1 |
=head1 NAME |
2 |
|
3 |
Guard - safe cleanup blocks |
4 |
|
5 |
=head1 SYNOPSIS |
6 |
|
7 |
use Guard; |
8 |
|
9 |
=head1 DESCRIPTION |
10 |
|
11 |
This module implements so-called "guards". A guard is something (usually |
12 |
an object) that "guards" a resource, ensuring that it is cleaned up when |
13 |
expected. |
14 |
|
15 |
Specifically, this module supports two different types of guards: guard |
16 |
objects, which execute a given code block when destroyed, and scoped |
17 |
guards, which are tied to the scope exit. |
18 |
|
19 |
=head1 FUNCTIONS |
20 |
|
21 |
This module currently exports the C<scope_guard> and C<guard> functions by |
22 |
default. |
23 |
|
24 |
=over 4 |
25 |
|
26 |
=cut |
27 |
|
28 |
package Guard; |
29 |
|
30 |
BEGIN { |
31 |
$VERSION = '0.01'; |
32 |
@ISA = qw(Exporter); |
33 |
@EXPORT = qw(guard scope_guard cancel); |
34 |
|
35 |
require Exporter; |
36 |
|
37 |
require XSLoader; |
38 |
XSLoader::load Guard, $VERSION; |
39 |
} |
40 |
|
41 |
our $DIED = sub { warn "$@" }; |
42 |
|
43 |
=item scope_guard BLOCK |
44 |
|
45 |
Registers a block that is executed when the current scope (block, |
46 |
function, method, eval etc.) is exited. |
47 |
|
48 |
The description below sounds a bit complicated, but that's just because |
49 |
C<scope_guard> tries to get even corner cases "right": the goal is to |
50 |
provide you with a rock solid clean up tool. |
51 |
|
52 |
This is similar to this code fragment: |
53 |
|
54 |
eval ... code following scope_guard ... |
55 |
{ |
56 |
local $@; |
57 |
eval BLOCK; |
58 |
eval { $Guard::DIED->() } if $@; |
59 |
} |
60 |
die if $@; |
61 |
|
62 |
Except it is much faster, and the whole thing gets executed even when the |
63 |
BLOCK calls C<exit>, C<goto>, C<last> or escapes via other means. |
64 |
|
65 |
See B<EXCEPTIONS>, below, for an explanation of exception handling |
66 |
(C<die>) within guard blocks. |
67 |
|
68 |
Example: Temporarily change the directory to F</etc> and make sure it's |
69 |
set back to F</> when the function returns: |
70 |
|
71 |
sub dosomething { |
72 |
scope_guard { chdir "/" }; |
73 |
chdir "/etc"; |
74 |
|
75 |
... |
76 |
} |
77 |
|
78 |
=item my $guard = guard BLOCK |
79 |
|
80 |
Behaves the same as C<scope_guard>, except that instead of executing |
81 |
the block on scope exit, it returns an object whose lifetime determines |
82 |
when the BLOCK gets executed: when the last reference to the object gets |
83 |
destroyed, the BLOCK gets executed as with C<scope_guard>. |
84 |
|
85 |
The returned object can be copied as many times as you want. |
86 |
|
87 |
See B<EXCEPTIONS>, below, for an explanation of exception handling |
88 |
(C<die>) within guard blocks. |
89 |
|
90 |
Example: acquire a Coro::Semaphore for a second by registering a |
91 |
timer. The timer callback references the guard used to unlock it again. |
92 |
|
93 |
use AnyEvent; |
94 |
use Coro::Semaphore; |
95 |
|
96 |
my $sem = new Coro::Semaphore; |
97 |
|
98 |
sub lock_1s { |
99 |
$sem->down; |
100 |
my $guard = guard { $sem->up }; |
101 |
|
102 |
my $timer; |
103 |
$timer = AnyEvent->timer (after => 1, sub { |
104 |
# do something |
105 |
undef $sem; |
106 |
undef $timer; |
107 |
}); |
108 |
} |
109 |
|
110 |
The 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, |
112 |
or your code can throw an exception before it can create the timer, or you |
113 |
can create multiple timers or other event watchers and only when the last |
114 |
one gets executed will the lock be unlocked. |
115 |
|
116 |
=item Guard::cancel $guard |
117 |
|
118 |
Calling this function will "disable" the guard object returned by the |
119 |
C<guard> function, i.e. it will free the BLOCK originally passed to |
120 |
C<guard >and will arrange for the BLOCK not to be executed. |
121 |
|
122 |
This can be useful when you use C<guard> to create a fatal cleanup handler |
123 |
and later decide it is no longer needed. |
124 |
|
125 |
=cut |
126 |
|
127 |
1; |
128 |
|
129 |
=back |
130 |
|
131 |
=head1 EXCEPTIONS |
132 |
|
133 |
Guard blocks should not normally throw exceptions (e.g. C<die>), after |
134 |
all, they are usually used to clean up after such exceptions. However, if |
135 |
something truly exceptional is happening, a guard block should be allowed |
136 |
to die. Also, programming errors are a large source of exceptions, and the |
137 |
programmer certainly wants to know about those. |
138 |
|
139 |
Since in most cases, the block executing when the guard gets executes does |
140 |
not know or does not care about the guard blocks, it makes little sense to |
141 |
let containing code handle the exception. |
142 |
|
143 |
Therefore, whenever a guard block throws an exception, it will be caught, |
144 |
and this module will call the code reference stored in C<$Guard::DIED> |
145 |
(with C<$@> set to the actual exception), which is similar to how most |
146 |
event loops handle this case. |
147 |
|
148 |
The code reference stored in C<$Guard::DIED> should not die (behaviour is |
149 |
not guaranteed, but right now, the exception will simply be ignored). |
150 |
|
151 |
The default for C<$Guard::DIED> is to call C<warn "$@">. |
152 |
|
153 |
=head1 AUTHOR |
154 |
|
155 |
Marc Lehmann <schmorp@schmorp.de> |
156 |
http://home.schmorp.de/ |
157 |
|
158 |
=head1 THANKS |
159 |
|
160 |
To Marco Maisenhelder, who reminded me of the C<$Guard::DIED> solution to |
161 |
the problem of exceptions. |
162 |
|
163 |
=cut |
164 |
|