Commit | Line | Data |
95bebf8c |
1 | |
2 | package Class::C3; |
3 | |
4 | use strict; |
5 | use warnings; |
6 | |
95bebf8c |
7 | use Scalar::Util 'blessed'; |
2ffffc6d |
8 | use Algorithm::C3; |
95bebf8c |
9 | |
f4a893b2 |
10 | our $VERSION = '0.13'; |
d401eda1 |
11 | |
12 | # this is our global stash of both |
13 | # MRO's and method dispatch tables |
14 | # the structure basically looks like |
15 | # this: |
16 | # |
17 | # $MRO{$class} = { |
18 | # MRO => [ <class precendence list> ], |
19 | # methods => { |
20 | # orig => <original location of method>, |
21 | # code => \&<ref to original method> |
680100b1 |
22 | # }, |
23 | # has_overload_fallback => (1 | 0) |
d401eda1 |
24 | # } |
25 | # |
f7facd7b |
26 | our %MRO; |
95bebf8c |
27 | |
d0e2efe5 |
28 | # use these for debugging ... |
d401eda1 |
29 | sub _dump_MRO_table { %MRO } |
d401eda1 |
30 | our $TURN_OFF_C3 = 0; |
31 | |
95bebf8c |
32 | sub import { |
33 | my $class = caller(); |
d401eda1 |
34 | # skip if the caller is main:: |
35 | # since that is clearly not relevant |
95bebf8c |
36 | return if $class eq 'main'; |
d401eda1 |
37 | return if $TURN_OFF_C3; |
38 | # make a note to calculate $class |
39 | # during INIT phase |
f7facd7b |
40 | $MRO{$class} = undef unless exists $MRO{$class}; |
95bebf8c |
41 | } |
42 | |
d401eda1 |
43 | ## initializers |
44 | |
d401eda1 |
45 | sub initialize { |
46 | # why bother if we don't have anything ... |
47 | return unless keys %MRO; |
48 | _calculate_method_dispatch_tables(); |
49 | _apply_method_dispatch_tables(); |
5d5c86d9 |
50 | %next::METHOD_CACHE = (); |
d401eda1 |
51 | } |
52 | |
d0e2efe5 |
53 | sub uninitialize { |
54 | # why bother if we don't have anything ... |
55 | return unless keys %MRO; |
56 | _remove_method_dispatch_tables(); |
5d5c86d9 |
57 | %next::METHOD_CACHE = (); |
d0e2efe5 |
58 | } |
59 | |
60 | sub reinitialize { |
61 | uninitialize(); |
62 | # clean up the %MRO before we re-initialize |
63 | $MRO{$_} = undef foreach keys %MRO; |
64 | initialize(); |
65 | } |
66 | |
d401eda1 |
67 | ## functions for applying C3 to classes |
68 | |
69 | sub _calculate_method_dispatch_tables { |
f4a893b2 |
70 | my %merge_cache; |
95bebf8c |
71 | foreach my $class (keys %MRO) { |
f4a893b2 |
72 | _calculate_method_dispatch_table($class, \%merge_cache); |
95bebf8c |
73 | } |
d401eda1 |
74 | } |
75 | |
76 | sub _calculate_method_dispatch_table { |
f4a893b2 |
77 | my ($class, $merge_cache) = @_; |
d401eda1 |
78 | no strict 'refs'; |
f4a893b2 |
79 | my @MRO = calculateMRO($class, $merge_cache); |
d401eda1 |
80 | $MRO{$class} = { MRO => \@MRO }; |
680100b1 |
81 | my $has_overload_fallback = 0; |
d401eda1 |
82 | my %methods; |
83 | # NOTE: |
84 | # we do @MRO[1 .. $#MRO] here because it |
85 | # makes no sense to interogate the class |
86 | # which you are calculating for. |
87 | foreach my $local (@MRO[1 .. $#MRO]) { |
680100b1 |
88 | # if overload has tagged this module to |
89 | # have use "fallback", then we want to |
90 | # grab that value |
91 | $has_overload_fallback = ${"${local}::()"} |
92 | if defined ${"${local}::()"}; |
d401eda1 |
93 | foreach my $method (grep { defined &{"${local}::$_"} } keys %{"${local}::"}) { |
94 | # skip if already overriden in local class |
95 | next unless !defined *{"${class}::$method"}{CODE}; |
96 | $methods{$method} = { |
97 | orig => "${local}::$method", |
98 | code => \&{"${local}::$method"} |
99 | } unless exists $methods{$method}; |
95bebf8c |
100 | } |
d401eda1 |
101 | } |
102 | # now stash them in our %MRO table |
680100b1 |
103 | $MRO{$class}->{methods} = \%methods; |
104 | $MRO{$class}->{has_overload_fallback} = $has_overload_fallback; |
d401eda1 |
