ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/EV/EV.pm
(Generate patch)

Comparing EV/EV.pm (file contents):
Revision 1.3 by root, Fri Oct 26 17:40:31 2007 UTC vs.
Revision 1.93 by root, Tue May 20 19:52:12 2008 UTC

1=head1 NAME 1=head1 NAME
2 2
3EV - perl interface to libevent, monkey.org/~provos/libevent/ 3EV - perl interface to libev, a high performance full-featured event loop
4 4
5=head1 SYNOPSIS 5=head1 SYNOPSIS
6 6
7 use EV; 7 use EV;
8 8
9 # TIMER 9 # TIMERS
10 10
11 my $w = EV::timer 2, 0, sub { 11 my $w = EV::timer 2, 0, sub {
12 warn "is called after 2s"; 12 warn "is called after 2s";
13 }; 13 };
14 14
15 my $w = EV::timer 2, 1, sub { 15 my $w = EV::timer 2, 2, sub {
16 warn "is called roughly every 2s (repeat = 1)"; 16 warn "is called roughly every 2s (repeat = 2)";
17 }; 17 };
18 18
19 undef $w; # destroy event watcher again 19 undef $w; # destroy event watcher again
20 20
21 # IO
22
23 my $w = EV::timer_abs 0, 60, sub { 21 my $w = EV::periodic 0, 60, 0, sub {
24 warn "is called every minute, on the minute, exactly"; 22 warn "is called every minute, on the minute, exactly";
25 }; 23 };
26 24
25 # IO
26
27 my $w = EV::io \*STDIN, EV::READ | EV::PERSIST, sub { 27 my $w = EV::io *STDIN, EV::READ, sub {
28 my ($w, $events) = @_; # all callbacks get the watcher object and event mask 28 my ($w, $revents) = @_; # all callbacks receive the watcher and event mask
29 if ($events & EV::TIMEOUT) {
30 warn "nothign received on stdin for 10 seconds, retrying";
31 } else {
32 warn "stdin is readable, you entered: ", <STDIN>; 29 warn "stdin is readable, you entered: ", <STDIN>;
33 }
34 }; 30 };
35 $w->timeout (10); 31
32 # SIGNALS
33
34 my $w = EV::signal 'QUIT', sub {
35 warn "sigquit received\n";
36 };
37
38 # CHILD/PID STATUS CHANGES
36 39
40 my $w = EV::child 666, 0, sub {
41 my ($w, $revents) = @_;
42 my $status = $w->rstatus;
43 };
44
45 # STAT CHANGES
46 my $w = EV::stat "/etc/passwd", 10, sub {
47 my ($w, $revents) = @_;
48 warn $w->path, " has changed somehow.\n";
49 };
50
37 # MAINLOOP 51 # MAINLOOP
38 EV::dispatch; # loop as long as watchers are active 52 EV::loop; # loop until EV::unloop is called or all watchers stop
39 EV::loop; # the same thing 53 EV::loop EV::LOOP_ONESHOT; # block until at least one event could be handled
40 EV::loop EV::LOOP_ONCE; 54 EV::loop EV::LOOP_NONBLOCK; # try to handle same events, but do not block
41 EV::loop EV::LOOP_ONSHOT;
42 55
43=head1 DESCRIPTION 56=head1 DESCRIPTION
44 57
45This module provides an interface to libevent 58This module provides an interface to libev
46(L<http://monkey.org/~provos/libevent/>). 59(L<http://software.schmorp.de/pkg/libev.html>). While the documentation
60below is comprehensive, one might also consult the documentation of libev
61itself (L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod>) for more
62subtle details on watcher semantics or some discussion on the available
63backends, or how to force a specific backend with C<LIBEV_FLAGS>, or just
64about in any case because it has much more detailed information.
65
66This module is very fast and scalable. It is actually so fast that you
67can use it through the L<AnyEvent> module, stay portable to other event
68loops (if you don't rely on any watcher types not available through it)
69and still be faster than with any other event loop currently supported in
70Perl.
47 71
48=cut 72=cut
49 73
50package EV; 74package EV;
51 75
52use strict; 76use strict;
53 77
54BEGIN { 78BEGIN {
55 our $VERSION = '0.01'; 79 our $VERSION = '3.4';
56 use XSLoader; 80 use XSLoader;
57 XSLoader::load "EV", $VERSION; 81 XSLoader::load "EV", $VERSION;
58} 82}
59 83
60=head1 FUNCTIONAL INTERFACE 84@EV::IO::ISA =
85@EV::Timer::ISA =
86@EV::Periodic::ISA =
87@EV::Signal::ISA =
88@EV::Child::ISA =
89@EV::Stat::ISA =
90@EV::Idle::ISA =
91@EV::Prepare::ISA =
92@EV::Check::ISA =
93@EV::Embed::ISA =
94@EV::Fork::ISA =
95@EV::Async::ISA =
96 "EV::Watcher";
61 97
62=over 4 98@EV::Loop::Default::ISA = "EV::Loop";
63 99
100=head1 EVENT LOOPS
101
102EV supports multiple event loops: There is a single "default event loop"
103that can handle everything including signals and child watchers, and any
104number of "dynamic event loops" that can use different backends (with
105various limitations), but no child and signal watchers.
