This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Bump File::Glob's VERSION to go with change 30103.
[perl5.git] / ext / File / Glob / Glob.pm
CommitLineData
72b16652
GS
1package File::Glob;
2
3use strict;
17f410f9
GS
4our($VERSION, @ISA, @EXPORT_OK, @EXPORT_FAIL, %EXPORT_TAGS,
5 $AUTOLOAD, $DEFAULT_FLAGS);
72b16652 6
9426adcd 7use XSLoader ();
72b16652 8
72f7b9a1 9@ISA = qw(Exporter);
72b16652 10
00c80938
GS
11# NOTE: The glob() export is only here for compatibility with 5.6.0.
12# csh_glob() should not be used directly, unless you know what you're doing.
13
72b16652 14@EXPORT_OK = qw(
72b16652 15 csh_glob
00c80938 16 bsd_glob
72b16652
GS
17 glob
18 GLOB_ABEND
2d5e9e5d 19 GLOB_ALPHASORT
72b16652
GS
20 GLOB_ALTDIRFUNC
21 GLOB_BRACE
220398a0 22 GLOB_CSH
72b16652
GS
23 GLOB_ERR
24 GLOB_ERROR
b8ef571c 25 GLOB_LIMIT
72b16652 26 GLOB_MARK
220398a0 27 GLOB_NOCASE
72b16652
GS
28 GLOB_NOCHECK
29 GLOB_NOMAGIC
30 GLOB_NOSORT
31 GLOB_NOSPACE
32 GLOB_QUOTE
33 GLOB_TILDE
34);
35
72b16652
GS
36%EXPORT_TAGS = (
37 'glob' => [ qw(
38 GLOB_ABEND
2d5e9e5d 39 GLOB_ALPHASORT
72b16652
GS
40 GLOB_ALTDIRFUNC
41 GLOB_BRACE
220398a0 42 GLOB_CSH
72b16652
GS
43 GLOB_ERR
44 GLOB_ERROR
b8ef571c 45 GLOB_LIMIT
72b16652 46 GLOB_MARK
220398a0 47 GLOB_NOCASE
72b16652
GS
48 GLOB_NOCHECK
49 GLOB_NOMAGIC
50 GLOB_NOSORT
51 GLOB_NOSPACE
52 GLOB_QUOTE
53 GLOB_TILDE
54 glob
00c80938 55 bsd_glob
72b16652
GS
56 ) ],
57);
58
8c7d55fc 59$VERSION = '1.06';
220398a0
PM
60
61sub import {
7d3fb230 62 require Exporter;
220398a0
PM
63 my $i = 1;
64 while ($i < @_) {
1e2c6ed7 65 if ($_[$i] =~ /^:(case|nocase|globally)$/) {
220398a0
PM
66 splice(@_, $i, 1);
67 $DEFAULT_FLAGS &= ~GLOB_NOCASE() if $1 eq 'case';
68 $DEFAULT_FLAGS |= GLOB_NOCASE() if $1 eq 'nocase';
69 if ($1 eq 'globally') {
7d3fb230 70 local $^W;
220398a0
PM
71 *CORE::GLOBAL::glob = \&File::Glob::csh_glob;
72 }
73 next;
74 }
75 ++$i;
72b16652 76 }
220398a0 77 goto &Exporter::import;
72b16652
GS
78}
79
80sub AUTOLOAD {
81 # This AUTOLOAD is used to 'autoload' constants from the constant()
82 # XS function. If a constant is not found then control is passed
83 # to the AUTOLOAD in AutoLoader.
84
85 my $constname;
86 ($constname = $AUTOLOAD) =~ s/.*:://;
72f7b9a1
NC
87 my ($error, $val) = constant($constname);
88 if ($error) {
89 require Carp;
90 Carp::croak($error);
72b16652
GS
91 }
92 eval "sub $AUTOLOAD { $val }";
93 goto &$AUTOLOAD;
94}
95
9426adcd 96XSLoader::load 'File::Glob', $VERSION;
72b16652
GS
97
98# Preloaded methods go here.
99
100sub GLOB_ERROR {
72f7b9a1 101 return (constant('GLOB_ERROR'))[1];
72b16652
GS
102}
103
2d5e9e5d
JH
104sub GLOB_CSH () {
105 GLOB_BRACE()
106 | GLOB_NOMAGIC()
107 | GLOB_QUOTE()
108 | GLOB_TILDE()
109 | GLOB_ALPHASORT()
110}
72b16652 111
220398a0
PM
112$DEFAULT_FLAGS = GLOB_CSH();
113if ($^O =~ /^(?:MSWin32|VMS|os2|dos|riscos|MacOS)$/) {
114 $DEFAULT_FLAGS |= GLOB_NOCASE();
115}
116
72b16652
GS
117# Autoload methods go after =cut, and are processed by the autosplit program.
118
00c80938 119sub bsd_glob {
f0963acb
GS
120 my ($pat,$flags) = @_;
121 $flags = $DEFAULT_FLAGS if @_ < 2;
122 return doglob($pat,$flags);
72b16652
GS
123}
124
00c80938
GS
125# File::Glob::glob() is deprecated because its prototype is different from
126# CORE::glob() (use bsd_glob() instead)
127sub glob {
e0e8a4dc 128 splice @_, 1; # don't pass PL_glob_index as flags!
00c80938
GS
129 goto &bsd_glob;
130}
131
72b16652
GS
132## borrowed heavily from gsar's File::DosGlob
133my %iter;
134my %entries;
135
136sub csh_glob {
137 my $pat = shift;
138 my $cxix = shift;
139 my @pat;
140
141 # glob without args defaults to $_
142 $pat = $_ unless defined $pat;
143
144 # extract patterns
be3174d2
GS
145 $pat =~ s/^\s+//; # Protect against empty elements in
146 $pat =~ s/\s+$//; # things like < *.c> and <*.c >.
147 # These alone shouldn't trigger ParseWords.
72b16652
GS
148 if ($pat =~ /\s/) {
149 # XXX this is needed for compatibility with the csh
150 # implementation in Perl. Need to support a flag
151 # to disable this behavior.
152 require Text::ParseWords;
153 @pat = Text::ParseWords::parse_line('\s+',0,$pat);
154 }
155
156 # assume global context if not provided one
157 $cxix = '_G_' unless defined $cxix;
158 $iter{$cxix} = 0 unless exists $iter{$cxix};
159
160 # if we're just beginning, do it all first
161 if ($iter{$cxix} == 0) {
162 if (@pat) {
220398a0 163 $entries{$cxix} = [ map { doglob($_, $DEFAULT_FLAGS) } @pat ];
72b16652
GS
164 }
165 else {
220398a0 166 $entries{$cxix} = [ doglob($pat, $DEFAULT_FLAGS) ];
72b16652
GS
167 }
168 }
169
170 # chuck it all out, quick or slow
171 if (wantarray) {
172 delete $iter{$cxix};
173 return @{delete $entries{$cxix}};
174 }
175 else {
176 if ($iter{$cxix} = scalar @{$entries{$cxix}}) {
177 return shift @{$entries{$cxix}};
178 }
179 else {
180 # return undef for EOL
181 delete $iter{$cxix};
182 delete $entries{$cxix};
183 return undef;
184 }
185 }
186}
187
1881;
189__END__
190
191=head1 NAME
192
193File::Glob - Perl extension for BSD glob routine
194
195=head1 SYNOPSIS
196
197 use File::Glob ':glob';
9d70ac1b 198
00c80938
GS
199 @list = bsd_glob('*.[ch]');
200 $homedir = bsd_glob('~gnat', GLOB_TILDE | GLOB_ERR);
9d70ac1b 201
72b16652
GS
202 if (GLOB_ERROR) {
203 # an error occurred reading $homedir
204 }
205
00c80938 206 ## override the core glob (CORE::glob() does this automatically
11fe14b1 207 ## by default anyway, since v5.6.0)
220398a0 208 use File::Glob ':globally';
6bd08436 209 my @sources = <*.{c,h,y}>;
220398a0
PM
210
211 ## override the core glob, forcing case sensitivity
212 use File::Glob qw(:globally :case);
6bd08436 213 my @sources = <*.{c,h,y}>;
220398a0
PM
214
215 ## override the core glob forcing case insensitivity
216 use File::Glob qw(:globally :nocase);
6bd08436 217 my @sources = <*.{c,h,y}>;
9d70ac1b 218
6bd08436
SS
219 ## glob on all files in home directory
220 use File::Glob ':globally';
221 my @sources = <~gnat/*>;
72b16652
GS
222
223=head1 DESCRIPTION
224
9d70ac1b
RGS
225The glob angle-bracket operator C<< <> >> is a pathname generator that
226implements the rules for file name pattern matching used by Unix-like shells
227such as the Bourne shell or C shell.