105 | } |
106 | |
107 | sub _apply_method_dispatch_tables { |
108 | foreach my $class (keys %MRO) { |
109 | _apply_method_dispatch_table($class); |
110 | } |
95bebf8c |
111 | } |
112 | |
d401eda1 |
113 | sub _apply_method_dispatch_table { |
114 | my $class = shift; |
115 | no strict 'refs'; |
680100b1 |
116 | ${"${class}::()"} = $MRO{$class}->{has_overload_fallback} |
117 | if $MRO{$class}->{has_overload_fallback}; |
d401eda1 |
118 | foreach my $method (keys %{$MRO{$class}->{methods}}) { |
119 | *{"${class}::$method"} = $MRO{$class}->{methods}->{$method}->{code}; |
120 | } |
121 | } |
122 | |
d0e2efe5 |
123 | sub _remove_method_dispatch_tables { |
124 | foreach my $class (keys %MRO) { |
125 | _remove_method_dispatch_table($class); |
126 | } |
127 | } |
128 | |
129 | sub _remove_method_dispatch_table { |
130 | my $class = shift; |
131 | no strict 'refs'; |
680100b1 |
132 | delete ${"${class}::"}{"()"} if $MRO{$class}->{has_overload_fallback}; |
d0e2efe5 |
133 | foreach my $method (keys %{$MRO{$class}->{methods}}) { |
5dd9299c |
134 | delete ${"${class}::"}{$method} |
135 | if defined *{"${class}::${method}"}{CODE} && |
136 | (*{"${class}::${method}"}{CODE} eq $MRO{$class}->{methods}->{$method}->{code}); |
d0e2efe5 |
137 | } |
138 | } |
139 | |
d401eda1 |
140 | ## functions for calculating C3 MRO |
141 | |
95bebf8c |
142 | sub calculateMRO { |
f4a893b2 |
143 | my ($class, $merge_cache) = @_; |
2ffffc6d |
144 | return Algorithm::C3::merge($class, sub { |
145 | no strict 'refs'; |
146 | @{$_[0] . '::ISA'}; |
f4a893b2 |
147 | }, $merge_cache); |
95bebf8c |
148 | } |
149 | |
5d5c86d9 |
150 | package # hide me from PAUSE |
151 | next; |
152 | |
153 | use strict; |
154 | use warnings; |
155 | |
156 | use Scalar::Util 'blessed'; |
157 | |
ac6b0914 |
158 | our $VERSION = '0.05'; |
5d5c86d9 |
159 | |
160 | our %METHOD_CACHE; |
161 | |
fa91a1c7 |
162 | sub method { |
163 | my $indirect = caller() =~ /^(?:next|maybe::next)$/; |
164 | my $level = $indirect ? 2 : 1; |
165 | |
7bb662d7 |
166 | my ($method_caller, $label, @label); |
ac6b0914 |
167 | while ($method_caller = (caller($level++))[3]) { |
7bb662d7 |
168 | @label = (split '::', $method_caller); |
169 | $label = pop @label; |
170 | last unless |
171 | $label eq '(eval)' || |
172 | $label eq '__ANON__'; |
ac6b0914 |
173 | } |
5d5c86d9 |
174 | my $caller = join '::' => @label; |
175 | my $self = $_[0]; |
176 | my $class = blessed($self) || $self; |
177 | |
fa91a1c7 |
178 | my $method = $METHOD_CACHE{"$class|$caller|$label"} ||= do { |
322a5920 |
179 | |
180 | my @MRO = Class::C3::calculateMRO($class); |
181 | |
182 | my $current; |
183 | while ($current = shift @MRO) { |
184 | last if $caller eq $current; |
185 | } |
186 | |
187 | no strict 'refs'; |
188 | my $found; |
189 | foreach my $class (@MRO) { |
190 | next if (defined $Class::C3::MRO{$class} && |
191 | defined $Class::C3::MRO{$class}{methods}{$label}); |
192 | last if (defined ($found = *{$class . '::' . $label}{CODE})); |
193 | } |
194 | |
322a5920 |
195 | $found; |
196 | }; |
fa91a1c7 |
197 | |
198 | return $method if $indirect; |
199 | |
200 | die "No next::method '$label' found for $self" if !$method; |
201 | |
202 | goto &{$method}; |
322a5920 |
203 | } |
5d5c86d9 |
204 | |
fa91a1c7 |
205 | sub can { method($_[0]) } |
5d5c86d9 |
206 | |
fa91a1c7 |
207 | package # hide me from PAUSE |
208 | maybe::next; |
209 | |
210 | use strict; |
211 | use warnings; |
212 | |
213 | our $VERSION = '0.01'; |
214 | |
215 | sub method { (next::method($_[0]) || return)->(@_) } |
5d5c86d9 |
216 | |
95bebf8c |
217 | 1; |
218 | |
219 | __END__ |
220 | |
221 | =pod |
222 | |
223 | =head1 NAME |
224 | |
225 | Class::C3 - A pragma to use the C3 method resolution order algortihm |
226 | |
227 | =head1 SYNOPSIS |
228 | |
229 | package A; |
230 | use Class::C3; |
231 | sub hello { 'A::hello' } |
232 | |
233 | package B; |
234 | use base 'A'; |
235 | use Class::C3; |
236 | |
237 | package C; |
238 | use base 'A'; |
239 | use Class::C3; |
240 | |
241 | sub hello { 'C::hello' } |
242 | |
243 | package D; |
244 | use base ('B', 'C'); |
245 | use Class::C3; |
246 | |
247 | # Classic Diamond MI pattern |
d401eda1 |
248 | # <A> |
249 | # / \ |
250 | # <B> <C> |
251 | # \ / |
252 | # <D> |
95bebf8c |
253 | |
254 | package main; |
2ffffc6d |
255 | |
256 | # initializez the C3 module |
257 | # (formerly called in INIT) |
258 | Class::C3::initialize(); |
95bebf8c |
259 | |
260 | print join ', ' => Class::C3::calculateMRO('Diamond_D') # prints D, B, C, A |
261 | |
262 | print D->hello() # prints 'C::hello' instead of the standard p5 'A::hello' |
263 | |
264 | D->can('hello')->(); # can() also works correctly |
265 | UNIVERSAL::can('D', 'hello'); # as does UNIVERSAL::can() |
266 | |
267 | =head1 DESCRIPTION |
268 | |
2ffffc6d |
269 | This is pragma to change Perl 5's standard method resolution order from depth-first left-to-right |
270 | (a.k.a - pre-order) to the more sophisticated C3 method resolution order. |
95bebf8c |
271 | |
272 | =head2 What is C3? |
273 | |
274 | C3 is the name of an algorithm which aims to provide a sane method resolution order under multiple |
275 | inheritence. It was first introduced in the langauge Dylan (see links in the L<SEE ALSO> section), |
276 | and then later adopted as the prefered MRO (Method Resolution Order) for the new-style classes in |
277 | Python 2.3. Most recently it has been adopted as the 'canonical' MRO for Perl 6 classes, and the |
278 | default MRO for Parrot objects as well. |
279 | |
280 | =head2 How does C3 work. |
281 | |
282 | C3 works by always preserving local precendence ordering. This essentially means that no class will |
283 | appear before any of it's subclasses. Take the classic diamond inheritence pattern for instance: |
284 | |
d401eda1 |
285 | <A> |
286 | / \ |
287 | <B> <C> |
288 | \ / |
289 | <D> |
95bebf8c |
290 | |
291 | The standard Perl 5 MRO would be (D, B, A, C). The result being that B<A> appears before B<C>, even |
292 | though B<C> is the subclass of B<A>. The C3 MRO algorithm however, produces the following MRO |
293 | (D, B, C, A), which does not have this same issue. |
294 | |
295 | This example is fairly trival, for more complex examples and a deeper explaination, see the links in |
296 | the L<SEE ALSO> section. |
297 | |
298 | =head2 How does this module work? |
299 | |
2ffffc6d |
300 | This module uses a technique similar to Perl 5's method caching. When C<Class::C3::initialize> is |
301 | called, this module calculates the MRO of all the classes which called C<use Class::C3>. It then |
302 | gathers information from the symbol tables of each of those classes, and builds a set of method |
303 | aliases for the correct dispatch ordering. Once all these C3-based method tables are created, it |
304 | then adds the method aliases into the local classes symbol table. |
95bebf8c |
305 | |
306 | The end result is actually classes with pre-cached method dispatch. However, this caching does not |
307 | do well if you start changing your C<@ISA> or messing with class symbol tables, so you should consider |
308 | your classes to be effectively closed. See the L<CAVEATS> section for more details. |
309 | |
d401eda1 |
310 | =head1 OPTIONAL LOWERCASE PRAGMA |
311 | |
312 | This release also includes an optional module B<c3> in the F<opt/> folder. I did not include this in |