106
107You do not have to do anything to create the default event loop: When
108the module is loaded a suitable backend is selected on the premise of
109selecting a working backend (which for example rules out kqueue on most
110BSDs). Modules should, unless they have "special needs" always use the
111default loop as this is fastest (perl-wise), best supported by other
112modules (e.g. AnyEvent or Coro) and most portable event loop.
113
114For specific programs you can create additional event loops dynamically.
115
116=over 4
117
118=item $loop = new EV::loop [$flags]
119
120Create a new event loop as per the specified flags. Please refer to the
121C<ev_loop_new ()> function description in the libev documentation
122(L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#GLOBAL_FUNCTIONS>)
123for more info.
124
125The loop will automatically be destroyed when it is no longer referenced
126by any watcher and the loop object goes out of scope.
127
128Using C<EV::FLAG_FORKCHECK> is recommended, as only the default event loop
129is protected by this module.
130
131=item $loop->loop_fork
132
133Must be called after a fork in the child, before entering or continuing
134the event loop. An alternative is to use C<EV::FLAG_FORKCHECK> which calls
135this fucntion automatically, at some performance loss (refer to the libev
136documentation).
137
138=item $loop = EV::default_loop [$flags]
139
140Return the default loop (which is a singleton object).
141
142=back
143
144
145=head1 BASIC INTERFACE
146
147=over 4
148
64=item $EV::NPRI 149=item $EV::DIED
65 150
66How many priority levels are available. 151Must contain a reference to a function that is called when a callback
152throws an exception (with $@ containing the error). The default prints an
153informative message and continues.
154
155If this callback throws an exception it will be silently ignored.
156
157=item $flags = EV::supported_backends
158
159=item $flags = EV::recommended_backends
160
161=item $flags = EV::embeddable_backends
162
163Returns the set (see C<EV::BACKEND_*> flags) of backends supported by this
164instance of EV, the set of recommended backends (supposed to be good) for
165this platform and the set of embeddable backends (see EMBED WATCHERS).
166
167=item EV::sleep $seconds
168
169Block the process for the given number of (fractional) seconds.
170
171=item $time = EV::time
172
173Returns the current time in (fractional) seconds since the epoch.
67 174
68=item $time = EV::now 175=item $time = EV::now
69 176
70Returns the time in (fractional) seconds since the epoch. 177=item $time = $loop->now
71 178
72=item $version = EV::version 179Returns the time the last event loop iteration has been started. This
180is the time that (relative) timers are based on, and refering to it is
181usually faster then calling EV::time.
73 182
74=item $method = EV::method 183=item $backend = EV::backend
75 184
76Return version string and event polling method used. 185=item $backend = $loop->backend
77 186
78=item EV::loop $flags # EV::LOOP_ONCE, EV::LOOP_ONESHOT 187Returns an integer describing the backend used by libev (EV::METHOD_SELECT
188or EV::METHOD_EPOLL).
79 189
80=item EV::loopexit $after 190=item EV::loop [$flags]
81 191
82Exit any active loop or dispatch after C<$after> seconds or immediately if 192=item $loop->loop ([$flags])
83C<$after> is missing or zero.
84 193
85=item EV::dispatch 194Begin checking for events and calling callbacks. It returns when a
195callback calls EV::unloop.
86 196
87Same as C<EV::loop 0>. 197The $flags argument can be one of the following:
88 198
89=item EV::event $callback 199 0 as above
200 EV::LOOP_ONESHOT block at most once (wait, but do not loop)
201 EV::LOOP_NONBLOCK do not block at all (fetch/handle events but do not wait)
90 202
91Creates a new event watcher waiting for nothing, calling the given callback. 203=item EV::unloop [$how]
92 204
205=item $loop->unloop ([$how])
206
207When called with no arguments or an argument of EV::UNLOOP_ONE, makes the
208innermost call to EV::loop return.
209
210When called with an argument of EV::UNLOOP_ALL, all calls to EV::loop will return as
211fast as possible.
212
213=item $count = EV::loop_count
214
215=item $count = $loop->loop_count
216
217Return the number of times the event loop has polled for new
218events. Sometiems useful as a generation counter.
219
220=item EV::once $fh_or_undef, $events, $timeout, $cb->($revents)
221
222=item $loop->once ($fh_or_undef, $events, $timeout, $cb->($revents))
223
224This function rolls together an I/O and a timer watcher for a single
225one-shot event without the need for managing a watcher object.
226
227If C<$fh_or_undef> is a filehandle or file descriptor, then C<$events>
228must be a bitset containing either C<EV::READ>, C<EV::WRITE> or C<EV::READ
229| EV::WRITE>, indicating the type of I/O event you want to wait for. If
230you do not want to wait for some I/O event, specify C<undef> for
231C<$fh_or_undef> and C<0> for C<$events>).
232
233If timeout is C<undef> or negative, then there will be no
234timeout. Otherwise a EV::timer with this value will be started.
235
236When an error occurs or either the timeout or I/O watcher triggers, then
237the callback will be called with the received event set (in general
238you can expect it to be a combination of C<EV::ERROR>, C<EV::READ>,
239C<EV::WRITE> and C<EV::TIMEOUT>).
240
241EV::once doesn't return anything: the watchers stay active till either
242of them triggers, then they will be stopped and freed, and the callback
243invoked.
244
245=item EV::feed_fd_event ($fd, $revents)
246
247=item $loop->feed_fd_event ($fd, $revents)
248
249Feed an event on a file descriptor into EV. EV will react to this call as
250if the readyness notifications specified by C<$revents> (a combination of
251C<EV::READ> and C<EV::WRITE>) happened on the file descriptor C<$fd>.
252
253=item EV::feed_signal_event ($signal)
254
255Feed a signal event into EV. EV will react to this call as if the signal
256specified by C<$signal> had occured.
257
258=item EV::set_io_collect_interval $time
259
260=item $loop->set_io_collect_interval ($time)
261
262=item EV::set_timeout_collect_interval $time
263
264=item $loop->set_timeout_collect_interval ($time)
265
266These advanced functions set the minimum block interval when polling for I/O events and the minimum
267wait interval for timer events. See the libev documentation at
268L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#FUNCTIONS_CONTROLLING_THE_EVENT_LOOP> for
269a more detailed discussion.
270
271=back
272
273
274=head1 WATCHER OBJECTS
275
276A watcher is an object that gets created to record your interest in some
277event. For instance, if you want to wait for STDIN to become readable, you
278would create an EV::io watcher for that:
279
280 my $watcher = EV::io *STDIN, EV::READ, sub {
281 my ($watcher, $revents) = @_;
282 warn "yeah, STDIN should now be readable without blocking!\n"
283 };
284
285All watchers can be active (waiting for events) or inactive (paused). Only
286active watchers will have their callbacks invoked. All callbacks will be
287called with at least two arguments: the watcher and a bitmask of received
288events.
289
290Each watcher type has its associated bit in revents, so you can use the
291same callback for multiple watchers. The event mask is named after the
292type, i..e. EV::child sets EV::CHILD, EV::prepare sets EV::PREPARE,
293EV::periodic sets EV::PERIODIC and so on, with the exception of I/O events
294(which can set both EV::READ and EV::WRITE bits), and EV::timer (which
295uses EV::TIMEOUT).
296
297In the rare case where one wants to create a watcher but not start it at
298the same time, each constructor has a variant with a trailing C<_ns> in
299its name, e.g. EV::io has a non-starting variant EV::io_ns and so on.
300
301Please note that a watcher will automatically be stopped when the watcher
302object is destroyed, so you I<need> to keep the watcher objects returned by
303the constructors.
304
305Also, all methods changing some aspect of a watcher (->set, ->priority,
306->fh and so on) automatically stop and start it again if it is active,
307which means pending events get lost.
308
309=head2 COMMON WATCHER METHODS
310
311This section lists methods common to all watchers.
312
313=over 4
314
315=item $w->start
316
317Starts a watcher if it isn't active already. Does nothing to an already
318active watcher. By default, all watchers start out in the active state
319(see the description of the C<_ns> variants if you need stopped watchers).
320
321=item $w->stop
322
323Stop a watcher if it is active. Also clear any pending events (events that
324have been received but that didn't yet result in a callback invocation),
325regardless of whether the watcher was active or not.
326
327=item $bool = $w->is_active
328
329Returns true if the watcher is active, false otherwise.
330
331=item $current_data = $w->data
332
333=item $old_data = $w->data ($new_data)
334
335Queries a freely usable data scalar on the watcher and optionally changes
336it. This is a way to associate custom data with a watcher:
337
338 my $w = EV::timer 60, 0, sub {
339 warn $_[0]->data;
340 };
341 $w->data ("print me!");
342
343=item $current_cb = $w->cb
344
345=item $old_cb = $w->cb ($new_cb)
346
347Queries the callback on the watcher and optionally changes it. You can do
348this at any time without the watcher restarting.
349
350=item $current_priority = $w->priority
351
352=item $old_priority = $w->priority ($new_priority)
353
354Queries the priority on the watcher and optionally changes it. Pending
355watchers with higher priority will be invoked first. The valid range of
356priorities lies between EV::MAXPRI (default 2) and EV::MINPRI (default
357-2). If the priority is outside this range it will automatically be
358normalised to the nearest valid priority.
359
360The default priority of any newly-created watcher is 0.
361
362Note that the priority semantics have not yet been fleshed out and are
363subject to almost certain change.
364
365=item $w->invoke ($revents)
366
367Call the callback *now* with the given event mask.
368
369=item $w->feed_event ($revents)
370
371Feed some events on this watcher into EV. EV will react to this call as if
372the watcher had received the given C<$revents> mask.
373
374=item $revents = $w->clear_pending
375
376If the watcher is pending, this function clears its pending status and
377returns its C<$revents> bitset (as if its callback was invoked). If the
378watcher isn't pending it does nothing and returns C<0>.
379
380=item $previous_state = $w->keepalive ($bool)
381
382Normally, C<EV::loop> will return when there are no active watchers
383(which is a "deadlock" because no progress can be made anymore). This is
384convinient because it allows you to start your watchers (and your jobs),
385call C<EV::loop> once and when it returns you know that all your jobs are
386finished (or they forgot to register some watchers for their task :).
