ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Guard/Guard.pm
(Generate patch)

Comparing Guard/Guard.pm (file contents):
Revision 1.9 by root, Sat Dec 13 18:49:22 2008 UTC vs.
Revision 1.21 by root, Sun Jul 19 05:44:10 2009 UTC

… …
10 # to go back to "/" no matter how myfun exits: 10 # to go back to "/" no matter how myfun exits:
11 sub myfun { 11 sub myfun {
12 scope_guard { chdir "/" }; 12 scope_guard { chdir "/" };
13 chdir "/etc"; 13 chdir "/etc";
14 14
15 call_function_that_might_die_or_other_fun_stuff; 15 code_that_might_die_or_does_other_fun_stuff;
16 } 16 }
17
18 # create an object that, when the last reference to it is gone,
19 # invokes the given codeblock:
20 my $guard = guard { print "destroyed!\n" };
21 undef $guard; # probably destroyed here
17 22
18=head1 DESCRIPTION 23=head1 DESCRIPTION
19 24
20This module implements so-called "guards". A guard is something (usually 25This module implements so-called "guards". A guard is something (usually
21an object) that "guards" a resource, ensuring that it is cleaned up when 26an object) that "guards" a resource, ensuring that it is cleaned up when
… …
34 39
35=cut 40=cut
36 41
37package Guard; 42package Guard;
38 43
44no warnings;
45
39BEGIN { 46BEGIN {
40 $VERSION = '0.1'; 47 $VERSION = '1.021';
41 @ISA = qw(Exporter); 48 @ISA = qw(Exporter);
42 @EXPORT = qw(guard scope_guard); 49 @EXPORT = qw(guard scope_guard);
43 50
44 require Exporter; 51 require Exporter;
45 52
… …
107 114
108See the EXCEPTIONS section for an explanation of how exceptions 115See the EXCEPTIONS section for an explanation of how exceptions
109(i.e. C<die>) are handled inside guard blocks. 116(i.e. C<die>) are handled inside guard blocks.
110 117
111Example: acquire a Coro::Semaphore for a second by registering a 118Example: acquire a Coro::Semaphore for a second by registering a
112timer. The timer callback references the guard used to unlock it again. 119timer. The timer callback references the guard used to unlock it
120again. (Please ignore the fact that C<Coro::Semaphore> has a C<guard>
121method that does this already):
113 122
114 use Guard; 123 use Guard;
115 use AnyEvent; 124 use AnyEvent;
116 use Coro::Semaphore; 125 use Coro::Semaphore;
117 126
… …
131 140
132The advantage of doing this with a guard instead of simply calling C<< 141The advantage of doing this with a guard instead of simply calling C<<
133$sem->down >> in the callback is that you can opt not to create the timer, 142$sem->down >> in the callback is that you can opt not to create the timer,
134or your code can throw an exception before it can create the timer, or you 143or your code can throw an exception before it can create the timer, or you
135can create multiple timers or other event watchers and only when the last 144can create multiple timers or other event watchers and only when the last
136one gets executed will the lock be unlocked. 145one gets executed will the lock be unlocked. Using the C<guard>, you do
146not have to worry about catching all the places where you have to unlock
147the semaphore.
137 148
138=item Guard::cancel $guard 149=item $guard->cancel
139 150
140Calling this function will "disable" the guard object returned by the 151Calling this function will "disable" the guard object returned by the
141C<guard> function, i.e. it will free the BLOCK originally passed to 152C<guard> function, i.e. it will free the BLOCK originally passed to
142C<guard >and will arrange for the BLOCK not to be executed. 153C<guard >and will arrange for the BLOCK not to be executed.
143 154
… …
156all, they are usually used to clean up after such exceptions. However, if 167all, they are usually used to clean up after such exceptions. However, if
157something truly exceptional is happening, a guard block should be allowed 168something truly exceptional is happening, a guard block should be allowed
158to die. Also, programming errors are a large source of exceptions, and the 169to die. Also, programming errors are a large source of exceptions, and the
159programmer certainly wants to know about those. 170programmer certainly wants to know about those.
160 171
161Since in most cases, the block executing when the guard gets executes does 172Since in most cases, the block executing when the guard gets executed does
162not know or does not care about the guard blocks, it makes little sense to 173not know or does not care about the guard blocks, it makes little sense to
163let containing code handle the exception. 174let containing code handle the exception.
164 175
165Therefore, whenever a guard block throws an exception, it will be caught, 176Therefore, whenever a guard block throws an exception, it will be caught,
166and this module will call the code reference stored in C<$Guard::DIED> 177followed by calling the code reference stored in C<$Guard::DIED> (with
167(with C<$@> set to the actual exception), which is similar to how most 178C<$@> set to the actual exception), which is similar to how most event
168event loops handle this case. 179loops handle this case.
180
181The default for C<$Guard::DIED> is to call C<warn "$@">.
182
183The C<$@> variable will be restored to its value before the guard call in
184all cases, so guards will not disturb C<$@> in any way.
169 185
170The code reference stored in C<$Guard::DIED> should not die (behaviour is 186The code reference stored in C<$Guard::DIED> should not die (behaviour is
171not guaranteed, but right now, the exception will simply be ignored). 187not guaranteed, but right now, the exception will simply be ignored).
172 188
173The default for C<$Guard::DIED> is to call C<warn "$@">.
174
175=head1 AUTHOR 189=head1 AUTHOR
176 190
177 Marc Lehmann <schmorp@schmorp.de> 191 Marc Lehmann <schmorp@schmorp.de>
178 http://home.schmorp.de/ 192 http://home.schmorp.de/
179 193
180=head1 THANKS 194=head1 THANKS
181 195
182Thanks to Marco Maisenhelder, who reminded me of the C<$Guard::DIED> 196Thanks to Marco Maisenhelder, who reminded me of the C<$Guard::DIED>
183solution to the problem of exceptions. 197solution to the problem of exceptions.
184 198
199=head1 SEE ALSO
200
201L<Scope::Guard> and L<Sub::ScopeFinalizer>, which actually implement
202dynamic, not scoped guards, and have a lot higher CPU, memory and typing
203overhead.
204
205L<Hook::Scope>, which has apparently never been finished and corrupts
206memory when used.
207
185=cut 208=cut
186 209

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines