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

Comparing Guard/Guard.pm (file contents):
Revision 1.2 by root, Sat Dec 13 17:49:12 2008 UTC vs.
Revision 1.22 by root, Sun Jul 26 08:07:11 2009 UTC

2 2
3Guard - safe cleanup blocks 3Guard - safe cleanup blocks
4 4
5=head1 SYNOPSIS 5=head1 SYNOPSIS
6 6
7 use Guard; 7 use Guard;
8
9 # temporarily chdir to "/etc" directory, but make sure
10 # to go back to "/" no matter how myfun exits:
11 sub myfun {
12 scope_guard { chdir "/" };
13 chdir "/etc";
14
15 code_that_might_die_or_does_other_fun_stuff;
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
8 22
9=head1 DESCRIPTION 23=head1 DESCRIPTION
10 24
11This module implements so-called "guards". A guard is something (usually 25This module implements so-called "guards". A guard is something (usually
12an 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
25 39
26=cut 40=cut
27 41
28package Guard; 42package Guard;
29 43
44no warnings;
45
30BEGIN { 46BEGIN {
31 $VERSION = '0.01'; 47 $VERSION = '1.021';
32 @ISA = qw(Exporter); 48 @ISA = qw(Exporter);
33 @EXPORT = qw(guard scope_guard cancel); 49 @EXPORT = qw(guard scope_guard);
34 50
35 require Exporter; 51 require Exporter;
36 52
37 require XSLoader; 53 require XSLoader;
38 XSLoader::load Guard, $VERSION; 54 XSLoader::load Guard, $VERSION;
43=item scope_guard BLOCK 59=item scope_guard BLOCK
44 60
45Registers a block that is executed when the current scope (block, 61Registers a block that is executed when the current scope (block,
46function, method, eval etc.) is exited. 62function, method, eval etc.) is exited.
47 63
64See the EXCEPTIONS section for an explanation of how exceptions
65(i.e. C<die>) are handled inside guard blocks.
66
48The description below sounds a bit complicated, but that's just because 67The description below sounds a bit complicated, but that's just because
49C<scope_guard> tries to get even corner cases "right": the goal is to 68C<scope_guard> tries to get even corner cases "right": the goal is to
50provide you with a rock solid clean up tool. 69provide you with a rock solid clean up tool.
51 70
52This is similar to this code fragment: 71The behaviour is similar to this code fragment:
53 72
54 eval ... code following scope_guard ... 73 eval ... code following scope_guard ...
55 { 74 {
56 local $@; 75 local $@;
57 eval BLOCK; 76 eval BLOCK;
60 die if $@; 79 die if $@;
61 80
62Except it is much faster, and the whole thing gets executed even when the 81Except it is much faster, and the whole thing gets executed even when the
63BLOCK calls C<exit>, C<goto>, C<last> or escapes via other means. 82BLOCK calls C<exit>, C<goto>, C<last> or escapes via other means.
64 83
65See B<EXCEPTIONS>, below, for an explanation of exception handling 84If multiple BLOCKs are registered to the same scope, they will be executed
66(C<die>) within guard blocks. 85in reverse order. Other scope-related things such as C<local> are managed
86via the same mechanism, so variables C<local>ised I<after> calling
87C<scope_guard> will be restored when the guard runs.
67 88
68Example: Temporarily change the directory to F</etc> and make sure it's 89Example: temporarily change the timezone for the current process,
69set back to F</> when the function returns: 90ensuring it will be reset when the C<if> scope is exited:
70 91
71 sub dosomething { 92 use Guard;
72 scope_guard { chdir "/" }; 93 use POSIX ();
73 chdir "/etc";
74 94
75 ... 95 if ($need_to_switch_tz) {
96 # make sure we call tzset after $ENV{TZ} has been restored
97 scope_guard { POSIX::tzset };
98
99 # localise after the scope_guard, so it gets undone in time
100 local $ENV{TZ} = "Europe/London";
101 POSIX::tzset;
102
103 # do something with the new timezone
76 } 104 }
77 105
78=item my $guard = guard BLOCK 106=item my $guard = guard BLOCK
79 107
80Behaves the same as C<scope_guard>, except that instead of executing 108Behaves the same as C<scope_guard>, except that instead of executing
81the block on scope exit, it returns an object whose lifetime determines 109the block on scope exit, it returns an object whose lifetime determines
82when the BLOCK gets executed: when the last reference to the object gets 110when the BLOCK gets executed: when the last reference to the object gets
83destroyed, the BLOCK gets executed as with C<scope_guard>. 111destroyed, the BLOCK gets executed as with C<scope_guard>.
84 112
85The returned object can be copied as many times as you want. 113See the EXCEPTIONS section for an explanation of how exceptions
86 114(i.e. C<die>) are handled inside guard blocks.
87See B<EXCEPTIONS>, below, for an explanation of exception handling
88(C<die>) within guard blocks.
89 115
90Example: acquire a Coro::Semaphore for a second by registering a 116Example: acquire a Coro::Semaphore for a second by registering a
91timer. The timer callback references the guard used to unlock it again. 117timer. The timer callback references the guard used to unlock it
118again. (Please ignore the fact that C<Coro::Semaphore> has a C<guard>
119method that does this already):
92 120
121 use Guard;
93 use AnyEvent; 122 use Coro::AnyEvent;
94 use Coro::Semaphore; 123 use Coro::Semaphore;
95 124
96 my $sem = new Coro::Semaphore; 125 my $sem = new Coro::Semaphore;
97 126
