1 |
=head1 NAME |
2 |
|
3 |
Guard - safe cleanup blocks |
4 |
|
5 |
=head1 SYNOPSIS |
6 |
|
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 |
22 |
|
23 |
=head1 DESCRIPTION |
24 |
|
25 |
This module implements so-called "guards". A guard is something (usually |
26 |
an object) that "guards" a resource, ensuring that it is cleaned up when |
27 |
expected. |
28 |
|
29 |
Specifically, this module supports two different types of guards: guard |
30 |
objects, which execute a given code block when destroyed, and scoped |
31 |
guards, which are tied to the scope exit. |
32 |
|
33 |
=head1 FUNCTIONS |
34 |
|
35 |
This module currently exports the C<scope_guard> and C<guard> functions by |
36 |
default. |
37 |
|
38 |
=over 4 |
39 |
|
40 |
=cut |
41 |
|
42 |
package Guard; |
43 |
|
44 |
no warnings; |
45 |
|
46 |
BEGIN { |
47 |
$VERSION = 1.023; |
48 |
@ISA = qw(Exporter); |
49 |
@EXPORT = qw(guard scope_guard); |
50 |
|
51 |
require Exporter; |
52 |
|
53 |
require XSLoader; |
54 |
XSLoader::load Guard, $VERSION; |
55 |
} |
56 |
|
57 |
our $DIED = sub { warn "$@" }; |
58 |
|
59 |
=item scope_guard BLOCK |
60 |
|
61 |
=item scope_guard ($coderef) |
62 |
|
63 |
Registers a block that is executed when the current scope (block, |
64 |
function, method, eval etc.) is exited. |
65 |
|
66 |
See the EXCEPTIONS section for an explanation of how exceptions |
67 |
(i.e. C<die>) are handled inside guard blocks. |
68 |
|
69 |
The description below sounds a bit complicated, but that's just because |
70 |
C<scope_guard> tries to get even corner cases "right": the goal is to |
71 |
provide you with a rock solid clean up tool. |
72 |
|
73 |
The behaviour is similar to this code fragment: |
74 |
|
75 |
eval ... code following scope_guard ... |
76 |
{ |
77 |
local $@; |
78 |
eval BLOCK; |
79 |
eval { $Guard::DIED->() } if $@; |
80 |
} |
81 |
die if $@; |
82 |
|
83 |
Except it is much faster, and the whole thing gets executed even when the |
84 |
BLOCK calls C<exit>, C<goto>, C<last> or escapes via other means. |
85 |
|
86 |
If multiple BLOCKs are registered to the same scope, they will be executed |
87 |
in reverse order. Other scope-related things such as C<local> are managed |
88 |
via the same mechanism, so variables C<local>ised I<after> calling |
89 |
C<scope_guard> will be restored I<before> the guard runs. |
90 |
|
91 |
Example: temporarily change the timezone for the current process, |
92 |
ensuring it will be reset when the C<if> scope is exited: |
93 |
|
94 |
use Guard; |
95 |
use POSIX (); |
96 |
|
97 |
if ($need_to_switch_tz) { |
98 |
# make sure we call tzset after $ENV{TZ} has been restored |
99 |
scope_guard { POSIX::tzset }; |
100 |
|
101 |
# localise after the scope_guard, so it gets undone in time |
102 |
local $ENV{TZ} = "Europe/London"; |
103 |
POSIX::tzset; |
104 |
|
105 |
# do something with the new timezone |
106 |
} |
107 |
|
108 |
=item my $guard = guard BLOCK |
109 |
|
110 |
=item my $guard = guard ($coderef) |
111 |
|
112 |
Behaves the same as C<scope_guard>, except that instead of executing |
113 |
the block on scope exit, it returns an object whose lifetime determines |
114 |
when the BLOCK gets executed: when the last reference to the object gets |
115 |
destroyed, the BLOCK gets executed as with C<scope_guard>. |
116 |
|
117 |
See the EXCEPTIONS section for an explanation of how exceptions |
118 |
(i.e. C<die>) are handled inside guard blocks. |
119 |
|
120 |
Example: acquire a Coro::Semaphore for a second by registering a |
121 |
timer. The timer callback references the guard used to unlock it |
122 |
again. (Please ignore the fact that C<Coro::Semaphore> has a C<guard> |
123 |
