1 |
=head1 NAME |
2 |
|
3 |
EV - perl interface to libev, a high performance full-featured event loop |
4 |
|
5 |
=head1 SYNOPSIS |
6 |
|
7 |
use EV; |
8 |
|
9 |
# TIMERS |
10 |
|
11 |
my $w = EV::timer 2, 0, sub { |
12 |
warn "is called after 2s"; |
13 |
}; |
14 |
|
15 |
my $w = EV::timer 2, 2, sub { |
16 |
warn "is called roughly every 2s (repeat = 2)"; |
17 |
}; |
18 |
|
19 |
undef $w; # destroy event watcher again |
20 |
|
21 |
my $w = EV::periodic 0, 60, 0, sub { |
22 |
warn "is called every minute, on the minute, exactly"; |
23 |
}; |
24 |
|
25 |
# IO |
26 |
|
27 |
my $w = EV::io *STDIN, EV::READ, sub { |
28 |
my ($w, $revents) = @_; # all callbacks receive the watcher and event mask |
29 |
warn "stdin is readable, you entered: ", <STDIN>; |
30 |
}; |
31 |
|
32 |
# SIGNALS |
33 |
|
34 |
my $w = EV::signal 'QUIT', sub { |
35 |
warn "sigquit received\n"; |
36 |
}; |
37 |
|
38 |
# CHILD/PID STATUS CHANGES |
39 |
|
40 |
my $w = EV::child 666, sub { |
41 |
my ($w, $revents) = @_; |
42 |
my $status = $w->rstatus; |
43 |
}; |
44 |
|
45 |
# MAINLOOP |
46 |
EV::loop; # loop until EV::unloop is called or all watchers stop |
47 |
EV::loop EV::LOOP_ONESHOT; # block until at least one event could be handled |
48 |
EV::loop EV::LOOP_NONBLOCK; # try to handle same events, but do not block |
49 |
|
50 |
=head1 DESCRIPTION |
51 |
|
52 |
This module provides an interface to libev |
53 |
(L<http://software.schmorp.de/pkg/libev.html>). While the documentation |
54 |
below is comprehensive, one might also consult the documentation of libev |
55 |
itself (L<http://cvs.schmorp.de/libev/ev.html>) for more subtle details on |
56 |
watcher semantics or some discussion on the available backends, or how to |
57 |
force a specific backend with C<LIBEV_FLAGS>. |
58 |
|
59 |
=cut |
60 |
|
61 |
package EV; |
62 |
|
63 |
use strict; |
64 |
|
65 |
BEGIN { |
66 |
our $VERSION = '1.3'; |
67 |
use XSLoader; |
68 |
XSLoader::load "EV", $VERSION; |
69 |
} |
70 |
|
71 |
@EV::IO::ISA = |
72 |
@EV::Timer::ISA = |
73 |
@EV::Periodic::ISA = |
74 |
@EV::Signal::ISA = |
75 |
@EV::Idle::ISA = |
76 |
@EV::Prepare::ISA = |
77 |
@EV::Check::ISA = |
78 |
@EV::Child::ISA = "EV::Watcher"; |
79 |
|
80 |
=head1 BASIC INTERFACE |
81 |
|
82 |
=over 4 |
83 |
|
84 |
=item $EV::DIED |
85 |
|
86 |
Must contain a reference to a function that is called when a callback |
87 |
throws an exception (with $@ containing thr error). The default prints an |
88 |
informative message and continues. |
89 |
|
90 |
If this callback throws an exception it will be silently ignored. |
91 |
|
92 |
=item $time = EV::time |
93 |
|
94 |
Returns the current time in (fractional) seconds since the epoch. |
95 |
|
96 |
=item $time = EV::now |
97 |
|
98 |
Returns the time the last event loop iteration has been started. This |
99 |
is the time that (relative) timers are based on, and refering to it is |
100 |
usually faster then calling EV::time. |
101 |
|
102 |
=item $method = EV::method |
103 |
|
104 |
Returns an integer describing the backend used by libev (EV::METHOD_SELECT |
105 |
or EV::METHOD_EPOLL). |
106 |
|
107 |
=item EV::loop [$flags] |
108 |
|
109 |
Begin checking for events and calling callbacks. It returns when a |
110 |
callback calls EV::unloop. |
111 |
|
112 |
The $flags argument can be one of the following: |
113 |
|
114 |
0 as above |
115 |
EV::LOOP_ONESHOT block at most once (wait, but do not loop) |
116 |
EV::LOOP_NONBLOCK do not block at all (fetch/handle events but do not wait) |
117 |
|
118 |
=item EV::unloop [$how] |
119 |
|
120 |
When called with no arguments or an argument of EV::UNLOOP_ONE, makes the |
121 |
innermost call to EV::loop return. |
122 |
|
123 |
When called with an argument of EV::UNLOOP_ALL, all calls to EV::loop will return as |
124 |
fast as possible. |
125 |
|
126 |
=item EV::once $fh_or_undef, $events, $timeout, $cb->($revents) |
127 |
|
128 |
This function rolls together an I/O and a timer watcher for a single |
129 |
one-shot event without the need for managing a watcher object. |
130 |
|
131 |
If C<$fh_or_undef> is a filehandle or file descriptor, then C<$events> |
132 |
must be a bitset containing either C<EV::READ>, C<EV::WRITE> or C<EV::READ |
133 |
| EV::WRITE>, indicating the type of I/O event you want to wait for. If |
134 |
you do not want to wait for some I/O event, specify C<undef> for |
135 |
C<$fh_or_undef> and C<0> for C<$events>). |
136 |
|
137 |
If timeout is C<undef> or negative, then there will be no |
138 |
timeout. Otherwise a EV::timer with this value will be started. |
139 |
|
140 |
When an error occurs or either the timeout or I/O watcher triggers, then |
141 |
the callback will be called with the received event set (in general |
142 |
you can expect it to be a combination of C<EV:ERROR>, C<EV::READ>, |
143 |
C<EV::WRITE> and C<EV::TIMEOUT>). |
144 |
|
145 |
EV::once doesn't return anything: the watchers stay active till either |
146 |
of them triggers, then they will be stopped and freed, and the callback |
147 |
invoked. |
148 |
|
149 |
=back |
150 |
|
151 |
=head2 WATCHER |
152 |
|
153 |
A watcher is an object that gets created to record your interest in some |
154 |
event. For instance, if you want to wait for STDIN to become readable, you |
155 |
would create an EV::io watcher for that: |
156 |
|
157 |
my $watcher = EV::io *STDIN, EV::READ, sub { |
158 |
my ($watcher, $revents) = @_; |
159 |
warn "yeah, STDIN should not be readable without blocking!\n" |
160 |
}; |
161 |
|
162 |
All watchers can be active (waiting for events) or inactive (paused). Only |
163 |
active watchers will have their callbacks invoked. All callbacks will be |
164 |
called with at least two arguments: the watcher and a bitmask of received |
165 |
events. |
166 |
|
167 |
Each watcher type has its associated bit in revents, so you can use the |
168 |
same callback for multiple watchers. The event mask is named after the |
169 |
type, i..e. EV::child sets EV::CHILD, EV::prepare sets EV::PREPARE, |
170 |
EV::periodic sets EV::PERIODIC and so on, with the exception of IO events |
171 |
(which can set both EV::READ and EV::WRITE bits), and EV::timer (which |
172 |
uses EV::TIMEOUT). |
173 |
|
174 |
In the rare case where one wants to create a watcher but not start it at |
175 |
the same time, each constructor has a variant with a trailing C<_ns> in |
176 |
its name, e.g. EV::io has a non-starting variant EV::io_ns and so on. |
177 |
|
178 |
Please note that a watcher will automatically be stopped when the watcher |
179 |
object is destroyed, so you I<need> to keep the watcher objects returned by |
180 |
the constructors. |
181 |
|
182 |
Also, all methods changing some aspect of a watcher (->set, ->priority, |
183 |
->fh and so on) automatically stop and start it again if it is active, |
184 |
which means pending events get lost. |
185 |
|
186 |
=head2 WATCHER TYPES |
187 |
|
188 |
Now lets move to the existing watcher types and asociated methods. |
189 |
|
190 |
The following methods are available for all watchers. Then followes a |
191 |
description of each watcher constructor (EV::io, EV::timer, EV::periodic, |
192 |
EV::signal, EV::child, EV::idle, EV::prepare and EV::check), followed by |
193 |
any type-specific methods (if any). |
194 |
|
195 |
=over 4 |
196 |
|
197 |
=item $w->start |
198 |
|
199 |
Starts a watcher if it isn't active already. Does nothing to an already |
200 |
active watcher. By default, all watchers start out in the active state |
201 |
(see the description of the C<_ns> variants if you need stopped watchers). |
202 |
|
203 |
=item $w->stop |
204 |
|
205 |
Stop a watcher if it is active. Also clear any pending events (events that |
206 |
have been received but that didn't yet result in a callback invocation), |
207 |
regardless of wether the watcher was active or not. |
208 |
|
209 |
=item $bool = $w->is_active |
210 |
|
211 |
Returns true if the watcher is active, false otherwise. |
212 |
|
213 |
=item $current_data = $w->data |
214 |
|
215 |
=item $old_data = $w->data ($new_data) |
216 |
|
217 |
Queries a freely usable data scalar on the watcher and optionally changes |
218 |
it. This is a way to associate custom data with a watcher: |
219 |
|
220 |
my $w = EV::timer 60, 0, sub { |
221 |
warn $_[0]->data; |
222 |
}; |
223 |
$w->data ("print me!"); |
224 |
|
225 |
=item $current_cb = $w->cb |
226 |
|
227 |
=item $old_cb = $w->cb ($new_cb) |
228 |
|
229 |
Queries the callback on the watcher and optionally changes it. You can do |
230 |
this at any time without the watcher restarting. |
231 |
|
232 |
=item $current_priority = $w->priority |
233 |
|
234 |
=item $old_priority = $w->priority ($new_priority) |
235 |
|
236 |
Queries the priority on the watcher and optionally changes it. Pending |
237 |
watchers with higher priority will be invoked first. The valid range of |
238 |
priorities lies between EV::MAXPRI (default 2) and EV::MINPRI (default |
239 |
-2). If the priority is outside this range it will automatically be |
240 |
normalised to the nearest valid priority. |
241 |
|
242 |
The default priority of any newly-created watcher is 0. |
243 |
|
244 |
Note that the priority semantics have not yet been fleshed out and are |
245 |
subject to almost certain change. |
246 |
|
247 |
=item $w->trigger ($revents) |
248 |
|
249 |
Call the callback *now* with the given event mask. |
250 |
|
251 |
=item $previous_state = $w->keepalive ($bool) |
252 |
|
253 |
Normally, C<EV::loop> will return when there are no active watchers |
254 |
(which is a "deadlock" because no progress can be made anymore). This is |
255 |
convinient because it allows you to start your watchers (and your jobs), |
256 |
call C<EV::loop> once and when it returns you know that all your jobs are |
257 |
finished (or they forgot to register some watchers for their task :). |
258 |
|
259 |
Sometimes, however, this gets in your way, for example when you the module |
260 |
that calls C<EV::loop> (usually the main program) is not the same module |
261 |
as a long-living watcher (for example a DNS client module written by |
262 |
somebody else even). Then you might want any outstanding requests to be |
263 |
handled, but you would not want to keep C<EV::loop> from returning just |
264 |
because you happen to have this long-running UDP port watcher. |
265 |
|
266 |
In this case you can clear the keepalive status, which means that even |
267 |
though your watcher is active, it won't keep C<EV::loop> from returning. |
268 |
|
269 |
The initial value for keepalive is true (enabled), and you cna change it |
270 |
any time. |
271 |
|
272 |
Example: Register an IO watcher for some UDP socket but do not keep the |
273 |
event loop from running just because of that watcher. |
274 |
|
275 |
my $udp_socket = ... |
276 |
my $udp_watcher = EV::io $udp_socket, EV::READ, sub { ... }; |
277 |
$udp_watcher->keepalive (0); |
278 |
|
279 |
=item $w = EV::io $fileno_or_fh, $eventmask, $callback |
280 |
|
281 |
=item $w = EV::io_ns $fileno_or_fh, $eventmask, $callback |
282 |
|
283 |
As long as the returned watcher object is alive, call the C<$callback> |
284 |
when the events specified in C<$eventmask>. |
285 |
|
286 |
The $eventmask can be one or more of these constants ORed together: |
287 |
|
288 |
EV::READ wait until read() wouldn't block anymore |
289 |
EV::WRITE wait until write() wouldn't block anymore |
290 |
|
291 |
The C<io_ns> variant doesn't start (activate) the newly created watcher. |
292 |
|
293 |
=item $w->set ($fileno_or_fh, $eventmask) |
294 |
|
295 |
Reconfigures the watcher, see the constructor above for details. Can be |
296 |
called at any time. |
297 |
|
298 |
=item $current_fh = $w->fh |
299 |
|
300 |
=item $old_fh = $w->fh ($new_fh) |
301 |
|
302 |
Returns the previously set filehandle and optionally set a new one. |
303 |
|
304 |
=item $current_eventmask = $w->events |
305 |
|
306 |
=item $old_eventmask = $w->events ($new_eventmask) |
307 |
|
308 |
Returns the previously set event mask and optionally set a new one. |
309 |
|
310 |
|
311 |
=item $w = EV::timer $after, $repeat, $callback |
312 |
|
313 |
=item $w = EV::timer_ns $after, $repeat, $callback |
314 |
|
315 |
Calls the callback after C<$after> seconds. If C<$repeat> is non-zero, |
316 |
the timer will be restarted (with the $repeat value as $after) after the |
317 |
callback returns. |
318 |
|
319 |
This means that the callback would be called roughly after C<$after> |
320 |
seconds, and then every C<$repeat> seconds. The timer does his best not |
321 |
to drift, but it will not invoke the timer more often then once per event |
322 |
loop iteration, and might drift in other cases. If that isn't acceptable, |
323 |
look at EV::periodic, which can provide long-term stable timers. |
324 |
|
325 |
The timer is based on a monotonic clock, that is, if somebody is sitting |
326 |
in front of the machine while the timer is running and changes the system |
327 |
clock, the timer will nevertheless run (roughly) the same time. |
328 |
|
329 |
The C<timer_ns> variant doesn't start (activate) the newly created watcher. |
330 |
|
331 |
=item $w->set ($after, $repeat) |
332 |
|
333 |
Reconfigures the watcher, see the constructor above for details. Can be at |
334 |
any time. |
335 |
|
336 |
=item $w->again |
337 |
|
338 |
Similar to the C<start> method, but has special semantics for repeating timers: |
339 |
|
340 |
If the timer is active and non-repeating, it will be stopped. |
341 |
|
342 |
If the timer is active and repeating, reset the timeout to occur |
343 |
C<$repeat> seconds after now. |
344 |
|
345 |
If the timer is inactive and repeating, start it using the repeat value. |
346 |
|
347 |
Otherwise do nothing. |
348 |
|
349 |
This behaviour is useful when you have a timeout for some IO |
350 |
operation. You create a timer object with the same value for C<$after> and |
351 |
C<$repeat>, and then, in the read/write watcher, run the C<again> method |
352 |
on the timeout. |
353 |
|
354 |
|
355 |
=item $w = EV::periodic $at, $interval, $reschedule_cb, $callback |
356 |
|
357 |
=item $w = EV::periodic_ns $at, $interval, $reschedule_cb, $callback |
358 |
|
359 |
Similar to EV::timer, but is not based on relative timeouts but on |