313 | the regular install since lowercase module names are considered I<"bad"> by some people. However I |
314 | think that code looks much nicer like this: |
315 | |
316 | package MyClass; |
317 | use c3; |
318 | |
319 | The the more clunky: |
320 | |
321 | package MyClass; |
322 | use Class::C3; |
323 | |
324 | But hey, it's your choice, thats why it is optional. |
325 | |
95bebf8c |
326 | =head1 FUNCTIONS |
327 | |
328 | =over 4 |
329 | |
330 | =item B<calculateMRO ($class)> |
331 | |
332 | Given a C<$class> this will return an array of class names in the proper C3 method resolution order. |
333 | |
d401eda1 |
334 | =item B<initialize> |
335 | |
2ffffc6d |
336 | This B<must be called> to initalize the C3 method dispatch tables, this module B<will not work> if |
5f01eb5f |
337 | you do not do this. It is advised to do this as soon as possible B<after> loading any classes which |
338 | use C3. Here is a quick code example: |
339 | |
340 | package Foo; |
341 | use Class::C3; |
342 | # ... Foo methods here |
343 | |
344 | package Bar; |
345 | use Class::C3; |
346 | use base 'Foo'; |
347 | # ... Bar methods here |
348 | |
349 | package main; |
350 | |
351 | Class::C3::initialize(); # now it is safe to use Foo and Bar |
2ffffc6d |
352 | |
353 | This function used to be called automatically for you in the INIT phase of the perl compiler, but |
354 | that lead to warnings if this module was required at runtime. After discussion with my user base |
355 | (the L<DBIx::Class> folks), we decided that calling this in INIT was more of an annoyance than a |
356 | convience. I apologize to anyone this causes problems for (although i would very suprised if I had |
357 | any other users other than the L<DBIx::Class> folks). The simplest solution of course is to define |
358 | your own INIT method which calls this function. |
d401eda1 |
359 | |
360 | NOTE: |
d0e2efe5 |
361 | This can B<not> be used to re-load the dispatch tables for all classes. Use C<reinitialize> for that. |
362 | |
363 | =item B<uninitialize> |
364 | |
365 | Calling this function results in the removal of all cached methods, and the restoration of the old Perl 5 |
366 | style dispatch order (depth-first, left-to-right). |
367 | |
368 | =item B<reinitialize> |
369 | |
370 | This effectively calls C<uninitialize> followed by C<initialize> the result of which is a reloading of |
371 | B<all> the calculated C3 dispatch tables. |
372 | |
373 | It should be noted that if you have a large class library, this could potentially be a rather costly |
374 | operation. |
d401eda1 |
375 | |
95bebf8c |
376 | =back |
377 | |
5d5c86d9 |
378 | =head1 METHOD REDISPATCHING |
379 | |
380 | It is always useful to be able to re-dispatch your method call to the "next most applicable method". This |
381 | module provides a pseudo package along the lines of C<SUPER::> or C<NEXT::> which will re-dispatch the |
382 | method along the C3 linearization. This is best show with an examples. |
383 | |
384 | # a classic diamond MI pattern ... |
385 | <A> |
386 | / \ |
387 | <B> <C> |
388 | \ / |
389 | <D> |
390 | |
391 | package A; |
392 | use c3; |
393 | sub foo { 'A::foo' } |
394 | |
395 | package B; |
396 | use base 'A'; |
397 | use c3; |
398 | sub foo { 'B::foo => ' . (shift)->next::method() } |
399 | |
400 | package B; |
401 | use base 'A'; |
402 | use c3; |
403 | sub foo { 'C::foo => ' . (shift)->next::method() } |
404 | |
405 | package D; |
406 | use base ('B', 'C'); |
407 | use c3; |
408 | sub foo { 'D::foo => ' . (shift)->next::method() } |
409 | |
410 | print D->foo; # prints out "D::foo => B::foo => C::foo => A::foo" |
411 | |
412 | A few things to note. First, we do not require you to add on the method name to the C<next::method> |
413 | call (this is unlike C<NEXT::> and C<SUPER::> which do require that). This helps to enforce the rule |
414 | that you cannot dispatch to a method of a different name (this is how C<NEXT::> behaves as well). |
415 | |
416 | The next thing to keep in mind is that you will need to pass all arguments to C<next::method> it can |
417 | not automatically use the current C<@_>. |
418 | |
322a5920 |
419 | If C<next::method> cannot find a next method to re-dispatch the call to, it will throw an exception. |
420 | You can use C<next::can> to see if C<next::method> will succeed before you call it like so: |
421 | |
422 | $self->next::method(@_) if $self->next::can; |
423 | |
fa91a1c7 |
424 | Additionally, you can use C<maybe::next::method> as a shortcut to only call the next method if it exists. |
425 | The previous example could be simply written as: |
426 | |
427 | $self->maybe::next::method(@_); |
322a5920 |
428 | |
2ffffc6d |
429 | There are some caveats about using C<next::method>, see below for those. |
95bebf8c |
430 | |
2ffffc6d |
431 | =head1 CAVEATS |
95bebf8c |
432 | |
2ffffc6d |
433 | This module used to be labeled as I<experimental>, however it has now been pretty heavily tested by |
434 | the good folks over at L<DBIx::Class> and I am confident this module is perfectly usable for |
435 | whatever your needs might be. |
95bebf8c |
436 | |
2ffffc6d |
437 | But there are still caveats, so here goes ... |
95bebf8c |
438 | |
439 | =over 4 |
440 | |
441 | =item Use of C<SUPER::>. |
442 | |
443 | The idea of C<SUPER::> under multiple inheritence is ambigious, and generally not recomended anyway. |
444 | However, it's use in conjuntion with this module is very much not recommended, and in fact very |
5d5c86d9 |
445 | discouraged. The recommended approach is to instead use the supplied C<next::method> feature, see |
446 | more details on it's usage above. |
95bebf8c |
447 | |
448 | =item Changing C<@ISA>. |
449 | |
450 | It is the author's opinion that changing C<@ISA> at runtime is pure insanity anyway. However, people |
451 | do it, so I must caveat. Any changes to the C<@ISA> will not be reflected in the MRO calculated by this |
d0e2efe5 |
452 | module, and therefor probably won't even show up. If you do this, you will need to call C<reinitialize> |
453 | in order to recalulate B<all> method dispatch tables. See the C<reinitialize> documentation and an example |
454 | in F<t/20_reinitialize.t> for more information. |
95bebf8c |
455 | |
456 | =item Adding/deleting methods from class symbol tables. |
457 | |
2ffffc6d |
458 | This module calculates the MRO for each requested class by interogatting the symbol tables of said classes. |
459 | So any symbol table manipulation which takes place after our INIT phase is run will not be reflected in |
460 | the calculated MRO. Just as with changing the C<@ISA>, you will need to call C<reinitialize> for any |
461 | changes you make to take effect. |
95bebf8c |
462 | |
2ffffc6d |
463 | =item Calling C<next::method> from methods defined outside the class |
95bebf8c |
464 | |
2ffffc6d |
465 | There is an edge case when using C<next::method> from within a subroutine which was created in a different |
466 | module than the one it is called from. It sounds complicated, but it really isn't. Here is an example which |
467 | will not work correctly: |
15eeb546 |
468 | |
2ffffc6d |
469 | *Foo::foo = sub { (shift)->next::method(@_) }; |
470 | |
471 | The problem exists because the anonymous subroutine being assigned to the glob C<*Foo::foo> will show up |