387
388Sometimes, however, this gets in your way, for example when the module
389that calls C<EV::loop> (usually the main program) is not the same module
390as a long-living watcher (for example a DNS client module written by
391somebody else even). Then you might want any outstanding requests to be
392handled, but you would not want to keep C<EV::loop> from returning just
393because you happen to have this long-running UDP port watcher.
394
395In this case you can clear the keepalive status, which means that even
396though your watcher is active, it won't keep C<EV::loop> from returning.
397
398The initial value for keepalive is true (enabled), and you cna change it
399any time.
400
401Example: Register an I/O watcher for some UDP socket but do not keep the
402event loop from running just because of that watcher.
403
404 my $udp_socket = ...
405 my $udp_watcher = EV::io $udp_socket, EV::READ, sub { ... };
406 $1000udp_watcher->keepalive (0);
407
408=item $loop = $w->loop
409
410Return the loop that this watcher is attached to.
411
412=back
413
414
415=head1 WATCHER TYPES
416
417Each of the following subsections describes a single watcher type.
418
419=head3 I/O WATCHERS - is this file descriptor readable or writable?
420
421=over 4
422
93=item my $w = EV::io $fileno_or_fh, $eventmask, $callback 423=item $w = EV::io $fileno_or_fh, $eventmask, $callback
94 424
95=item my $w = EV::io_ns $fileno_or_fh, $eventmask, $callback 425=item $w = EV::io_ns $fileno_or_fh, $eventmask, $callback
426
427=item $w = $loop->io ($fileno_or_fh, $eventmask, $callback)
428
429=item $w = $loop->io_ns ($fileno_or_fh, $eventmask, $callback)
96 430
97As long as the returned watcher object is alive, call the C<$callback> 431As long as the returned watcher object is alive, call the C<$callback>
98when the events specified in C<$eventmask> happen. Initially, the timeout 432when at least one of events specified in C<$eventmask> occurs.
99is disabled.
100 433
101The C<io_ns> variant doesn't add/start the newly created watcher.
102
103Eventmask can be one or more of these constants ORed together: 434The $eventmask can be one or more of these constants ORed together:
104 435
105 EV::READ wait until read() wouldn't block anymore 436 EV::READ wait until read() wouldn't block anymore
106 EV::WRITE wait until write() wouldn't block anymore 437 EV::WRITE wait until write() wouldn't block anymore
107 EV::PERSIST stay active after an event occured
108 438
109=item my $w = EV::timer $after, $repeat, $callback
110
111=item my $w = EV::timer_ns $after, $repeat, $callback
112
113Calls the callback after C<$after> seconds. If C<$repeat> is true, the
114timer will be restarted after the callback returns. This means that the
115callback would be called roughly every C<$after> seconds, prolonged by the
116time the callback takes.
117
118The C<timer_ns> variant doesn't add/start the newly created watcher. 439The C<io_ns> variant doesn't start (activate) the newly created watcher.
119 440
120=item my $w = EV::timer_abs $at, $interval, $callback 441=item $w->set ($fileno_or_fh, $eventmask)
121 442
122=item my $w = EV::timer_abs_ns $at, $interval, $callback 443Reconfigures the watcher, see the constructor above for details. Can be
123 444called at any time.
124Similar to EV::timer, but the time is given as an absolute point in time
125(C<$at>), plus an optional C<$interval>.
126
127If the C<$interval> is zero, then the callback will be called at the time
128C<$at> if that is in the future, or as soon as possible if its in the
129past. It will not automatically repeat.
130
131If the C<$interval> is nonzero, then the watcher will always be scheduled
132to time out at the next C<$at + integer * $interval> time.
133
134This can be used to schedule a callback to run at very regular intervals,
135as long as the processing time is less then the interval (otherwise
136obviously events will be skipped).
137
138The C<timer_abs_ns> variant doesn't add/start the newly created watcher.
139
140=item my $w = EV::signal $signum, $callback
141
142=item my $w = EV::signal_ns $signum, $callback
143
144Call the callback when signal $signum is received.
145
146The C<signal_ns> variant doesn't add/start the newly created watcher.
147
148=back
149
150=head1 THE EV::Event CLASS
151
152All EV functions creating an event watcher (designated by C<my $w =>
153above) support the following methods on the returned watcher object:
154
155=over 4
156
157=item $w->add ($timeout)
158
159Stops and (re-)starts the event watcher, setting the optional timeout to
160the given value, or clearing the timeout if none is given.
161
162=item $w->start
163
164Stops and (re-)starts the event watcher without touching the timeout.
165
166=item $w->del
167
168=item $w->stop
169
170Stop the event watcher if it was started.
171
172=item $current_callback = $w->cb
173
174=item $old_callback = $w->cb ($new_callback)
175
176Return the previously set callback and optionally set a new one.
177 445
178=item $current_fh = $w->fh 446=item $current_fh = $w->fh
179 447
180=item $old_fh = $w->fh ($new_fh) 448=item $old_fh = $w->fh ($new_fh)
181 449
185 453
186=item $old_eventmask = $w->events ($new_eventmask) 454=item $old_eventmask = $w->events ($new_eventmask)
187 455
188Returns the previously set event mask and optionally set a new one. 456Returns the previously set event mask and optionally set a new one.
189 457
458=back
459
460
461=head3 TIMER WATCHERS - relative and optionally repeating timeouts
462
463=over 4
464
465=item $w = EV::timer $after, $repeat, $callback
466
467=item $w = EV::timer_ns $after, $repeat, $callback
468
469=item $w = $loop->timer ($after, $repeat, $callback)
470
471=item $w = $loop->timer_ns ($after, $repeat, $callback)
472
473Calls the callback after C<$after> seconds (which may be fractional). If
474C<$repeat> is non-zero, the timer will be restarted (with the $repeat
475value as $after) after the callback returns.
476
477This means that the callback would be called roughly after C<$after>
478seconds, and then every C<$repeat> seconds. The timer does his best not
479to drift, but it will not invoke the timer more often then once per event
480loop iteration, and might drift in other cases. If that isn't acceptable,
481look at EV::periodic, which can provide long-term stable timers.
482
483The timer is based on a monotonic clock, that is, if somebody is sitting
484in front of the machine while the timer is running and changes the system
485clock, the timer will nevertheless run (roughly) the same time.
486
487The C<timer_ns> variant doesn't start (activate) the newly created watcher.
488
190=item $w->timeout ($after, $repeat) 489=item $w->set ($after, $repeat)
191 490
192Resets the timeout (see C<EV::timer> for details). 491Reconfigures the watcher, see the constructor above for details. Can be called at
492any time.
193 493
494=item $w->again
495
496Similar to the C<start> method, but has special semantics for repeating timers:
497
498If the timer is active and non-repeating, it will be stopped.
499
500If the timer is active and repeating, reset the timeout to occur
501C<$repeat> seconds after now.
502
503If the timer is inactive and repeating, start it using the repeat value.
504
505Otherwise do nothing.
506
507This behaviour is useful when you have a timeout for some IO
508operation. You create a timer object with the same value for C<$after> and
509C<$repeat>, and then, in the read/write watcher, run the C<again> method
510on the timeout.
511
512=back
513
514
515=head3 PERIODIC WATCHERS - to cron or not to cron?
516
517=over 4
518
519=item $w = EV::periodic $at, $interval, $reschedule_cb, $callback
520
521=item $w = EV::periodic_ns $at, $interval, $reschedule_cb, $callback
522
523=item $w = $loop->periodic ($at, $interval, $reschedule_cb, $callback)
524
525=item $w = $loop->periodic_ns ($at, $interval, $reschedule_cb, $callback)
526
527Similar to EV::timer, but is not based on relative timeouts but on
528absolute times. Apart from creating "simple" timers that trigger "at" the
529specified time, it can also be used for non-drifting absolute timers and
530more complex, cron-like, setups that are not adversely affected by time
531jumps (i.e. when the system clock is changed by explicit date -s or other
532means such as ntpd). It is also the most complex watcher type in EV.
533
534It has three distinct "modes":
535
536=over 4
537
538=item * absolute timer ($interval = $reschedule_cb = 0)
539
540This time simply fires at the wallclock time C<$at> and doesn't repeat. It
541will not adjust when a time jump occurs, that is, if it is to be run
542at January 1st 2011 then it will run when the system time reaches or
543surpasses this time.
544
545=item * repeating interval timer ($interval > 0, $reschedule_cb = 0)
546
547In this mode the watcher will always be scheduled to time out at the
548next C<$at + N * $interval> time (for some integer N) and then repeat,
549regardless of any time jumps.
550
551This can be used to create timers that do not drift with respect to system
552time:
553
554 my $hourly = EV::periodic 0, 3600, 0, sub { print "once/hour\n" };
555
556That doesn't mean there will always be 3600 seconds in between triggers,
557but only that the the clalback will be called when the system time shows a
558full hour (UTC).
559
560Another way to think about it (for the mathematically inclined) is that
561EV::periodic will try to run the callback in this mode at the next
562possible time where C<$time = $at (mod $interval)>, regardless of any time
563jumps.
564
565=item * manual reschedule mode ($reschedule_cb = coderef)
566
567In this mode $interval and $at are both being ignored. Instead, each
568time the periodic watcher gets scheduled, the reschedule callback
569($reschedule_cb) will be called with the watcher as first, and the current
570time as second argument.
571
572I<This callback MUST NOT stop or destroy this or any other periodic
573watcher, ever>. If you need to stop it, return 1e30 and stop it
574afterwards.
575
576It must return the next time to trigger, based on the passed time value
577(that is, the lowest time value larger than to the second argument). It
578will usually be called just before the callback will be triggered, but
579might be called at other times, too.
580
581This can be used to create very complex timers, such as a timer that
582triggers on each midnight, local time (actually 24 hours after the last
583midnight, to keep the example simple. If you know a way to do it correctly
584in about the same space (without requiring elaborate modules), drop me a
585note :):
586
587 my $daily = EV::periodic 0, 0, sub {
588 my ($w, $now) = @_;
589
590 use Time::Local ();
591 my (undef, undef, undef, $d, $m, $y) = localtime $now;
592 86400 + Time::Local::timelocal 0, 0, 0, $d, $m, $y
593 }, sub {
594 print "it's midnight or likely shortly after, now\n";
595 };
596
597=back
598
599The C<periodic_ns> variant doesn't start (activate) the newly created watcher.
600
601=item $w->set ($at, $interval, $reschedule_cb)
602
603Reconfigures the watcher, see the constructor above for details. Can be called at
604any time.
605
606=item $w->again
607
608Simply stops and starts the watcher again.
609
610=item $time = $w->at
611
612Return the time that the watcher is expected to trigger next.
613
614=back
615
616
617=head3 SIGNAL WATCHERS - signal me when a signal gets signalled!
618
619=over 4
620
621=item $w = EV::signal $signal, $callback
622
623=item $w = EV::signal_ns $signal, $callback
624
625Call the callback when $signal is received (the signal can be specified by
626number or by name, just as with C<kill> or C<%SIG>).
627
628EV will grab the signal for the process (the kernel only allows one
629component to receive a signal at a time) when you start a signal watcher,
630and removes it again when you stop it. Perl does the same when you
631add/remove callbacks to C<%SIG>, so watch out.
632
633You can have as many signal watchers per signal as you want.
634
635The C<signal_ns> variant doesn't start (activate) the newly created watcher.
636
637=item $w->set ($signal)
638
639Reconfigures the watcher, see the constructor above for details. Can be
640called at any time.
641
642=item $current_signum = $w->signal
643
644=item $old_signum = $w->signal ($new_signal)
645
646Returns the previously set signal (always as a number not name) and
647optionally set a new one.
648
649=back
650
651
652=head3 CHILD WATCHERS - watch out for process status changes
653
654=over 4
655
656=item $w = EV::child $pid, $trace, $callback
657
658=item $w = EV::child_ns $pid, $trace, $callback
659
660=item $w = $loop->child ($pid, $trace, $callback)
661
662=item $w = $loop->child_ns ($pid, $trace, $callback)
663
664Call the callback when a status change for pid C<$pid> (or any pid
665if C<$pid> is 0) has been received (a status change happens when the
666process terminates or is killed, or, when trace is true, additionally when
667it is stopped or continued). More precisely: when the process receives
668a C<SIGCHLD>, EV will fetch the outstanding exit/wait status for all
669changed/zombie children and call the callback.
670
671It is valid (and fully supported) to install a child watcher after a child
672has exited but before the event loop has started its next iteration (for
673example, first you C<fork>, then the new child process might exit, and
674only then do you install a child watcher in the parent for the new pid).
675
676You can access both exit (or tracing) status and pid by using the
677C<rstatus> and C<rpid> methods on the watcher object.
678
679You can have as many pid watchers per pid as you want, they will all be
680called.
681
682The C<child_ns> variant doesn't start (activate) the newly created watcher.
683
684=item $w->set ($pid, $trace)
685
686Reconfigures the watcher, see the constructor above for details. Can be called at
687any time.
688
689=item $current_pid = $w->pid
690
691Returns the previously set process id and optionally set a new one.
692
693=item $exit_status = $w->rstatus
694
695Return the exit/wait status (as returned by waitpid, see the waitpid entry
696in perlfunc).
697
698=item $pid = $w->rpid
699
700Return the pid of the awaited child (useful when you have installed a
701watcher for all pids).
702
703=back
704
705
706=head3 STAT WATCHERS - did the file attributes just change?
707
708=over 4
709
710=item $w = EV::stat $path, $interval, $callback
711
712=item $w = EV::stat_ns $path, $interval, $callback
713
714=item $w = $loop->stat ($path, $interval, $callback)
715
716=item $w = $loop->stat_ns ($path, $interval, $callback)
717
718Call the callback when a file status change has been detected on
719C<$path>. The C<$path> does not need to exist, changing from "path exists"
720to "path does not exist" is a status change like any other.
721
722The C<$interval> is a recommended polling interval for systems where
723OS-supported change notifications don't exist or are not supported. If
724you use C<0> then an unspecified default is used (which is highly
725recommended!), which is to be expected to be around five seconds usually.
726
727This watcher type is not meant for massive numbers of stat watchers,
728as even with OS-supported change notifications, this can be
729resource-intensive.
730
731The C<stat_ns> variant doesn't start (activate) the newly created watcher.
732
733=item ... = $w->stat
734
735This call is very similar to the perl C<stat> built-in: It stats (using
736C<lstat>) the path specified in the watcher and sets perls stat cache (as
737well as EV's idea of the current stat values) to the values found.
738
739In scalar context, a boolean is return indicating success or failure of
740the stat. In list context, the same 13-value list as with stat is returned
741(except that the blksize and blocks fields are not reliable).
742
743In the case of an error, errno is set to C<ENOENT> (regardless of the
744actual error value) and the C<nlink> value is forced to zero (if the stat
745was successful then nlink is guaranteed to be non-zero).
746
747See also the next two entries for more info.
748
749=item ... = $w->attr
750
751Just like C<< $w->stat >>, but without the initial stat'ing: this returns
752the values most recently detected by EV. See the next entry for more info.
753
754=item ... = $w->prev
755
756Just like C<< $w->stat >>, but without the initial stat'ing: this returns
757the previous set of values, before the change.
758
759That is, when the watcher callback is invoked, C<< $w->prev >> will be set
760to the values found I<before> a change was detected, while C<< $w->attr >>
761returns the values found leading to the change detection. The difference (if any)
762between C<prev> and C<attr> is what triggered the callback.
763
764If you did something to the filesystem object and do not want to trigger
765yet another change, you can call C<stat> to update EV's idea of what the
766current attributes are.
767
194=item $w->timeout_abs ($at, $interval) 768=item $w->set ($path, $interval)
195 769
196Resets the timeout (see C<EV::timer_abs> for details). 770Reconfigures the watcher, see the constructor above for details. Can be
771called at any time.
197 772
198=item $w->priority_set ($priority) 773=item $current_path = $w->path
199 774
200Set the priority of the watcher to C<$priority> (0 <= $priority < $EV::NPRI). 775=item $old_path = $w->path ($new_path)
201 776
202=back 777Returns the previously set path and optionally set a new one.
203 778
204=head1 BUGS 779=item $current_interval = $w->interval
205 780
206Lots. Libevent itself isn't well tested and rather buggy, and this module 781=item $old_interval = $w->interval ($new_interval)
207is quite new at the moment. 782
783Returns the previously set interval and optionally set a new one. Can be
784used to query the actual interval used.
785
786=back
787
788
789=head3 IDLE WATCHERS - when you've got nothing better to do...
790
791=over 4
792
793=item $w = EV::idle $callback
794
795=item $w = EV::idle_ns $callback
796
797=item $w = $loop->idle ($callback)
798
799=item $w = $loop->idle_ns ($callback)
800
801Call the callback when there are no other pending watchers of the same or
802higher priority (excluding check, prepare and other idle watchers of the
803same or lower priority, of course). They are called idle watchers because
804when the watcher is the highest priority pending event in the process, the
805process is considered to be idle at that priority.
806
807If you want a watcher that is only ever called when I<no> other events are
808outstanding you have to set the priority to C<EV::MINPRI>.
809
810The process will not block as long as any idle watchers are active, and
811they will be called repeatedly until stopped.
812
813For example, if you have idle watchers at priority C<0> and C<1>, and
814an I/O watcher at priority C<0>, then the idle watcher at priority C<1>
815and the I/O watcher will always run when ready. Only when the idle watcher
816at priority C<1> is stopped and the I/O watcher at priority C<0> is not
817pending with the C<0>-priority idle watcher be invoked.
818
819The C<idle_ns> variant doesn't start (activate) the newly created watcher.
820
821=back
822
823
824=head3 PREPARE WATCHERS - customise your event loop!
825
826=over 4
827
828=item $w = EV::prepare $callback
829
830=item $w = EV::prepare_ns $callback
831
832=item $w = $loop->prepare ($callback)
833
834=item $w = $loop->prepare_ns ($callback)
835
836Call the callback just before the process would block. You can still
837create/modify any watchers at this point.
838
839See the EV::check watcher, below, for explanations and an example.
840
841The C<prepare_ns> variant doesn't start (activate) the newly created watcher.
842
843=back
844
845
846=head3 CHECK WATCHERS - customise your event loop even more!
847
848=over 4
849
850=item $w = EV::check $callback
851
852=item $w = EV::check_ns $callback
853
854=item $w = $loop->check ($callback)
855
856=item $w = $loop->check_ns ($callback)
857
858Call the callback just after the process wakes up again (after it has
859gathered events), but before any other callbacks have been invoked.
860
861This is used to integrate other event-based software into the EV
862mainloop: You register a prepare callback and in there, you create io and
863timer watchers as required by the other software. Here is a real-world
864example of integrating Net::SNMP (with some details left out):
865
866 our @snmp_watcher;
867
868 our $snmp_prepare = EV::prepare sub {
869 # do nothing unless active
870 $dispatcher->{_event_queue_h}
871 or return;
872
873 # make the dispatcher handle any outstanding stuff
874 ... not shown
875
876 # create an I/O watcher for each and every socket
877 @snmp_watcher = (
878 (map { EV::io $_, EV::READ, sub { } }
879 keys %{ $dispatcher->{_descriptors} }),
880
881 EV::timer +($event->[Net::SNMP::Dispatcher::_ACTIVE]
882 ? $event->[Net::SNMP::Dispatcher::_TIME] - EV::now : 0),
883 0, sub { },
884 );
885 };
886
887The callbacks are irrelevant (and are not even being called), the
888only purpose of those watchers is to wake up the process as soon as
889one of those events occurs (socket readable, or timer timed out). The
890corresponding EV::check watcher will then clean up:
891
892 our $snmp_check = EV::check sub {
893 # destroy all watchers
894 @snmp_watcher = ();
895
896 # make the dispatcher handle any new stuff
897 ... not shown
898 };
899
900The callbacks of the created watchers will not be called as the watchers
901are destroyed before this cna happen (remember EV::check gets called
902first).
903
904The C<check_ns> variant doesn't start (activate) the newly created watcher.
905
906=back
907
908
909=head3 FORK WATCHERS - the audacity to resume the event loop after a fork
910
911Fork watchers are called when a C<fork ()> was detected. The invocation
912is done before the event loop blocks next and before C<check> watchers
913are being called, and only in the child after the fork.
914
915=over 4
916
917=item $w = EV::fork $callback
918
919=item $w = EV::fork_ns $callback
920
921=item $w = $loop->fork ($callback)
922
923=item $w = $loop->fork_ns ($callback)
924
925Call the callback before the event loop is resumed in the child process
926after a fork.
927
928The C<fork_ns> variant doesn't start (activate) the newly created watcher.
929
930=back
931
932
933=head3 EMBED WATCHERS - when one backend isn't enough...
934
935This is a rather advanced watcher type that lets you embed one event loop
936into another (currently only IO events are supported in the embedded
937loop, other types of watchers might be handled in a delayed or incorrect
938fashion and must not be used).
939
940See the libev documentation at
941L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#code_ev_embed_code_when_one_backend_>
942for more details.
943
944In short, this watcher is most useful on BSD systems without working
945kqueue to still be able to handle a large number of sockets:
946
947 my $socket_loop;
948
949 # check wether we use SELECT or POLL _and_ KQUEUE is supported
950 if (
951 (EV::backend & (EV::BACKEND_POLL | EV::BACKEND_SELECT))
952 && (EV::supported_backends & EV::embeddable_backends & EV::BACKEND_KQUEUE)
953 ) {
954 # use kqueue for sockets
955 $socket_loop = new EV::Loop EV::BACKEND_KQUEUE | EV::FLAG_NOENV;
956 }
957
958 # use the default loop otherwise
959 $socket_loop ||= EV::default_loop;
960
961=over 4
962
963=item $w = EV::embed $otherloop, $callback
964
965=item $w = EV::embed_ns $otherloop, $callback
966
967=item $w = $loop->embed ($otherloop, $callback)
968
969=item $w = $loop->embed_ns ($otherloop, $callback)
970
971Call the callback when the embedded event loop (C<$otherloop>) has any
972I/O activity. The C<$callback> should alwas be specified as C<undef> in
973this version of EV, which means the embedded event loop will be managed
974automatically.
975
976The C<embed_ns> variant doesn't start (activate) the newly created watcher.
977
978=back
979
980=head3 ASYNC WATCHERS - how to wake up another event loop
981
982Async watchers are provided by EV, but have little use in perl directly, as perl
983neither supports threads nor direct access to signal handlers or other
984contexts where they could be of value.
985
986It is, however, possible to use them from the XS level.
987
988Please see the libev documentation for further details.
989
990=over 4
991
992=item $w = EV::async $callback
993
994=item $w = EV::async_ns $callback
995
996=item $w->send
997
998=item $bool = $w->async_pending
999
1000=back
1001
1002
1003=head1 PERL SIGNALS
1004
1005While Perl signal handling (C<%SIG>) is not affected by EV, the behaviour
1006with EV is as the same as any other C library: Perl-signals will only be
1007handled when Perl runs, which means your signal handler might be invoked
1008only the next time an event callback is invoked.
1009
1010The solution is to use EV signal watchers (see C<EV::signal>), which will
1011ensure proper operations with regards to other event watchers.
1012
1013If you cannot do this for whatever reason, you can also force a watcher
1014to be called on every event loop iteration by installing a C<EV::check>
1015watcher:
1016
1017 my $async_check = EV::check sub { };
1018
1019This ensures that perl gets into control for a short time to handle any
1020pending signals, and also ensures (slightly) slower overall operation.
1021
1022=head1 THREADS
1023
1024Threads are not supported by this module in any way. Perl pseudo-threads
1025is evil stuff and must die. As soon as Perl gains real threads I will work
1026on thread support for it.
1027
1028=head1 FORK
1029
1030Most of the "improved" event delivering mechanisms of modern operating
1031systems have quite a few problems with fork(2) (to put it bluntly: it is
1032not supported and usually destructive). Libev makes it possible to work
1033around this by having a function that recreates the kernel state after
1034fork in the child.
1035
1036On non-win32 platforms, this module requires the pthread_atfork
1037functionality to do this automatically for you. This function is quite
1038buggy on most BSDs, though, so YMMV. The overhead for this is quite
1039negligible, because everything the function currently does is set a flag
1040that is checked only when the event loop gets used the next time, so when
1041you do fork but not use EV, the overhead is minimal.
1042
1043On win32, there is no notion of fork so all this doesn't apply, of course.
208 1044
209=cut 1045=cut
210 1046
211our $NPRI = 4; 1047our $DIED = sub {
212our $BASE = init; 1048 warn "EV: error in callback (ignoring): $@";
213priority_init $NPRI; 1049};
1050
1051default_loop
1052 or die 'EV: cannot initialise libev backend. bad $ENV{LIBEV_FLAGS}?';
214 1053
2151; 10541;
216 1055
217=head1 SEE ALSO 1056=head1 SEE ALSO
218 1057
219 L<EV::DNS>, L<event(3)>, L<event.h>, L<evdns.h>. 1058L<EV::ADNS> (asynchronous DNS), L<Glib::EV> (makes Glib/Gtk2 use EV as
1059event loop), L<EV::Glib> (embed Glib into EV), L<Coro::EV> (efficient
1060coroutines with EV), L<Net::SNMP::EV> (asynchronous SNMP), L<AnyEvent> for
1061event-loop agnostic and portable event driven programming.
220 1062
221=head1 AUTHOR 1063=head1 AUTHOR
222 1064
223 Marc Lehmann <schmorp@schmorp.de> 1065 Marc Lehmann <schmorp@schmorp.de>
224 http://home.schmorp.de/ 1066 http://home.schmorp.de/

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines