1;
805 break;
806 }
807
220398a0 808 if (!match(pathend, pattern, restpattern, nocase)) {
3e5d0dec 809 *pathend = BG_EOS;
72b16652
GS
810 continue;
811 }
b8ef571c
JH
812 err = glob2(pathbuf, pathbuf_last, --dc, pathend_last,
813 restpattern, restpattern_last, pglob, limitp);
72b16652
GS
814 if (err)
815 break;
816 }
817
818 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
819 (*pglob->gl_closedir)(dirp);
820 else
4f49e16e 821 PerlDir_close(dirp);
72b16652
GS
822 return(err);
823}
824
825
826/*
b7b1e41b 827 * Extend the gl_pathv member of a glob_t structure to accommodate a new item,
72b16652
GS
828 * add the new item, and update gl_pathc.
829 *
830 * This assumes the BSD realloc, which only copies the block when its size
831 * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
832 * behavior.
833 *
834 * Return 0 if new item added, error code if memory couldn't be allocated.
835 *
836 * Invariant of the glob_t structure:
837 * Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
838 * gl_pathv points to (gl_offs + gl_pathc + 1) items.
839 */
840static int
b8ef571c 841globextend(const Char *path, glob_t *pglob, size_t *limitp)
72b16652
GS
842{
843 register char **pathv;
844 register int i;
b8ef571c 845 STRLEN newsize, len;
72b16652
GS
846 char *copy;
847 const Char *p;
848
849#ifdef GLOB_DEBUG
850 printf("Adding ");
851 for (p = path; *p; p++)
852 (void)printf("%c", CHAR(*p));
853 printf("\n");
3e5d0dec 854#endif /* GLOB_DEBUG */
72b16652 855
b8ef571c 856 newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
4f49e16e 857 if (pglob->gl_pathv)
b8ef571c 858 pathv = Renew(pglob->gl_pathv,newsize,char*);
4f49e16e 859 else
a02a5408 860 Newx(pathv,newsize,char*);
b8ef571c
JH
861 if (pathv == NULL) {
862 if (pglob->gl_pathv) {
863 Safefree(pglob->gl_pathv);
864 pglob->gl_pathv = NULL;
865 }
72b16652 866 return(GLOB_NOSPACE);
b8ef571c 867 }
72b16652
GS
868
869 if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
870 /* first time around -- clear initial gl_offs items */
871 pathv += pglob->gl_offs;
872 for (i = pglob->gl_offs; --i >= 0; )
873 *--pathv = NULL;
874 }
875 pglob->gl_pathv = pathv;
876
877 for (p = path; *p++;)
b8ef571c
JH
878 ;
879 len = (STRLEN)(p - path);
bac331f5 880 *limitp += len;
a02a5408 881 Newx(copy, p-path, char);
4f49e16e 882 if (copy != NULL) {
b8ef571c
JH
883 if (g_Ctoc(path, copy, len)) {
884 Safefree(copy);
885 return(GLOB_NOSPACE);
886 }
72b16652
GS
887 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
888 }
889 pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
b8ef571c
JH
890
891 if ((pglob->gl_flags & GLOB_LIMIT) &&
892 newsize + *limitp >= ARG_MAX) {
893 errno = 0;
894 return(GLOB_NOSPACE);
895 }
bac331f5 896
72b16652
GS
897 return(copy == NULL ? GLOB_NOSPACE : 0);
898}
899
900
901/*
902 * pattern matching function for filenames. Each occurrence of the *
903 * pattern causes a recursion level.