472 | in the call stack as being called C<__ANON__> and not C<foo> as you might expect. Since C<next::method> |
473 | uses C<caller> to find the name of the method it was called in, it will fail in this case. |
15eeb546 |
474 | |
2ffffc6d |
475 | But fear not, there is a simple solution. The module C<Sub::Name> will reach into the perl internals and |
476 | assign a name to an anonymous subroutine for you. Simply do this: |
477 | |
478 | use Sub::Name 'subname'; |
479 | *Foo::foo = subname 'Foo::foo' => sub { (shift)->next::method(@_) }; |
15eeb546 |
480 | |
2ffffc6d |
481 | and things will Just Work. Of course this is not always possible to do, but to be honest, I just can't |
482 | manage to find a workaround for it, so until someone gives me a working patch this will be a known |
483 | limitation of this module. |
15eeb546 |
484 | |
5d5c86d9 |
485 | =back |
15eeb546 |
486 | |
5d5c86d9 |
487 | =head1 CODE COVERAGE |
15eeb546 |
488 | |
ac6b0914 |
489 | I use B<Devel::Cover> to test the code coverage of my tests, below is the B<Devel::Cover> report on this |
490 | module's test suite. |
5d5c86d9 |
491 | |
492 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
493 | File stmt bran cond sub pod time total |
494 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
58f0eafe |
495 | Class/C3.pm 98.3 84.4 80.0 96.2 100.0 98.4 94.4 |
5d5c86d9 |
496 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
58f0eafe |
497 | Total 98.3 84.4 80.0 96.2 100.0 98.4 94.4 |
5d5c86d9 |
498 | ---------------------------- ------ ------ ------ ------ ------ ------ ------ |
15eeb546 |
499 | |
95bebf8c |
500 | =head1 SEE ALSO |
501 | |
502 | =head2 The original Dylan paper |
503 | |
504 | =over 4 |
505 | |
506 | =item L<http://www.webcom.com/haahr/dylan/linearization-oopsla96.html> |
507 | |
508 | =back |
509 | |
510 | =head2 The prototype Perl 6 Object Model uses C3 |
511 | |
512 | =over 4 |
513 | |
514 | =item L<http://svn.openfoundry.org/pugs/perl5/Perl6-MetaModel/> |
515 | |
516 | =back |
517 | |
518 | =head2 Parrot now uses C3 |
519 | |
520 | =over 4 |
521 | |
522 | =item L<http://aspn.activestate.com/ASPN/Mail/Message/perl6-internals/2746631> |
523 | |
524 | =item L<http://use.perl.org/~autrijus/journal/25768> |
525 | |
526 | =back |
527 | |
528 | =head2 Python 2.3 MRO related links |
529 | |
530 | =over 4 |
531 | |
532 | =item L<http://www.python.org/2.3/mro.html> |
533 | |
534 | =item L<http://www.python.org/2.2.2/descrintro.html#mro> |
535 | |
536 | =back |
537 | |
538 | =head2 C3 for TinyCLOS |
539 | |
540 | =over 4 |
541 | |
542 | =item L<http://www.call-with-current-continuation.org/eggs/c3.html> |
543 | |
544 | =back |
545 | |
bad9dc59 |
546 | =head1 ACKNOWLEGEMENTS |
547 | |
548 | =over 4 |
549 | |
550 | =item Thanks to Matt S. Trout for using this module in his module L<DBIx::Class> |
551 | and finding many bugs and providing fixes. |
552 | |
553 | =item Thanks to Justin Guenther for making C<next::method> more robust by handling |
554 | calls inside C<eval> and anon-subs. |
555 | |
f480cda1 |
556 | =item Thanks to Robert Norris for adding support for C<next::can> and |
557 | C<maybe::next::method>. |
558 | |
bad9dc59 |
559 | =back |
560 | |
95bebf8c |
561 | =head1 AUTHOR |
562 | |
d401eda1 |
563 | Stevan Little, E<lt>stevan@iinteractive.comE<gt> |
95bebf8c |
564 | |
565 | =head1 COPYRIGHT AND LICENSE |
566 | |
08c29211 |
567 | Copyright 2005, 2006 by Infinity Interactive, Inc. |
95bebf8c |
568 | |
569 | L<http://www.iinteractive.com> |
570 | |
571 | This library is free software; you can redistribute it and/or modify |
572 | it under the same terms as Perl itself. |
573 | |
f4a893b2 |
574 | =cut |