method that does this already): |
124 |
|
125 |
use Guard; |
126 |
use Coro::AnyEvent; |
127 |
use Coro::Semaphore; |
128 |
|
129 |
my $sem = new Coro::Semaphore; |
130 |
|
131 |
sub lock_for_a_second { |
132 |
$sem->down; |
133 |
my $guard = guard { $sem->up }; |
134 |
|
135 |
Coro::AnyEvent::sleep 1; |
136 |
|
137 |
# $sem->up gets executed when returning |
138 |
} |
139 |
|
140 |
The advantage of doing this with a guard instead of simply calling C<< |
141 |
$sem->down >> in the callback is that you can opt not to create the timer, |
142 |
or your code can throw an exception before it can create the timer (or |
143 |
the thread gets canceled), or you can create multiple timers or other |
144 |
event watchers and only when the last one gets executed will the lock be |
145 |
unlocked. Using the C<guard>, you do not have to worry about catching all |
146 |
the places where you have to unlock the semaphore. |
147 |
|
148 |
=item $guard->cancel |
149 |
|
150 |
Calling this function will "disable" the guard object returned by the |
151 |
C<guard> function, i.e. it will free the BLOCK originally passed to |
152 |
C<guard >and will arrange for the BLOCK not to be executed. |
153 |
|
154 |
This can be useful when you use C<guard> to create a cleanup handler to be |
155 |
called under fatal conditions and later decide it is no longer needed. |
156 |
|
157 |
=cut |
158 |
|
159 |
1; |
160 |
|
161 |
=back |
162 |
|
163 |
=head1 EXCEPTIONS |
164 |
|
165 |
Guard blocks should not normally throw exceptions (that is, C<die>). After |
166 |
all, they are usually used to clean up after such exceptions. However, |
167 |
if something truly exceptional is happening, a guard block should of |
168 |
course be allowed to die. Also, programming errors are a large source of |
169 |
exceptions, and the programmer certainly wants to know about those. |
170 |
|
171 |
Since in most cases, the block executing when the guard gets executed does |
172 |
not know or does not care about the guard blocks, it makes little sense to |
173 |
let containing code handle the exception. |
174 |
|
175 |
Therefore, whenever a guard block throws an exception, it will be caught |
176 |
by Guard, followed by calling the code reference stored in C<$Guard::DIED> |
177 |
(with C<$@> set to the actual exception), which is similar to how most |
178 |
event loops handle this case. |
179 |
|
180 |
The default for C<$Guard::DIED> is to call C<warn "$@">, i.e. the error is |
181 |
printed as a warning and the program continues. |
182 |
|
183 |
The C<$@> variable will be restored to its value before the guard call in |
184 |
all cases, so guards will not disturb C<$@> in any way. |
185 |
|
186 |
The code reference stored in C<$Guard::DIED> should not die (behaviour is |
187 |
not guaranteed, but right now, the exception will simply be ignored). |
188 |
|
189 |
=head1 AUTHOR |
190 |
|
191 |
Marc Lehmann <schmorp@schmorp.de> |
192 |
http://home.schmorp.de/ |
193 |
|
194 |
=head1 THANKS |
195 |
|
196 |
Thanks to Marco Maisenhelder, who reminded me of the C<$Guard::DIED> |
197 |
solution to the problem of exceptions. |
198 |
|
199 |
=head1 SEE ALSO |
200 |
|
201 |
L<Scope::Guard> and L<Sub::ScopeFinalizer>, which actually implement |
202 |
dynamically scoped guards only, not the lexically scoped guards that their |
203 |
documentation promises, and have a lot higher CPU, memory and typing |
204 |
overhead. |
205 |
|
206 |
L<Hook::Scope>, which has apparently never been finished and can corrupt |
207 |
memory when used. |
208 |
|
209 |
L<Scope::Guard> seems to have a big SEE ALSO section for even more |
210 |
modules like it. |
211 |
|
212 |
=cut |
213 |
|