1 |
root |
1.1 |
NAME |
2 |
root |
1.2 |
Guard - safe cleanup blocks |
3 |
root |
1.1 |
|
4 |
|
|
SYNOPSIS |
5 |
root |
1.3 |
use Guard; |
6 |
|
|
|
7 |
|
|
# temporarily chdir to "/etc" directory, but make sure |
8 |
|
|
# to go back to "/" no matter how myfun exits: |
9 |
|
|
sub myfun { |
10 |
|
|
scope_guard { chdir "/" }; |
11 |
|
|
chdir "/etc"; |
12 |
|
|
|
13 |
|
|
call_function_that_might_die_or_other_fun_stuff; |
14 |
|
|
} |
15 |
root |
1.1 |
|
16 |
|
|
DESCRIPTION |
17 |
root |
1.2 |
This module implements so-called "guards". A guard is something (usually |
18 |
|
|
an object) that "guards" a resource, ensuring that it is cleaned up when |
19 |
|
|
expected. |
20 |
|
|
|
21 |
|
|
Specifically, this module supports two different types of guards: guard |
22 |
|
|
objects, which execute a given code block when destroyed, and scoped |
23 |
|
|
guards, which are tied to the scope exit. |
24 |
|
|
|
25 |
|
|
FUNCTIONS |
26 |
|
|
This module currently exports the "scope_guard" and "guard" functions by |
27 |
|
|
default. |
28 |
|
|
|
29 |
|
|
scope_guard BLOCK |
30 |
|
|
Registers a block that is executed when the current scope (block, |
31 |
|
|
function, method, eval etc.) is exited. |
32 |
|
|
|
33 |
root |
1.3 |
See the EXCEPTIONS section for an explanation of how exceptions |
34 |
|
|
(i.e. "die") are handled inside guard blocks. |
35 |
|
|
|
36 |
root |
1.2 |
The description below sounds a bit complicated, but that's just |
37 |
|
|
because "scope_guard" tries to get even corner cases "right": the |
38 |
|
|
goal is to provide you with a rock solid clean up tool. |
39 |
|
|
|
40 |
root |
1.3 |
The behaviour is similar to this code fragment: |
41 |
root |
1.2 |
|
42 |
|
|
eval ... code following scope_guard ... |
43 |
|
|
{ |
44 |
|
|
local $@; |
45 |
|
|
eval BLOCK; |
46 |
|
|
eval { $Guard::DIED->() } if $@; |
47 |
|
|
} |
48 |
|
|
die if $@; |
49 |
|
|
|
50 |
|
|
Except it is much faster, and the whole thing gets executed even |
51 |
|
|
when the BLOCK calls "exit", "goto", "last" or escapes via other |
52 |
|
|
means. |
53 |
|
|
|
54 |
root |
1.3 |
If multiple BLOCKs are registered to the same scope, they will be |
55 |
|
|
executed in reverse order. Other scope-related things such as |
56 |
|
|
"local" are managed via the same mechanism, so variables "local"ised |
57 |
|
|
*after* calling "scope_guard" will be restored when the guard runs. |
58 |
|
|
|
59 |
|
|
Example: temporarily change the timezone for the current process, |
60 |
|
|
ensuring it will be reset when the "if" scope is exited: |
61 |
|
|
|
62 |
|
|
use Guard; |
63 |
|
|
use POSIX (); |
64 |
|
|
|
65 |
|
|
if ($need_to_switch_tz) { |
66 |
|
|
# make sure we call tzset after $ENV{TZ} has been restored |
67 |
|
|
scope_guard { POSIX::tzset }; |
68 |
|
|
|
69 |
|
|
# localise after the scope_guard, so it gets undone in time |
70 |
|
|
local $ENV{TZ} = "Europe/London"; |
71 |
|
|
POSIX::tzset; |
72 |
root |
1.2 |
|
73 |
root |
1.3 |
# do something with the new timezone |
74 |
root |
1.2 |
} |
75 |
|
|
|
76 |
|
|
my $guard = guard BLOCK |
77 |
|
|
Behaves the same as "scope_guard", except that instead of executing |
78 |
|
|
the block on scope exit, it returns an object whose lifetime |
79 |
|
|
determines when the BLOCK gets executed: when the last reference to |
80 |
|
|
the object gets destroyed, the BLOCK gets executed as with |
81 |
|
|
"scope_guard". |
82 |
|
|
|
83 |
|
|
The returned object can be copied as many times as you want. |
84 |
|
|
|
85 |
root |
1.3 |
See the EXCEPTIONS section for an explanation of how exceptions |
86 |
|
|
(i.e. "die") are handled inside guard blocks. |
87 |
root |
1.2 |
|
88 |
|
|
Example: acquire a Coro::Semaphore for a second by registering a |
89 |
|
|
timer. The timer callback references the guard used to unlock it |
90 |
root |
1.3 |
again. (Please ignore the fact that "Coro::Semaphore" has a "guard" |
91 |
|
|
method that does this already): |
92 |
root |
1.2 |
|
93 |
root |
1.3 |
use Guard; |
94 |
root |
1.2 |
use AnyEvent; |
95 |
|
|
use Coro::Semaphore; |
96 |
|
|
|
97 |
|
|
my $sem = new Coro::Semaphore; |
98 |
|
|
|
99 |
root |
1.3 |
sub lock_for_a_second { |
100 |
root |
1.2 |
$sem->down; |
101 |
|
|
my $guard = guard { $sem->up }; |
102 |
|
|
|
103 |
|
|
my $timer; |
104 |
|
|
$timer = AnyEvent->timer (after => 1, sub { |
105 |
|
|
# do something |
106 |
|
|
undef $sem; |
107 |
|
|
undef $timer; |
108 |
|
|
}); |
109 |
|
|
} |
110 |
|
|
|
111 |
|
|
The advantage of doing this with a guard instead of simply calling |
112 |
|
|
"$sem->down" in the callback is that you can opt not to create the |
113 |
|
|
timer, or your code can throw an exception before it can create the |
114 |
|
|
timer, or you can create multiple timers or other event watchers and |
115 |
|
|
only when the last one gets executed will the lock be unlocked. |
116 |
root |
1.3 |
Using the "guard", you do not have to worry about catching all the |
117 |
|
|
places where you have to unlock the semaphore. |
118 |
root |
1.2 |
|
119 |
root |
1.3 |
$guard->cancel |
120 |
root |
1.2 |
Calling this function will "disable" the guard object returned by |
121 |
|
|
the "guard" function, i.e. it will free the BLOCK originally passed |
122 |
|
|
to "guard "and will arrange for the BLOCK not to be executed. |
123 |
|
|
|
124 |
|
|
This can be useful when you use "guard" to create a fatal cleanup |
125 |
|
|
handler and later decide it is no longer needed. |
126 |
|
|
|
127 |
|
|
EXCEPTIONS |
128 |
root |
1.3 |
Guard blocks should not normally throw exceptions (that is, "die"). |
129 |
|
|
After all, they are usually used to clean up after such exceptions. |
130 |
|
|
However, if something truly exceptional is happening, a guard block |
131 |
|
|
should be allowed to die. Also, programming errors are a large source of |
132 |
root |
1.2 |
exceptions, and the programmer certainly wants to know about those. |
133 |
|
|
|
134 |
|
|
Since in most cases, the block executing when the guard gets executes |
135 |
|
|
does not know or does not care about the guard blocks, it makes little |
136 |
|
|
sense to let containing code handle the exception. |
137 |
|
|
|
138 |
|
|
Therefore, whenever a guard block throws an exception, it will be |
139 |
|
|
caught, and this module will call the code reference stored in |
140 |
|
|
$Guard::DIED (with $@ set to the actual exception), which is similar to |
141 |
|
|
how most event loops handle this case. |
142 |
|
|
|
143 |
root |
1.3 |
The default for $Guard::DIED is to call "warn "$@"". |
144 |
|
|
|
145 |
|
|
The $@ variable will be restored to its value before the guard call in |
146 |
|
|
all cases, so guards will not disturb $@ in any way. |
147 |
|
|
|
148 |
root |
1.2 |
The code reference stored in $Guard::DIED should not die (behaviour is |
149 |
|
|
not guaranteed, but right now, the exception will simply be ignored). |
150 |
|
|
|
151 |
root |
1.1 |
AUTHOR |
152 |
|
|
Marc Lehmann <schmorp@schmorp.de> |
153 |
|
|
http://home.schmorp.de/ |
154 |
|
|
|
155 |
root |
1.2 |
THANKS |
156 |
root |
1.3 |
Thanks to Marco Maisenhelder, who reminded me of the $Guard::DIED |
157 |
|
|
solution to the problem of exceptions. |
158 |
root |
1.2 |
|