xref: /onnv-gate/usr/src/cmd/perl/5.8.4/distrib/ext/File/Glob/Glob.pm (revision 0:68f95e015346)
1package File::Glob;
2
3use strict;
4our($VERSION, @ISA, @EXPORT_OK, @EXPORT_FAIL, %EXPORT_TAGS,
5    $AUTOLOAD, $DEFAULT_FLAGS);
6
7use XSLoader ();
8
9@ISA = qw(Exporter);
10
11# NOTE: The glob() export is only here for compatibility with 5.6.0.
12# csh_glob() should not be used directly, unless you know what you're doing.
13
14@EXPORT_OK   = qw(
15    csh_glob
16    bsd_glob
17    glob
18    GLOB_ABEND
19    GLOB_ALPHASORT
20    GLOB_ALTDIRFUNC
21    GLOB_BRACE
22    GLOB_CSH
23    GLOB_ERR
24    GLOB_ERROR
25    GLOB_LIMIT
26    GLOB_MARK
27    GLOB_NOCASE
28    GLOB_NOCHECK
29    GLOB_NOMAGIC
30    GLOB_NOSORT
31    GLOB_NOSPACE
32    GLOB_QUOTE
33    GLOB_TILDE
34);
35
36%EXPORT_TAGS = (
37    'glob' => [ qw(
38        GLOB_ABEND
39	GLOB_ALPHASORT
40        GLOB_ALTDIRFUNC
41        GLOB_BRACE
42        GLOB_CSH
43        GLOB_ERR
44        GLOB_ERROR
45        GLOB_LIMIT
46        GLOB_MARK
47        GLOB_NOCASE
48        GLOB_NOCHECK
49        GLOB_NOMAGIC
50        GLOB_NOSORT
51        GLOB_NOSPACE
52        GLOB_QUOTE
53        GLOB_TILDE
54        glob
55        bsd_glob
56    ) ],
57);
58
59$VERSION = '1.02';
60
61sub import {
62    require Exporter;
63    my $i = 1;
64    while ($i < @_) {
65	if ($_[$i] =~ /^:(case|nocase|globally)$/) {
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') {
70		local $^W;
71		*CORE::GLOBAL::glob = \&File::Glob::csh_glob;
72	    }
73	    next;
74	}
75	++$i;
76    }
77    goto &Exporter::import;
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 = $AUTOLOAD) =~ s/.*:://;
87    my ($error, $val) = constant($constname);
88    if ($error) {
89	require Carp;
90	Carp::croak($error);
91    }
92    eval "sub $AUTOLOAD { $val }";
93    goto &$AUTOLOAD;
94}
95
96XSLoader::load 'File::Glob', $VERSION;
97
98# Preloaded methods go here.
99
100sub GLOB_ERROR {
101    return (constant('GLOB_ERROR'))[1];
102}
103
104sub GLOB_CSH () {
105    GLOB_BRACE()
106	| GLOB_NOMAGIC()
107	| GLOB_QUOTE()
108	| GLOB_TILDE()
109	| GLOB_ALPHASORT()
110}
111
112$DEFAULT_FLAGS = GLOB_CSH();
113if ($^O =~ /^(?:MSWin32|VMS|os2|dos|riscos|MacOS)$/) {
114    $DEFAULT_FLAGS |= GLOB_NOCASE();
115}
116
117# Autoload methods go after =cut, and are processed by the autosplit program.
118
119sub bsd_glob {
120    my ($pat,$flags) = @_;
121    $flags = $DEFAULT_FLAGS if @_ < 2;
122    return doglob($pat,$flags);
123}
124
125# File::Glob::glob() is deprecated because its prototype is different from
126# CORE::glob() (use bsd_glob() instead)
127sub glob {
128    splice @_, 1; # don't pass PL_glob_index as flags!
129    goto &bsd_glob;
130}
131
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
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.
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) {
163	    $entries{$cxix} = [ map { doglob($_, $DEFAULT_FLAGS) } @pat ];
164	}
165	else {
166	    $entries{$cxix} = [ doglob($pat, $DEFAULT_FLAGS) ];
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';
198  @list = bsd_glob('*.[ch]');
199  $homedir = bsd_glob('~gnat', GLOB_TILDE | GLOB_ERR);
200  if (GLOB_ERROR) {
201    # an error occurred reading $homedir
202  }
203
204  ## override the core glob (CORE::glob() does this automatically
205  ## by default anyway, since v5.6.0)
206  use File::Glob ':globally';
207  my @sources = <*.{c,h,y}>
208
209  ## override the core glob, forcing case sensitivity
210  use File::Glob qw(:globally :case);
211  my @sources = <*.{c,h,y}>
212
213  ## override the core glob forcing case insensitivity
214  use File::Glob qw(:globally :nocase);
215  my @sources = <*.{c,h,y}>
216
217=head1 DESCRIPTION
218
219File::Glob::bsd_glob() implements the FreeBSD glob(3) routine, which is
220a superset of the POSIX glob() (described in IEEE Std 1003.2 "POSIX.2").
221bsd_glob() takes a mandatory C<pattern> argument, and an optional
222C<flags> argument, and returns a list of filenames matching the
223pattern, with interpretation of the pattern modified by the C<flags>
224variable.
225
226Since v5.6.0, Perl's CORE::glob() is implemented in terms of bsd_glob().
227Note that they don't share the same prototype--CORE::glob() only accepts
228a single argument.  Due to historical reasons, CORE::glob() will also
229split its argument on whitespace, treating it as multiple patterns,
230whereas bsd_glob() considers them as one pattern.
231
232The POSIX defined flags for bsd_glob() are:
233
234=over 4
235
236=item C<GLOB_ERR>
237
238Force bsd_glob() to return an error when it encounters a directory it
239cannot open or read.  Ordinarily bsd_glob() continues to find matches.
240
241=item C<GLOB_LIMIT>
242
243Make bsd_glob() return an error (GLOB_NOSPACE) when the pattern expands
244to a size bigger than the system constant C<ARG_MAX> (usually found in
245limits.h).  If your system does not define this constant, bsd_glob() uses
246C<sysconf(_SC_ARG_MAX)> or C<_POSIX_ARG_MAX> where available (in that
247order).  You can inspect these values using the standard C<POSIX>
248extension.
249
250=item C<GLOB_MARK>
251
252Each pathname that is a directory that matches the pattern has a slash
253appended.
254
255=item C<GLOB_NOCASE>
256
257By default, file names are assumed to be case sensitive; this flag
258makes bsd_glob() treat case differences as not significant.
259
260=item C<GLOB_NOCHECK>
261
262If the pattern does not match any pathname, then bsd_glob() returns a list
263consisting of only the pattern.  If C<GLOB_QUOTE> is set, its effect
264is present in the pattern returned.
265
266=item C<GLOB_NOSORT>
267
268By default, the pathnames are sorted in ascending ASCII order; this
269flag prevents that sorting (speeding up bsd_glob()).
270
271=back
272
273The FreeBSD extensions to the POSIX standard are the following flags:
274
275=over 4
276
277=item C<GLOB_BRACE>
278
279Pre-process the string to expand C<{pat,pat,...}> strings like csh(1).
280The pattern '{}' is left unexpanded for historical reasons (and csh(1)
281does the same thing to ease typing of find(1) patterns).
282
283=item C<GLOB_NOMAGIC>
284
285Same as C<GLOB_NOCHECK> but it only returns the pattern if it does not
286contain any of the special characters "*", "?" or "[".  C<NOMAGIC> is
287provided to simplify implementing the historic csh(1) globbing
288behaviour and should probably not be used anywhere else.
289
290=item C<GLOB_QUOTE>
291
292Use the backslash ('\') character for quoting: every occurrence of a
293backslash followed by a character in the pattern is replaced by that
294character, avoiding any special interpretation of the character.
295(But see below for exceptions on DOSISH systems).
296
297=item C<GLOB_TILDE>
298
299Expand patterns that start with '~' to user name home directories.
300
301=item C<GLOB_CSH>
302
303For convenience, C<GLOB_CSH> is a synonym for
304C<GLOB_BRACE | GLOB_NOMAGIC | GLOB_QUOTE | GLOB_TILDE | GLOB_ALPHASORT>.
305
306=back
307
308The POSIX provided C<GLOB_APPEND>, C<GLOB_DOOFFS>, and the FreeBSD
309extensions C<GLOB_ALTDIRFUNC>, and C<GLOB_MAGCHAR> flags have not been
310implemented in the Perl version because they involve more complex
311interaction with the underlying C structures.
312
313The following flag has been added in the Perl implementation for
314csh compatibility:
315
316=over 4
317
318=item C<GLOB_ALPHASORT>
319
320If C<GLOB_NOSORT> is not in effect, sort filenames is alphabetical
321order (case does not matter) rather than in ASCII order.
322
323=back
324
325=head1 DIAGNOSTICS
326
327bsd_glob() returns a list of matching paths, possibly zero length.  If an
328error occurred, &File::Glob::GLOB_ERROR will be non-zero and C<$!> will be
329set.  &File::Glob::GLOB_ERROR is guaranteed to be zero if no error occurred,
330or one of the following values otherwise:
331
332=over 4
333
334=item C<GLOB_NOSPACE>
335
336An attempt to allocate memory failed.
337
338=item C<GLOB_ABEND>
339
340The glob was stopped because an error was encountered.
341
342=back
343
344In the case where bsd_glob() has found some matching paths, but is
345interrupted by an error, it will return a list of filenames B<and>
346set &File::Glob::ERROR.
347
348Note that bsd_glob() deviates from POSIX and FreeBSD glob(3) behaviour
349by not considering C<ENOENT> and C<ENOTDIR> as errors - bsd_glob() will
350continue processing despite those errors, unless the C<GLOB_ERR> flag is
351set.
352
353Be aware that all filenames returned from File::Glob are tainted.
354
355=head1 NOTES
356
357=over 4
358
359=item *
360
361If you want to use multiple patterns, e.g. C<bsd_glob "a* b*">, you should
362probably throw them in a set as in C<bsd_glob "{a*,b*}">.  This is because
363the argument to bsd_glob() isn't subjected to parsing by the C shell.
364Remember that you can use a backslash to escape things.
365
366=item *
367
368On DOSISH systems, backslash is a valid directory separator character.
369In this case, use of backslash as a quoting character (via GLOB_QUOTE)
370interferes with the use of backslash as a directory separator. The
371best (simplest, most portable) solution is to use forward slashes for
372directory separators, and backslashes for quoting. However, this does
373not match "normal practice" on these systems. As a concession to user
374expectation, therefore, backslashes (under GLOB_QUOTE) only quote the
375glob metacharacters '[', ']', '{', '}', '-', '~', and backslash itself.
376All other backslashes are passed through unchanged.
377
378=item *
379
380Win32 users should use the real slash.  If you really want to use
381backslashes, consider using Sarathy's File::DosGlob, which comes with
382the standard Perl distribution.
383
384=item *
385
386Mac OS (Classic) users should note a few differences. Since
387Mac OS is not Unix, when the glob code encounters a tilde glob (e.g.
388~user) and the C<GLOB_TILDE> flag is used, it simply returns that
389pattern without doing any expansion.
390
391Glob on Mac OS is case-insensitive by default (if you don't use any
392flags). If you specify any flags at all and still want glob
393to be case-insensitive, you must include C<GLOB_NOCASE> in the flags.
394
395The path separator is ':' (aka colon), not '/' (aka slash). Mac OS users
396should be careful about specifying relative pathnames. While a full path
397always begins with a volume name, a relative pathname should always
398begin with a ':'.  If specifying a volume name only, a trailing ':' is
399required.
400
401The specification of pathnames in glob patterns adheres to the usual Mac
402OS conventions: The path separator is a colon ':', not a slash '/'. A
403full path always begins with a volume name. A relative pathname on Mac
404OS must always begin with a ':', except when specifying a file or
405directory name in the current working directory, where the leading colon
406is optional. If specifying a volume name only, a trailing ':' is
407required. Due to these rules, a glob like E<lt>*:E<gt> will find all
408mounted volumes, while a glob like E<lt>*E<gt> or E<lt>:*E<gt> will find
409all files and directories in the current directory.
410
411Note that updirs in the glob pattern are resolved before the matching begins,
412i.e. a pattern like "*HD:t?p::a*" will be matched as "*HD:a*". Note also,
413that a single trailing ':' in the pattern is ignored (unless it's a volume
414name pattern like "*HD:"), i.e. a glob like E<lt>:*:E<gt> will find both
415directories I<and> files (and not, as one might expect, only directories).
416You can, however, use the C<GLOB_MARK> flag to distinguish (without a file
417test) directory names from file names.
418
419If the C<GLOB_MARK> flag is set, all directory paths will have a ':' appended.
420Since a directory like 'lib:' is I<not> a valid I<relative> path on Mac OS,
421both a leading and a trailing colon will be added, when the directory name in
422question doesn't contain any colons (e.g. 'lib' becomes ':lib:').
423
424=back
425
426=head1 AUTHOR
427
428The Perl interface was written by Nathan Torkington E<lt>gnat@frii.comE<gt>,
429and is released under the artistic license.  Further modifications were
430made by Greg Bacon E<lt>gbacon@cs.uah.eduE<gt>, Gurusamy Sarathy
431E<lt>gsar@activestate.comE<gt>, and Thomas Wegner
432E<lt>wegner_thomas@yahoo.comE<gt>.  The C glob code has the
433following copyright:
434
435    Copyright (c) 1989, 1993 The Regents of the University of California.
436    All rights reserved.
437
438    This code is derived from software contributed to Berkeley by
439    Guido van Rossum.
440
441    Redistribution and use in source and binary forms, with or without
442    modification, are permitted provided that the following conditions
443    are met:
444
445    1. Redistributions of source code must retain the above copyright
446       notice, this list of conditions and the following disclaimer.
447    2. Redistributions in binary form must reproduce the above copyright
448       notice, this list of conditions and the following disclaimer in the
449       documentation and/or other materials provided with the distribution.
450    3. Neither the name of the University nor the names of its contributors
451       may be used to endorse or promote products derived from this software
452       without specific prior written permission.
453
454    THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
455    ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
456    IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
457    ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
458    FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
459    DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
460    OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
461    HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
462    LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
463    OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
464    SUCH DAMAGE.
465
466=cut
467