360 |
absolute times. Apart from creating "simple" timers that trigger "at" the |
361 |
specified time, it can also be used for non-drifting absolute timers and |
362 |
more complex, cron-like, setups that are not adversely affected by time |
363 |
jumps (i.e. when the system clock is changed by explicit date -s or other |
364 |
means such as ntpd). It is also the most complex watcher type in EV. |
365 |
|
366 |
It has three distinct "modes": |
367 |
|
368 |
=over 4 |
369 |
|
370 |
=item * absolute timer ($interval = $reschedule_cb = 0) |
371 |
|
372 |
This time simply fires at the wallclock time C<$at> and doesn't repeat. It |
373 |
will not adjust when a time jump occurs, that is, if it is to be run |
374 |
at January 1st 2011 then it will run when the system time reaches or |
375 |
surpasses this time. |
376 |
|
377 |
=item * non-repeating interval timer ($interval > 0, $reschedule_cb = 0) |
378 |
|
379 |
In this mode the watcher will always be scheduled to time out at the |
380 |
next C<$at + N * $interval> time (for some integer N) and then repeat, |
381 |
regardless of any time jumps. |
382 |
|
383 |
This can be used to create timers that do not drift with respect to system |
384 |
time: |
385 |
|
386 |
my $hourly = EV::periodic 0, 3600, 0, sub { print "once/hour\n" }; |
387 |
|
388 |
That doesn't mean there will always be 3600 seconds in between triggers, |
389 |
but only that the the clalback will be called when the system time shows a |
390 |
full hour (UTC). |
391 |
|
392 |
Another way to think about it (for the mathematically inclined) is that |
393 |
EV::periodic will try to run the callback in this mode at the next |
394 |
possible time where C<$time = $at (mod $interval)>, regardless of any time |
395 |
jumps. |
396 |
|
397 |
=item * manual reschedule mode ($reschedule_cb = coderef) |
398 |
|
399 |
In this mode $interval and $at are both being ignored. Instead, each |
400 |
time the periodic watcher gets scheduled, the reschedule callback |
401 |
($reschedule_cb) will be called with the watcher as first, and the current |
402 |
time as second argument. |
403 |
|
404 |
I<This callback MUST NOT stop or destroy this or any other periodic |
405 |
watcher, ever>. If you need to stop it, return 1e30 and stop it |
406 |
afterwards. |
407 |
|
408 |
It must return the next time to trigger, based on the passed time value |
409 |
(that is, the lowest time value larger than to the second argument). It |
410 |
will usually be called just before the callback will be triggered, but |
411 |
might be called at other times, too. |
412 |
|
413 |
This can be used to create very complex timers, such as a timer that |
414 |
triggers on each midnight, local time (actually 24 hours after the last |
415 |
midnight, to keep the example simple. If you know a way to do it correctly |
416 |
in about the same space (without requiring elaborate modules), drop me a |
417 |
note :): |
418 |
|
419 |
my $daily = EV::periodic 0, 0, sub { |
420 |
my ($w, $now) = @_; |
421 |
|
422 |
use Time::Local (); |
423 |
my (undef, undef, undef, $d, $m, $y) = localtime $now; |
424 |
86400 + Time::Local::timelocal 0, 0, 0, $d, $m, $y |
425 |
}, sub { |
426 |
print "it's midnight or likely shortly after, now\n"; |
427 |
}; |
428 |
|
429 |
=back |
430 |
|
431 |
The C<periodic_ns> variant doesn't start (activate) the newly created watcher. |
432 |
|
433 |
=item $w->set ($at, $interval, $reschedule_cb) |
434 |
|
435 |
Reconfigures the watcher, see the constructor above for details. Can be at |
436 |
any time. |
437 |
|
438 |
=item $w->again |
439 |
|
440 |
Simply stops and starts the watcher again. |
441 |
|
442 |
|
443 |
=item $w = EV::signal $signal, $callback |
444 |
|
445 |
=item $w = EV::signal_ns $signal, $callback |
446 |
|
447 |
Call the callback when $signal is received (the signal can be specified |
448 |
by number or by name, just as with kill or %SIG). |
449 |
|
450 |
EV will grab the signal for the process (the kernel only allows one |
451 |
component to receive a signal at a time) when you start a signal watcher, |
452 |
and removes it again when you stop it. Perl does the same when you |
453 |
add/remove callbacks to %SIG, so watch out. |
454 |
|
455 |
You can have as many signal watchers per signal as you want. |
456 |
|
457 |
The C<signal_ns> variant doesn't start (activate) the newly created watcher. |
458 |
|
459 |
=item $w->set ($signal) |
460 |
|
461 |
Reconfigures the watcher, see the constructor above for details. Can be at |
462 |
any time. |
463 |
|
464 |
=item $current_signum = $w->signal |
465 |
|
466 |
=item $old_signum = $w->signal ($new_signal) |
467 |
|
468 |
Returns the previously set signal (always as a number not name) and |
469 |
optionally set a new one. |
470 |
|
471 |
|
472 |
=item $w = EV::child $pid, $callback |
473 |
|
474 |
=item $w = EV::child_ns $pid, $callback |
475 |
|
476 |
Call the callback when a status change for pid C<$pid> (or any pid |
477 |
if C<$pid> is 0) has been received. More precisely: when the process |
478 |
receives a SIGCHLD, EV will fetch the outstanding exit/wait status for all |
479 |
changed/zombie children and call the callback. |
480 |
|
481 |
You can access both status and pid by using the C<rstatus> and C<rpid> |
482 |
methods on the watcher object. |
483 |
|
484 |
You can have as many pid watchers per pid as you want. |
485 |
|
486 |
The C<child_ns> variant doesn't start (activate) the newly created watcher. |
487 |
|
488 |
=item $w->set ($pid) |
489 |
|
490 |
Reconfigures the watcher, see the constructor above for details. Can be at |
491 |
any time. |
492 |
|
493 |
=item $current_pid = $w->pid |
494 |
|
495 |
=item $old_pid = $w->pid ($new_pid) |
496 |
|
497 |
Returns the previously set process id and optionally set a new one. |
498 |
|
499 |
=item $exit_status = $w->rstatus |
500 |
|
501 |
Return the exit/wait status (as returned by waitpid, see the waitpid entry |
502 |
in perlfunc). |
503 |
|
504 |
=item $pid = $w->rpid |
505 |
|
506 |
Return the pid of the awaited child (useful when you have installed a |
507 |
watcher for all pids). |
508 |
|
509 |
|
510 |
=item $w = EV::idle $callback |
511 |
|
512 |
=item $w = EV::idle_ns $callback |
513 |
|
514 |
Call the callback when there are no pending io, timer/periodic, signal or |
515 |
child events, i.e. when the process is idle. |
516 |
|
517 |
The process will not block as long as any idle watchers are active, and |
518 |
they will be called repeatedly until stopped. |
519 |
|
520 |
The C<idle_ns> variant doesn't start (activate) the newly created watcher. |
521 |
|
522 |
|
523 |
=item $w = EV::prepare $callback |
524 |
|
525 |
=item $w = EV::prepare_ns $callback |
526 |
|
527 |
Call the callback just before the process would block. You can still |
528 |
create/modify any watchers at this point. |
529 |
|
530 |
See the EV::check watcher, below, for explanations and an example. |
531 |
|
532 |
The C<prepare_ns> variant doesn't start (activate) the newly created watcher. |
533 |
|
534 |
|
535 |
=item $w = EV::check $callback |
536 |
|
537 |
=item $w = EV::check_ns $callback |
538 |
|
539 |
Call the callback just after the process wakes up again (after it has |
540 |
gathered events), but before any other callbacks have been invoked. |
541 |
|
542 |
This is used to integrate other event-based software into the EV |
543 |
mainloop: You register a prepare callback and in there, you create io and |
544 |
timer watchers as required by the other software. Here is a real-world |
545 |
example of integrating Net::SNMP (with some details left out): |
546 |
|
547 |
our @snmp_watcher; |
548 |
|
549 |
our $snmp_prepare = EV::prepare sub { |
550 |
# do nothing unless active |
551 |
$dispatcher->{_event_queue_h} |
552 |
or return; |
553 |
|
554 |
# make the dispatcher handle any outstanding stuff |
555 |
... not shown |
556 |
|
557 |
# create an IO watcher for each and every socket |
558 |
@snmp_watcher = ( |
559 |
(map { EV::io $_, EV::READ, sub { } } |
560 |
keys %{ $dispatcher->{_descriptors} }), |
561 |
|
562 |
EV::timer +($event->[Net::SNMP::Dispatcher::_ACTIVE] |
563 |
? $event->[Net::SNMP::Dispatcher::_TIME] - EV::now : 0), |
564 |
0, sub { }, |
565 |
); |
566 |
}; |
567 |
|
568 |
The callbacks are irrelevant (and are not even being called), the |
569 |
only purpose of those watchers is to wake up the process as soon as |
570 |
one of those events occurs (socket readable, or timer timed out). The |
571 |
corresponding EV::check watcher will then clean up: |
572 |
|
573 |
our $snmp_check = EV::check sub { |
574 |
# destroy all watchers |
575 |
@snmp_watcher = (); |
576 |
|
577 |
# make the dispatcher handle any new stuff |
578 |
... not shown |
579 |
}; |
580 |
|
581 |
The callbacks of the created watchers will not be called as the watchers |
582 |
are destroyed before this cna happen (remember EV::check gets called |
583 |
first). |
584 |
|
585 |
The C<check_ns> variant doesn't start (activate) the newly created watcher. |
586 |
|
587 |
=back |
588 |
|
589 |
=head1 THREADS |
590 |
|
591 |
Threads are not supported by this module in any way. Perl pseudo-threads |
592 |
is evil stuff and must die. As soon as Perl gains real threads I will work |
593 |
on thread support for it. |
594 |
|
595 |
=head1 FORK |
596 |
|
597 |
Most of the "improved" event delivering mechanisms of modern operating |
598 |
systems have quite a few problems with fork(2) (to put it bluntly: it is |
599 |
not supported and usually destructive). Libev makes it possible to work |
600 |
around this by having a function that recreates the kernel state after |
601 |
fork in the child. |
602 |
|
603 |
On non-win32 platforms, this module requires the pthread_atfork |
604 |
functionality to do this automatically for you. This function is quite |
605 |
buggy on most BSDs, though, so YMMV. The overhead for this is quite |
606 |
negligible, because everything the function currently does is set a flag |
607 |
that is checked only when the event loop gets used the next time, so when |
608 |
you do fork but not use EV, the overhead is minimal. |
609 |
|
610 |
On win32, there is no notion of fork so all this doesn't apply, of course. |
611 |
|
612 |
=cut |
613 |
|
614 |
our $DIED = sub { |
615 |
warn "EV: error in callback (ignoring): $@"; |
616 |
}; |
617 |
|
618 |
default_loop |
619 |
or die 'EV: cannot initialise libev backend. bad $ENV{LIBEV_METHODS}?'; |
620 |
|
621 |
1; |
622 |
|
623 |
=head1 SEE ALSO |
624 |
|
625 |
L<EV::DNS>. |
626 |
|
627 |
=head1 AUTHOR |
628 |
|
629 |
Marc Lehmann <schmorp@schmorp.de> |
630 |
http://home.schmorp.de/ |
631 |
|
632 |
=cut |
633 |
|