Print this page
1097 glob(3c) needs to support non-POSIX options
| Split |
Close |
| Expand all |
| Collapse all |
--- old/usr/src/lib/libc/port/regex/glob.c
+++ new/usr/src/lib/libc/port/regex/glob.c
1 1 /*
2 2 * CDDL HEADER START
3 3 *
4 4 * The contents of this file are subject to the terms of the
5 5 * Common Development and Distribution License (the "License").
6 6 * You may not use this file except in compliance with the License.
7 7 *
8 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9 9 * or http://www.opensolaris.org/os/licensing.
10 10 * See the License for the specific language governing permissions
11 11 * and limitations under the License.
12 12 *
|
↓ open down ↓ |
12 lines elided |
↑ open up ↑ |
13 13 * When distributing Covered Code, include this CDDL HEADER in each
14 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15 15 * If applicable, add the following below this CDDL HEADER, with the
16 16 * fields enclosed by brackets "[]" replaced with your own identifying
17 17 * information: Portions Copyright [yyyy] [name of copyright owner]
18 18 *
19 19 * CDDL HEADER END
20 20 */
21 21
22 22 /*
23 - * Copyright 2008 Sun Microsystems, Inc. All rights reserved.
24 - * Use is subject to license terms.
23 + * Copyright (c) 2012 Gary Mills
25 24 */
26 25
26 +/* $OpenBSD: glob.c,v 1.39 2012/01/20 07:09:42 tedu Exp $ */
27 27 /*
28 - * This code is MKS code ported to Solaris originally with minimum
29 - * modifications so that upgrades from MKS would readily integrate.
30 - * The MKS basis for this modification was:
28 + * Copyright (c) 1989, 1993
29 + * The Regents of the University of California. All rights reserved.
31 30 *
32 - * $Id: glob.c 1.31 1994/04/07 22:50:43 mark
31 + * This code is derived from software contributed to Berkeley by
32 + * Guido van Rossum.
33 33 *
34 - * Additional modifications have been made to this code to make it
35 - * 64-bit clean.
34 + * Redistribution and use in source and binary forms, with or without
35 + * modification, are permitted provided that the following conditions
36 + * are met:
37 + * 1. Redistributions of source code must retain the above copyright
38 + * notice, this list of conditions and the following disclaimer.
39 + * 2. Redistributions in binary form must reproduce the above copyright
40 + * notice, this list of conditions and the following disclaimer in the
41 + * documentation and/or other materials provided with the distribution.
42 + * 3. Neither the name of the University nor the names of its contributors
43 + * may be used to endorse or promote products derived from this software
44 + * without specific prior written permission.
45 + *
46 + * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
47 + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
48 + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
49 + * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
50 + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
51 + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
52 + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
53 + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
54 + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
55 + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
56 + * SUCH DAMAGE.
36 57 */
37 58
38 59 /*
39 - * glob, globfree -- POSIX.2 compatible file name expansion routines.
60 + * glob(3) -- a superset of the one defined in POSIX 1003.2.
40 61 *
41 - * Copyright 1985, 1991 by Mortice Kern Systems Inc. All rights reserved.
62 + * The [!...] convention to negate a range is supported (SysV, Posix, ksh).
42 63 *
43 - * Written by Eric Gisin.
64 + * Optional extra services, controlled by flags not defined by POSIX:
65 + *
66 + * GLOB_QUOTE:
67 + * Escaping convention: \ inhibits any special meaning the following
68 + * character might have (except \ at end of string is retained).
69 + * GLOB_MAGCHAR:
70 + * Set in gl_flags if pattern contained a globbing character.
71 + * GLOB_NOMAGIC:
72 + * Same as GLOB_NOCHECK, but it will only append pattern if it did
73 + * not contain any magic characters. [Used in csh style globbing]
74 + * GLOB_ALTDIRFUNC:
75 + * Use alternately specified directory access functions.
76 + * GLOB_TILDE:
77 + * expand ~user/foo to the /home/dir/of/user/foo
78 + * GLOB_BRACE:
79 + * expand {1,2}{a,b} to 1a 1b 2a 2b
80 + * gl_matchc:
81 + * Number of matches in the current invocation of glob.
44 82 */
45 83
46 -#pragma ident "%Z%%M% %I% %E% SMI"
84 +#include <sys/param.h>
85 +#include <sys/stat.h>
47 86
48 -#pragma weak _glob = glob
49 -#pragma weak _globfree = globfree
50 -
51 -#include "lint.h"
52 -#include <stdio.h>
53 -#include <unistd.h>
87 +#include <ctype.h>
88 +#include <dirent.h>
89 +#include <errno.h>
90 +#include <glob.h>
54 91 #include <limits.h>
92 +#include <pwd.h>
93 +#include <stdio.h>
55 94 #include <stdlib.h>
56 95 #include <string.h>
57 -#include <dirent.h>
58 -#include <sys/stat.h>
59 -#include <glob.h>
60 -#include <errno.h>
61 -#include <fnmatch.h>
96 +#include <unistd.h>
62 97
63 -#define GLOB__CHECK 0x80 /* stat generated paths */
98 +#include "charclass.h"
64 99
65 -#define INITIAL 8 /* initial pathv allocation */
66 -#define NULLCPP ((char **)0) /* Null char ** */
67 -#define NAME_MAX 1024 /* something large */
100 +#define DOLLAR '$'
101 +#define DOT '.'
102 +#define EOS '\0'
103 +#define LBRACKET '['
104 +#define NOT '!'
105 +#define QUESTION '?'
106 +#define QUOTE '\\'
107 +#define RANGE '-'
108 +#define RBRACKET ']'
109 +#define SEP '/'
110 +#define STAR '*'
111 +#define TILDE '~'
112 +#define UNDERSCORE '_'
113 +#define LBRACE '{'
114 +#define RBRACE '}'
115 +#define SLASH '/'
116 +#define COMMA ','
68 117
69 -static int globit(size_t, const char *, glob_t *, int,
70 - int (*)(const char *, int), char **);
71 -static int pstrcmp(const void *, const void *);
72 -static int append(glob_t *, const char *);
118 +#ifndef DEBUG
73 119
120 +#define M_QUOTE 0x8000
121 +#define M_PROTECT 0x4000
122 +#define M_MASK 0xffff
123 +#define M_ASCII 0x00ff
124 +
125 +typedef ushort_t Char;
126 +
127 +#else
128 +
129 +#define M_QUOTE 0x80
130 +#define M_PROTECT 0x40
131 +#define M_MASK 0xff
132 +#define M_ASCII 0x7f
133 +
134 +typedef char Char;
135 +
136 +#endif
137 +
138 +
139 +#define CHAR(c) ((Char)((c)&M_ASCII))
140 +#define META(c) ((Char)((c)|M_QUOTE))
141 +#define M_ALL META('*')
142 +#define M_END META(']')
143 +#define M_NOT META('!')
144 +#define M_ONE META('?')
145 +#define M_RNG META('-')
146 +#define M_SET META('[')
147 +#define M_CLASS META(':')
148 +#define ismeta(c) (((c)&M_QUOTE) != 0)
149 +
150 +#define GLOB_LIMIT_MALLOC 65536
151 +#define GLOB_LIMIT_STAT 2048
152 +#define GLOB_LIMIT_READDIR 16384
153 +
154 +/* Limit of recursion during matching attempts. */
155 +#define GLOB_LIMIT_RECUR 64
156 +
157 +struct glob_lim {
158 + size_t glim_malloc;
159 + size_t glim_stat;
160 + size_t glim_readdir;
161 +};
162 +
163 +struct glob_path_stat {
164 + char *gps_path;
165 + struct stat *gps_stat;
166 +};
167 +
168 +static int compare(const void *, const void *);
169 +static int compare_gps(const void *, const void *);
170 +static int g_Ctoc(const Char *, char *, uint_t);
171 +static int g_lstat(Char *, struct stat *, glob_t *);
172 +static DIR *g_opendir(Char *, glob_t *);
173 +static Char *g_strchr(const Char *, int);
174 +static int g_strncmp(const Char *, const char *, size_t);
175 +static int g_stat(Char *, struct stat *, glob_t *);
176 +static int glob0(const Char *, glob_t *, struct glob_lim *,
177 + int (*)(const char *, int));
178 +static int glob1(Char *, Char *, glob_t *, struct glob_lim *,
179 + int (*)(const char *, int));
180 +static int glob2(Char *, Char *, Char *, Char *, Char *, Char *,
181 + glob_t *, struct glob_lim *,
182 + int (*)(const char *, int));
183 +static int glob3(Char *, Char *, Char *, Char *, Char *,
184 + Char *, Char *, glob_t *, struct glob_lim *,
185 + int (*)(const char *, int));
186 +static int globextend(const Char *, glob_t *, struct glob_lim *,
187 + struct stat *);
188 +static
189 +const Char *globtilde(const Char *, Char *, size_t, glob_t *);
190 +static int globexp1(const Char *, glob_t *, struct glob_lim *,
191 + int (*)(const char *, int));
192 +static int globexp2(const Char *, const Char *, glob_t *,
193 + struct glob_lim *, int (*)(const char *, int));
194 +static int match(Char *, Char *, Char *, int);
195 +#ifdef DEBUG
196 +static void qprintf(const char *, Char *);
197 +#endif
198 +
199 +int
200 +glob(const char *pattern, int flags, int (*errfunc)(const char *, int),
201 + glob_t *pglob)
202 +{
203 + const uchar_t *patnext;
204 + int c;
205 + Char *bufnext, *bufend, patbuf[MAXPATHLEN];
206 + struct glob_lim limit = { 0, 0, 0 };
207 +
208 + if (strnlen(pattern, PATH_MAX) == PATH_MAX)
209 + return (GLOB_NOMATCH);
210 +
211 + patnext = (uchar_t *)pattern;
212 + if (!(flags & GLOB_APPEND)) {
213 + pglob->gl_pathc = 0;
214 + pglob->gl_pathv = NULL;
215 + if ((flags & GLOB_KEEPSTAT) != 0)
216 + pglob->gl_statv = NULL;
217 + if (!(flags & GLOB_DOOFFS))
218 + pglob->gl_offs = 0;
219 + }
220 + pglob->gl_flags = flags & ~GLOB_MAGCHAR;
221 + pglob->gl_matchc = 0;
222 +
223 + if (pglob->gl_offs < 0 || pglob->gl_pathc < 0 ||
224 + pglob->gl_offs >= INT_MAX || pglob->gl_pathc >= INT_MAX ||
225 + pglob->gl_pathc >= INT_MAX - pglob->gl_offs - 1)
226 + return (GLOB_NOSPACE);
227 +
228 + bufnext = patbuf;
229 + bufend = bufnext + MAXPATHLEN - 1;
230 + if (flags & GLOB_NOESCAPE)
231 + while (bufnext < bufend && (c = *patnext++) != EOS)
232 + *bufnext++ = c;
233 + else {
234 + /* Protect the quoted characters. */
235 + while (bufnext < bufend && (c = *patnext++) != EOS)
236 + if (c == QUOTE) {
237 + if ((c = *patnext++) == EOS) {
238 + c = QUOTE;
239 + --patnext;
240 + }
241 + *bufnext++ = c | M_PROTECT;
242 + } else
243 + *bufnext++ = c;
244 + }
245 + *bufnext = EOS;
246 +
247 + if (flags & GLOB_BRACE)
248 + return (globexp1(patbuf, pglob, &limit, errfunc));
249 + else
250 + return (glob0(patbuf, pglob, &limit, errfunc));
251 +}
252 +
74 253 /*
75 - * Free all space consumed by glob.
254 + * Expand recursively a glob {} pattern. When there is no more expansion
255 + * invoke the standard globbing routine to glob the rest of the magic
256 + * characters
76 257 */
77 -void
78 -globfree(glob_t *gp)
258 +static int
259 +globexp1(const Char *pattern, glob_t *pglob, struct glob_lim *limitp,
260 + int (*errfunc)(const char *, int))
79 261 {
80 - size_t i;
262 + const Char* ptr = pattern;
81 263
82 - if (gp->gl_pathv == 0)
83 - return;
264 + /* Protect a single {}, for find(1), like csh */
265 + if (pattern[0] == LBRACE && pattern[1] == RBRACE && pattern[2] == EOS)
266 + return (glob0(pattern, pglob, limitp, errfunc));
84 267
85 - for (i = gp->gl_offs; i < gp->gl_offs + gp->gl_pathc; ++i)
86 - free(gp->gl_pathv[i]);
87 - free((void *)gp->gl_pathv);
268 + if ((ptr = (const Char *) g_strchr(ptr, LBRACE)) != NULL)
269 + return (globexp2(ptr, pattern, pglob, limitp, errfunc));
88 270
89 - gp->gl_pathc = 0;
90 - gp->gl_pathv = NULLCPP;
271 + return (glob0(pattern, pglob, limitp, errfunc));
91 272 }
92 273
274 +
93 275 /*
94 - * Do filename expansion.
276 + * Recursive brace globbing helper. Tries to expand a single brace.
277 + * If it succeeds then it invokes globexp1 with the new pattern.
278 + * If it fails then it tries to glob the rest of the pattern and returns.
95 279 */
96 -int
97 -glob(const char *pattern, int flags,
98 - int (*errfn)(const char *, int), glob_t *gp)
280 +static int
281 +globexp2(const Char *ptr, const Char *pattern, glob_t *pglob,
282 + struct glob_lim *limitp, int (*errfunc)(const char *, int))
99 283 {
100 - int rv;
101 - size_t i;
102 - size_t ipathc;
103 - char *path;
284 + int i, rv;
285 + Char *lm, *ls;
286 + const Char *pe, *pm, *pl;
287 + Char patbuf[MAXPATHLEN];
104 288
105 - if ((flags & GLOB_DOOFFS) == 0)
106 - gp->gl_offs = 0;
289 + /* copy part up to the brace */
290 + for (lm = patbuf, pm = pattern; pm != ptr; *lm++ = *pm++)
291 + ;
292 + *lm = EOS;
293 + ls = lm;
107 294
108 - if (!(flags & GLOB_APPEND)) {
109 - gp->gl_pathc = 0;
110 - gp->gl_pathn = gp->gl_offs + INITIAL;
111 - gp->gl_pathv = (char **)malloc(sizeof (char *) * gp->gl_pathn);
295 + /* Find the balanced brace */
296 + for (i = 0, pe = ++ptr; *pe; pe++)
297 + if (*pe == LBRACKET) {
298 + /* Ignore everything between [] */
299 + for (pm = pe++; *pe != RBRACKET && *pe != EOS; pe++)
300 + ;
301 + if (*pe == EOS) {
302 + /*
303 + * We could not find a matching RBRACKET.
304 + * Ignore and just look for RBRACE
305 + */
306 + pe = pm;
307 + }
308 + } else if (*pe == LBRACE)
309 + i++;
310 + else if (*pe == RBRACE) {
311 + if (i == 0)
312 + break;
313 + i--;
314 + }
112 315
113 - if (gp->gl_pathv == NULLCPP)
114 - return (GLOB_NOSPACE);
115 - gp->gl_pathp = gp->gl_pathv + gp->gl_offs;
316 + /* Non matching braces; just glob the pattern */
317 + if (i != 0 || *pe == EOS)
318 + return (glob0(patbuf, pglob, limitp, errfunc));
116 319
117 - for (i = 0; i < gp->gl_offs; ++i)
118 - gp->gl_pathv[i] = NULL;
320 + for (i = 0, pl = pm = ptr; pm <= pe; pm++) {
321 + switch (*pm) {
322 + case LBRACKET:
323 + /* Ignore everything between [] */
324 + for (pl = pm++; *pm != RBRACKET && *pm != EOS; pm++)
325 + ;
326 + if (*pm == EOS) {
327 + /*
328 + * We could not find a matching RBRACKET.
329 + * Ignore and just look for RBRACE
330 + */
331 + pm = pl;
332 + }
333 + break;
334 +
335 + case LBRACE:
336 + i++;
337 + break;
338 +
339 + case RBRACE:
340 + if (i) {
341 + i--;
342 + break;
343 + }
344 + /* FALLTHROUGH */
345 + case COMMA:
346 + if (i && *pm == COMMA)
347 + break;
348 + else {
349 + /* Append the current string */
350 + for (lm = ls; (pl < pm); *lm++ = *pl++)
351 + ;
352 +
353 + /*
354 + * Append the rest of the pattern after the
355 + * closing brace
356 + */
357 + for (pl = pe + 1; (*lm++ = *pl++) != EOS; )
358 + ;
359 +
360 + /* Expand the current pattern */
361 +#ifdef DEBUG
362 + qprintf("globexp2:", patbuf);
363 +#endif
364 + rv = globexp1(patbuf, pglob, limitp, errfunc);
365 + if (rv && rv != GLOB_NOMATCH)
366 + return (rv);
367 +
368 + /* move after the comma, to the next string */
369 + pl = pm + 1;
370 + }
371 + break;
372 +
373 + default:
374 + break;
375 + }
119 376 }
377 + return (0);
378 +}
120 379
121 - if ((path = malloc(strlen(pattern)+1)) == NULL)
122 - return (GLOB_NOSPACE);
123 380
124 - ipathc = gp->gl_pathc;
125 - rv = globit(0, pattern, gp, flags, errfn, &path);
126 381
127 - if (rv == GLOB_ABORTED) {
382 +/*
383 + * expand tilde from the passwd file.
384 + */
385 +static const Char *
386 +globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob)
387 +{
388 + struct passwd *pwd;
389 + char *h;
390 + const Char *p;
391 + Char *b, *eb;
392 +
393 + if (*pattern != TILDE || !(pglob->gl_flags & GLOB_TILDE))
394 + return (pattern);
395 +
396 + /* Copy up to the end of the string or / */
397 + eb = &patbuf[patbuf_len - 1];
398 + for (p = pattern + 1, h = (char *)patbuf;
399 + h < (char *)eb && *p && *p != SLASH; *h++ = *p++)
400 + ;
401 +
402 + *h = EOS;
403 +
404 +#if 0
405 + if (h == (char *)eb)
406 + return (what);
407 +#endif
408 +
409 + if (((char *)patbuf)[0] == EOS) {
128 410 /*
129 - * User's error function returned non-zero, or GLOB_ERR was
130 - * set, and we encountered a directory we couldn't search.
411 + * handle a plain ~ or ~/ by expanding $HOME
412 + * first and then trying the password file
131 413 */
132 - free(path);
133 - return (GLOB_ABORTED);
134 - }
135 -
136 - i = gp->gl_pathc - ipathc;
137 - if (i >= 1 && !(flags & GLOB_NOSORT)) {
138 - qsort((char *)(gp->gl_pathp+ipathc), i, sizeof (char *),
139 - pstrcmp);
140 - }
141 - if (i == 0) {
142 - if (flags & GLOB_NOCHECK)
143 - (void) append(gp, pattern);
414 + if (issetugid() != 0 || (h = getenv("HOME")) == NULL) {
415 + if ((pwd = getpwuid(getuid())) == NULL)
416 + return (pattern);
417 + else
418 + h = pwd->pw_dir;
419 + }
420 + } else {
421 + /*
422 + * Expand a ~user
423 + */
424 + if ((pwd = getpwnam((char *)patbuf)) == NULL)
425 + return (pattern);
144 426 else
145 - rv = GLOB_NOMATCH;
427 + h = pwd->pw_dir;
146 428 }
147 - gp->gl_pathp[gp->gl_pathc] = NULL;
148 - free(path);
149 429
430 + /* Copy the home directory */
431 + for (b = patbuf; b < eb && *h; *b++ = *h++)
432 + ;
433 +
434 + /* Append the rest of the pattern */
435 + while (b < eb && (*b++ = *p++) != EOS)
436 + ;
437 + *b = EOS;
438 +
439 + return (patbuf);
440 +}
441 +
442 +static int
443 +g_strncmp(const Char *s1, const char *s2, size_t n)
444 +{
445 + int rv = 0;
446 +
447 + while (n--) {
448 + rv = *(Char *)s1 - *(const unsigned char *)s2++;
449 + if (rv)
450 + break;
451 + if (*s1++ == '\0')
452 + break;
453 + }
150 454 return (rv);
151 455 }
152 456
457 +static int
458 +g_charclass(const Char **patternp, Char **bufnextp)
459 +{
460 + const Char *pattern = *patternp + 1;
461 + Char *bufnext = *bufnextp;
462 + const Char *colon;
463 + struct cclass *cc;
464 + size_t len;
153 465
466 + if ((colon = g_strchr(pattern, ':')) == NULL || colon[1] != ']')
467 + return (1); /* not a character class */
468 +
469 + len = (size_t)(colon - pattern);
470 + for (cc = cclasses; cc->name != NULL; cc++) {
471 + if (!g_strncmp(pattern, cc->name, len) && cc->name[len] == '\0')
472 + break;
473 + }
474 + if (cc->name == NULL)
475 + return (-1); /* invalid character class */
476 + *bufnext++ = M_CLASS;
477 + *bufnext++ = (Char)(cc - &cclasses[0]);
478 + *bufnextp = bufnext;
479 + *patternp += len + 3;
480 +
481 + return (0);
482 +}
483 +
154 484 /*
155 - * Recursive routine to match glob pattern, and walk directories.
485 + * The main glob() routine: compiles the pattern (optionally processing
486 + * quotes), calls glob1() to do the real pattern matching, and finally
487 + * sorts the list (unless unsorted operation is requested). Returns 0
488 + * if things went well, nonzero if errors occurred. It is not an error
489 + * to find no matches.
156 490 */
157 -int
158 -globit(size_t dend, const char *sp, glob_t *gp, int flags,
159 - int (*errfn)(const char *, int), char **path)
491 +static int
492 +glob0(const Char *pattern, glob_t *pglob, struct glob_lim *limitp,
493 + int (*errfunc)(const char *, int))
160 494 {
161 - size_t n;
162 - size_t m;
163 - ssize_t end = 0; /* end of expanded directory */
164 - char *pat = (char *)sp; /* pattern component */
165 - char *dp = (*path) + dend;
166 - int expand = 0; /* path has pattern */
167 - char *cp;
168 - struct stat64 sb;
169 - DIR *dirp;
170 - struct dirent64 *d;
171 - int err;
495 + const Char *qpatnext;
496 + int c, err, oldpathc;
497 + Char *bufnext, patbuf[MAXPATHLEN];
172 498
173 - for (;;)
174 - switch (*dp++ = *(unsigned char *)sp++) {
175 - case '\0': /* end of source path */
176 - if (expand)
177 - goto Expand;
178 - else {
179 - if (!(flags & GLOB_NOCHECK) ||
180 - flags & (GLOB__CHECK|GLOB_MARK))
181 - if (stat64(*path, &sb) < 0) {
182 - return (0);
183 - }
184 - if (flags & GLOB_MARK && S_ISDIR(sb.st_mode)) {
185 - *dp = '\0';
186 - *--dp = '/';
499 + qpatnext = globtilde(pattern, patbuf, MAXPATHLEN, pglob);
500 + oldpathc = pglob->gl_pathc;
501 + bufnext = patbuf;
502 +
503 + /* We don't need to check for buffer overflow any more. */
504 + while ((c = *qpatnext++) != EOS) {
505 + switch (c) {
506 + case LBRACKET:
507 + c = *qpatnext;
508 + if (c == NOT)
509 + ++qpatnext;
510 + if (*qpatnext == EOS ||
511 + g_strchr(qpatnext+1, RBRACKET) == NULL) {
512 + *bufnext++ = LBRACKET;
513 + if (c == NOT)
514 + --qpatnext;
515 + break;
516 + }
517 + *bufnext++ = M_SET;
518 + if (c == NOT)
519 + *bufnext++ = M_NOT;
520 + c = *qpatnext++;
521 + do {
522 + if (c == LBRACKET && *qpatnext == ':') {
523 + do {
524 + err = g_charclass(&qpatnext,
525 + &bufnext);
526 + if (err)
527 + break;
528 + c = *qpatnext++;
529 + } while (c == LBRACKET &&
530 + *qpatnext == ':');
531 + if (err == -1 &&
532 + !(pglob->gl_flags & GLOB_NOCHECK))
533 + return (GLOB_NOMATCH);
534 + if (c == RBRACKET)
535 + break;
187 536 }
188 - if (append(gp, *path) < 0) {
189 - return (GLOB_NOSPACE);
537 + *bufnext++ = CHAR(c);
538 + if (*qpatnext == RANGE &&
539 + (c = qpatnext[1]) != RBRACKET) {
540 + *bufnext++ = M_RNG;
541 + *bufnext++ = CHAR(c);
542 + qpatnext += 2;
190 543 }
191 - return (0);
192 - }
193 - /*NOTREACHED*/
194 -
195 - case '*':
196 - case '?':
197 - case '[':
198 - case '\\':
199 - ++expand;
544 + } while ((c = *qpatnext++) != RBRACKET);
545 + pglob->gl_flags |= GLOB_MAGCHAR;
546 + *bufnext++ = M_END;
200 547 break;
201 -
202 - case '/':
203 - if (expand)
204 - goto Expand;
205 - end = dp - *path;
206 - pat = (char *)sp;
548 + case QUESTION:
549 + pglob->gl_flags |= GLOB_MAGCHAR;
550 + *bufnext++ = M_ONE;
207 551 break;
552 + case STAR:
553 + pglob->gl_flags |= GLOB_MAGCHAR;
554 + /*
555 + * collapse adjacent stars to one,
556 + * to avoid exponential behavior
557 + */
558 + if (bufnext == patbuf || bufnext[-1] != M_ALL)
559 + *bufnext++ = M_ALL;
560 + break;
561 + default:
562 + *bufnext++ = CHAR(c);
563 + break;
564 + }
565 + }
566 + *bufnext = EOS;
567 +#ifdef DEBUG
568 + qprintf("glob0:", patbuf);
569 +#endif
208 570
209 - Expand:
210 - /* determine directory and open it */
211 - (*path)[end] = '\0';
212 - dirp = opendir(**path == '\0' ? "." : *path);
213 - if (dirp == NULL) {
214 - if (errfn != 0 && errfn(*path, errno) != 0 ||
215 - flags&GLOB_ERR) {
216 - return (GLOB_ABORTED);
217 - }
218 - return (0);
219 - }
571 + if ((err = glob1(patbuf, patbuf+MAXPATHLEN-1, pglob, limitp, errfunc))
572 + != 0)
573 + return (err);
220 574
221 - /* extract pattern component */
222 - n = sp - pat;
223 - if ((cp = malloc(n)) == NULL) {
224 - (void) closedir(dirp);
575 + /*
576 + * If there was no match we are going to append the pattern
577 + * if GLOB_NOCHECK was specified or if GLOB_NOMAGIC was specified
578 + * and the pattern did not contain any magic characters
579 + * GLOB_NOMAGIC is there just for compatibility with csh.
580 + */
581 + if (pglob->gl_pathc == oldpathc) {
582 + if ((pglob->gl_flags & GLOB_NOCHECK) ||
583 + ((pglob->gl_flags & GLOB_NOMAGIC) &&
584 + !(pglob->gl_flags & GLOB_MAGCHAR)))
585 + return (globextend(pattern, pglob, limitp, NULL));
586 + else
587 + return (GLOB_NOMATCH);
588 + }
589 + if (!(pglob->gl_flags & GLOB_NOSORT)) {
590 + if ((pglob->gl_flags & GLOB_KEEPSTAT)) {
591 + /* Keep the paths and stat info synced during sort */
592 + struct glob_path_stat *path_stat;
593 + int i;
594 + int n = pglob->gl_pathc - oldpathc;
595 + int o = pglob->gl_offs + oldpathc;
596 +
597 + if ((path_stat = calloc(n, sizeof (*path_stat))) ==
598 + NULL)
225 599 return (GLOB_NOSPACE);
600 + for (i = 0; i < n; i++) {
601 + path_stat[i].gps_path = pglob->gl_pathv[o + i];
602 + path_stat[i].gps_stat = pglob->gl_statv[o + i];
226 603 }
227 - pat = memcpy(cp, pat, n);
228 - pat[n-1] = '\0';
229 - if (*--sp != '\0')
230 - flags |= GLOB__CHECK;
604 + qsort(path_stat, n, sizeof (*path_stat), compare_gps);
605 + for (i = 0; i < n; i++) {
606 + pglob->gl_pathv[o + i] = path_stat[i].gps_path;
607 + pglob->gl_statv[o + i] = path_stat[i].gps_stat;
608 + }
609 + free(path_stat);
610 + } else {
611 + qsort(pglob->gl_pathv + pglob->gl_offs + oldpathc,
612 + pglob->gl_pathc - oldpathc, sizeof (char *),
613 + compare);
614 + }
615 + }
616 + return (0);
617 +}
231 618
232 - /* expand path to max. expansion */
233 - n = dp - *path;
234 - *path = realloc(*path,
235 - strlen(*path) + NAME_MAX + strlen(sp) + 1);
236 - if (*path == NULL) {
237 - (void) closedir(dirp);
238 - free(pat);
619 +static int
620 +compare(const void *p, const void *q)
621 +{
622 + return (strcmp(*(char **)p, *(char **)q));
623 +}
624 +
625 +static int
626 +compare_gps(const void *_p, const void *_q)
627 +{
628 + const struct glob_path_stat *p = (const struct glob_path_stat *)_p;
629 + const struct glob_path_stat *q = (const struct glob_path_stat *)_q;
630 +
631 + return (strcmp(p->gps_path, q->gps_path));
632 +}
633 +
634 +static int
635 +glob1(Char *pattern, Char *pattern_last, glob_t *pglob,
636 + struct glob_lim *limitp, int (*errfunc)(const char *, int))
637 +{
638 + Char pathbuf[MAXPATHLEN];
639 +
640 + /* A null pathname is invalid -- POSIX 1003.1 sect. 2.4. */
641 + if (*pattern == EOS)
642 + return (0);
643 + return (glob2(pathbuf, pathbuf+MAXPATHLEN-1,
644 + pathbuf, pathbuf+MAXPATHLEN-1,
645 + pattern, pattern_last, pglob, limitp, errfunc));
646 +}
647 +
648 +/*
649 + * The functions glob2 and glob3 are mutually recursive; there is one level
650 + * of recursion for each segment in the pattern that contains one or more
651 + * meta characters.
652 + */
653 +static int
654 +glob2(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
655 + Char *pattern, Char *pattern_last, glob_t *pglob,
656 + struct glob_lim *limitp, int (*errfunc)(const char *, int))
657 +{
658 + struct stat sb;
659 + Char *p, *q;
660 + int anymeta;
661 +
662 + /*
663 + * Loop over pattern segments until end of pattern or until
664 + * segment with meta character found.
665 + */
666 + for (anymeta = 0; ; ) {
667 + if (*pattern == EOS) { /* End of pattern? */
668 + *pathend = EOS;
669 +
670 + if ((pglob->gl_flags & GLOB_LIMIT) &&
671 + limitp->glim_stat++ >= GLOB_LIMIT_STAT) {
672 + errno = 0;
673 + *pathend++ = SEP;
674 + *pathend = EOS;
239 675 return (GLOB_NOSPACE);
240 676 }
241 - dp = (*path) + n;
677 + if (g_lstat(pathbuf, &sb, pglob))
678 + return (0);
242 679
243 - /* read directory and match entries */
244 - err = 0;
245 - while ((d = readdir64(dirp)) != NULL) {
246 - cp = d->d_name;
247 - if ((flags&GLOB_NOESCAPE)
248 - ? fnmatch(pat, cp, FNM_PERIOD|FNM_NOESCAPE)
249 - : fnmatch(pat, cp, FNM_PERIOD))
250 - continue;
680 + if (((pglob->gl_flags & GLOB_MARK) &&
681 + pathend[-1] != SEP) && (S_ISDIR(sb.st_mode) ||
682 + (S_ISLNK(sb.st_mode) &&
683 + (g_stat(pathbuf, &sb, pglob) == 0) &&
684 + S_ISDIR(sb.st_mode)))) {
685 + if (pathend+1 > pathend_last)
686 + return (1);
687 + *pathend++ = SEP;
688 + *pathend = EOS;
689 + }
690 + ++pglob->gl_matchc;
691 + return (globextend(pathbuf, pglob, limitp, &sb));
692 + }
251 693
252 - n = strlen(cp);
253 - (void) memcpy((*path) + end, cp, n);
254 - m = dp - *path;
255 - err = globit(end+n, sp, gp, flags, errfn, path);
256 - dp = (*path) + m; /* globit can move path */
257 - if (err != 0)
258 - break;
694 + /* Find end of next segment, copy tentatively to pathend. */
695 + q = pathend;
696 + p = pattern;
697 + while (*p != EOS && *p != SEP) {
698 + if (ismeta(*p))
699 + anymeta = 1;
700 + if (q+1 > pathend_last)
701 + return (1);
702 + *q++ = *p++;
703 + }
704 +
705 + if (!anymeta) { /* No expansion, do next segment. */
706 + pathend = q;
707 + pattern = p;
708 + while (*pattern == SEP) {
709 + if (pathend+1 > pathend_last)
710 + return (1);
711 + *pathend++ = *pattern++;
259 712 }
713 + } else
714 + /* Need expansion, recurse. */
715 + return (glob3(pathbuf, pathbuf_last, pathend,
716 + pathend_last, pattern, p, pattern_last,
717 + pglob, limitp, errfunc));
718 + }
719 + /* NOTREACHED */
720 +}
260 721
261 - (void) closedir(dirp);
262 - free(pat);
263 - return (err);
722 +static int
723 +glob3(Char *pathbuf, Char *pathbuf_last, Char *pathend, Char *pathend_last,
724 + Char *pattern, Char *restpattern, Char *restpattern_last, glob_t *pglob,
725 + struct glob_lim *limitp, int (*errfunc)(const char *, int))
726 +{
727 + struct dirent *dp;
728 + DIR *dirp;
729 + int err;
730 + char buf[MAXPATHLEN];
731 + struct dirent *(*readdirfunc)(DIR *);
732 +
733 + if (pathend > pathend_last)
734 + return (1);
735 + *pathend = EOS;
736 + errno = 0;
737 +
738 + if ((dirp = g_opendir(pathbuf, pglob)) == NULL) {
739 + /* TODO: don't call for ENOENT or ENOTDIR? */
740 + if (errfunc) {
741 + if (g_Ctoc(pathbuf, buf, sizeof (buf)))
742 + return (GLOB_ABORTED);
743 + if (errfunc(buf, errno) ||
744 + pglob->gl_flags & GLOB_ERR)
745 + return (GLOB_ABORTED);
264 746 }
265 - /* NOTREACHED */
747 + return (0);
748 + }
749 +
750 + err = 0;
751 +
752 + /* Search directory for matching names. */
753 + if (pglob->gl_flags & GLOB_ALTDIRFUNC)
754 + readdirfunc = pglob->gl_readdir;
755 + else
756 + readdirfunc = readdir;
757 + while ((dp = (*readdirfunc)(dirp))) {
758 + uchar_t *sc;
759 + Char *dc;
760 +
761 + if ((pglob->gl_flags & GLOB_LIMIT) &&
762 + limitp->glim_readdir++ >= GLOB_LIMIT_READDIR) {
763 + errno = 0;
764 + *pathend++ = SEP;
765 + *pathend = EOS;
766 + err = GLOB_NOSPACE;
767 + break;
768 + }
769 +
770 + /* Initial DOT must be matched literally. */
771 + if (dp->d_name[0] == DOT && *pattern != DOT)
772 + continue;
773 + dc = pathend;
774 + sc = (uchar_t *)dp->d_name;
775 + while (dc < pathend_last && (*dc++ = *sc++) != EOS)
776 + ;
777 + if (dc >= pathend_last) {
778 + *dc = EOS;
779 + err = 1;
780 + break;
781 + }
782 +
783 + if (!match(pathend, pattern, restpattern, GLOB_LIMIT_RECUR)) {
784 + *pathend = EOS;
785 + continue;
786 + }
787 + err = glob2(pathbuf, pathbuf_last, --dc, pathend_last,
788 + restpattern, restpattern_last, pglob, limitp,
789 + errfunc);
790 + if (err)
791 + break;
792 + }
793 +
794 + if (pglob->gl_flags & GLOB_ALTDIRFUNC)
795 + (*pglob->gl_closedir)(dirp);
796 + else
797 + closedir(dirp);
798 + return (err);
266 799 }
267 800
801 +
268 802 /*
269 - * Comparison routine for two name arguments, called by qsort.
803 + * Extend the gl_pathv member of a glob_t structure to accommodate a new item,
804 + * add the new item, and update gl_pathc.
805 + *
806 + * This assumes the BSD realloc, which only copies the block when its size
807 + * crosses a power-of-two boundary; for v7 realloc, this would cause quadratic
808 + * behavior.
809 + *
810 + * Return 0 if new item added, error code if memory couldn't be allocated.
811 + *
812 + * Invariant of the glob_t structure:
813 + * Either gl_pathc is zero and gl_pathv is NULL; or gl_pathc > 0 and
814 + * gl_pathv points to (gl_offs + gl_pathc + 1) items.
270 815 */
271 -int
272 -pstrcmp(const void *npp1, const void *npp2)
816 +static int
817 +globextend(const Char *path, glob_t *pglob, struct glob_lim *limitp,
818 + struct stat *sb)
273 819 {
274 - return (strcoll(*(char **)npp1, *(char **)npp2));
820 + char **pathv;
821 + ssize_t i;
822 + size_t newn, len;
823 + char *copy = NULL;
824 + const Char *p;
825 + struct stat **statv;
826 +
827 + newn = 2 + pglob->gl_pathc + pglob->gl_offs;
828 + if (pglob->gl_offs >= INT_MAX ||
829 + pglob->gl_pathc >= INT_MAX ||
830 + newn >= INT_MAX ||
831 + SIZE_MAX / sizeof (*pathv) <= newn ||
832 + SIZE_MAX / sizeof (*statv) <= newn) {
833 + nospace:
834 + for (i = pglob->gl_offs; i < (ssize_t)(newn - 2); i++) {
835 + if (pglob->gl_pathv && pglob->gl_pathv[i])
836 + free(pglob->gl_pathv[i]);
837 + if ((pglob->gl_flags & GLOB_KEEPSTAT) != 0 &&
838 + pglob->gl_pathv && pglob->gl_pathv[i])
839 + free(pglob->gl_statv[i]);
840 + }
841 + if (pglob->gl_pathv) {
842 + free(pglob->gl_pathv);
843 + pglob->gl_pathv = NULL;
844 + }
845 + if ((pglob->gl_flags & GLOB_KEEPSTAT) != 0 &&
846 + pglob->gl_statv) {
847 + free(pglob->gl_statv);
848 + pglob->gl_statv = NULL;
849 + }
850 + return (GLOB_NOSPACE);
851 + }
852 +
853 + pathv = realloc(pglob->gl_pathv, newn * sizeof (*pathv));
854 + if (pathv == NULL)
855 + goto nospace;
856 + if (pglob->gl_pathv == NULL && pglob->gl_offs > 0) {
857 + /* first time around -- clear initial gl_offs items */
858 + pathv += pglob->gl_offs;
859 + for (i = pglob->gl_offs; --i >= 0; )
860 + *--pathv = NULL;
861 + }
862 + pglob->gl_pathv = pathv;
863 +
864 + if ((pglob->gl_flags & GLOB_KEEPSTAT) != 0) {
865 + statv = realloc(pglob->gl_statv, newn * sizeof (*statv));
866 + if (statv == NULL)
867 + goto nospace;
868 + if (pglob->gl_statv == NULL && pglob->gl_offs > 0) {
869 + /* first time around -- clear initial gl_offs items */
870 + statv += pglob->gl_offs;
871 + for (i = pglob->gl_offs; --i >= 0; )
872 + *--statv = NULL;
873 + }
874 + pglob->gl_statv = statv;
875 + if (sb == NULL)
876 + statv[pglob->gl_offs + pglob->gl_pathc] = NULL;
877 + else {
878 + limitp->glim_malloc += sizeof (**statv);
879 + if ((pglob->gl_flags & GLOB_LIMIT) &&
880 + limitp->glim_malloc >= GLOB_LIMIT_MALLOC) {
881 + errno = 0;
882 + return (GLOB_NOSPACE);
883 + }
884 + if ((statv[pglob->gl_offs + pglob->gl_pathc] =
885 + malloc(sizeof (**statv))) == NULL)
886 + goto copy_error;
887 + memcpy(statv[pglob->gl_offs + pglob->gl_pathc], sb,
888 + sizeof (*sb));
889 + }
890 + statv[pglob->gl_offs + pglob->gl_pathc + 1] = NULL;
891 + }
892 +
893 + for (p = path; *p++; )
894 + ;
895 + len = (size_t)(p - path);
896 + limitp->glim_malloc += len;
897 + if ((copy = malloc(len)) != NULL) {
898 + if (g_Ctoc(path, copy, len)) {
899 + free(copy);
900 + return (GLOB_NOSPACE);
901 + }
902 + pathv[pglob->gl_offs + pglob->gl_pathc++] = copy;
903 + }
904 + pathv[pglob->gl_offs + pglob->gl_pathc] = NULL;
905 +
906 + if ((pglob->gl_flags & GLOB_LIMIT) &&
907 + (newn * sizeof (*pathv)) + limitp->glim_malloc >
908 + GLOB_LIMIT_MALLOC) {
909 + errno = 0;
910 + return (GLOB_NOSPACE);
911 + }
912 + copy_error:
913 + return (copy == NULL ? GLOB_NOSPACE : 0);
275 914 }
276 915
916 +
277 917 /*
278 - * Add a new matched filename to the glob_t structure, increasing the
279 - * size of that array, as required.
918 + * pattern matching function for filenames. Each occurrence of the *
919 + * pattern causes a recursion level.
280 920 */
281 -int
282 -append(glob_t *gp, const char *str)
921 +static int
922 +match(Char *name, Char *pat, Char *patend, int recur)
283 923 {
284 - char *cp;
924 + int ok, negate_range;
925 + Char c, k;
285 926
286 - if ((cp = malloc(strlen(str)+1)) == NULL)
927 + if (recur-- == 0)
287 928 return (GLOB_NOSPACE);
288 - gp->gl_pathp[gp->gl_pathc++] = strcpy(cp, str);
289 929
290 - if ((gp->gl_pathc + gp->gl_offs) >= gp->gl_pathn) {
291 - gp->gl_pathn *= 2;
292 - gp->gl_pathv = (char **)realloc((void *)gp->gl_pathv,
293 - gp->gl_pathn * sizeof (char *));
294 - if (gp->gl_pathv == NULLCPP)
295 - return (GLOB_NOSPACE);
296 - gp->gl_pathp = gp->gl_pathv + gp->gl_offs;
930 + while (pat < patend) {
931 + c = *pat++;
932 + switch (c & M_MASK) {
933 + case M_ALL:
934 + while (pat < patend && (*pat & M_MASK) == M_ALL)
935 + pat++; /* eat consecutive '*' */
936 + if (pat == patend)
937 + return (1);
938 + do {
939 + if (match(name, pat, patend, recur))
940 + return (1);
941 + } while (*name++ != EOS);
942 + return (0);
943 + case M_ONE:
944 + if (*name++ == EOS)
945 + return (0);
946 + break;
947 + case M_SET:
948 + ok = 0;
949 + if ((k = *name++) == EOS)
950 + return (0);
951 + if ((negate_range = ((*pat & M_MASK) == M_NOT)) != EOS)
952 + ++pat;
953 + while (((c = *pat++) & M_MASK) != M_END) {
954 + if ((c & M_MASK) == M_CLASS) {
955 + Char idx = *pat & M_MASK;
956 + if (idx < NCCLASSES &&
957 + cclasses[idx].isctype(k))
958 + ok = 1;
959 + ++pat;
960 + }
961 + if ((*pat & M_MASK) == M_RNG) {
962 + if (c <= k && k <= pat[1])
963 + ok = 1;
964 + pat += 2;
965 + } else if (c == k)
966 + ok = 1;
967 + }
968 + if (ok == negate_range)
969 + return (0);
970 + break;
971 + default:
972 + if (*name++ != c)
973 + return (0);
974 + break;
975 + }
297 976 }
298 - return (0);
977 + return (*name == EOS);
299 978 }
979 +
980 +/* Free allocated data belonging to a glob_t structure. */
981 +void
982 +globfree(glob_t *pglob)
983 +{
984 + int i;
985 + char **pp;
986 +
987 + if (pglob->gl_pathv != NULL) {
988 + pp = pglob->gl_pathv + pglob->gl_offs;
989 + for (i = pglob->gl_pathc; i--; ++pp)
990 + if (*pp)
991 + free(*pp);
992 + free(pglob->gl_pathv);
993 + pglob->gl_pathv = NULL;
994 + }
995 + if ((pglob->gl_flags & GLOB_KEEPSTAT) != 0 &&
996 + pglob->gl_statv != NULL) {
997 + for (i = 0; i < pglob->gl_pathc; i++) {
998 + if (pglob->gl_statv[i] != NULL)
999 + free(pglob->gl_statv[i]);
1000 + }
1001 + free(pglob->gl_statv);
1002 + pglob->gl_statv = NULL;
1003 + }
1004 +}
1005 +
1006 +static DIR *
1007 +g_opendir(Char *str, glob_t *pglob)
1008 +{
1009 + char buf[MAXPATHLEN];
1010 +
1011 + if (!*str)
1012 + strlcpy(buf, ".", sizeof (buf));
1013 + else {
1014 + if (g_Ctoc(str, buf, sizeof (buf)))
1015 + return (NULL);
1016 + }
1017 +
1018 + if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1019 + return ((*pglob->gl_opendir)(buf));
1020 +
1021 + return (opendir(buf));
1022 +}
1023 +
1024 +static int
1025 +g_lstat(Char *fn, struct stat *sb, glob_t *pglob)
1026 +{
1027 + char buf[MAXPATHLEN];
1028 +
1029 + if (g_Ctoc(fn, buf, sizeof (buf)))
1030 + return (-1);
1031 + if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1032 + return ((*pglob->gl_lstat)(buf, sb));
1033 + return (lstat(buf, sb));
1034 +}
1035 +
1036 +static int
1037 +g_stat(Char *fn, struct stat *sb, glob_t *pglob)
1038 +{
1039 + char buf[MAXPATHLEN];
1040 +
1041 + if (g_Ctoc(fn, buf, sizeof (buf)))
1042 + return (-1);
1043 + if (pglob->gl_flags & GLOB_ALTDIRFUNC)
1044 + return ((*pglob->gl_stat)(buf, sb));
1045 + return (stat(buf, sb));
1046 +}
1047 +
1048 +static Char *
1049 +g_strchr(const Char *str, int ch)
1050 +{
1051 + do {
1052 + if (*str == ch)
1053 + return ((Char *)str);
1054 + } while (*str++);
1055 + return (NULL);
1056 +}
1057 +
1058 +static int
1059 +g_Ctoc(const Char *str, char *buf, uint_t len)
1060 +{
1061 +
1062 + while (len--) {
1063 + if ((*buf++ = *str++) == EOS)
1064 + return (0);
1065 + }
1066 + return (1);
1067 +}
1068 +
1069 +#ifdef DEBUG
1070 +static void
1071 +qprintf(const char *str, Char *s)
1072 +{
1073 + Char *p;
1074 +
1075 + (void) printf("%s:\n", str);
1076 + for (p = s; *p; p++)
1077 + (void) printf("%c", CHAR(*p));
1078 + (void) printf("\n");
1079 + for (p = s; *p; p++)
1080 + (void) printf("%c", *p & M_PROTECT ? '"' : ' ');
1081 + (void) printf("\n");
1082 + for (p = s; *p; p++)
1083 + (void) printf("%c", ismeta(*p) ? '_' : ' ');
1084 + (void) printf("\n");
1085 +}
1086 +#endif
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX