ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/EV/README
Revision: 1.27
Committed: Mon May 26 05:37:18 2008 UTC (15 years, 11 months ago) by root
Branch: MAIN
CVS Tags: rel-3_42
Changes since 1.26: +10 -2 lines
Log Message:
3.42

File Contents

# Content
1 NAME
2 EV - perl interface to libev, a high performance full-featured event
3 loop
4
5 SYNOPSIS
6 use EV;
7
8 # TIMERS
9
10 my $w = EV::timer 2, 0, sub {
11 warn "is called after 2s";
12 };
13
14 my $w = EV::timer 2, 2, sub {
15 warn "is called roughly every 2s (repeat = 2)";
16 };
17
18 undef $w; # destroy event watcher again
19
20 my $w = EV::periodic 0, 60, 0, sub {
21 warn "is called every minute, on the minute, exactly";
22 };
23
24 # IO
25
26 my $w = EV::io *STDIN, EV::READ, sub {
27 my ($w, $revents) = @_; # all callbacks receive the watcher and event mask
28 warn "stdin is readable, you entered: ", <STDIN>;
29 };
30
31 # SIGNALS
32
33 my $w = EV::signal 'QUIT', sub {
34 warn "sigquit received\n";
35 };
36
37 # CHILD/PID STATUS CHANGES
38
39 my $w = EV::child 666, 0, sub {
40 my ($w, $revents) = @_;
41 my $status = $w->rstatus;
42 };
43
44 # STAT CHANGES
45 my $w = EV::stat "/etc/passwd", 10, sub {
46 my ($w, $revents) = @_;
47 warn $w->path, " has changed somehow.\n";
48 };
49
50 # MAINLOOP
51 EV::loop; # loop until EV::unloop is called or all watchers stop
52 EV::loop EV::LOOP_ONESHOT; # block until at least one event could be handled
53 EV::loop EV::LOOP_NONBLOCK; # try to handle same events, but do not block
54
55 DESCRIPTION
56 This module provides an interface to libev
57 (<http://software.schmorp.de/pkg/libev.html>). While the documentation
58 below is comprehensive, one might also consult the documentation of
59 libev itself (<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>)
60 for more subtle details on watcher semantics or some discussion on the
61 available backends, or how to force a specific backend with
62 "LIBEV_FLAGS", or just about in any case because it has much more
63 detailed information.
64
65 This module is very fast and scalable. It is actually so fast that you
66 can use it through the AnyEvent module, stay portable to other event
67 loops (if you don't rely on any watcher types not available through it)
68 and still be faster than with any other event loop currently supported
69 in Perl.
70
71 EVENT LOOPS
72 EV supports multiple event loops: There is a single "default event loop"
73 that can handle everything including signals and child watchers, and any
74 number of "dynamic event loops" that can use different backends (with
75 various limitations), but no child and signal watchers.
76
77 You do not have to do anything to create the default event loop: When
78 the module is loaded a suitable backend is selected on the premise of
79 selecting a working backend (which for example rules out kqueue on most
80 BSDs). Modules should, unless they have "special needs" always use the
81 default loop as this is fastest (perl-wise), best supported by other
82 modules (e.g. AnyEvent or Coro) and most portable event loop.
83
84 For specific programs you can create additional event loops dynamically.
85
86 $loop = new EV::loop [$flags]
87 Create a new event loop as per the specified flags. Please refer to
88 the "ev_loop_new ()" function description in the libev documentation
89 (<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#GLOBAL_FUNCTI
90 ONS>) for more info.
91
92 The loop will automatically be destroyed when it is no longer
93 referenced by any watcher and the loop object goes out of scope.
94
95 Using "EV::FLAG_FORKCHECK" is recommended, as only the default event
96 loop is protected by this module.
97
98 $loop->loop_fork
99 Must be called after a fork in the child, before entering or
100 continuing the event loop. An alternative is to use
101 "EV::FLAG_FORKCHECK" which calls this function automatically, at
102 some performance loss (refer to the libev documentation).
103
104 $loop->loop_verify
105 Calls "ev_verify" to make internal consistency checks (for debugging
106 libev) and abort the program if any data structures wree found to be
107 corrupted.
108
109 $loop = EV::default_loop [$flags]
110 Return the default loop (which is a singleton object). Since this
111 module already creates the default loop with default flags,
112 specifying flags here will not have any effect unless you destroy
113 the default loop.
114
115 BASIC INTERFACE
116 $EV::DIED
117 Must contain a reference to a function that is called when a
118 callback throws an exception (with $@ containing the error). The
119 default prints an informative message and continues.
120
121 If this callback throws an exception it will be silently ignored.
122
123 $flags = EV::supported_backends
124 $flags = EV::recommended_backends
125 $flags = EV::embeddable_backends
126 Returns the set (see "EV::BACKEND_*" flags) of backends supported by
127 this instance of EV, the set of recommended backends (supposed to be
128 good) for this platform and the set of embeddable backends (see
129 EMBED WATCHERS).
130
131 EV::sleep $seconds
132 Block the process for the given number of (fractional) seconds.
133
134 $time = EV::time
135 Returns the current time in (fractional) seconds since the epoch.
136
137 $time = EV::now
138 $time = $loop->now
139 Returns the time the last event loop iteration has been started.
140 This is the time that (relative) timers are based on, and refering
141 to it is usually faster then calling EV::time.
142
143 $backend = EV::backend
144 $backend = $loop->backend
145 Returns an integer describing the backend used by libev
146 (EV::METHOD_SELECT or EV::METHOD_EPOLL).
147
148 EV::loop [$flags]
149 $loop->loop ([$flags])
150 Begin checking for events and calling callbacks. It returns when a
151 callback calls EV::unloop.
152
153 The $flags argument can be one of the following:
154
155 0 as above
156 EV::LOOP_ONESHOT block at most once (wait, but do not loop)
157 EV::LOOP_NONBLOCK do not block at all (fetch/handle events but do not wait)
158
159 EV::unloop [$how]
160 $loop->unloop ([$how])
161 When called with no arguments or an argument of EV::UNLOOP_ONE,
162 makes the innermost call to EV::loop return.
163
164 When called with an argument of EV::UNLOOP_ALL, all calls to
165 EV::loop will return as fast as possible.
166
167 $count = EV::loop_count
168 $count = $loop->loop_count
169 Return the number of times the event loop has polled for new events.
170 Sometiems useful as a generation counter.
171
172 EV::once $fh_or_undef, $events, $timeout, $cb->($revents)
173 $loop->once ($fh_or_undef, $events, $timeout, $cb->($revents))
174 This function rolls together an I/O and a timer watcher for a single
175 one-shot event without the need for managing a watcher object.
176
177 If $fh_or_undef is a filehandle or file descriptor, then $events
178 must be a bitset containing either "EV::READ", "EV::WRITE" or
179 "EV::READ | EV::WRITE", indicating the type of I/O event you want to
180 wait for. If you do not want to wait for some I/O event, specify
181 "undef" for $fh_or_undef and 0 for $events).
182
183 If timeout is "undef" or negative, then there will be no timeout.
184 Otherwise a EV::timer with this value will be started.
185
186 When an error occurs or either the timeout or I/O watcher triggers,
187 then the callback will be called with the received event set (in
188 general you can expect it to be a combination of "EV::ERROR",
189 "EV::READ", "EV::WRITE" and "EV::TIMEOUT").
190
191 EV::once doesn't return anything: the watchers stay active till
192 either of them triggers, then they will be stopped and freed, and
193 the callback invoked.
194
195 EV::feed_fd_event ($fd, $revents)
196 $loop->feed_fd_event ($fd, $revents)
197 Feed an event on a file descriptor into EV. EV will react to this
198 call as if the readyness notifications specified by $revents (a
199 combination of "EV::READ" and "EV::WRITE") happened on the file
200 descriptor $fd.
201
202 EV::feed_signal_event ($signal)
203 Feed a signal event into EV. EV will react to this call as if the
204 signal specified by $signal had occured.
205
206 EV::set_io_collect_interval $time
207 $loop->set_io_collect_interval ($time)
208 EV::set_timeout_collect_interval $time
209 $loop->set_timeout_collect_interval ($time)
210 These advanced functions set the minimum block interval when polling
211 for I/O events and the minimum wait interval for timer events. See
212 the libev documentation at
213 <http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#FUNCTIONS_CONT
214 ROLLING_THE_EVENT_LOOP> for a more detailed discussion.
215
216 WATCHER OBJECTS
217 A watcher is an object that gets created to record your interest in some
218 event. For instance, if you want to wait for STDIN to become readable,
219 you would create an EV::io watcher for that:
220
221 my $watcher = EV::io *STDIN, EV::READ, sub {
222 my ($watcher, $revents) = @_;
223 warn "yeah, STDIN should now be readable without blocking!\n"
224 };
225
226 All watchers can be active (waiting for events) or inactive (paused).
227 Only active watchers will have their callbacks invoked. All callbacks
228 will be called with at least two arguments: the watcher and a bitmask of
229 received events.
230
231 Each watcher type has its associated bit in revents, so you can use the
232 same callback for multiple watchers. The event mask is named after the
233 type, i..e. EV::child sets EV::CHILD, EV::prepare sets EV::PREPARE,
234 EV::periodic sets EV::PERIODIC and so on, with the exception of I/O
235 events (which can set both EV::READ and EV::WRITE bits), and EV::timer
236 (which uses EV::TIMEOUT).
237
238 In the rare case where one wants to create a watcher but not start it at
239 the same time, each constructor has a variant with a trailing "_ns" in
240 its name, e.g. EV::io has a non-starting variant EV::io_ns and so on.
241
242 Please note that a watcher will automatically be stopped when the
243 watcher object is destroyed, so you *need* to keep the watcher objects
244 returned by the constructors.
245
246 Also, all methods changing some aspect of a watcher (->set, ->priority,
247 ->fh and so on) automatically stop and start it again if it is active,
248 which means pending events get lost.
249
250 COMMON WATCHER METHODS
251 This section lists methods common to all watchers.
252
253 $w->start
254 Starts a watcher if it isn't active already. Does nothing to an
255 already active watcher. By default, all watchers start out in the
256 active state (see the description of the "_ns" variants if you need
257 stopped watchers).
258
259 $w->stop
260 Stop a watcher if it is active. Also clear any pending events
261 (events that have been received but that didn't yet result in a
262 callback invocation), regardless of whether the watcher was active
263 or not.
264
265 $bool = $w->is_active
266 Returns true if the watcher is active, false otherwise.
267
268 $current_data = $w->data
269 $old_data = $w->data ($new_data)
270 Queries a freely usable data scalar on the watcher and optionally
271 changes it. This is a way to associate custom data with a watcher:
272
273 my $w = EV::timer 60, 0, sub {
274 warn $_[0]->data;
275 };
276 $w->data ("print me!");
277
278 $current_cb = $w->cb
279 $old_cb = $w->cb ($new_cb)
280 Queries the callback on the watcher and optionally changes it. You
281 can do this at any time without the watcher restarting.
282
283 $current_priority = $w->priority
284 $old_priority = $w->priority ($new_priority)
285 Queries the priority on the watcher and optionally changes it.
286 Pending watchers with higher priority will be invoked first. The
287 valid range of priorities lies between EV::MAXPRI (default 2) and
288 EV::MINPRI (default -2). If the priority is outside this range it
289 will automatically be normalised to the nearest valid priority.
290
291 The default priority of any newly-created watcher is 0.
292
293 Note that the priority semantics have not yet been fleshed out and
294 are subject to almost certain change.
295
296 $w->invoke ($revents)
297 Call the callback *now* with the given event mask.
298
299 $w->feed_event ($revents)
300 Feed some events on this watcher into EV. EV will react to this call
301 as if the watcher had received the given $revents mask.
302
303 $revents = $w->clear_pending
304 If the watcher is pending, this function clears its pending status
305 and returns its $revents bitset (as if its callback was invoked). If
306 the watcher isn't pending it does nothing and returns 0.
307
308 $previous_state = $w->keepalive ($bool)
309 Normally, "EV::loop" will return when there are no active watchers
310 (which is a "deadlock" because no progress can be made anymore).
311 This is convinient because it allows you to start your watchers (and
312 your jobs), call "EV::loop" once and when it returns you know that
313 all your jobs are finished (or they forgot to register some watchers
314 for their task :).
315
316 Sometimes, however, this gets in your way, for example when the
317 module that calls "EV::loop" (usually the main program) is not the
318 same module as a long-living watcher (for example a DNS client
319 module written by somebody else even). Then you might want any
320 outstanding requests to be handled, but you would not want to keep
321 "EV::loop" from returning just because you happen to have this
322 long-running UDP port watcher.
323
324 In this case you can clear the keepalive status, which means that
325 even though your watcher is active, it won't keep "EV::loop" from
326 returning.
327
328 The initial value for keepalive is true (enabled), and you cna
329 change it any time.
330
331 Example: Register an I/O watcher for some UDP socket but do not keep
332 the event loop from running just because of that watcher.
333
334 my $udp_socket = ...
335 my $udp_watcher = EV::io $udp_socket, EV::READ, sub { ... };
336 $1000udp_watcher->keepalive (0);
337
338 $loop = $w->loop
339 Return the loop that this watcher is attached to.
340
341 WATCHER TYPES
342 Each of the following subsections describes a single watcher type.
343
344 I/O WATCHERS - is this file descriptor readable or writable?
345 $w = EV::io $fileno_or_fh, $eventmask, $callback
346 $w = EV::io_ns $fileno_or_fh, $eventmask, $callback
347 $w = $loop->io ($fileno_or_fh, $eventmask, $callback)
348 $w = $loop->io_ns ($fileno_or_fh, $eventmask, $callback)
349 As long as the returned watcher object is alive, call the $callback
350 when at least one of events specified in $eventmask occurs.
351
352 The $eventmask can be one or more of these constants ORed together:
353
354 EV::READ wait until read() wouldn't block anymore
355 EV::WRITE wait until write() wouldn't block anymore
356
357 The "io_ns" variant doesn't start (activate) the newly created
358 watcher.
359
360 $w->set ($fileno_or_fh, $eventmask)
361 Reconfigures the watcher, see the constructor above for details. Can
362 be called at any time.
363
364 $current_fh = $w->fh
365 $old_fh = $w->fh ($new_fh)
366 Returns the previously set filehandle and optionally set a new one.
367
368 $current_eventmask = $w->events
369 $old_eventmask = $w->events ($new_eventmask)
370 Returns the previously set event mask and optionally set a new one.
371
372 TIMER WATCHERS - relative and optionally repeating timeouts
373 $w = EV::timer $after, $repeat, $callback
374 $w = EV::timer_ns $after, $repeat, $callback
375 $w = $loop->timer ($after, $repeat, $callback)
376 $w = $loop->timer_ns ($after, $repeat, $callback)
377 Calls the callback after $after seconds (which may be fractional).
378 If $repeat is non-zero, the timer will be restarted (with the
379 $repeat value as $after) after the callback returns.
380
381 This means that the callback would be called roughly after $after
382 seconds, and then every $repeat seconds. The timer does his best not
383 to drift, but it will not invoke the timer more often then once per
384 event loop iteration, and might drift in other cases. If that isn't
385 acceptable, look at EV::periodic, which can provide long-term stable
386 timers.
387
388 The timer is based on a monotonic clock, that is, if somebody is
389 sitting in front of the machine while the timer is running and
390 changes the system clock, the timer will nevertheless run (roughly)
391 the same time.
392
393 The "timer_ns" variant doesn't start (activate) the newly created
394 watcher.
395
396 $w->set ($after, $repeat)
397 Reconfigures the watcher, see the constructor above for details. Can
398 be called at any time.
399
400 $w->again
401 Similar to the "start" method, but has special semantics for
402 repeating timers:
403
404 If the timer is active and non-repeating, it will be stopped.
405
406 If the timer is active and repeating, reset the timeout to occur
407 $repeat seconds after now.
408
409 If the timer is inactive and repeating, start it using the repeat
410 value.
411
412 Otherwise do nothing.
413
414 This behaviour is useful when you have a timeout for some IO
415 operation. You create a timer object with the same value for $after
416 and $repeat, and then, in the read/write watcher, run the "again"
417 method on the timeout.
418
419 PERIODIC WATCHERS - to cron or not to cron?
420 $w = EV::periodic $at, $interval, $reschedule_cb, $callback
421 $w = EV::periodic_ns $at, $interval, $reschedule_cb, $callback
422 $w = $loop->periodic ($at, $interval, $reschedule_cb, $callback)
423 $w = $loop->periodic_ns ($at, $interval, $reschedule_cb, $callback)
424 Similar to EV::timer, but is not based on relative timeouts but on
425 absolute times. Apart from creating "simple" timers that trigger
426 "at" the specified time, it can also be used for non-drifting
427 absolute timers and more complex, cron-like, setups that are not
428 adversely affected by time jumps (i.e. when the system clock is
429 changed by explicit date -s or other means such as ntpd). It is also
430 the most complex watcher type in EV.
431
432 It has three distinct "modes":
433
434 * absolute timer ($interval = $reschedule_cb = 0)
435
436 This time simply fires at the wallclock time $at and doesn't
437 repeat. It will not adjust when a time jump occurs, that is, if
438 it is to be run at January 1st 2011 then it will run when the
439 system time reaches or surpasses this time.
440
441 * repeating interval timer ($interval > 0, $reschedule_cb = 0)
442
443 In this mode the watcher will always be scheduled to time out at
444 the next "$at + N * $interval" time (for some integer N) and
445 then repeat, regardless of any time jumps.
446
447 This can be used to create timers that do not drift with respect
448 to system time:
449
450 my $hourly = EV::periodic 0, 3600, 0, sub { print "once/hour\n" };
451
452 That doesn't mean there will always be 3600 seconds in between
453 triggers, but only that the the clalback will be called when the
454 system time shows a full hour (UTC).
455
456 Another way to think about it (for the mathematically inclined)
457 is that EV::periodic will try to run the callback in this mode
458 at the next possible time where "$time = $at (mod $interval)",
459 regardless of any time jumps.
460
461 * manual reschedule mode ($reschedule_cb = coderef)
462
463 In this mode $interval and $at are both being ignored. Instead,
464 each time the periodic watcher gets scheduled, the reschedule
465 callback ($reschedule_cb) will be called with the watcher as
466 first, and the current time as second argument.
467
468 *This callback MUST NOT stop or destroy this or any other
469 periodic watcher, ever, and MUST NOT call any event loop
470 functions or methods*. If you need to stop it, return 1e30 and
471 stop it afterwards. You may create and start a "EV::prepare"
472 watcher for this task.
473
474 It must return the next time to trigger, based on the passed
475 time value (that is, the lowest time value larger than or equal
476 to to the second argument). It will usually be called just
477 before the callback will be triggered, but might be called at
478 other times, too.
479
480 This can be used to create very complex timers, such as a timer
481 that triggers on each midnight, local time (actually 24 hours
482 after the last midnight, to keep the example simple. If you know
483 a way to do it correctly in about the same space (without
484 requiring elaborate modules), drop me a note :):
485
486 my $daily = EV::periodic 0, 0, sub {
487 my ($w, $now) = @_;
488
489 use Time::Local ();
490 my (undef, undef, undef, $d, $m, $y) = localtime $now;
491 86400 + Time::Local::timelocal 0, 0, 0, $d, $m, $y
492 }, sub {
493 print "it's midnight or likely shortly after, now\n";
494 };
495
496 The "periodic_ns" variant doesn't start (activate) the newly created
497 watcher.
498
499 $w->set ($at, $interval, $reschedule_cb)
500 Reconfigures the watcher, see the constructor above for details. Can
501 be called at any time.
502
503 $w->again
504 Simply stops and starts the watcher again.
505
506 $time = $w->at
507 Return the time that the watcher is expected to trigger next.
508
509 SIGNAL WATCHERS - signal me when a signal gets signalled!
510 $w = EV::signal $signal, $callback
511 $w = EV::signal_ns $signal, $callback
512 Call the callback when $signal is received (the signal can be
513 specified by number or by name, just as with "kill" or %SIG).
514
515 EV will grab the signal for the process (the kernel only allows one
516 component to receive a signal at a time) when you start a signal
517 watcher, and removes it again when you stop it. Perl does the same
518 when you add/remove callbacks to %SIG, so watch out.
519
520 You can have as many signal watchers per signal as you want.
521
522 The "signal_ns" variant doesn't start (activate) the newly created
523 watcher.
524
525 $w->set ($signal)
526 Reconfigures the watcher, see the constructor above for details. Can
527 be called at any time.
528
529 $current_signum = $w->signal
530 $old_signum = $w->signal ($new_signal)
531 Returns the previously set signal (always as a number not name) and
532 optionally set a new one.
533
534 CHILD WATCHERS - watch out for process status changes
535 $w = EV::child $pid, $trace, $callback
536 $w = EV::child_ns $pid, $trace, $callback
537 $w = $loop->child ($pid, $trace, $callback)
538 $w = $loop->child_ns ($pid, $trace, $callback)
539 Call the callback when a status change for pid $pid (or any pid if
540 $pid is 0) has been received (a status change happens when the
541 process terminates or is killed, or, when trace is true,
542 additionally when it is stopped or continued). More precisely: when
543 the process receives a "SIGCHLD", EV will fetch the outstanding
544 exit/wait status for all changed/zombie children and call the
545 callback.
546
547 It is valid (and fully supported) to install a child watcher after a
548 child has exited but before the event loop has started its next
549 iteration (for example, first you "fork", then the new child process
550 might exit, and only then do you install a child watcher in the
551 parent for the new pid).
552
553 You can access both exit (or tracing) status and pid by using the
554 "rstatus" and "rpid" methods on the watcher object.
555
556 You can have as many pid watchers per pid as you want, they will all
557 be called.
558
559 The "child_ns" variant doesn't start (activate) the newly created
560 watcher.
561
562 $w->set ($pid, $trace)
563 Reconfigures the watcher, see the constructor above for details. Can
564 be called at any time.
565
566 $current_pid = $w->pid
567 Returns the previously set process id and optionally set a new one.
568
569 $exit_status = $w->rstatus
570 Return the exit/wait status (as returned by waitpid, see the waitpid
571 entry in perlfunc).
572
573 $pid = $w->rpid
574 Return the pid of the awaited child (useful when you have installed
575 a watcher for all pids).
576
577 STAT WATCHERS - did the file attributes just change?
578 $w = EV::stat $path, $interval, $callback
579 $w = EV::stat_ns $path, $interval, $callback
580 $w = $loop->stat ($path, $interval, $callback)
581 $w = $loop->stat_ns ($path, $interval, $callback)
582 Call the callback when a file status change has been detected on
583 $path. The $path does not need to exist, changing from "path exists"
584 to "path does not exist" is a status change like any other.
585
586 The $interval is a recommended polling interval for systems where
587 OS-supported change notifications don't exist or are not supported.
588 If you use 0 then an unspecified default is used (which is highly
589 recommended!), which is to be expected to be around five seconds
590 usually.
591
592 This watcher type is not meant for massive numbers of stat watchers,
593 as even with OS-supported change notifications, this can be
594 resource-intensive.
595
596 The "stat_ns" variant doesn't start (activate) the newly created
597 watcher.
598
599 ... = $w->stat
600 This call is very similar to the perl "stat" built-in: It stats
601 (using "lstat") the path specified in the watcher and sets perls
602 stat cache (as well as EV's idea of the current stat values) to the
603 values found.
604
605 In scalar context, a boolean is return indicating success or failure
606 of the stat. In list context, the same 13-value list as with stat is
607 returned (except that the blksize and blocks fields are not
608 reliable).
609
610 In the case of an error, errno is set to "ENOENT" (regardless of the
611 actual error value) and the "nlink" value is forced to zero (if the
612 stat was successful then nlink is guaranteed to be non-zero).
613
614 See also the next two entries for more info.
615
616 ... = $w->attr
617 Just like "$w->stat", but without the initial stat'ing: this returns
618 the values most recently detected by EV. See the next entry for more
619 info.
620
621 ... = $w->prev
622 Just like "$w->stat", but without the initial stat'ing: this returns
623 the previous set of values, before the change.
624
625 That is, when the watcher callback is invoked, "$w->prev" will be
626 set to the values found *before* a change was detected, while
627 "$w->attr" returns the values found leading to the change detection.
628 The difference (if any) between "prev" and "attr" is what triggered
629 the callback.
630
631 If you did something to the filesystem object and do not want to
632 trigger yet another change, you can call "stat" to update EV's idea
633 of what the current attributes are.
634
635 $w->set ($path, $interval)
636 Reconfigures the watcher, see the constructor above for details. Can
637 be called at any time.