98 sub lock_1s { 127 sub lock_for_a_second {
99 $sem->down; 128 $sem->down;
100 my $guard = guard { $sem->up }; 129 my $guard = guard { $sem->up };
101 130
102 my $timer; 131 Coro::AnyEvent::sleep 1;
103 $timer = AnyEvent->timer (after => 1, sub { 132
104 # do something 133 # $sem->up gets executed when returning
105 undef $sem;
106 undef $timer;
107 });
108 } 134 }
109 135
110The advantage of doing this with a guard instead of simply calling C<< 136The advantage of doing this with a guard instead of simply calling C<<
111$sem->down >> in the callback is that you can opt not to create the timer, 137$sem->down >> in the callback is that you can opt not to create the timer,
112or your code can throw an exception before it can create the timer, or you 138or your code can throw an exception before it can create the timer (or
113can create multiple timers or other event watchers and only when the last 139the thread gets canceled), or you can create multiple timers or other
114one gets executed will the lock be unlocked. 140event watchers and only when the last one gets executed will the lock be
141unlocked. Using the C<guard>, you do not have to worry about catching all
142the places where you have to unlock the semaphore.
115 143
116=item Guard::cancel $guard 144=item $guard->cancel
117 145
118Calling this function will "disable" the guard object returned by the 146Calling this function will "disable" the guard object returned by the
119C<guard> function, i.e. it will free the BLOCK originally passed to 147C<guard> function, i.e. it will free the BLOCK originally passed to
120C<guard >and will arrange for the BLOCK not to be executed. 148C<guard >and will arrange for the BLOCK not to be executed.
121 149
122This can be useful when you use C<guard> to create a fatal cleanup handler 150This can be useful when you use C<guard> to create a cleanup handler to be
123and later decide it is no longer needed. 151called under fatal conditions and later decide it is no longer needed.
124 152
125=cut 153=cut
126 154
1271; 1551;
128 156
129=back 157=back
130 158
131=head1 EXCEPTIONS 159=head1 EXCEPTIONS
132 160
133Guard blocks should not normally throw exceptions (e.g. C<die>), after 161Guard blocks should not normally throw exceptions (that is, C<die>). After
134all, they are usually used to clean up after such exceptions. However, if 162all, they are usually used to clean up after such exceptions. However,
135something truly exceptional is happening, a guard block should be allowed 163if something truly exceptional is happening, a guard block should of
136to die. Also, programming errors are a large source of exceptions, and the 164course be allowed to die. Also, programming errors are a large source of
137programmer certainly wants to know about those. 165exceptions, and the programmer certainly wants to know about those.
138 166
139Since in most cases, the block executing when the guard gets executes does 167Since in most cases, the block executing when the guard gets executed does
140not know or does not care about the guard blocks, it makes little sense to 168not know or does not care about the guard blocks, it makes little sense to
141let containing code handle the exception. 169let containing code handle the exception.
142 170
143Therefore, whenever a guard block throws an exception, it will be caught, 171Therefore, whenever a guard block throws an exception, it will be caught
144and this module will call the code reference stored in C<$Guard::DIED> 172by Guard, followed by calling the code reference stored in C<$Guard::DIED>
145(with C<$@> set to the actual exception), which is similar to how most 173(with C<$@> set to the actual exception), which is similar to how most
146event loops handle this case. 174event loops handle this case.
147 175
176The default for C<$Guard::DIED> is to call C<warn "$@">, i.e. the error is
177printed as a warning and the program continues.
178
179The C<$@> variable will be restored to its value before the guard call in
180all cases, so guards will not disturb C<$@> in any way.
181
148The code reference stored in C<$Guard::DIED> should not die (behaviour is 182The code reference stored in C<$Guard::DIED> should not die (behaviour is
149not guaranteed, but right now, the exception will simply be ignored). 183not guaranteed, but right now, the exception will simply be ignored).
150 184
151The default for C<$Guard::DIED> is to call C<warn "$@">.
152
153=head1 AUTHOR 185=head1 AUTHOR
154 186
155 Marc Lehmann <schmorp@schmorp.de> 187 Marc Lehmann <schmorp@schmorp.de>
156 http://home.schmorp.de/ 188 http://home.schmorp.de/
157 189
158=head1 THANKS 190=head1 THANKS
159 191
160To Marco Maisenhelder, who reminded me of the C<$Guard::DIED> solution to 192Thanks to Marco Maisenhelder, who reminded me of the C<$Guard::DIED>
161the problem of exceptions. 193solution to the problem of exceptions.
194
195=head1 SEE ALSO
196
197L<Scope::Guard> and L<Sub::ScopeFinalizer>, which actually implement
198dynamic guards only, not scoped guards, and have a lot higher CPU, memory
199and typing overhead.
200
201L<Hook::Scope>, which has apparently never been finished and can corrupt
202memory when used.
162 203
163=cut 204=cut
164 205

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines