1 |
NAME |
2 |
Guard - safe cleanup blocks |
3 |
|
4 |
SYNOPSIS |
5 |
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 |
code_that_might_die_or_does_other_fun_stuff; |
14 |
} |
15 |
|
16 |
# create an object that, when the last reference to it is gone, |
17 |
# invokes the given codeblock: |
18 |
my $guard = guard { print "destroyed!\n" }; |
19 |
undef $guard; # probably destroyed here |
20 |
|
21 |
DESCRIPTION |
22 |
This module implements so-called "guards". A guard is something (usually |
23 |
an object) that "guards" a resource, ensuring that it is cleaned up when |
24 |
expected. |
25 |
|
26 |
Specifically, this module supports two different types of guards: guard |
27 |
objects, which execute a given code block when destroyed, and scoped |
28 |
guards, which are tied to the scope exit. |
29 |
|
30 |
FUNCTIONS |
31 |
This module currently exports the "scope_guard" and "guard" functions by |
32 |
default. |
33 |
|
34 |
scope_guard BLOCK |
35 |
scope_guard ($coderef) |
36 |
Registers a block that is executed when the current scope (block, |
37 |
function, method, eval etc.) is exited. |
38 |
|
39 |
See the EXCEPTIONS section for an explanation of how exceptions |
40 |
(i.e. "die") are handled inside guard blocks. |
41 |
|
42 |
The description below sounds a bit complicated, but that's just |
43 |
because "scope_guard" tries to get even corner cases "right": the |
44 |
goal is to provide you with a rock solid clean up tool. |
45 |
|
46 |
The behaviour is similar to this code fragment: |
47 |
|
48 |
eval ... code following scope_guard ... |
49 |
{ |
50 |
local $@; |
51 |
eval BLOCK; |
52 |
eval { $Guard::DIED->() } if $@; |
53 |
} |
54 |
die if $@; |
55 |
|
56 |
Except it is much faster, and the whole thing gets executed even |
57 |
when the BLOCK calls "exit", "goto", "last" or escapes via other |
58 |
means. |
59 |
|
60 |
If multiple BLOCKs are registered to the same scope, they will be |
61 |
executed in reverse order. Other scope-related things such as |
62 |
"local" are managed via the same mechanism, so variables "local"ised |
63 |
*after* calling "scope_guard" will be restored when the guard runs. |
64 |
|
65 |
Example: temporarily change the timezone for the current process, |
66 |
ensuring it will be reset when the "if" scope is exited: |
67 |
|
68 |
use Guard; |
69 |
use POSIX (); |
70 |
|
71 |
if ($need_to_switch_tz) { |
72 |
# make sure we call tzset after $ENV{TZ} has been restored |
73 |
scope_guard { POSIX::tzset }; |
74 |
|
75 |
# localise after the scope_guard, so it gets undone in time |
76 |
local $ENV{TZ} = "Europe/London"; |
77 |
POSIX::tzset; |
78 |
|
79 |
# do something with the new timezone |
80 |
} |
81 |
|
82 |
my $guard = guard BLOCK |
83 |
my $guard = guard ($coderef) |
84 |
Behaves the same as "scope_guard", except that instead of executing |
85 |
the block on scope exit, it returns an object whose lifetime |
86 |
determines when the BLOCK gets executed: when the last reference to |
87 |
the object gets destroyed, the BLOCK gets executed as with |
88 |
"scope_guard". |
89 |
|
90 |
See the EXCEPTIONS section for an explanation of how exceptions |
91 |
(i.e. "die") are handled inside guard blocks. |
92 |
|
93 |
Example: acquire a Coro::Semaphore for a second by registering a |
94 |
timer. The timer callback references the guard used to unlock it |
95 |
again. (Please ignore the fact that "Coro::Semaphore" has a "guard" |
96 |
method that does this already): |
97 |
|
98 |
use Guard; |
99 |
use Coro::AnyEvent; |
100 |
use Coro::Semaphore; |
101 |
|
102 |
my $sem = new Coro::Semaphore; |
103 |
|
104 |
sub lock_for_a_second { |
105 |
$sem->down; |
106 |
my $guard = guard { $sem->up }; |
107 |
|
108 |
Coro::AnyEvent::sleep 1; |
109 |
|
110 |
# $sem->up gets executed when returning |
111 |
} |
112 |
|
113 |
The advantage of doing this with a guard instead of simply calling |
114 |
"$sem->down" in the callback is that you can opt not to create the |
115 |
timer, or your code can throw an exception before it can create the |
116 |
timer (or the thread gets canceled), or you can create multiple |
117 |
timers or other event watchers and only when the last one gets |
118 |
executed will the lock be unlocked. Using the "guard", you do not |
119 |
have to worry about catching all the places where you have to unlock |
120 |
the semaphore. |
121 |
|
122 |
$guard->cancel |
123 |
Calling this function will "disable" the guard object returned by |
124 |
the "guard" function, i.e. it will free the BLOCK originally passed |
125 |
to "guard "and will arrange for the BLOCK not to be executed. |
126 |
|
127 |
This can be useful when you use "guard" to create a cleanup handler |
128 |
to be called under fatal conditions and later decide it is no longer |
129 |
needed. |
130 |
|
131 |
EXCEPTIONS |
132 |
Guard blocks should not normally throw exceptions (that is, "die"). |
133 |
After all, they are usually used to clean up after such exceptions. |
134 |
However, if something truly exceptional is happening, a guard block |
135 |
should of course be allowed to die. Also, programming errors are a large |
136 |
source of exceptions, and the programmer certainly wants to know about |
137 |
those. |
138 |
|
139 |
Since in most cases, the block executing when the guard gets executed |
140 |
does not know or does not care about the guard blocks, it makes little |
141 |
sense to let containing code handle the exception. |
142 |
|
143 |
Therefore, whenever a guard block throws an exception, it will be caught |
144 |
by Guard, followed by calling the code reference stored in $Guard::DIED |
145 |
(with $@ set to the actual exception), which is similar to how most |
146 |
event loops handle this case. |
147 |
|
148 |
The default for $Guard::DIED is to call "warn "$@"", i.e. the error is |
149 |
printed as a warning and the program continues. |
150 |
|
151 |
The $@ variable will be restored to its value before the guard call in |
152 |
all cases, so guards will not disturb $@ in any way. |
153 |
|
154 |
The code reference stored in $Guard::DIED should not die (behaviour is |
155 |
not guaranteed, but right now, the exception will simply be ignored). |
156 |
|
157 |
AUTHOR |
158 |
Marc Lehmann <schmorp@schmorp.de> |
159 |
http://home.schmorp.de/ |
160 |
|
161 |
THANKS |
162 |
Thanks to Marco Maisenhelder, who reminded me of the $Guard::DIED |
163 |
solution to the problem of exceptions. |
164 |
|
165 |
SEE ALSO |
166 |
Scope::Guard and Sub::ScopeFinalizer, which actually implement |
167 |
dynamically scoped guards only, not the lexically scoped guards that |
168 |
their documentation promises, and have a lot higher CPU, memory and |
169 |
typing overhead. |
170 |
|
171 |
Hook::Scope, which has apparently never been finished and can corrupt |
172 |
memory when used. |
173 |
|
174 |
Scope::Guard seems to have a big SEE ALSO section for even more modules |
175 |
like it. |
176 |
|