904 */
905static int
220398a0 906match(register Char *name, register Char *pat, register Char *patend, int nocase)
72b16652
GS
907{
908 int ok, negate_range;
909 Char c, k;
910
911 while (pat < patend) {
912 c = *pat++;
913 switch (c & M_MASK) {
914 case M_ALL:
915 if (pat == patend)
916 return(1);
917 do
220398a0 918 if (match(name, pat, patend, nocase))
72b16652 919 return(1);
b8ef571c
JH
920 while (*name++ != BG_EOS)
921 ;
72b16652
GS
922 return(0);
923 case M_ONE:
3e5d0dec 924 if (*name++ == BG_EOS)
72b16652
GS
925 return(0);
926 break;
927 case M_SET:
928 ok = 0;
3e5d0dec 929 if ((k = *name++) == BG_EOS)
72b16652 930 return(0);
3e5d0dec 931 if ((negate_range = ((*pat & M_MASK) == M_NOT)) != BG_EOS)
72b16652
GS
932 ++pat;
933 while (((c = *pat++) & M_MASK) != M_END)
934 if ((*pat & M_MASK) == M_RNG) {
220398a0
PM
935 if (nocase) {
936 if (tolower(c) <= tolower(k) && tolower(k) <= tolower(pat[1]))
937 ok = 1;
938 } else {
939 if (c <= k && k <= pat[1])
940 ok = 1;
941 }
72b16652 942 pat += 2;
220398a0 943 } else if (nocase ? (tolower(c) == tolower(k)) : (c == k))
72b16652
GS
944 ok = 1;
945 if (ok == negate_range)
946 return(0);
947 break;
948 default:
220398a0
PM
949 k = *name++;
950 if (nocase ? (tolower(k) != tolower(c)) : (k != c))
72b16652
GS
951 return(0);
952 break;
953 }
954 }
3e5d0dec 955 return(*name == BG_EOS);
72b16652
GS
956}
957
958/* Free allocated data belonging to a glob_t structure. */
959void
960bsd_globfree(glob_t *pglob)
961{
962 register int i;
963 register char **pp;
964
965 if (pglob->gl_pathv != NULL) {
966 pp = pglob->gl_pathv + pglob->gl_offs;
967 for (i = pglob->gl_pathc; i--; ++pp)
968 if (*pp)
4f49e16e
GS
969 Safefree(*pp);
970 Safefree(pglob->gl_pathv);
b8ef571c 971 pglob->gl_pathv = NULL;
72b16652
GS
972 }
973}
974
975static DIR *
976g_opendir(register Char *str, glob_t *pglob)
977{
978 char buf[MAXPATHLEN];
979
7369a524 980 if (!*str) {
28f0d0ec 981 my_strlcpy(buf, ".", sizeof(buf));
7369a524 982 } else {
b8ef571c
JH
983 if (g_Ctoc(str, buf, sizeof(buf)))
984 return(NULL);
7369a524 985 }
72b16652
GS
986
987 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
8f77bfdb 988 return((DIR*)(*pglob->gl_opendir)(buf));
bac331f5
JH
989
990 return(PerlDir_open(buf));
72b16652
GS
991}
992
72b16652
GS
993static int
994g_lstat(register Char *fn, Stat_t *sb, glob_t *pglob)
995{
996 char buf[MAXPATHLEN];
997
b8ef571c
JH
998 if (g_Ctoc(fn, buf, sizeof(buf)))
999 return(-1);
72b16652
GS
1000 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1001 return((*pglob->gl_lstat)(buf, sb));
4f49e16e
GS
1002#ifdef HAS_LSTAT
1003 return(PerlLIO_lstat(buf, sb));
1004#else
1005 return(PerlLIO_stat(buf, sb));
72b16652 1006#endif /* HAS_LSTAT */
4f49e16e 1007}
72b16652
GS
1008
1009static int
1010g_stat(register Char *fn, Stat_t *sb, glob_t *pglob)
1011{
1012 char buf[MAXPATHLEN];
1013
b8ef571c
JH
1014 if (g_Ctoc(fn, buf, sizeof(buf)))
1015 return(-1);
72b16652
GS
1016 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1017 return((*pglob->gl_stat)(buf, sb));
4f49e16e 1018 return(PerlLIO_stat(buf, sb));
72b16652
GS
1019}
1020
1021static Char *
1022g_strchr(Char *str, int ch)
1023{
1024 do {
1025 if (*str == ch)
1026 return (str);
1027 } while (*str++);
1028 return (NULL);
1029}
1030
b8ef571c
JH
1031static int
1032g_Ctoc(register const Char *str, char *buf, STRLEN len)
72b16652 1033{
b8ef571c 1034 while (len--) {
7c436af3 1035 if ((*buf++ = (char)*str++) == BG_EOS)
b8ef571c
JH
1036 return (0);
1037 }
1038 return (1);
72b16652
GS
1039}
1040
1041#ifdef GLOB_DEBUG
1042static void
1043qprintf(const char *str, register Char *s)
1044{
1045 register Char *p;
1046
1047 (void)printf("%s:\n", str);
1048 for (p = s; *p; p++)
1049 (void)printf("%c", CHAR(*p));
1050 (void)printf("\n");
1051 for (p = s; *p; p++)
1052 (void)printf("%c", *p & M_PROTECT ? '"' : ' ');
1053 (void)printf("\n");
1054 for (p = s; *p; p++)
1055 (void)printf("%c", ismeta(*p) ? '_' : ' ');
1056 (void)printf("\n");
1057}
1058#endif /* GLOB_DEBUG */