This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
ext/ + -Wall
[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
7369a524
CN
85# ifdef MACOS_TRADITIONAL
86# define MAXPATHLEN 255
87# else
88# define MAXPATHLEN 1024
89# endif
72b16652
GS
90# endif
91#endif
92
b8ef571c
JH
93#ifdef I_LIMITS
94#include <limits.h>
95#endif
96
97#ifndef ARG_MAX
6eb630b7
CN
98# ifdef MACOS_TRADITIONAL
99# define ARG_MAX 65536 /* Mac OS is actually unlimited */
b8ef571c 100# else
6eb630b7
CN
101# ifdef _SC_ARG_MAX
102# define ARG_MAX (sysconf(_SC_ARG_MAX))
b8ef571c 103# else
6eb630b7
CN
104# ifdef _POSIX_ARG_MAX
105# define ARG_MAX _POSIX_ARG_MAX
b8ef571c 106# else
6eb630b7
CN
107# ifdef WIN32
108# define ARG_MAX 14500 /* from VC's limits.h */
109# else
110# define ARG_MAX 4096 /* from POSIX, be conservative */
111# endif
b8ef571c
JH
112# endif
113# endif
114# endif
115#endif
116
3e5d0dec
GS
117#define BG_DOLLAR '$'
118#define BG_DOT '.'
119#define BG_EOS '\0'
120#define BG_LBRACKET '['
121#define BG_NOT '!'
122#define BG_QUESTION '?'
123#define BG_QUOTE '\\'
124#define BG_RANGE '-'
125#define BG_RBRACKET ']'
7369a524
CN
126#ifdef MACOS_TRADITIONAL
127# define BG_SEP ':'
128#else
129# define BG_SEP '/'
130#endif
220398a0
PM
131#ifdef DOSISH
132#define BG_SEP2 '\\'
133#endif
3e5d0dec
GS
134#define BG_STAR '*'
135#define BG_TILDE '~'
136#define BG_UNDERSCORE '_'
137#define BG_LBRACE '{'
138#define BG_RBRACE '}'
139#define BG_SLASH '/'
140#define BG_COMMA ','
72b16652
GS
141
142#ifndef GLOB_DEBUG
143
144#define M_QUOTE 0x8000
145#define M_PROTECT 0x4000
146#define M_MASK 0xffff
147#define M_ASCII 0x00ff
148
149typedef U16 Char;
150
151#else
152
153#define M_QUOTE 0x80
154#define M_PROTECT 0x40
155#define M_MASK 0xff
156#define M_ASCII 0x7f
157
158typedef U8 Char;
159
160#endif /* !GLOB_DEBUG */
161
162
163#define CHAR(c) ((Char)((c)&M_ASCII))
164#define META(c) ((Char)((c)|M_QUOTE))
165#define M_ALL META('*')
166#define M_END META(']')
167#define M_NOT META('!')
168#define M_ONE META('?')
169#define M_RNG META('-')
170#define M_SET META('[')
171#define ismeta(c) (((c)&M_QUOTE) != 0)
172
173
174static int compare(const void *, const void *);
220398a0 175static int ci_compare(const void *, const void *);
b8ef571c 176static int g_Ctoc(const Char *, char *, STRLEN);
72b16652
GS
177static int g_lstat(Char *, Stat_t *, glob_t *);
178static DIR *g_opendir(Char *, glob_t *);
179static Char *g_strchr(Char *, int);
72b16652
GS
180static int g_stat(Char *, Stat_t *, glob_t *);
181static int glob0(const Char *, glob_t *);
b8ef571c
JH
182static int glob1(Char *, Char *, glob_t *, size_t *);
183static int glob2(Char *, Char *, Char *, Char *, Char *, Char *,
184 glob_t *, size_t *);
185static int glob3(Char *, Char *, Char *, Char *, Char *, Char *,
186 Char *, Char *, glob_t *, size_t *);
187static int globextend(const Char *, glob_t *, size_t *);
188static const Char *
189 globtilde(const Char *, Char *, size_t, glob_t *);
72b16652
GS
190static int globexp1(const Char *, glob_t *);
191static int globexp2(const Char *, const Char *, glob_t *, int *);
220398a0 192static int match(Char *, Char *, Char *, int);
72b16652
GS
193#ifdef GLOB_DEBUG
194static void qprintf(const char *, Char *);
195#endif /* GLOB_DEBUG */
196
4f49e16e
GS
197#ifdef PERL_IMPLICIT_CONTEXT
198static Direntry_t * my_readdir(DIR*);
199
200static Direntry_t *
201my_readdir(DIR *d)
202{
203 return PerlDir_read(d);
204}
205#else
206#define my_readdir readdir
207#endif
208
72b16652
GS
209int
210bsd_glob(const char *pattern, int flags,
211 int (*errfunc)(const char *, int), glob_t *pglob)
212{
213 const U8 *patnext;
214 int c;
b8ef571c 215 Char *bufnext, *bufend, patbuf[MAXPATHLEN];
72b16652
GS
216
217 patnext = (U8 *) pattern;
218 if (!(flags & GLOB_APPEND)) {
219 pglob->gl_pathc = 0;
220 pglob->gl_pathv = NULL;
221 if (!(flags & GLOB_DOOFFS))
222 pglob->gl_offs = 0;
223 }
224 pglob->gl_flags = flags & ~GLOB_MAGCHAR;
225 pglob->gl_errfunc = errfunc;
226 pglob->gl_matchc = 0;
227
228 bufnext = patbuf;
b8ef571c 229 bufend = bufnext + MAXPATHLEN - 1;
220398a0
PM
230#ifdef DOSISH
231 /* Nasty hack to treat patterns like "C:*" correctly. In this
232 * case, the * should match any file in the current directory
233 * on the C: drive. However, the glob code does not treat the
234 * colon specially, so it looks for files beginning "C:" in
235 * the current directory. To fix this, change the pattern to
236 * add an explicit "./" at the start (just after the drive
c6c619a9 237 * letter and colon - ie change to "C:./").
220398a0
PM
238 */
239 if (isalpha(pattern[0]) && pattern[1] == ':' &&
240 pattern[2] != BG_SEP && pattern[2] != BG_SEP2 &&
241 bufend - bufnext > 4) {
242 *bufnext++ = pattern[0];
243 *bufnext++ = ':';
244 *bufnext++ = '.';
245 *bufnext++ = BG_SEP;
246 patnext += 2;
247 }
248#endif
72b16652
GS
249 if (flags & GLOB_QUOTE) {
250 /* Protect the quoted characters. */
3e5d0dec
GS
251 while (bufnext < bufend && (c = *patnext++) != BG_EOS)
252 if (c == BG_QUOTE) {
220398a0
PM
253#ifdef DOSISH
254 /* To avoid backslashitis on Win32,
255 * we only treat \ as a quoting character
256 * if it precedes one of the
257 * metacharacters []-{}~\
258 */
259 if ((c = *patnext++) != '[' && c != ']' &&
260 c != '-' && c != '{' && c != '}' &&
261 c != '~' && c != '\\') {
262#else
3e5d0dec 263 if ((c = *patnext++) == BG_EOS) {
220398a0 264#endif
3e5d0dec 265 c = BG_QUOTE;
72b16652
GS
266 --patnext;
267 }
268 *bufnext++ = c | M_PROTECT;
b8ef571c 269 } else
72b16652 270 *bufnext++ = c;
b8ef571c
JH
271 } else
272 while (bufnext < bufend && (c = *patnext++) != BG_EOS)
273 *bufnext++ = c;
3e5d0dec 274 *bufnext = BG_EOS;
72b16652
GS
275
276 if (flags & GLOB_BRACE)
277 return globexp1(patbuf, pglob);
278 else
279 return glob0(patbuf, pglob);
280}
281
282/*
283 * Expand recursively a glob {} pattern. When there is no more expansion
284 * invoke the standard globbing routine to glob the rest of the magic
285 * characters
286 */
b8ef571c
JH
287static int
288globexp1(const Char *pattern, glob_t *pglob)
72b16652
GS
289{
290 const Char* ptr = pattern;
291 int rv;
292
293 /* Protect a single {}, for find(1), like csh */
3e5d0dec 294 if (pattern[0] == BG_LBRACE && pattern[1] == BG_RBRACE && pattern[2] == BG_EOS)
72b16652
GS
295 return glob0(pattern, pglob);
296
3e5d0dec 297 while ((ptr = (const Char *) g_strchr((Char *) ptr, BG_LBRACE)) != NULL)
72b16652
GS
298 if (!globexp2(ptr, pattern, pglob, &rv))
299 return rv;
300
301 return glob0(pattern, pglob);
302}
303
304
305/*
306 * Recursive brace globbing helper. Tries to expand a single brace.
307 * If it succeeds then it invokes globexp1 with the new pattern.
308 * If it fails then it tries to glob the rest of the pattern and returns.
309 */
b8ef571c
JH
310static int
311globexp2(const Char *ptr, const Char *pattern,
312 glob_t *pglob, int *rv)
72b16652
GS
313{
314 int i;
315 Char *lm, *ls;
316 const Char *pe, *pm, *pl;
b8ef571c 317 Char patbuf[MAXPATHLEN];
72b16652
GS
318
319 /* copy part up to the brace */
320 for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
b8ef571c
JH
321 ;
322 *lm = BG_EOS;
72b16652
GS
323 ls = lm;
324
325 /* Find the balanced brace */
326 for (i = 0, pe = ++ptr; *pe; pe++)
3e5d0dec 327 if (*pe == BG_LBRACKET) {
72b16652 328 /* Ignore everything between [] */
3e5d0dec 329 for (pm = pe++; *pe != BG_RBRACKET && *pe != BG_EOS; pe++)
b8ef571c 330 ;
3e5d0dec 331 if (*pe == BG_EOS) {
72b16652 332 /*
3e5d0dec
GS
333 * We could not find a matching BG_RBRACKET.
334 * Ignore and just look for BG_RBRACE
72b16652
GS
335 */
336 pe = pm;
337 }
b8ef571c 338 } else if (*pe == BG_LBRACE)
72b16652 339 i++;
3e5d0dec 340 else if (*pe == BG_RBRACE) {
72b16652
GS
341 if (i == 0)
342 break;
343 i--;
344 }
345
346 /* Non matching braces; just glob the pattern */
3e5d0dec 347 if (i != 0 || *pe == BG_EOS) {
72b16652
GS
348 *rv = glob0(patbuf, pglob);
349 return 0;
350 }
351
b8ef571c 352 for (i = 0, pl = pm = ptr; pm <= pe; pm++) {
72b16652 353 switch (*pm) {
3e5d0dec 354 case BG_LBRACKET:
72b16652 355 /* Ignore everything between [] */
3e5d0dec 356 for (pl = pm++; *pm != BG_RBRACKET && *pm != BG_EOS; pm++)
b8ef571c 357 ;
3e5d0dec 358 if (*pm == BG_EOS) {
72b16652 359 /*
3e5d0dec
GS
360 * We could not find a matching BG_RBRACKET.
361 * Ignore and just look for BG_RBRACE
72b16652
GS
362 */
363 pm = pl;
364 }
365 break;
366
3e5d0dec 367 case BG_LBRACE:
72b16652
GS
368 i++;
369 break;
370
3e5d0dec 371 case BG_RBRACE:
72b16652 372 if (i) {
b8ef571c
JH
373 i--;
374 break;
72b16652
GS
375 }
376 /* FALLTHROUGH */
3e5d0dec
GS
377 case BG_COMMA:
378 if (i && *pm == BG_COMMA)
72b16652
GS
379 break;
380 else {
381 /* Append the current string */
382 for (lm = ls; (pl < pm); *lm++ = *pl++)
b8ef571c
JH
383 ;
384
72b16652
GS
385 /*
386 * Append the rest of the pattern after the
387 * closing brace
388 */
b8ef571c
JH
389 for (pl = pe + 1; (*lm++ = *pl++) != BG_EOS; )
390 ;
72b16652
GS
391
392 /* Expand the current pattern */
393#ifdef GLOB_DEBUG
394 qprintf("globexp2:", patbuf);
395#endif /* GLOB_DEBUG */
396 *rv = globexp1(patbuf, pglob);
397
398 /* move after the comma, to the next string */
399 pl = pm + 1;
400 }
401 break;
402
403 default:
404 break;
405 }
b8ef571c 406 }
72b16652
GS
407 *rv = 0;
408 return 0;
409}
410
411
412
413/*
414 * expand tilde from the passwd file.
415 */
416static const Char *
bac331f5 417globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob)
72b16652
GS
418{
419 struct passwd *pwd;
420 char *h;
421 const Char *p;
bac331f5 422 Char *b, *eb;
72b16652 423
3e5d0dec 424 if (*pattern != BG_TILDE || !(pglob->gl_flags & GLOB_TILDE))
72b16652
GS
425 return pattern;
426
427 /* Copy up to the end of the string or / */
bac331f5
JH
428 eb = &patbuf[patbuf_len - 1];
429 for (p = pattern + 1, h = (char *) patbuf;
430 h < (char*)eb && *p && *p != BG_SLASH; *h++ = *p++)
b8ef571c 431 ;
72b16652 432
3e5d0dec 433 *h = BG_EOS;
72b16652 434
b8ef571c
JH
435#if 0
436 if (h == (char *)eb)
437 return what;
438#endif
439
3e5d0dec 440 if (((char *) patbuf)[0] == BG_EOS) {
72b16652
GS
441 /*
442 * handle a plain ~ or ~/ by expanding $HOME
443 * first and then trying the password file
444 */
445 if ((h = getenv("HOME")) == NULL) {
446#ifdef HAS_PASSWD
447 if ((pwd = getpwuid(getuid())) == NULL)
448 return pattern;
449 else
450 h = pwd->pw_dir;
451#else
452 return pattern;
453#endif
454 }
b8ef571c 455 } else {
72b16652
GS
456 /*
457 * Expand a ~user
458 */
459#ifdef HAS_PASSWD
460 if ((pwd = getpwnam((char*) patbuf)) == NULL)
461 return pattern;
462 else
463 h = pwd->pw_dir;
464#else
465 return pattern;
466#endif
467 }
468
469 /* Copy the home directory */
bac331f5 470 for (b = patbuf; b < eb && *h; *b++ = *h++)
b8ef571c 471 ;
72b16652
GS
472
473 /* Append the rest of the pattern */
bac331f5 474 while (b < eb && (*b++ = *p++) != BG_EOS)
b8ef571c 475 ;
bac331f5 476 *b = BG_EOS;
72b16652
GS
477
478 return patbuf;
479}
480
481
482/*
483 * The main glob() routine: compiles the pattern (optionally processing
484 * quotes), calls glob1() to do the real pattern matching, and finally
485 * sorts the list (unless unsorted operation is requested). Returns 0
486 * if things went well, nonzero if errors occurred. It is not an error
487 * to find no matches.
488 */
489static int
490glob0(const Char *pattern, glob_t *pglob)
491{
492 const Char *qpat, *qpatnext;
493 int c, err, oldflags, oldpathc;
b8ef571c
JH
494 Char *bufnext, patbuf[MAXPATHLEN];
495 size_t limit = 0;
72b16652 496
7369a524
CN
497#ifdef MACOS_TRADITIONAL
498 if ( (*pattern == BG_TILDE) && (pglob->gl_flags & GLOB_TILDE) ) {
6eb630b7 499 return(globextend(pattern, pglob, &limit));
7369a524
CN
500 }
501#endif
502
b8ef571c 503 qpat = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
72b16652
GS
504 qpatnext = qpat;
505 oldflags = pglob->gl_flags;
506 oldpathc = pglob->gl_pathc;
507 bufnext = patbuf;
508
509 /* We don't need to check for buffer overflow any more. */
3e5d0dec 510 while ((c = *qpatnext++) != BG_EOS) {
72b16652 511 switch (c) {
3e5d0dec 512 case BG_LBRACKET:
72b16652 513 c = *qpatnext;
3e5d0dec 514 if (c == BG_NOT)
72b16652 515 ++qpatnext;
3e5d0dec
GS
516 if (*qpatnext == BG_EOS ||
517 g_strchr((Char *) qpatnext+1, BG_RBRACKET) == NULL) {
518 *bufnext++ = BG_LBRACKET;
519 if (c == BG_NOT)
72b16652
GS
520 --qpatnext;
521 break;
522 }
523 *bufnext++ = M_SET;
3e5d0dec 524 if (c == BG_NOT)
72b16652
GS
525 *bufnext++ = M_NOT;
526 c = *qpatnext++;
527 do {
528 *bufnext++ = CHAR(c);
3e5d0dec
GS
529 if (*qpatnext == BG_RANGE &&
530 (c = qpatnext[1]) != BG_RBRACKET) {
72b16652
GS
531 *bufnext++ = M_RNG;
532 *bufnext++ = CHAR(c);
533 qpatnext += 2;
534 }
3e5d0dec 535 } while ((c = *qpatnext++) != BG_RBRACKET);
72b16652
GS
536 pglob->gl_flags |= GLOB_MAGCHAR;
537 *bufnext++ = M_END;
538 break;
3e5d0dec 539 case BG_QUESTION:
72b16652
GS
540 pglob->gl_flags |= GLOB_MAGCHAR;
541 *bufnext++ = M_ONE;
542 break;
3e5d0dec 543 case BG_STAR:
72b16652
GS
544 pglob->gl_flags |= GLOB_MAGCHAR;
545 /* collapse adjacent stars to one,
546 * to avoid exponential behavior
547 */
548 if (bufnext == patbuf || bufnext[-1] != M_ALL)
b8ef571c 549 *bufnext++ = M_ALL;
72b16652
GS
550 break;
551 default:
552 *bufnext++ = CHAR(c);
553 break;
554 }
555 }
3e5d0dec 556 *bufnext = BG_EOS;
72b16652
GS
557#ifdef GLOB_DEBUG
558 qprintf("glob0:", patbuf);
559#endif /* GLOB_DEBUG */
560
b8ef571c 561 if ((err = glob1(patbuf, patbuf+MAXPATHLEN-1, pglob, &limit)) != 0) {
72b16652
GS
562 pglob->gl_flags = oldflags;
563 return(err);
564 }
565
566 /*
567 * If there was no match we are going to append the pattern
568 * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
569 * and the pattern did not contain any magic characters
570 * GLOB_NOMAGIC is there just for compatibility with csh.
571 */
572 if (pglob->gl_pathc == oldpathc &&
573 ((pglob->gl_flags & GLOB_NOCHECK) ||
574 ((pglob->gl_flags & GLOB_NOMAGIC) &&
575 !(pglob->gl_flags & GLOB_MAGCHAR))))
576 {
577#ifdef GLOB_DEBUG
578 printf("calling globextend from glob0\n");
579#endif /* GLOB_DEBUG */
580 pglob->gl_flags = oldflags;
b8ef571c 581 return(globextend(qpat, pglob, &limit));
72b16652
GS
582 }
583 else if (!(pglob->gl_flags & GLOB_NOSORT))
584 qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
220398a0 585 pglob->gl_pathc - oldpathc, sizeof(char *),
2d5e9e5d
JH
586 (pglob->gl_flags & (GLOB_ALPHASORT|GLOB_NOCASE))
587 ? ci_compare : compare);
72b16652
GS
588 pglob->gl_flags = oldflags;
589 return(0);
590}
591
592static int
220398a0
PM
593ci_compare(const void *p, const void *q)
594{
b8ef571c
JH
595 const char *pp = *(const char **)p;
596 const char *qq = *(const char **)q;
597 int ci;
598 while (*pp && *qq) {
3f939f22 599 if (toLOWER(*pp) != toLOWER(*qq))
b8ef571c
JH
600 break;
601 ++pp;
602 ++qq;
603 }
3f939f22 604 ci = toLOWER(*pp) - toLOWER(*qq);
b8ef571c
JH
605 if (ci == 0)
606 return compare(p, q);
607 return ci;
220398a0
PM
608}
609
610static int
72b16652
GS
611compare(const void *p, const void *q)
612{
613 return(strcmp(*(char **)p, *(char **)q));
614}
615
616static int
b8ef571c 617glob1(Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp)
72b16652 618{
b8ef571c 619 Char pathbuf[MAXPATHLEN];
72b16652
GS
620
621 /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
3e5d0dec 622 if (*pattern == BG_EOS)
72b16652 623 return(0);
b8ef571c
JH
624 return(glob2(pathbuf, pathbuf+MAXPATHLEN-1,
625 pathbuf, pathbuf+MAXPATHLEN-1,
626 pattern, pattern_last, pglob, limitp));
72b16652
GS
627}
628
629/*
630 * The functions glob2 and glob3 are mutually recursive; there is one level
631 * of recursion for each segment in the pattern that contains one or more
632 * meta characters.
633 */
634static int
b8ef571c
JH
635glob2(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
636 Char *pattern, Char *pattern_last, glob_t *pglob, size_t *limitp)
72b16652
GS
637{
638 Stat_t sb;
639 Char *p, *q;
640 int anymeta;
641
642 /*
643 * Loop over pattern segments until end of pattern or until
644 * segment with meta character found.
