This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Add :bsd_glob export tag to File::Glob [perl #96116]
[perl5.git] / ext / File-Glob / Glob.pm
CommitLineData
72b16652
GS
1package File::Glob;
2
3use strict;
7f39e0ae 4our($VERSION, @ISA, @EXPORT_OK, @EXPORT_FAIL, %EXPORT_TAGS, $DEFAULT_FLAGS);
72b16652 5
da4061d3 6require XSLoader;
22bc907a 7use feature 'switch';
72b16652 8
72f7b9a1 9@ISA = qw(Exporter);
72b16652 10
00c80938
GS
11# NOTE: The glob() export is only here for compatibility with 5.6.0.
12# csh_glob() should not be used directly, unless you know what you're doing.
13
72b16652
GS
14%EXPORT_TAGS = (
15 'glob' => [ qw(
16 GLOB_ABEND
2d5e9e5d 17 GLOB_ALPHASORT
72b16652
GS
18 GLOB_ALTDIRFUNC
19 GLOB_BRACE
220398a0 20 GLOB_CSH
72b16652
GS
21 GLOB_ERR
22 GLOB_ERROR
b8ef571c 23 GLOB_LIMIT
72b16652 24 GLOB_MARK
220398a0 25 GLOB_NOCASE
72b16652
GS
26 GLOB_NOCHECK
27 GLOB_NOMAGIC
28 GLOB_NOSORT
29 GLOB_NOSPACE
30 GLOB_QUOTE
31 GLOB_TILDE
00c80938 32 bsd_glob
f4cbf990 33 glob
72b16652
GS
34 ) ],
35);
f4cbf990
FC
36$EXPORT_TAGS{bsd_glob} = [@{$EXPORT_TAGS{glob}}];
37pop @{$EXPORT_TAGS{bsd_glob}}; # no "glob"
72b16652 38
aa0c903b
NC
39@EXPORT_OK = (@{$EXPORT_TAGS{'glob'}}, 'csh_glob');
40
0b0e6d70 41$VERSION = '1.14';
220398a0
PM
42
43sub import {
7d3fb230 44 require Exporter;
df5a3819
NC
45 local $Exporter::ExportLevel = $Exporter::ExportLevel + 1;
46 Exporter::import(grep {
47 my $passthrough;
48 given ($_) {
22bc907a
NC
49 $DEFAULT_FLAGS &= ~GLOB_NOCASE() when ':case';
50 $DEFAULT_FLAGS |= GLOB_NOCASE() when ':nocase';
51 when (':globally') {
52 no warnings 'redefine';
220398a0
PM
53 *CORE::GLOBAL::glob = \&File::Glob::csh_glob;
54 }
f4cbf990
FC
55 if ($_ eq ':bsd_glob') {
56 no strict; *{caller."::glob"} = \&bsd_glob_override;
57 }
df5a3819 58 $passthrough = 1;
220398a0 59 }
df5a3819
NC
60 $passthrough;
61 } @_);
72b16652
GS
62}
63
da4061d3 64XSLoader::load();
72b16652 65
220398a0 66$DEFAULT_FLAGS = GLOB_CSH();
862f843b 67if ($^O =~ /^(?:MSWin32|VMS|os2|dos|riscos)$/) {
220398a0
PM
68 $DEFAULT_FLAGS |= GLOB_NOCASE();
69}
70
00c80938
GS
71# File::Glob::glob() is deprecated because its prototype is different from
72# CORE::glob() (use bsd_glob() instead)
73sub glob {
e0e8a4dc 74 splice @_, 1; # don't pass PL_glob_index as flags!
