Profile of Module/Runtime.pm

Filename	/usr/share/perl5/Module/Runtime.pm
Statements	Executed 0 statements in 0s

Line	State ments	Time on line	Calls	Time in subs	Code
1					=head1 NAME
2
3					Module::Runtime - runtime module handling
4
5					=head1 SYNOPSIS
6
7					use Module::Runtime qw(
8					$module_name_rx is_module_name check_module_name
9					module_notional_filename require_module
10					);
11
12					if($module_name =~ /\A$module_name_rx\z/o) { ...
13					if(is_module_name($module_name)) { ...
14					check_module_name($module_name);
15
16					$notional_filename = module_notional_filename($module_name);
17					require_module($module_name);
18
19					use Module::Runtime qw(use_module use_package_optimistically);
20
21					$bi = use_module("Math::BigInt", 1.31)->new("1_234");
22					$widget = use_package_optimistically("Local::Widget")->new;
23
24					use Module::Runtime qw(
25					$top_module_spec_rx $sub_module_spec_rx
26					is_module_spec check_module_spec
27					compose_module_name
28					);
29
30					if($spec =~ /\A$top_module_spec_rx\z/o) { ...
31					if($spec =~ /\A$sub_module_spec_rx\z/o) { ...
32					if(is_module_spec("Standard::Prefix", $spec)) { ...
33					check_module_spec("Standard::Prefix", $spec);
34
35					$module_name =
36					compose_module_name("Standard::Prefix", $spec);
37
38					=head1 DESCRIPTION
39
40					The functions exported by this module deal with runtime handling of
41					Perl modules, which are normally handled at compile time. This module
42					avoids using any other modules, so that it can be used in low-level
43					infrastructure.
44
45					The parts of this module that work with module names apply the same syntax
46					that is used for barewords in Perl source. In principle this syntax
47					can vary between versions of Perl, and this module applies the syntax of
48					the Perl on which it is running. In practice the usable syntax hasn't
49					changed yet. There's some intent for Unicode module names to be supported
50					in the future, but this hasn't yet amounted to any consistent facility.
51
52					The functions of this module whose purpose is to load modules include
53					workarounds for three old Perl core bugs regarding C<require>. These
54					workarounds are applied on any Perl version where the bugs exist, except
55					for a case where one of the bugs cannot be adequately worked around in
56					pure Perl.
57
58					=head2 Module name syntax
59
60					The usable module name syntax has not changed from Perl 5.000 up to
61					Perl 5.19.8. The syntax is composed entirely of ASCII characters.
62					From Perl 5.6 onwards there has been some attempt to allow the use of
63					non-ASCII Unicode characters in Perl source, but it was fundamentally
64					broken (like the entirety of Perl 5.6's Unicode handling) and remained
65					pretty much entirely unusable until it got some attention in the Perl
66					5.15 series. Although Unicode is now consistently accepted by the
67					parser in some places, it remains broken for module names. Furthermore,
68					there has not yet been any work on how to map Unicode module names into
69					filenames, so in that respect also Unicode module names are unusable.
70
71					The module name syntax is, precisely: the string must consist of one or
72					more segments separated by C<::>; each segment must consist of one or more
73					identifier characters (ASCII alphanumerics plus "_"); the first character
74					of the string must not be a digit. Thus "C<IO::File>", "C<warnings>",
75					and "C<foo::123::x_0>" are all valid module names, whereas "C<IO::>"
76					and "C<1foo::bar>" are not. C<'> separators are not permitted by this
77					module, though they remain usable in Perl source, being translated to
78					C<::> in the parser.
79
80					=head2 Core bugs worked around
81
82					The first bug worked around is core bug [perl #68590], which causes
83					lexical state in one file to leak into another that is C<require>d/C<use>d
84					from it. This bug is present from Perl 5.6 up to Perl 5.10, and is
85					fixed in Perl 5.11.0. From Perl 5.9.4 up to Perl 5.10.0 no satisfactory
86					workaround is possible in pure Perl. The workaround means that modules
87					loaded via this module don't suffer this pollution of their lexical
88					state. Modules loaded in other ways, or via this module on the Perl
89					versions where the pure Perl workaround is impossible, remain vulnerable.