638
639 $current_path = $w->path
640 $old_path = $w->path ($new_path)
641 Returns the previously set path and optionally set a new one.
642
643 $current_interval = $w->interval
644 $old_interval = $w->interval ($new_interval)
645 Returns the previously set interval and optionally set a new one.
646 Can be used to query the actual interval used.
647
648 IDLE WATCHERS - when you've got nothing better to do...
649 $w = EV::idle $callback
650 $w = EV::idle_ns $callback
651 $w = $loop->idle ($callback)
652 $w = $loop->idle_ns ($callback)
653 Call the callback when there are no other pending watchers of the
654 same or higher priority (excluding check, prepare and other idle
655 watchers of the same or lower priority, of course). They are called
656 idle watchers because when the watcher is the highest priority
657 pending event in the process, the process is considered to be idle
658 at that priority.
659
660 If you want a watcher that is only ever called when *no* other
661 events are outstanding you have to set the priority to "EV::MINPRI".
662
663 The process will not block as long as any idle watchers are active,
664 and they will be called repeatedly until stopped.
665
666 For example, if you have idle watchers at priority 0 and 1, and an
667 I/O watcher at priority 0, then the idle watcher at priority 1 and
668 the I/O watcher will always run when ready. Only when the idle
669 watcher at priority 1 is stopped and the I/O watcher at priority 0
670 is not pending with the 0-priority idle watcher be invoked.
671
672 The "idle_ns" variant doesn't start (activate) the newly created
673 watcher.
674
675 PREPARE WATCHERS - customise your event loop!
676 $w = EV::prepare $callback
677 $w = EV::prepare_ns $callback
678 $w = $loop->prepare ($callback)
679 $w = $loop->prepare_ns ($callback)
680 Call the callback just before the process would block. You can still
681 create/modify any watchers at this point.
682
683 See the EV::check watcher, below, for explanations and an example.
684
685 The "prepare_ns" variant doesn't start (activate) the newly created
686 watcher.
687
688 CHECK WATCHERS - customise your event loop even more!
689 $w = EV::check $callback
690 $w = EV::check_ns $callback
691 $w = $loop->check ($callback)
692 $w = $loop->check_ns ($callback)
693 Call the callback just after the process wakes up again (after it
694 has gathered events), but before any other callbacks have been
695 invoked.
696
697 This is used to integrate other event-based software into the EV
698 mainloop: You register a prepare callback and in there, you create
699 io and timer watchers as required by the other software. Here is a
700 real-world example of integrating Net::SNMP (with some details left
701 out):
702
703 our @snmp_watcher;
704
705 our $snmp_prepare = EV::prepare sub {
706 # do nothing unless active
707 $dispatcher->{_event_queue_h}
708 or return;
709
710 # make the dispatcher handle any outstanding stuff
711 ... not shown
712
713 # create an I/O watcher for each and every socket
714 @snmp_watcher = (
715 (map { EV::io $_, EV::READ, sub { } }
716 keys %{ $dispatcher->{_descriptors} }),
717
718 EV::timer +($event->[Net::SNMP::Dispatcher::_ACTIVE]
719 ? $event->[Net::SNMP::Dispatcher::_TIME] - EV::now : 0),
720 0, sub { },
721 );
722 };
723
724 The callbacks are irrelevant (and are not even being called), the
725 only purpose of those watchers is to wake up the process as soon as
726 one of those events occurs (socket readable, or timer timed out).
727 The corresponding EV::check watcher will then clean up:
728
729 our $snmp_check = EV::check sub {
730 # destroy all watchers
731 @snmp_watcher = ();
732
733 # make the dispatcher handle any new stuff
734 ... not shown
735 };
736
737 The callbacks of the created watchers will not be called as the
738 watchers are destroyed before this cna happen (remember EV::check
739 gets called first).
740
741 The "check_ns" variant doesn't start (activate) the newly created
742 watcher.
743
744 FORK WATCHERS - the audacity to resume the event loop after a fork
745 Fork watchers are called when a "fork ()" was detected. The invocation
746 is done before the event loop blocks next and before "check" watchers
747 are being called, and only in the child after the fork.
748
749 $w = EV::fork $callback
750 $w = EV::fork_ns $callback
751 $w = $loop->fork ($callback)
752 $w = $loop->fork_ns ($callback)
753 Call the callback before the event loop is resumed in the child
754 process after a fork.
755
756 The "fork_ns" variant doesn't start (activate) the newly created
757 watcher.
758
759 EMBED WATCHERS - when one backend isn't enough...
760 This is a rather advanced watcher type that lets you embed one event
761 loop into another (currently only IO events are supported in the
762 embedded loop, other types of watchers might be handled in a delayed or
763 incorrect fashion and must not be used).
764
765 See the libev documentation at
766 <http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#code_ev_embed_code
767 _when_one_backend_> for more details.
768
769 In short, this watcher is most useful on BSD systems without working
770 kqueue to still be able to handle a large number of sockets:
771
772 my $socket_loop;
773
774 # check wether we use SELECT or POLL _and_ KQUEUE is supported
775 if (
776 (EV::backend & (EV::BACKEND_POLL | EV::BACKEND_SELECT))
777 && (EV::supported_backends & EV::embeddable_backends & EV::BACKEND_KQUEUE)
778 ) {
779 # use kqueue for sockets
780 $socket_loop = new EV::Loop EV::BACKEND_KQUEUE | EV::FLAG_NOENV;
781 }
782
783 # use the default loop otherwise
784 $socket_loop ||= EV::default_loop;
785
786 $w = EV::embed $otherloop, $callback
787 $w = EV::embed_ns $otherloop, $callback
788 $w = $loop->embed ($otherloop, $callback)
789 $w = $loop->embed_ns ($otherloop, $callback)
790 Call the callback when the embedded event loop ($otherloop) has any
791 I/O activity. The $callback should alwas be specified as "undef" in
792 this version of EV, which means the embedded event loop will be
793 managed automatically.
794
795 The "embed_ns" variant doesn't start (activate) the newly created
796 watcher.
797
798 ASYNC WATCHERS - how to wake up another event loop
799 Async watchers are provided by EV, but have little use in perl directly,
800 as perl neither supports threads nor direct access to signal handlers or
801 other contexts where they could be of value.
802
803 It is, however, possible to use them from the XS level.
804
805 Please see the libev documentation for further details.
806
807 $w = EV::async $callback
808 $w = EV::async_ns $callback
809 $w->send
810 $bool = $w->async_pending
811
812 PERL SIGNALS
813 While Perl signal handling (%SIG) is not affected by EV, the behaviour
814 with EV is as the same as any other C library: Perl-signals will only be
815 handled when Perl runs, which means your signal handler might be invoked
816 only the next time an event callback is invoked.
817
818 The solution is to use EV signal watchers (see "EV::signal"), which will
819 ensure proper operations with regards to other event watchers.
820
821 If you cannot do this for whatever reason, you can also force a watcher
822 to be called on every event loop iteration by installing a "EV::check"
823 watcher:
824
825 my $async_check = EV::check sub { };
826
827 This ensures that perl gets into control for a short time to handle any
828 pending signals, and also ensures (slightly) slower overall operation.
829
830 THREADS
831 Threads are not supported by this module in any way. Perl pseudo-threads
832 is evil stuff and must die. As soon as Perl gains real threads I will
833 work on thread support for it.
834
835 FORK
836 Most of the "improved" event delivering mechanisms of modern operating
837 systems have quite a few problems with fork(2) (to put it bluntly: it is
838 not supported and usually destructive). Libev makes it possible to work
839 around this by having a function that recreates the kernel state after
840 fork in the child.
841
842 On non-win32 platforms, this module requires the pthread_atfork
843 functionality to do this automatically for you. This function is quite
844 buggy on most BSDs, though, so YMMV. The overhead for this is quite
845 negligible, because everything the function currently does is set a flag
846 that is checked only when the event loop gets used the next time, so
847 when you do fork but not use EV, the overhead is minimal.
848
849 On win32, there is no notion of fork so all this doesn't apply, of
850 course.
851
852 SEE ALSO
853 EV::ADNS (asynchronous DNS), Glib::EV (makes Glib/Gtk2 use EV as event
854 loop), EV::Glib (embed Glib into EV), Coro::EV (efficient coroutines
855 with EV), Net::SNMP::EV (asynchronous SNMP), AnyEvent for event-loop
856 agnostic and portable event driven programming.
857
858 AUTHOR
859 Marc Lehmann <schmorp@schmorp.de>
860 http://home.schmorp.de/
861