00c80938
GS
75 goto &bsd_glob;
76}
77
72b16652
GS
781;
79__END__
80
81=head1 NAME
82
83File::Glob - Perl extension for BSD glob routine
84
85=head1 SYNOPSIS
86
87 use File::Glob ':glob';
9d70ac1b 88
00c80938
GS
89 @list = bsd_glob('*.[ch]');
90 $homedir = bsd_glob('~gnat', GLOB_TILDE | GLOB_ERR);
9d70ac1b 91
72b16652
GS
92 if (GLOB_ERROR) {
93 # an error occurred reading $homedir
94 }
95
00c80938 96 ## override the core glob (CORE::glob() does this automatically
11fe14b1 97 ## by default anyway, since v5.6.0)
220398a0 98 use File::Glob ':globally';
6bd08436 99 my @sources = <*.{c,h,y}>;
220398a0
PM
100
101 ## override the core glob, forcing case sensitivity
102 use File::Glob qw(:globally :case);
6bd08436 103 my @sources = <*.{c,h,y}>;
220398a0
PM
104
105 ## override the core glob forcing case insensitivity
106 use File::Glob qw(:globally :nocase);
6bd08436 107 my @sources = <*.{c,h,y}>;
9d70ac1b 108
6bd08436
SS
109 ## glob on all files in home directory
110 use File::Glob ':globally';
111 my @sources = <~gnat/*>;
72b16652
GS
112
113=head1 DESCRIPTION
114
9d70ac1b
RGS
115The glob angle-bracket operator C<< <> >> is a pathname generator that
116implements the rules for file name pattern matching used by Unix-like shells
117such as the Bourne shell or C shell.
6bd08436 118
00c80938
GS
119File::Glob::bsd_glob() implements the FreeBSD glob(3) routine, which is
120a superset of the POSIX glob() (described in IEEE Std 1003.2 "POSIX.2").
121bsd_glob() takes a mandatory C<pattern> argument, and an optional
72b16652
GS
122C<flags> argument, and returns a list of filenames matching the
123pattern, with interpretation of the pattern modified by the C<flags>
00c80938
GS
124variable.
125
126Since v5.6.0, Perl's CORE::glob() is implemented in terms of bsd_glob().
127Note that they don't share the same prototype--CORE::glob() only accepts
128a single argument. Due to historical reasons, CORE::glob() will also
129split its argument on whitespace, treating it as multiple patterns,
130whereas bsd_glob() considers them as one pattern.
131
6bd08436
SS
132=head2 META CHARACTERS
133
9d70ac1b
RGS
134 \ Quote the next metacharacter
135 [] Character class
136 {} Multiple pattern
137 * Match any string of characters
138 ? Match any single character
139 ~ User name home directory
140
141The metanotation C<a{b,c,d}e> is a shorthand for C<abe ace ade>. Left to
142right order is preserved, with results of matches being sorted separately
143at a low level to preserve this order. As a special case C<{>, C<}>, and
144C<{}> are passed undisturbed.
6bd08436
SS
145
146=head2 POSIX FLAGS
147
00c80938 148The POSIX defined flags for bsd_glob() are:
72b16652
GS
149
150=over 4
151
152=item C<GLOB_ERR>
153
00c80938
GS
154Force bsd_glob() to return an error when it encounters a directory it
155cannot open or read. Ordinarily bsd_glob() continues to find matches.
72b16652 156
b8ef571c
JH
157=item C<GLOB_LIMIT>
158
159Make bsd_glob() return an error (GLOB_NOSPACE) when the pattern expands
160to a size bigger than the system constant C<ARG_MAX> (usually found in
161limits.h). If your system does not define this constant, bsd_glob() uses
162C<sysconf(_SC_ARG_MAX)> or C<_POSIX_ARG_MAX> where available (in that
163order). You can inspect these values using the standard C<POSIX>
164extension.
165
72b16652
GS
166=item C<GLOB_MARK>
167
168Each pathname that is a directory that matches the pattern has a slash
169appended.
170
220398a0
PM
171=item C<GLOB_NOCASE>
172
173By default, file names are assumed to be case sensitive; this flag
00c80938 174makes bsd_glob() treat case differences as not significant.
220398a0 175
72b16652
GS
176=item C<GLOB_NOCHECK>
177
00c80938 178If the pattern does not match any pathname, then bsd_glob() returns a list
72b16652
GS
179consisting of only the pattern. If C<GLOB_QUOTE> is set, its effect
180is present in the pattern returned.
181
182=item C<GLOB_NOSORT>
183
184By default, the pathnames are sorted in ascending ASCII order; this
00c80938 185flag prevents that sorting (speeding up bsd_glob()).
72b16652
GS
186
187=back
188
189The FreeBSD extensions to the POSIX standard are the following flags:
190
191=over 4
192
193=item C<GLOB_BRACE>
194
a45bd81d 195Pre-process the string to expand C<{pat,pat,...}> strings like csh(1).
72b16652
GS
196The pattern '{}' is left unexpanded for historical reasons (and csh(1)
197does the same thing to ease typing of find(1) patterns).
198
199=item C<GLOB_NOMAGIC>
200
201Same as C<GLOB_NOCHECK> but it only returns the pattern if it does not
202contain any of the special characters "*", "?" or "[". C<NOMAGIC> is
203provided to simplify implementing the historic csh(1) globbing
204behaviour and should probably not be used anywhere else.
205
206=item C<GLOB_QUOTE>
207
208Use the backslash ('\') character for quoting: every occurrence of a
209backslash followed by a character in the pattern is replaced by that
210character, avoiding any special interpretation of the character.
220398a0 211(But see below for exceptions on DOSISH systems).
72b16652
GS
212
213=item C<GLOB_TILDE>
214
215Expand patterns that start with '~' to user name home directories.
216
217=item C<GLOB_CSH>
218
219For convenience, C<GLOB_CSH> is a synonym for
2d5e9e5d 220C<GLOB_BRACE | GLOB_NOMAGIC | GLOB_QUOTE | GLOB_TILDE | GLOB_ALPHASORT>.
72b16652
GS
221
222=back
223
224The POSIX provided C<GLOB_APPEND>, C<GLOB_DOOFFS>, and the FreeBSD
225extensions C<GLOB_ALTDIRFUNC>, and C<GLOB_MAGCHAR> flags have not been
226implemented in the Perl version because they involve more complex
227interaction with the underlying C structures.
228
2d5e9e5d
JH
229The following flag has been added in the Perl implementation for
230csh compatibility:
231
232=over 4
233
234=item C<GLOB_ALPHASORT>
235
236If C<GLOB_NOSORT> is not in effect, sort filenames is alphabetical
237order (case does not matter) rather than in ASCII order.
238
239=back
240
72b16652
GS
241=head1 DIAGNOSTICS
242
00c80938 243bsd_glob() returns a list of matching paths, possibly zero length. If an
72b16652
GS
244error occurred, &File::Glob::GLOB_ERROR will be non-zero and C<$!> will be
245set. &File::Glob::GLOB_ERROR is guaranteed to be zero if no error occurred,
246or one of the following values otherwise:
247
248=over 4
249
250=item C<GLOB_NOSPACE>
251
252An attempt to allocate memory failed.
253
254=item C<GLOB_ABEND>
255
256The glob was stopped because an error was encountered.
257
258=back
259
00c80938
GS
260In the case where bsd_glob() has found some matching paths, but is
261interrupted by an error, it will return a list of filenames B<and>
72b16652
GS
262set &File::Glob::ERROR.
263
00c80938
GS
264Note that bsd_glob() deviates from POSIX and FreeBSD glob(3) behaviour
265by not considering C<ENOENT> and C<ENOTDIR> as errors - bsd_glob() will
72b16652
GS
266continue processing despite those errors, unless the C<GLOB_ERR> flag is
267set.
268
269Be aware that all filenames returned from File::Glob are tainted.
270
271=head1 NOTES
272
273=over 4
274
275=item *
276
9d70ac1b
RGS
277If you want to use multiple patterns, e.g. C<bsd_glob("a* b*")>, you should
278probably throw them in a set as in C<bsd_glob("{a*,b*}")>. This is because
150b260b
GS
279the argument to bsd_glob() isn't subjected to parsing by the C shell.
280Remember that you can use a backslash to escape things.
72b16652
GS
281
282=item *
283
220398a0
PM
284On DOSISH systems, backslash is a valid directory separator character.
285In this case, use of backslash as a quoting character (via GLOB_QUOTE)
286interferes with the use of backslash as a directory separator. The
287best (simplest, most portable) solution is to use forward slashes for
288directory separators, and backslashes for quoting. However, this does
289not match "normal practice" on these systems. As a concession to user
290expectation, therefore, backslashes (under GLOB_QUOTE) only quote the
291glob metacharacters '[', ']', '{', '}', '-', '~', and backslash itself.
292All other backslashes are passed through unchanged.
293
294=item *
295
72b16652
GS
296Win32 users should use the real slash. If you really want to use
297backslashes, consider using Sarathy's File::DosGlob, which comes with
298the standard Perl distribution.
299
7369a524
CN
300=item *
301
302Mac OS (Classic) users should note a few differences. Since
303Mac OS is not Unix, when the glob code encounters a tilde glob (e.g.
be708cc0 304~user) and the C<GLOB_TILDE> flag is used, it simply returns that
7369a524
CN
305pattern without doing any expansion.
306
307Glob on Mac OS is case-insensitive by default (if you don't use any
308flags). If you specify any flags at all and still want glob
309to be case-insensitive, you must include C<GLOB_NOCASE> in the flags.
310
311The path separator is ':' (aka colon), not '/' (aka slash). Mac OS users
312should be careful about specifying relative pathnames. While a full path
313always begins with a volume name, a relative pathname should always
314begin with a ':'. If specifying a volume name only, a trailing ':' is
315required.
316
be708cc0
JH
317The specification of pathnames in glob patterns adheres to the usual Mac
318OS conventions: The path separator is a colon ':', not a slash '/'. A
319full path always begins with a volume name. A relative pathname on Mac
320OS must always begin with a ':', except when specifying a file or
321directory name in the current working directory, where the leading colon
322is optional. If specifying a volume name only, a trailing ':' is
323required. Due to these rules, a glob like E<lt>*:E<gt> will find all
324mounted volumes, while a glob like E<lt>*E<gt> or E<lt>:*E<gt> will find
325all files and directories in the current directory.
326
327Note that updirs in the glob pattern are resolved before the matching begins,
328i.e. a pattern like "*HD:t?p::a*" will be matched as "*HD:a*". Note also,
329that a single trailing ':' in the pattern is ignored (unless it's a volume
330name pattern like "*HD:"), i.e. a glob like E<lt>:*:E<gt> will find both
331directories I<and> files (and not, as one might expect, only directories).
332You can, however, use the C<GLOB_MARK> flag to distinguish (without a file
333test) directory names from file names.
334
335If the C<GLOB_MARK> flag is set, all directory paths will have a ':' appended.
336Since a directory like 'lib:' is I<not> a valid I<relative> path on Mac OS,
337both a leading and a trailing colon will be added, when the directory name in
338question doesn't contain any colons (e.g. 'lib' becomes ':lib:').
339
a45bd81d
GS
340=back
341
6bd08436
SS
342=head1 SEE ALSO
343
344L<perlfunc/glob>, glob(3)
345
72b16652
GS
346=head1 AUTHOR
347
0e950d83 348The Perl interface was written by Nathan Torkington E<lt>gnat@frii.comE<gt>,
72b16652 349and is released under the artistic license. Further modifications were
7369a524
CN
350made by Greg Bacon E<lt>gbacon@cs.uah.eduE<gt>, Gurusamy Sarathy
351E<lt>gsar@activestate.comE<gt>, and Thomas Wegner
352E<lt>wegner_thomas@yahoo.comE<gt>. The C glob code has the
72b16652
GS
353following copyright:
354
0e950d83
GS
355 Copyright (c) 1989, 1993 The Regents of the University of California.
356 All rights reserved.
3cb6de81 357
0e950d83
GS
358 This code is derived from software contributed to Berkeley by
359 Guido van Rossum.
360
361 Redistribution and use in source and binary forms, with or without
362 modification, are permitted provided that the following conditions
363 are met:
364
365 1. Redistributions of source code must retain the above copyright
366 notice, this list of conditions and the following disclaimer.
367 2. Redistributions in binary form must reproduce the above copyright
368 notice, this list of conditions and the following disclaimer in the
369 documentation and/or other materials provided with the distribution.
370 3. Neither the name of the University nor the names of its contributors
371 may be used to endorse or promote products derived from this software
372 without specific prior written permission.
373
374 THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
375 ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
376 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
377 ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
378 FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
379 DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
380 OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
381 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
382 LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
383 OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
384 SUCH DAMAGE.
72b16652
GS
385
386=cut