xref: /openbsd/gnu/usr.bin/perl/ext/File-Glob/Glob.pm (revision 898184e3)
1package File::Glob;
2
3use strict;
4our($VERSION, @ISA, @EXPORT_OK, @EXPORT_FAIL, %EXPORT_TAGS, $DEFAULT_FLAGS);
5
6require XSLoader;
7use feature 'switch';
8
9@ISA = qw(Exporter);
10
11# NOTE: The glob() export is only here for compatibility with 5.6.0.
12# csh_glob() should not be used directly, unless you know what you're doing.
13
14%EXPORT_TAGS = (
15    'glob' => [ qw(
16        GLOB_ABEND
17	GLOB_ALPHASORT
18        GLOB_ALTDIRFUNC
19        GLOB_BRACE
20        GLOB_CSH
21        GLOB_ERR
22        GLOB_ERROR
23        GLOB_LIMIT
24        GLOB_MARK
25        GLOB_NOCASE
26        GLOB_NOCHECK
27        GLOB_NOMAGIC
28        GLOB_NOSORT
29        GLOB_NOSPACE
30        GLOB_QUOTE
31        GLOB_TILDE
32        bsd_glob
33        glob
34    ) ],
35);
36$EXPORT_TAGS{bsd_glob} = [@{$EXPORT_TAGS{glob}}];
37pop @{$EXPORT_TAGS{bsd_glob}}; # no "glob"
38
39@EXPORT_OK   = (@{$EXPORT_TAGS{'glob'}}, 'csh_glob');
40
41$VERSION = '1.17';
42
43sub import {
44    require Exporter;
45    local $Exporter::ExportLevel = $Exporter::ExportLevel + 1;
46    Exporter::import(grep {
47	my $passthrough;
48	given ($_) {
49	    $DEFAULT_FLAGS &= ~GLOB_NOCASE() when ':case';
50	    $DEFAULT_FLAGS |= GLOB_NOCASE() when ':nocase';
51	    when (':globally') {
52		no warnings 'redefine';
53		*CORE::GLOBAL::glob = \&File::Glob::csh_glob;
54	    }
55	    if ($_ eq ':bsd_glob') {
56		no strict; *{caller."::glob"} = \&bsd_glob_override;
57	    }
58	    $passthrough = 1;
59	}
60	$passthrough;
61    } @_);
62}
63
64XSLoader::load();
65
66$DEFAULT_FLAGS = GLOB_CSH();
67if ($^O =~ /^(?:MSWin32|VMS|os2|dos|riscos)$/) {
68    $DEFAULT_FLAGS |= GLOB_NOCASE();
69}
70
71# File::Glob::glob() is deprecated because its prototype is different from
72# CORE::glob() (use bsd_glob() instead)
73sub glob {
74    splice @_, 1; # don't pass PL_glob_index as flags!
75    goto &bsd_glob;
76}
77
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 ':bsd_glob';
88
89  @list = bsd_glob('*.[ch]');
90  $homedir = bsd_glob('~gnat', GLOB_TILDE | GLOB_ERR);
91
92  if (GLOB_ERROR) {
93    # an error occurred reading $homedir
94  }
95
96  ## override the core glob (CORE::glob() does this automatically
97  ## by default anyway, since v5.6.0)
98  use File::Glob ':globally';
99  my @sources = <*.{c,h,y}>;
100
101  ## override the core glob, forcing case sensitivity
102  use File::Glob qw(:globally :case);
103  my @sources = <*.{c,h,y}>;
104
105  ## override the core glob forcing case insensitivity
106  use File::Glob qw(:globally :nocase);
107  my @sources = <*.{c,h,y}>;
108
109  ## glob on all files in home directory
110  use File::Glob ':globally';
111  my @sources = <~gnat/*>;
112
113=head1 DESCRIPTION
114
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.
118
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
122C<flags> argument, and returns a list of filenames matching the
123pattern, with interpretation of the pattern modified by the C<flags>
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.  But see C<:bsd_glob>
131under L</EXPORTS>, below.
132
133=head2 META CHARACTERS
134
135  \       Quote the next metacharacter
136  []      Character class
137  {}      Multiple pattern
138  *       Match any string of characters
139  ?       Match any single character
140  ~       User name home directory
141
142The metanotation C<a{b,c,d}e> is a shorthand for C<abe ace ade>.  Left to
143right order is preserved, with results of matches being sorted separately
144at a low level to preserve this order.  As a special case C<{>, C<}>, and
145C<{}> are passed undisturbed.
146
147=head2 EXPORTS
148
149See also the L</POSIX FLAGS> below, which can be exported individually.
150
151=head3 C<:bsd_glob>
152
153The C<:bsd_glob> export tag exports bsd_glob() and the constants listed
154below.  It also overrides glob() in the calling package with one that
155behaves like bsd_glob() with regard to spaces (the space is treated as part
156of a file name), but supports iteration in scalar context; i.e., it
157preserves the core function's feature of returning the next item each time
158it is called.
159
160=head3 C<:glob>
161
162The C<:glob> tag, now discouraged, is the old version of C<:bsd_glob>.  It
163exports the same constants and functions, but its glob() override does not
164support iteration; it returns the last file name in scalar context.  That
165means this will loop forever:
166
167    use File::Glob ':glob';
168    while (my $file = <* copy.txt>) {
169	...
170    }
171
172=head3 C<bsd_glob>
173
174This function, which is included in the two export tags listed above,
175takes one or two arguments.  The first is the glob pattern.  The second is
176a set of flags ORed together.  The available flags are listed below under
177L</POSIX FLAGS>.  If the second argument is omitted, C<GLOB_CSH> (or
178C<GLOB_CSH|GLOB_NOCASE> on VMS and DOSish systems) is used by default.
179
180=head3 C<:nocase> and C<:case>
181
182These two export tags globally modify the default flags that bsd_glob()
183and, except on VMS, Perl's built-in C<glob> operator use.  C<GLOB_NOCASE>
184is turned on or off, respectively.
185
186=head3 C<csh_glob>
187
188The csh_glob() function can also be exported, but you should not use it
189directly unless you really know what you are doing.  It splits the pattern
190into words and feeds each one to bsd_glob().  Perl's own glob() function
191uses this internally.
192
193=head2 POSIX FLAGS
194
195The POSIX defined flags for bsd_glob() are:
196
197=over 4
198
199=item C<GLOB_ERR>
200
201Force bsd_glob() to return an error when it encounters a directory it
202cannot open or read.  Ordinarily bsd_glob() continues to find matches.
203
204=item C<GLOB_LIMIT>
205
206Make bsd_glob() return an error (GLOB_NOSPACE) when the pattern expands
207to a size bigger than the system constant C<ARG_MAX> (usually found in
208limits.h).  If your system does not define this constant, bsd_glob() uses
209C<sysconf(_SC_ARG_MAX)> or C<_POSIX_ARG_MAX> where available (in that
210order).  You can inspect these values using the standard C<POSIX>
211extension.
212
213=item C<GLOB_MARK>
214
215Each pathname that is a directory that matches the pattern has a slash
216appended.
217
218=item C<GLOB_NOCASE>
219
220By default, file names are assumed to be case sensitive; this flag
221makes bsd_glob() treat case differences as not significant.
222
223=item C<GLOB_NOCHECK>
224
225If the pattern does not match any pathname, then bsd_glob() returns a list
226consisting of only the pattern.  If C<GLOB_QUOTE> is set, its effect
227is present in the pattern returned.
228
229=item C<GLOB_NOSORT>
230
231By default, the pathnames are sorted in ascending ASCII order; this
232flag prevents that sorting (speeding up bsd_glob()).
233
234=back
235
236The FreeBSD extensions to the POSIX standard are the following flags:
237
238=over 4
239
240=item C<GLOB_BRACE>
241
242Pre-process the string to expand C<{pat,pat,...}> strings like csh(1).
243The pattern '{}' is left unexpanded for historical reasons (and csh(1)
244does the same thing to ease typing of find(1) patterns).
245
246=item C<GLOB_NOMAGIC>
247
248Same as C<GLOB_NOCHECK> but it only returns the pattern if it does not
249contain any of the special characters "*", "?" or "[".  C<NOMAGIC> is
250provided to simplify implementing the historic csh(1) globbing
251behaviour and should probably not be used anywhere else.
252
253=item C<GLOB_QUOTE>
254
255Use the backslash ('\') character for quoting: every occurrence of a
256backslash followed by a character in the pattern is replaced by that
257character, avoiding any special interpretation of the character.
258(But see below for exceptions on DOSISH systems).
259
260=item C<GLOB_TILDE>
261
262Expand patterns that start with '~' to user name home directories.
263
264=item C<GLOB_CSH>
265
266For convenience, C<GLOB_CSH> is a synonym for
267C<GLOB_BRACE | GLOB_NOMAGIC | GLOB_QUOTE | GLOB_TILDE | GLOB_ALPHASORT>.
268
269=back
270
271The POSIX provided C<GLOB_APPEND>, C<GLOB_DOOFFS>, and the FreeBSD
272extensions C<GLOB_ALTDIRFUNC>, and C<GLOB_MAGCHAR> flags have not been
273implemented in the Perl version because they involve more complex
274interaction with the underlying C structures.
275
276The following flag has been added in the Perl implementation for
277csh compatibility:
278
279=over 4
280
281=item C<GLOB_ALPHASORT>
282
283If C<GLOB_NOSORT> is not in effect, sort filenames is alphabetical
284order (case does not matter) rather than in ASCII order.
285
286=back
287
288=head1 DIAGNOSTICS
289
290bsd_glob() returns a list of matching paths, possibly zero length.  If an
291error occurred, &File::Glob::GLOB_ERROR will be non-zero and C<$!> will be
292set.  &File::Glob::GLOB_ERROR is guaranteed to be zero if no error occurred,
293or one of the following values otherwise:
294
295=over 4
296
297=item C<GLOB_NOSPACE>
298
299An attempt to allocate memory failed.
300
301=item C<GLOB_ABEND>
302
303The glob was stopped because an error was encountered.
304
305=back
306
307In the case where bsd_glob() has found some matching paths, but is
308interrupted by an error, it will return a list of filenames B<and>
309set &File::Glob::ERROR.
310
311Note that bsd_glob() deviates from POSIX and FreeBSD glob(3) behaviour
312by not considering C<ENOENT> and C<ENOTDIR> as errors - bsd_glob() will
313continue processing despite those errors, unless the C<GLOB_ERR> flag is
314set.
315
316Be aware that all filenames returned from File::Glob are tainted.
317
318=head1 NOTES
319
320=over 4
321
322=item *
323
324If you want to use multiple patterns, e.g. C<bsd_glob("a* b*")>, you should
325probably throw them in a set as in C<bsd_glob("{a*,b*}")>.  This is because
326the argument to bsd_glob() isn't subjected to parsing by the C shell.
327Remember that you can use a backslash to escape things.
328
329=item *
330
331On DOSISH systems, backslash is a valid directory separator character.
332In this case, use of backslash as a quoting character (via GLOB_QUOTE)
333interferes with the use of backslash as a directory separator.  The
334best (simplest, most portable) solution is to use forward slashes for
335directory separators, and backslashes for quoting.  However, this does
336not match "normal practice" on these systems.  As a concession to user
337expectation, therefore, backslashes (under GLOB_QUOTE) only quote the
338glob metacharacters '[', ']', '{', '}', '-', '~', and backslash itself.
339All other backslashes are passed through unchanged.
340
341=item *
342
343Win32 users should use the real slash.  If you really want to use
344backslashes, consider using Sarathy's File::DosGlob, which comes with
345the standard Perl distribution.
346
347=back
348
349=head1 SEE ALSO
350
351L<perlfunc/glob>, glob(3)
352
353=head1 AUTHOR
354
355The Perl interface was written by Nathan Torkington E<lt>gnat@frii.comE<gt>,
356and is released under the artistic license.  Further modifications were
357made by Greg Bacon E<lt>gbacon@cs.uah.eduE<gt>, Gurusamy Sarathy
358E<lt>gsar@activestate.comE<gt>, and Thomas Wegner
359E<lt>wegner_thomas@yahoo.comE<gt>.  The C glob code has the
360following copyright:
361
362    Copyright (c) 1989, 1993 The Regents of the University of California.
363    All rights reserved.
364
365    This code is derived from software contributed to Berkeley by
366    Guido van Rossum.
367
368    Redistribution and use in source and binary forms, with or without
369    modification, are permitted provided that the following conditions
370    are met:
371
372    1. Redistributions of source code must retain the above copyright
373       notice, this list of conditions and the following disclaimer.
374    2. Redistributions in binary form must reproduce the above copyright
375       notice, this list of conditions and the following disclaimer in the
376       documentation and/or other materials provided with the distribution.
377    3. Neither the name of the University nor the names of its contributors
378       may be used to endorse or promote products derived from this software
379       without specific prior written permission.
380
381    THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS" AND
382    ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
383    IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
384    ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
385    FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
386    DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
387    OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
388    HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
389    LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
390    OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
391    SUCH DAMAGE.
392
393=cut
394