645 */
646 for (anymeta = 0;;) {
3e5d0dec
GS
647 if (*pattern == BG_EOS) { /* End of pattern? */
648 *pathend = BG_EOS;
72b16652
GS
649 if (g_lstat(pathbuf, &sb, pglob))
650 return(0);
72b16652
GS
651
652 if (((pglob->gl_flags & GLOB_MARK) &&
220398a0
PM
653 pathend[-1] != BG_SEP
654#ifdef DOSISH
655 && pathend[-1] != BG_SEP2
656#endif
b8ef571c
JH
657 ) && (S_ISDIR(sb.st_mode) ||
658 (S_ISLNK(sb.st_mode) &&
72b16652
GS
659 (g_stat(pathbuf, &sb, pglob) == 0) &&
660 S_ISDIR(sb.st_mode)))) {
b8ef571c
JH
661 if (pathend+1 > pathend_last)
662 return (1);
3e5d0dec
GS
663 *pathend++ = BG_SEP;
664 *pathend = BG_EOS;
72b16652
GS
665 }
666 ++pglob->gl_matchc;
667#ifdef GLOB_DEBUG
668 printf("calling globextend from glob2\n");
669#endif /* GLOB_DEBUG */
b8ef571c 670 return(globextend(pathbuf, pglob, limitp));
72b16652
GS
671 }
672
673 /* Find end of next segment, copy tentatively to pathend. */
674 q = pathend;
675 p = pattern;
220398a0
PM
676 while (*p != BG_EOS && *p != BG_SEP
677#ifdef DOSISH
678 && *p != BG_SEP2
679#endif
680 ) {
72b16652
GS
681 if (ismeta(*p))
682 anymeta = 1;
b8ef571c
JH
683 if (q+1 > pathend_last)
684 return (1);
72b16652
GS
685 *q++ = *p++;
686 }
687
688 if (!anymeta) { /* No expansion, do next segment. */
689 pathend = q;
690 pattern = p;
220398a0
PM
691 while (*pattern == BG_SEP
692#ifdef DOSISH
693 || *pattern == BG_SEP2
694#endif
b8ef571c
JH
695 ) {
696 if (pathend+1 > pathend_last)
697 return (1);
72b16652 698 *pathend++ = *pattern++;
b8ef571c
JH
699 }
700 } else
701 /* Need expansion, recurse. */
702 return(glob3(pathbuf, pathbuf_last, pathend,
703 pathend_last, pattern, pattern_last,
704 p, pattern_last, pglob, limitp));
72b16652
GS
705 }
706 /* NOTREACHED */
707}
708
709static int
b8ef571c
JH
710glob3(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
711 Char *pattern, Char *pattern_last,
712 Char *restpattern, Char *restpattern_last, glob_t *pglob, size_t *limitp)
72b16652
GS
713{
714 register Direntry_t *dp;
715 DIR *dirp;
716 int err;
220398a0 717 int nocase;
72b16652
GS
718 char buf[MAXPATHLEN];
719
720 /*
721 * The readdirfunc declaration can't be prototyped, because it is
722 * assigned, below, to two functions which are prototyped in glob.h
723 * and dirent.h as taking pointers to differently typed opaque
724 * structures.
725 */
cb359b41 726 Direntry_t *(*readdirfunc)(DIR*);
72b16652 727
b8ef571c
JH
728 if (pathend > pathend_last)
729 return (1);
3e5d0dec 730 *pathend = BG_EOS;
72b16652
GS
731 errno = 0;
732
f0963acb
GS
733#ifdef VMS
734 {
bac331f5
JH
735 Char *q = pathend;
736 if (q - pathbuf > 5) {
737 q -= 5;
738 if (q[0] == '.' &&
739 tolower(q[1]) == 'd' && tolower(q[2]) == 'i' &&
740 tolower(q[3]) == 'r' && q[4] == '/')
741 {
742 q[0] = '/';
743 q[1] = BG_EOS;
744 pathend = q+1;
745 }
746 }
f0963acb
GS
747 }
748#endif
72b16652
GS
749 if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
750 /* TODO: don't call for ENOENT or ENOTDIR? */
751 if (pglob->gl_errfunc) {
b8ef571c
JH
752 if (g_Ctoc(pathbuf, buf, sizeof(buf)))
753 return (GLOB_ABEND);
72b16652
GS
754 if (pglob->gl_errfunc(buf, errno) ||
755 (pglob->gl_flags & GLOB_ERR))
756 return (GLOB_ABEND);
757 }
758 return(0);
759 }
760
761 err = 0;
220398a0 762 nocase = ((pglob->gl_flags & GLOB_NOCASE) != 0);
72b16652
GS
763
764 /* Search directory for matching names. */
765 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
b8ef571c 766 readdirfunc = (Direntry_t *(*)(DIR *))pglob->gl_readdir;
72b16652 767 else
4f49e16e 768 readdirfunc = my_readdir;
72b16652
GS
769 while ((dp = (*readdirfunc)(dirp))) {
770 register U8 *sc;
771 register Char *dc;
772
3e5d0dec
GS
773 /* Initial BG_DOT must be matched literally. */
774 if (dp->d_name[0] == BG_DOT && *pattern != BG_DOT)
72b16652 775 continue;
b8ef571c
JH
776 dc = pathend;
777 sc = (U8 *) dp->d_name;
778 while (dc < pathend_last && (*dc++ = *sc++) != BG_EOS)
779 ;
780 if (dc >= pathend_last) {
781 *dc = BG_EOS;
782 err = 1;
783 break;
784 }
785
220398a0 786 if (!match(pathend, pattern, restpattern, nocase)) {
3e5d0dec 787 *pathend = BG_EOS;
72b16652
GS
788 continue;
789 }
b8ef571c
JH
790 err = glob2(pathbuf, pathbuf_last, --dc, pathend_last,
791 restpattern, restpattern_last, pglob, limitp);
72b16652
GS
792 if (err)
793 break;
794 }
795
796 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
797 (*pglob->gl_closedir)(dirp);
798 else
4f49e16e 799 PerlDir_close(dirp);
72b16652
GS
800 return(err);
801}
802
803
804/*
805 * Extend the gl_pathv member of a glob_t structure to accomodate a new item,
806 * add the new item, and update gl_pathc.
807 *
808 * This assumes the BSD realloc, which only copies the block when its size
809 * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
810 * behavior.
811 *
812 * Return 0 if new item added, error code if memory couldn't be allocated.
813 *
814 * Invariant of the glob_t structure:
815 * Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
816 * gl_pathv points to (gl_offs + gl_pathc + 1) items.
817 */
818static int
b8ef571c 819globextend(const Char *path, glob_t *pglob, size_t *limitp)
72b16652
GS
820{
821 register char **pathv;
822 register int i;
b8ef571c 823 STRLEN newsize, len;
72b16652
GS
824 char *copy;
825 const Char *p;
826
827#ifdef GLOB_DEBUG
828 printf("Adding ");
829 for (p = path; *p; p++)
830 (void)printf("%c", CHAR(*p));
831 printf("\n");
3e5d0dec 832#endif /* GLOB_DEBUG */
72b16652 833
b8ef571c 834 newsize = sizeof(*pathv) * (2 + pglob->gl_pathc + pglob->gl_offs);
4f49e16e 835 if (pglob->gl_pathv)
b8ef571c 836 pathv = Renew(pglob->gl_pathv,newsize,char*);
4f49e16e 837 else
b8ef571c
JH
838 New(0,pathv,newsize,char*);
839 if (pathv == NULL) {
840 if (pglob->gl_pathv) {
841 Safefree(pglob->gl_pathv);
842 pglob->gl_pathv = NULL;
843 }
72b16652 844 return(GLOB_NOSPACE);
b8ef571c 845 }
72b16652
GS
846
847 if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
848 /* first time around -- clear initial gl_offs items */
849 pathv += pglob->gl_offs;
850 for (i = pglob->gl_offs; --i >= 0; )
851 *--pathv = NULL;
852 }
853 pglob->gl_pathv = pathv;
854
855 for (p = path; *p++;)
b8ef571c
JH
856 ;
857 len = (STRLEN)(p - path);
bac331f5 858 *limitp += len;
4f49e16e
GS
859 New(0, copy, p-path, char);
860 if (copy != NULL) {
b8ef571c
JH
861 if (g_Ctoc(path, copy, len)) {
862 Safefree(copy);
863 return(GLOB_NOSPACE);
864 }
72b16652
GS
865 pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
866 }
867 pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
b8ef571c
JH
868
869 if ((pglob->gl_flags & GLOB_LIMIT) &&
870 newsize + *limitp >= ARG_MAX) {
871 errno = 0;
872 return(GLOB_NOSPACE);
873 }
bac331f5 874
72b16652
GS
875 return(copy == NULL ? GLOB_NOSPACE : 0);
876}
877
878
879/*
880 * pattern matching function for filenames. Each occurrence of the *
881 * pattern causes a recursion level.
882 */
883static int
220398a0 884match(register Char *name, register Char *pat, register Char *patend, int nocase)
72b16652
GS
885{
886 int ok, negate_range;
887 Char c, k;
888
889 while (pat < patend) {
890 c = *pat++;
891 switch (c & M_MASK) {
892 case M_ALL:
893 if (pat == patend)
894 return(1);
895 do
220398a0 896 if (match(name, pat, patend, nocase))
72b16652 897 return(1);
b8ef571c
JH
898 while (*name++ != BG_EOS)
899 ;
72b16652
GS
900 return(0);
901 case M_ONE:
3e5d0dec 902 if (*name++ == BG_EOS)
72b16652
GS
903 return(0);
904 break;
905 case M_SET:
906 ok = 0;
3e5d0dec 907 if ((k = *name++) == BG_EOS)
72b16652 908 return(0);
3e5d0dec 909 if ((negate_range = ((*pat & M_MASK) == M_NOT)) != BG_EOS)
72b16652
GS
910 ++pat;
911 while (((c = *pat++) & M_MASK) != M_END)
912 if ((*pat & M_MASK) == M_RNG) {
220398a0
PM
913 if (nocase) {
914 if (tolower(c) <= tolower(k) && tolower(k) <= tolower(pat[1]))
915 ok = 1;
916 } else {
917 if (c <= k && k <= pat[1])
918 ok = 1;
919 }
72b16652 920 pat += 2;
220398a0 921 } else if (nocase ? (tolower(c) == tolower(k)) : (c == k))
72b16652
GS
922 ok = 1;
923 if (ok == negate_range)
924 return(0);
925 break;
926 default:
220398a0
PM
927 k = *name++;
928 if (nocase ? (tolower(k) != tolower(c)) : (k != c))
72b16652
GS
929 return(0);
930 break;
931 }
932 }
3e5d0dec 933 return(*name == BG_EOS);
72b16652
GS
934}
935
936/* Free allocated data belonging to a glob_t structure. */
937void
938bsd_globfree(glob_t *pglob)
939{
940 register int i;
941 register char **pp;
942
943 if (pglob->gl_pathv != NULL) {
944 pp = pglob->gl_pathv + pglob->gl_offs;
945 for (i = pglob->gl_pathc; i--; ++pp)
946 if (*pp)
4f49e16e
GS
947 Safefree(*pp);
948 Safefree(pglob->gl_pathv);
b8ef571c 949 pglob->gl_pathv = NULL;
72b16652
GS
950 }
951}
952
953static DIR *
954g_opendir(register Char *str, glob_t *pglob)
955{
956 char buf[MAXPATHLEN];
957
7369a524
CN
958 if (!*str) {
959#ifdef MACOS_TRADITIONAL
960 strcpy(buf, ":");
961#else
72b16652 962 strcpy(buf, ".");
7369a524
CN
963#endif
964 } else {
b8ef571c
JH
965 if (g_Ctoc(str, buf, sizeof(buf)))
966 return(NULL);
7369a524 967 }
72b16652
GS
968
969 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
970 return((*pglob->gl_opendir)(buf));
bac331f5
JH
971
972 return(PerlDir_open(buf));
72b16652
GS
973}
974
72b16652
GS
975static int
976g_lstat(register Char *fn, Stat_t *sb, glob_t *pglob)
977{
978 char buf[MAXPATHLEN];
979
b8ef571c
JH
980 if (g_Ctoc(fn, buf, sizeof(buf)))
981 return(-1);
72b16652
GS
982 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
983 return((*pglob->gl_lstat)(buf, sb));
4f49e16e
GS
984#ifdef HAS_LSTAT
985 return(PerlLIO_lstat(buf, sb));
986#else
987 return(PerlLIO_stat(buf, sb));
72b16652 988#endif /* HAS_LSTAT */
4f49e16e 989}
72b16652
GS
990
991static int
992g_stat(register Char *fn, Stat_t *sb, glob_t *pglob)
993{
994 char buf[MAXPATHLEN];
995
b8ef571c
JH
996 if (g_Ctoc(fn, buf, sizeof(buf)))
997 return(-1);
72b16652
GS
998 if (pglob->gl_flags & GLOB_ALTDIRFUNC)
999 return((*pglob->gl_stat)(buf, sb));
4f49e16e 1000 return(PerlLIO_stat(buf, sb));
72b16652
GS
1001}
1002
1003static Char *
1004g_strchr(Char *str, int ch)
1005{
1006 do {
1007 if (*str == ch)
1008 return (str);
1009 } while (*str++);
1010 return (NULL);
1011}
1012
b8ef571c
JH
1013static int
1014g_Ctoc(register const Char *str, char *buf, STRLEN len)
72b16652 1015{
b8ef571c
JH
1016 while (len--) {
1017 if ((*buf++ = *str++) == BG_EOS)
1018 return (0);
1019 }
1020 return (1);
72b16652
GS
1021}
1022
1023#ifdef GLOB_DEBUG
1024static void
1025qprintf(const char *str, register Char *s)
1026{
1027 register Char *p;
1028
1029 (void)printf("%s:\n", str);
1030 for (p = s; *p; p++)
1031 (void)printf("%c", CHAR(*p));
1032 (void)printf("\n");
1033 for (p = s; *p; p++)
1034 (void)printf("%c", *p & M_PROTECT ? '"' : ' ');
1035 (void)printf("\n");
1036 for (p = s; *p; p++)
1037 (void)printf("%c", ismeta(*p) ? '_' : ' ');
1038 (void)printf("\n");
1039}
1040#endif /* GLOB_DEBUG */