90					The module L<Lexical::SealRequireHints> provides a complete workaround
91					for this bug.
92
93					The second bug worked around causes some kinds of failure in module
94					loading, principally compilation errors in the loaded module, to be
95					recorded in C<%INC> as if they were successful, so later attempts to load
96					the same module immediately indicate success. This bug is present up
97					to Perl 5.8.9, and is fixed in Perl 5.9.0. The workaround means that a
98					compilation error in a module loaded via this module won't be cached as
99					a success. Modules loaded in other ways remain liable to produce bogus
100					C<%INC> entries, and if a bogus entry exists then it will mislead this
101					module if it is used to re-attempt loading.
102
103					The third bug worked around causes the wrong context to be seen at
104					file scope of a loaded module, if C<require> is invoked in a location
105					that inherits context from a higher scope. This bug is present up to
106					Perl 5.11.2, and is fixed in Perl 5.11.3. The workaround means that
107					a module loaded via this module will always see the correct context.
108					Modules loaded in other ways remain vulnerable.
109
110					=cut
111
112					package Module::Runtime;
113
114					# Don't "use 5.006" here, because Perl 5.15.6 will load feature.pm if
115					# the version check is done that way.
116					BEGIN { require 5.006; }
117					# Don't "use warnings" here, to avoid dependencies. Do standardise the
118					# warning status by lexical override; unfortunately the only safe bitset
119					# to build in is the empty set, equivalent to "no warnings".
120					BEGIN { ${^WARNING_BITS} = ""; }
121					# Don't "use strict" here, to avoid dependencies.
122
123					our $VERSION = "0.014";
124
125					# Don't use Exporter here, to avoid dependencies.
126					our @EXPORT_OK = qw(
127					$module_name_rx is_module_name is_valid_module_name check_module_name
128					module_notional_filename require_module
129					use_module use_package_optimistically
130					$top_module_spec_rx $sub_module_spec_rx
131					is_module_spec is_valid_module_spec check_module_spec
132					compose_module_name
133					);
134					my %export_ok = map { ($_ => undef) } @EXPORT_OK;
135					sub import {
136					my $me = shift;
137					my $callpkg = caller(0);
138					my $errs = "";
139					foreach(@_) {
140					if(exists $export_ok{$_}) {
141					# We would need to do "no strict 'refs'" here
142					# if we had enabled strict at file scope.
143					if(/\A\$(.*)\z/s) {
144					*{$callpkg."::".$1} = \$$1;
145					} else {
146					*{$callpkg."::".$_} = \&$_;
147					}
148					} else {
149					$errs .= "\"$_\" is not exported by the $me module\n";
150					}
151					}
152					if($errs ne "") {
153					die "${errs}Can't continue after import errors ".
154					"at @{[(caller(0))[1]]} line @{[(caller(0))[2]]}.\n";
155					}
156					}
157
158					# Logic duplicated from Params::Classify. Duplicating it here avoids
159					# an extensive and potentially circular dependency graph.
160					sub _is_string($) {
161					my($arg) = @_;
162					return defined($arg) && ref(\$arg) eq "SCALAR";
163					}
164
165					=head1 REGULAR EXPRESSIONS
166
167					These regular expressions do not include any anchors, so to check
168					whether an entire string matches a syntax item you must supply the
169					anchors yourself.
170
171					=over
172
173					=item $module_name_rx
174
175					Matches a valid Perl module name in bareword syntax.
176
177					=cut
178
179					our $module_name_rx = qr/[A-Z_a-z][0-9A-Z_a-z](?:::[0-9A-Z_a-z]+)/;
180
181					=item $top_module_spec_rx
182
183					Matches a module specification for use with L</compose_module_name>,
184					where no prefix is being used.
185
186					=cut
187
188					my $qual_module_spec_rx =
189					qr#(?:/\|::)[A-Z_a-z][0-9A-Z_a-z](?:(?:/\|::)[0-9A-Z_a-z]+)#;
190
191					my $unqual_top_module_spec_rx =
192					qr#[A-Z_a-z][0-9A-Z_a-z](?:(?:/\|::)[0-9A-Z_a-z]+)#;
193
194					our $top_module_spec_rx = qr/$qual_module_spec_rx\|$unqual_top_module_spec_rx/o;
195
196					=item $sub_module_spec_rx
197
198					Matches a module specification for use with L</compose_module_name>,
199					where a prefix is being used.
200
201					=cut
202
203					my $unqual_sub_module_spec_rx = qr#[0-9A-Z_a-z]+(?:(?:/\|::)[0-9A-Z_a-z]+)*#;
204
205					our $sub_module_spec_rx = qr/$qual_module_spec_rx\|$unqual_sub_module_spec_rx/o;
206
207					=back
208
209					=head1 FUNCTIONS
210
211					=head2 Basic module handling
212
213					=over
214
215					=item is_module_name(ARG)
216
217					Returns a truth value indicating whether I<ARG> is a plain string
218					satisfying Perl module name syntax as described for L</$module_name_rx>.
219
220					=cut
221
222					sub is_module_name($) { _is_string($_[0]) && $_[0] =~ /\A$module_name_rx\z/o }
223
224					=item is_valid_module_name(ARG)
225
226					Deprecated alias for L</is_module_name>.
227
228					=cut
229
230					*is_valid_module_name = \&is_module_name;
231
232					=item check_module_name(ARG)
233
234					Check whether I<ARG> is a plain string
235					satisfying Perl module name syntax as described for L</$module_name_rx>.
236					Return normally if it is, or C<die> if it is not.
237
238					=cut
239
240					sub check_module_name($) {
241					unless(&is_module_name) {
242					die +(_is_string($_[0]) ? "`$_[0]'" : "argument").
243					" is not a module name\n";
244					}
245					}
246
247					=item module_notional_filename(NAME)
248
249					Generates a notional relative filename for a module, which is used in
250					some Perl core interfaces.
251					The I<NAME> is a string, which should be a valid module name (one or
252					more C<::>-separated segments). If it is not a valid name, the function
253					C<die>s.
254
255					The notional filename for the named module is generated and returned.
256					This filename is always in Unix style, with C</> directory separators
257					and a C<.pm> suffix. This kind of filename can be used as an argument to
258					C<require>, and is the key that appears in C<%INC> to identify a module,
259					regardless of actual local filename syntax.
260
261					=cut
262
263					sub module_notional_filename($) {
264					&check_module_name;
265					my($name) = @_;
266					$name =~ s!::!/!g;
267					return $name.".pm";
268					}
269
270					=item require_module(NAME)
271
272					This is essentially the bareword form of C<require>, in runtime form.
273					The I<NAME> is a string, which should be a valid module name (one or
274					more C<::>-separated segments). If it is not a valid name, the function
275					C<die>s.
276
277					The module specified by I<NAME> is loaded, if it hasn't been already,
278					in the manner of the bareword form of C<require>. That means that a
279					search through C<@INC> is performed, and a byte-compiled form of the
280					module will be used if available.
281
282					The return value is as for C<require>. That is, it is the value returned
283					by the module itself if the module is loaded anew, or C<1> if the module
284					was already loaded.
285
286					=cut
287
288					# Don't "use constant" here, to avoid dependencies.
289					BEGIN {
290					*_WORK_AROUND_HINT_LEAKAGE =
291					"$]" < 5.011 && !("$]" >= 5.009004 && "$]" < 5.010001)
292					? sub(){1} : sub(){0};
293					*_WORK_AROUND_BROKEN_MODULE_STATE = "$]" < 5.009 ? sub(){1} : sub(){0};
294					}
295
296					BEGIN { if(_WORK_AROUND_BROKEN_MODULE_STATE) { eval q{
297					sub Module::Runtime::__GUARD__::DESTROY {
298					delete $INC{$_[0]->[0]} if @{$_[0]};
299					}
300					1;
301					}; die $@ if $@ ne ""; } }
302
303					sub require_module($) {
304					# Localise %^H to work around [perl #68590], where the bug exists
305					# and this is a satisfactory workaround. The bug consists of
306					# %^H state leaking into each required module, polluting the
307					# module's lexical state.
308					local %^H if _WORK_AROUND_HINT_LEAKAGE;
309					if(_WORK_AROUND_BROKEN_MODULE_STATE) {
310					my $notional_filename = &module_notional_filename;
311					my $guard = bless([ $notional_filename ],
312					"Module::Runtime::__GUARD__");
313					my $result = CORE::require($notional_filename);
314					pop @$guard;
315					return $result;
316					} else {
317					return scalar(CORE::require(&module_notional_filename));
318					}
319					}
320
321					=back
322
323					=head2 Structured module use
324
325					=over
326
327					=item use_module(NAME[, VERSION])
328
329					This is essentially C<use> in runtime form, but without the importing
330					feature (which is fundamentally a compile-time thing). The I<NAME> is
331					handled just like in C<require_module> above: it must be a module name,
332					and the named module is loaded as if by the bareword form of C<require>.
333
334					If a I<VERSION> is specified, the C<VERSION> method of the loaded module is
335					called with the specified I<VERSION> as an argument. This normally serves to
336					ensure that the version loaded is at least the version required. This is
337					the same functionality provided by the I<VERSION> parameter of C<use>.
338
339					On success, the name of the module is returned. This is unlike
340					L</require_module>, and is done so that the entire call to L</use_module>
341					can be used as a class name to call a constructor, as in the example in
342					the synopsis.
343
344					=cut
345
346					sub use_module($;$) {
347					my($name, $version) = @_;
348					require_module($name);
349					$name->VERSION($version) if @_ >= 2;
350					return $name;
351					}
352
353					=item use_package_optimistically(NAME[, VERSION])
354
355					This is an analogue of L</use_module> for the situation where there is
356					uncertainty as to whether a package/class is defined in its own module
357					or by some other means. It attempts to arrange for the named package to
358					be available, either by loading a module or by doing nothing and hoping.
359
360					An attempt is made to load the named module (as if by the bareword form
361					of C<require>). If the module cannot be found then it is assumed that
362					the package was actually already loaded by other means, and no error
363					is signalled. That's the optimistic bit.
364
365					This is mostly the same operation that is performed by the L<base> pragma
366					to ensure that the specified base classes are available. The behaviour
367					of L<base> was simplified in version 2.18, and later improved in version
368					2.20, and on both occasions this function changed to match.
369
370					If a I<VERSION> is specified, the C<VERSION> method of the loaded package is
371					called with the specified I<VERSION> as an argument. This normally serves
372					to ensure that the version loaded is at least the version required.
373					On success, the name of the package is returned. These aspects of the
374					function work just like L</use_module>.
375
376					=cut
377
378					sub use_package_optimistically($;$) {
379					my($name, $version) = @_;
380					my $fn = module_notional_filename($name);
381					eval { local $SIG{__DIE__}; require_module($name); };
382					die $@ if $@ ne "" &&
383					($@ !~ /\ACan't locate \Q$fn\E .+ at \Q@{[__FILE__]}\E line/s \|\|
384					$@ =~ /^Compilation\ failed\ in\ require
385					\ at\ \Q@{[__FILE__]}\E\ line/xm);
386					$name->VERSION($version) if @_ >= 2;
387					return $name;
388					}
389
390					=back
391
392					=head2 Module name composition
393
394					=over
395
396					=item is_module_spec(PREFIX, SPEC)
397
398					Returns a truth value indicating
399					whether I<SPEC> is valid input for L</compose_module_name>.
400					See below for what that entails. Whether a I<PREFIX> is supplied affects
401					the validity of I<SPEC>, but the exact value of the prefix is unimportant,
402					so this function treats I<PREFIX> as a truth value.
403
404					=cut
405
406					sub is_module_spec($$) {
407					my($prefix, $spec) = @_;
408					return _is_string($spec) &&
409					$spec =~ ($prefix ? qr/\A$sub_module_spec_rx\z/o :
410					qr/\A$top_module_spec_rx\z/o);
411					}
412
413					=item is_valid_module_spec(PREFIX, SPEC)
414
415					Deprecated alias for L</is_module_spec>.
416
417					=cut
418
419					*is_valid_module_spec = \&is_module_spec;
420
421					=item check_module_spec(PREFIX, SPEC)
422
423					Check whether I<SPEC> is valid input for L</compose_module_name>.
424					Return normally if it is, or C<die> if it is not.
425
426					=cut
427
428					sub check_module_spec($$) {
429					unless(&is_module_spec) {
430					die +(_is_string($_[1]) ? "`$_[1]'" : "argument").
431					" is not a module specification\n";
432					}
433					}
434
435					=item compose_module_name(PREFIX, SPEC)
436
437					This function is intended to make it more convenient for a user to specify
438					a Perl module name at runtime. Users have greater need for abbreviations
439					and context-sensitivity than programmers, and Perl module names get a
440					little unwieldy. I<SPEC> is what the user specifies, and this function
441					translates it into a module name in standard form, which it returns.
442
443					I<SPEC> has syntax approximately that of a standard module name: it
444					should consist of one or more name segments, each of which consists
445					of one or more identifier characters. However, C</> is permitted as a
446					separator, in addition to the standard C<::>. The two separators are
447					entirely interchangeable.
448
449					Additionally, if I<PREFIX> is not C<undef> then it must be a module
450					name in standard form, and it is prefixed to the user-specified name.
451					The user can inhibit the prefix addition by starting I<SPEC> with a
452					separator (either C</> or C<::>).
453
454					=cut
455
456					sub compose_module_name($$) {
457					my($prefix, $spec) = @_;
458					check_module_name($prefix) if defined $prefix;
459					&check_module_spec;
460					if($spec =~ s#\A(?:/\|::)##) {
461					# OK
462					} else {
463					$spec = $prefix."::".$spec if defined $prefix;
464					}
465					$spec =~ s#/#::#g;
466					return $spec;
467					}
468
469					=back
470
471					=head1 BUGS
472
473					On Perl versions 5.7.2 to 5.8.8, if C<require> is overridden by the
474					C<CORE::GLOBAL> mechanism, it is likely to break the heuristics used by
475					L</use_package_optimistically>, making it signal an error for a missing
476					module rather than assume that it was already loaded. From Perl 5.8.9
477					onwards, and on 5.7.1 and earlier, this module can avoid being confused
478					by such an override. On the affected versions, a C<require> override
479					might be installed by L<Lexical::SealRequireHints>, if something requires
480					its bugfix but for some reason its XS implementation isn't available.
481
482					=head1 SEE ALSO
483
484					L<Lexical::SealRequireHints>,
485					L<base>,
486					L<perlfunc/require>,
487					L<perlfunc/use>
488
489					=head1 AUTHOR
490
491					Andrew Main (Zefram) <zefram@fysh.org>
492
493					=head1 COPYRIGHT
494
495					Copyright (C) 2004, 2006, 2007, 2009, 2010, 2011, 2012, 2014
496					Andrew Main (Zefram) <zefram@fysh.org>
497
498					=head1 LICENSE
499
500					This module is free software; you can redistribute it and/or modify it
501					under the same terms as Perl itself.
502
503					=cut
504
505					1;