6bd08436 228
00c80938
GS
229File::Glob::bsd_glob() implements the FreeBSD glob(3) routine, which is
230a superset of the POSIX glob() (described in IEEE Std 1003.2 "POSIX.2").
231bsd_glob() takes a mandatory C<pattern> argument, and an optional
72b16652
GS
232C<flags> argument, and returns a list of filenames matching the
233pattern, with interpretation of the pattern modified by the C<flags>
00c80938
GS
234variable.
235
236Since v5.6.0, Perl's CORE::glob() is implemented in terms of bsd_glob().
237Note that they don't share the same prototype--CORE::glob() only accepts
238a single argument. Due to historical reasons, CORE::glob() will also
239split its argument on whitespace, treating it as multiple patterns,
240whereas bsd_glob() considers them as one pattern.
241
6bd08436
SS
242=head2 META CHARACTERS
243
9d70ac1b
RGS
244 \ Quote the next metacharacter
245 [] Character class
246 {} Multiple pattern
247 * Match any string of characters
248 ? Match any single character
249 ~ User name home directory
250
251The metanotation C<a{b,c,d}e> is a shorthand for C<abe ace ade>. Left to
252right order is preserved, with results of matches being sorted separately
253at a low level to preserve this order. As a special case C<{>, C<}>, and
254C<{}> are passed undisturbed.
6bd08436
SS
255
256=head2 POSIX FLAGS
257
00c80938 258The POSIX defined flags for bsd_glob() are:
72b16652
GS
259
260=over 4
261
262=item C<GLOB_ERR>
263
00c80938
GS
264Force bsd_glob() to return an error when it encounters a directory it
265cannot open or read. Ordinarily bsd_glob() continues to find matches.
72b16652 266
b8ef571c
JH
267=item C<GLOB_LIMIT>
268
269Make bsd_glob() return an error (GLOB_NOSPACE) when the pattern expands
270to a size bigger than the system constant C<ARG_MAX> (usually found in
271limits.h). If your system does not define this constant, bsd_glob() uses
272C<sysconf(_SC_ARG_MAX)> or C<_POSIX_ARG_MAX> where available (in that
273order). You can inspect these values using the standard C<POSIX>
274extension.
275
72b16652
GS
276=item C<GLOB_MARK>
277
278Each pathname that is a directory that matches the pattern has a slash
279appended.
280
220398a0
PM
281=item C<GLOB_NOCASE>
282
283By default, file names are assumed to be case sensitive; this flag
00c80938 284makes bsd_glob() treat case differences as not significant.
220398a0 285
72b16652
GS
286=item C<GLOB_NOCHECK>
287
00c80938 288If the pattern does not match any pathname, then bsd_glob() returns a list
72b16652
GS
289consisting of only the pattern. If C<GLOB_QUOTE> is set, its effect
290is present in the pattern returned.
291
292=item C<GLOB_NOSORT>
293
294By default, the pathnames are sorted in ascending ASCII order; this
00c80938 295flag prevents that sorting (speeding up bsd_glob()).
72b16652
GS
296
297=back
298
299The FreeBSD extensions to the POSIX standard are the following flags:
300
301=over 4
302
303=item C<GLOB_BRACE>
304
a45bd81d 305Pre-process the string to expand C<{pat,pat,...}> strings like csh(1).
72b16652
GS
306The pattern '{}' is left unexpanded for historical reasons (and csh(1)
307does the same thing to ease typing of find(1) patterns).
308
309=item C<GLOB_NOMAGIC>
310
311Same as C<GLOB_NOCHECK> but it only returns the pattern if it does not
312contain any of the special characters "*", "?" or "[". C<NOMAGIC> is
313provided to simplify implementing the historic csh(1) globbing
314behaviour and should probably not be used anywhere else.
315
316=item C<GLOB_QUOTE>
317
318Use the backslash ('\') character for quoting: every occurrence of a
319backslash followed by a character in the pattern is replaced by that
320character, avoiding any special interpretation of the character.
220398a0 321(But see below for exceptions on DOSISH systems).
72b16652
GS
322
323=item C<GLOB_TILDE>
324
325Expand patterns that start with '~' to user name home directories.
326
327=item C<GLOB_CSH>
328
329For convenience, C<GLOB_CSH> is a synonym for
2d5e9e5d 330C<GLOB_BRACE | GLOB_NOMAGIC | GLOB_QUOTE | GLOB_TILDE | GLOB_ALPHASORT>.
72b16652
GS
331
332=back
333
334The POSIX provided C<GLOB_APPEND>, C<GLOB_DOOFFS>, and the FreeBSD
335extensions C<GLOB_ALTDIRFUNC>, and C<GLOB_MAGCHAR> flags have not been
336implemented in the Perl version because they involve more complex
337interaction with the underlying C structures.
338
2d5e9e5d
JH
339The following flag has been added in the Perl implementation for
340csh compatibility:
341
342=over 4
343
344=item C<GLOB_ALPHASORT>
345
346If C<GLOB_NOSORT> is not in effect, sort filenames is alphabetical
347order (case does not matter) rather than in ASCII order.
348
349=back
350
72b16652
GS
351=head1 DIAGNOSTICS
352
00c80938 353bsd_glob() returns a list of matching paths, possibly zero length. If an
72b16652
GS
354error occurred, &File::Glob::GLOB_ERROR will be non-zero and C<$!> will be
355set. &File::Glob::GLOB_ERROR is guaranteed to be zero if no error occurred,
356or one of the following values otherwise:
357
358=over 4
359
360=item C<GLOB_NOSPACE>
361
362An attempt to allocate memory failed.
363
364=item C<GLOB_ABEND>
365
366The glob was stopped because an error was encountered.
367
368=back
369
00c80938
GS
370In the case where bsd_glob() has found some matching paths, but is
371interrupted by an error, it will return a list of filenames B<and>
72b16652
GS
372set &File::Glob::ERROR.
373
00c80938
GS
374Note that bsd_glob() deviates from POSIX and FreeBSD glob(3) behaviour
375by not considering C<ENOENT> and C<ENOTDIR> as errors - bsd_glob() will
72b16652
GS
376continue processing despite those errors, unless the C<GLOB_ERR> flag is
377set.
378
379Be aware that all filenames returned from File::Glob are tainted.
380
381=head1 NOTES
382
383=over 4
384
385=item *
386
9d70ac1b
RGS
387If you want to use multiple patterns, e.g. C<bsd_glob("a* b*")>, you should
388probably throw them in a set as in C<bsd_glob("{a*,b*}")>. This is because
150b260b
GS
389the argument to bsd_glob() isn't subjected to parsing by the C shell.
390Remember that you can use a backslash to escape things.
72b16652
GS
391
392=item *
393
220398a0
PM
394On DOSISH systems, backslash is a valid directory separator character.
395In this case, use of backslash as a quoting character (via GLOB_QUOTE)
396interferes with the use of backslash as a directory separator. The
397best (simplest, most portable) solution is to use forward slashes for
398directory separators, and backslashes for quoting. However, this does
399not match "normal practice" on these systems. As a concession to user
400expectation, therefore, backslashes (under GLOB_QUOTE) only quote the
401glob metacharacters '[', ']', '{', '}', '-', '~', and backslash itself.
402All other backslashes are passed through unchanged.
403
404=item *
405
72b16652
GS
406Win32 users should use the real slash. If you really want to use
407backslashes, consider using Sarathy's File::DosGlob, which comes with
408the standard Perl distribution.
409
7369a524
CN
410=item *
411
412Mac OS (Classic) users should note a few differences. Since
413Mac OS is not Unix, when the glob code encounters a tilde glob (e.g.
be708cc0 414~user) and the C<GLOB_TILDE> flag is used, it simply returns that
7369a524
CN
415pattern without doing any expansion.
416
417Glob on Mac OS is case-insensitive by default (if you don't use any
418flags). If you specify any flags at all and still want glob
419to be case-insensitive, you must include C<GLOB_NOCASE> in the flags.
420
421The path separator is ':' (aka colon), not '/' (aka slash). Mac OS users
422should be careful about specifying relative pathnames. While a full path
423always begins with a volume name, a relative pathname should always
424begin with a ':'. If specifying a volume name only, a trailing ':' is
425required.
426
be708cc0
JH
427The specification of pathnames in glob patterns adheres to the usual Mac
428OS conventions: The path separator is a colon ':', not a slash '/'. A
429full path always begins with a volume name. A relative pathname on Mac
430OS must always begin with a ':', except when specifying a file or
431directory name in the current working directory, where the leading colon
432is optional. If specifying a volume name only, a trailing ':' is
433required. Due to these rules, a glob like E<lt>*:E<gt> will find all
434mounted volumes, while a glob like E<lt>*E<gt> or E<lt>:*E<gt> will find
435all files and directories in the current directory.
436
437Note that updirs in the glob pattern are resolved before the matching begins,
438i.e. a pattern like "*HD:t?p::a*" will be matched as "*HD:a*". Note also,
439that a single trailing ':' in the pattern is ignored (unless it's a volume
440name pattern like "*HD:"), i.e. a glob like E<lt>:*:E<gt> will find both
441directories I<and> files (and not, as one might expect, only directories).
442You can, however, use the C<GLOB_MARK> flag to distinguish (without a file
443test) directory names from file names.
444
445If the C<GLOB_MARK> flag is set, all directory paths will have a ':' appended.
446Since a directory like 'lib:' is I<not> a valid I<relative> path on Mac OS,
447both a leading and a trailing colon will be added, when the directory name in
448question doesn't contain any colons (e.g. 'lib' becomes ':lib:').
449
a45bd81d
GS
450=back
451
6bd08436
SS
452=head1 SEE ALSO
453
454L<perlfunc/glob>, glob(3)
455
72b16652
GS
456=head1 AUTHOR
457
0e950d83 458The Perl interface was written by Nathan Torkington E<lt>gnat@frii.comE<gt>,
72b16652 459and is released under the artistic license. Further modifications were
7369a524
CN
460made by Greg Bacon E<lt>gbacon@cs.uah.eduE<gt>, Gurusamy Sarathy
461E<lt>gsar@activestate.comE<gt>, and Thomas Wegner
462E<lt>wegner_thomas@yahoo.comE<gt>. The C glob code has the
72b16652
GS
463following copyright:
464
0e950d83
GS
465 Copyright (c) 1989, 1993 The Regents of the University of California.
466 All rights reserved.
3cb6de81 467
0e950d83
GS
468 This code is derived from software contributed to Berkeley by
469 Guido van Rossum.
470
471 Redistribution and use in source and binary forms, with or without
472 modification, are permitted provided that the following conditions
473 are met:
474
475 1. Redistributions of source code must retain the above copyright
476 notice, this list of conditions and the following disclaimer.
477 2. Redistributions in binary form must reproduce the above copyright
478 notice, this list of conditions and the following disclaimer in the
479 documentation and/or other materials provided with the distribution.
480 3. Neither the name of the University nor the names of its contributors
481 may be used to endorse or promote products derived from this software
482 without specific prior written permission.
483
484 THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
485 ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
486 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
487 ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
488 FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
489 DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
490 OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
491 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
492 LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
493 OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
494 SUCH DAMAGE.
72b16652
GS
495
496=cut