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

Comparing EV/EV.pm (file contents):
Revision 1.19 by root, Thu Nov 1 11:43:10 2007 UTC vs.
Revision 1.78 by root, Sat Dec 22 11:50:04 2007 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 my $w = EV::timer_abs 0, 60, sub { 21 my $w = EV::periodic 0, 60, 0, sub {
22 warn "is called every minute, on the minute, exactly"; 22 warn "is called every minute, on the minute, exactly";
23 }; 23 };
24 24
25 # IO 25 # IO
26 26
27 my $w = EV::io \*STDIN, EV::READ | EV::PERSIST, sub { 27 my $w = EV::io *STDIN, EV::READ, sub {
28 my ($w, $revents) = @_; # all callbacks get the watcher object and event mask 28 my ($w, $revents) = @_; # all callbacks receive the watcher and event mask
29 if ($revents & EV::TIMEOUT) {
30 warn "nothing 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 };
35 $w->timeout (10);
36
37 my $w = EV::timed_io \*STDIN, EV::READ, 30, sub {
38 my ($w, $revents) = @_;
39 if ($revents & EV::TIMEOUT) {
40 warn "nothing entered within 30 seconds, bye bye.\n";
41 $w->stop;
42 } else {
43 my $line = <STDIN>;
44 warn "you entered something, you again have 30 seconds.\n";
45 }
46 }; 30 };
47 31
48 # SIGNALS 32 # SIGNALS
49 33
50 my $w = EV::signal 'QUIT', sub { 34 my $w = EV::signal 'QUIT', sub {
51 warn "sigquit received\n"; 35 warn "sigquit received\n";
52 }; 36 };
53 37
38 # CHILD/PID STATUS CHANGES
39
54 my $w = EV::signal 3, sub { 40 my $w = EV::child 666, sub {
55 warn "sigquit received (this is GNU/Linux, right?)\n"; 41 my ($w, $revents) = @_;
42 my $status = $w->rstatus;
56 }; 43 };
57 44
58 # CHILD/PID STATUS CHANGES 45 # STAT CHANGES
59 46 my $w = EV::stat "/etc/passwd", 10, sub {
60 my $w = EV::child 666, sub {
61 my ($w, $revents, $status) = @_; 47 my ($w, $revents) = @_;
48 warn $w->path, " has changed somehow.\n";
62 }; 49 };
63 50
64 # MAINLOOP 51 # MAINLOOP
65 EV::dispatch; # loop as long as watchers are active 52 EV::loop; # loop until EV::unloop is called or all watchers stop
66 EV::loop; # the same thing
67 EV::loop EV::LOOP_ONESHOT; # block until some events could be handles 53 EV::loop EV::LOOP_ONESHOT; # block until at least one event could be handled
68 EV::loop EV::LOOP_NONBLOCK; # check and handle some events, but do not wait 54 EV::loop EV::LOOP_NONBLOCK; # try to handle same events, but do not block
69 55
70=head1 DESCRIPTION 56=head1 DESCRIPTION
71 57
72This module provides an interface to libev 58This module provides an interface to libev
73(L<http://software.schmorp.de/pkg/libev.html>). You probably should 59(L<http://software.schmorp.de/pkg/libev.html>). While the documentation
74acquaint yourself with its documentation and source code to be able to use 60below is comprehensive, one might also consult the documentation of libev
75this module fully. 61itself (L<http://cvs.schmorp.de/libev/ev.html>) for more subtle details on
62watcher semantics or some discussion on the available backends, or how to
63force a specific backend with C<LIBEV_FLAGS>, or just about in any case
64because it has much more detailed information.
76 65
77=cut 66=cut
78 67
79package EV; 68package EV;
80 69
81use strict; 70use strict;
82 71
83BEGIN { 72BEGIN {
84 our $VERSION = '0.1'; 73 our $VERSION = '2.0';
85 use XSLoader; 74 use XSLoader;
86 XSLoader::load "EV", $VERSION; 75 XSLoader::load "EV", $VERSION;
87} 76}
88 77
89@EV::Io::ISA = 78@EV::IO::ISA =
90@EV::Timer::ISA = 79@EV::Timer::ISA =
91@EV::Periodic::ISA = 80@EV::Periodic::ISA =
92@EV::Signal::ISA = 81@EV::Signal::ISA =
82@EV::Child::ISA =
83@EV::Stat::ISA =
93@EV::Idle::ISA = 84@EV::Idle::ISA =
94@EV::Prepare::ISA = 85@EV::Prepare::ISA =
95@EV::Check::ISA = 86@EV::Check::ISA =
96@EV::Child::ISA = "EV::Watcher"; 87@EV::Embed::ISA =
88@EV::Fork::ISA =
89 "EV::Watcher";
90
91@EV::Loop::Default::ISA = "EV::Loop";
92
93=head1 EVENT LOOPS
94
95EV supports multiple event loops: There is a single "default event loop"
96that can handle everything including signals and child watchers, and any
97number of "dynamic event loops" that can use different backends (with
98various limitations), but no child and signal watchers.
99
100You do not have to do anything to create the default event loop: When
101the module is loaded a suitable backend is selected on the premise of
102selecting a working backend (which for example rules out kqueue on most
103BSDs). Modules should, unless they have "special needs" always use the
104default loop as this is fastest (perl-wise), best supported by other
105modules (e.g. AnyEvent or Coro) and most portable event loop.
106
107For specific programs you cna create additional event loops dynamically.
108
109=over 4
110
111=item $loop = new EV::loop [$flags]
112
113Create a new event loop as per the specified flags. Please refer to the
114C<ev_loop_new ()> function description in the libev documentation
115(L<http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#GLOBAL_FUNCTIONS>)
116for more info.
117
118The loop will automatically be destroyed when it is no longer referenced
119by any watcher and the loop object goes out of scope.
120
121Using C<EV::FLAG_FORKCHECK> is recommended, as only the default event loop
122is protected by this module.
123
124=item $loop->loop_fork
125
126Must be called after a fork in the child, before entering or continuing
127the event loop. An alternative is to use C<EV::FLAG_FORKCHECK> which calls
128this fucntion automatically, at some performance loss (refer to the libev
129documentation).
130
131=back
132
97 133
98=head1 BASIC INTERFACE 134=head1 BASIC INTERFACE
99 135
100=over 4 136=over 4
101 137
102=item $EV::NPRI
103
104How many priority levels are available.
105
106=item $EV::DIED 138=item $EV::DIED
107 139
108Must contain a reference to a function that is called when a callback 140Must contain a reference to a function that is called when a callback
109throws an exception (with $@ containing thr error). The default prints an 141throws an exception (with $@ containing the error). The default prints an
110informative message and continues. 142informative message and continues.
111 143
112If this callback throws an exception it will be silently ignored. 144If this callback throws an exception it will be silently ignored.
113 145
146=item $time = EV::time
147
148Returns the current time in (fractional) seconds since the epoch.
149
114=item $time = EV::now 150=item $time = EV::now
115 151
116Returns the time in (fractional) seconds since the epoch. 152=item $time = $loop->now
117 153
118=item $version = EV::version 154Returns the time the last event loop iteration has been started. This
155is the time that (relative) timers are based on, and refering to it is
156usually faster then calling EV::time.
119 157
120=item $method = EV::method 158=item $backend = EV::backend
121 159
122Return version string and event polling method used. 160=item $backend = $loop->backend
123 161
124=item EV::loop $flags # EV::LOOP_ONCE, EV::LOOP_ONESHOT 162Returns an integer describing the backend used by libev (EV::METHOD_SELECT
163or EV::METHOD_EPOLL).
125 164
126=item EV::loopexit $after 165=item EV::loop [$flags]
127 166
128Exit any active loop or dispatch after C<$after> seconds or immediately if 167=item $loop->loop ([$flags])
129C<$after> is missing or zero.
130 168
131=item EV::dispatch 169Begin checking for events and calling callbacks. It returns when a
170callback calls EV::unloop.
132 171
133Same as C<EV::loop 0>. 172The $flags argument can be one of the following:
134 173
135=item EV::event $callback 174 0 as above
175 EV::LOOP_ONESHOT block at most once (wait, but do not loop)
176 EV::LOOP_NONBLOCK do not block at all (fetch/handle events but do not wait)
136 177
137Creates a new event watcher waiting for nothing, calling the given callback. 178=item EV::unloop [$how]
138 179
180=item $loop->unloop ([$how])
181
182When called with no arguments or an argument of EV::UNLOOP_ONE, makes the
183innermost call to EV::loop return.
184
185When called with an argument of EV::UNLOOP_ALL, all calls to EV::loop will return as
186fast as possible.
187
188=item $count = EV::loop_count
189
190=item $count = $loop->loop_count
191
192Return the number of times the event loop has polled for new
193events. Sometiems useful as a generation counter.
194
195=item EV::once $fh_or_undef, $events, $timeout, $cb->($revents)
196
197=item $loop->once ($fh_or_undef, $events, $timeout, $cb->($revents))
198
199This function rolls together an I/O and a timer watcher for a single
200one-shot event without the need for managing a watcher object.
201
202If C<$fh_or_undef> is a filehandle or file descriptor, then C<$events>
203must be a bitset containing either C<EV::READ>, C<EV::WRITE> or C<EV::READ
204| EV::WRITE>, indicating the type of I/O event you want to wait for. If
205you do not want to wait for some I/O event, specify C<undef> for
206C<$fh_or_undef> and C<0> for C<$events>).
207
208If timeout is C<undef> or negative, then there will be no
209timeout. Otherwise a EV::timer with this value will be started.
210
211When an error occurs or either the timeout or I/O watcher triggers, then
212the callback will be called with the received event set (in general
213you can expect it to be a combination of C<EV::ERROR>, C<EV::READ>,
214C<EV::WRITE> and C<EV::TIMEOUT>).
215
216EV::once doesn't return anything: the watchers stay active till either
217of them triggers, then they will be stopped and freed, and the callback
218invoked.
219
220=item EV::feed_fd_event ($fd, $revents)
221
222=item $loop->feed_fd_event ($fd, $revents)
223
224Feed an event on a file descriptor into EV. EV will react to this call as
225if the readyness notifications specified by C<$revents> (a combination of
226C<EV::READ> and C<EV::WRITE>) happened on the file descriptor C<$fd>.
227
228=item EV::feed_signal_event ($signal)
229
230Feed a signal event into EV. EV will react to this call as if the signal
231specified by C<$signal> had occured.
232
233=back
234
235
236=head1 WATCHER OBJECTS
237
238A watcher is an object that gets created to record your interest in some
239event. For instance, if you want to wait for STDIN to become readable, you
240would create an EV::io watcher for that:
241
242 my $watcher = EV::io *STDIN, EV::READ, sub {
243 my ($watcher, $revents) = @_;
244 warn "yeah, STDIN should now be readable without blocking!\n"
245 };
246
247All watchers can be active (waiting for events) or inactive (paused). Only
248active watchers will have their callbacks invoked. All callbacks will be
249called with at least two arguments: the watcher and a bitmask of received
250events.
251
252Each watcher type has its associated bit in revents, so you can use the
253same callback for multiple watchers. The event mask is named after the
254type, i..e. EV::child sets EV::CHILD, EV::prepare sets EV::PREPARE,
255EV::periodic sets EV::PERIODIC and so on, with the exception of I/O events
256(which can set both EV::READ and EV::WRITE bits), and EV::timer (which
257uses EV::TIMEOUT).
258
259In the rare case where one wants to create a watcher but not start it at
260the same time, each constructor has a variant with a trailing C<_ns> in
261its name, e.g. EV::io has a non-starting variant EV::io_ns and so on.
262
263Please note that a watcher will automatically be stopped when the watcher
264object is destroyed, so you I<need> to keep the watcher objects returned by
265the constructors.
266
267Also, all methods changing some aspect of a watcher (->set, ->priority,
268->fh and so on) automatically stop and start it again if it is active,
269which means pending events get lost.
270
271=head2 COMMON WATCHER METHODS
272
273This section lists methods common to all watchers.
274
275=over 4
276
277=item $w->start
278
279Starts a watcher if it isn't active already. Does nothing to an already
280active watcher. By default, all watchers start out in the active state
281(see the description of the C<_ns> variants if you need stopped watchers).
282
283=item $w->stop
284
285Stop a watcher if it is active. Also clear any pending events (events that
286have been received but that didn't yet result in a callback invocation),
287regardless of whether the watcher was active or not.
288
289=item $bool = $w->is_active
290
291Returns true if the watcher is active, false otherwise.
292
293=item $current_data = $w->data
294
295=item $old_data = $w->data ($new_data)
296
297Queries a freely usable data scalar on the watcher and optionally changes
298it. This is a way to associate custom data with a watcher:
299
300 my $w = EV::timer 60, 0, sub {
301 warn $_[0]->data;
302 };
303 $w->data ("print me!");
304
305=item $current_cb = $w->cb
306
307=item $old_cb = $w->cb ($new_cb)
308
309Queries the callback on the watcher and optionally changes it. You can do
310this at any time without the watcher restarting.
311
312=item $current_priority = $w->priority
313
314=item $old_priority = $w->priority ($new_priority)
315
316Queries the priority on the watcher and optionally changes it. Pending
317watchers with higher priority will be invoked first. The valid range of
318priorities lies between EV::MAXPRI (default 2) and EV::MINPRI (default
319-2). If the priority is outside this range it will automatically be
320normalised to the nearest valid priority.
321
322The default priority of any newly-created watcher is 0.
323
324Note that the priority semantics have not yet been fleshed out and are
325subject to almost certain change.
326
327=item $w->invoke ($revents)
328
329Call the callback *now* with the given event mask.
330
331=item $w->feed_event ($revents)
332
333Feed some events on this watcher into EV. EV will react to this call as if
334the watcher had received the given C<$revents> mask.
335
336=item $revents = $w->clear_pending
337
338If the watcher is pending, this function clears its pending status and
339returns its C<$revents> bitset (as if its callback was invoked). If the
340watcher isn't pending it does nothing and returns C<0>.
341
342=item $previous_state = $w->keepalive ($bool)
343
344Normally, C<EV::loop> will return when there are no active watchers
345(which is a "deadlock" because no progress can be made anymore). This is
346convinient because it allows you to start your watchers (and your jobs),
347call C<EV::loop> once and when it returns you know that all your jobs are
348finished (or they forgot to register some watchers for their task :).
349
350Sometimes, however, this gets in your way, for example when the module
351that calls C<EV::loop> (usually the main program) is not the same module
352as a long-living watcher (for example a DNS client module written by
353somebody else even). Then you might want any outstanding requests to be
354handled, but you would not want to keep C<EV::loop> from returning just
355because you happen to have this long-running UDP port watcher.
356
357In this case you can clear the keepalive status, which means that even
358though your watcher is active, it won't keep C<EV::loop> from returning.
359
360The initial value for keepalive is true (enabled), and you cna change it
361any time.
362
363Example: Register an I/O watcher for some UDP socket but do not keep the
364event loop from running just because of that watcher.
365
366 my $udp_socket = ...
367 my $udp_watcher = EV::io $udp_socket, EV::READ, sub { ... };
368 $1000udp_watcher->keepalive (0);
369
370=item $loop = $w->loop
371
372Return the loop that this watcher is attached to.
373
374=back
375
376
377=head1 WATCHER TYPES
378
379Each of the following subsections describes a single watcher type.
380
381=head3 I/O WATCHERS - is this file descriptor readable or writable?
382
383=over 4
384
139=item my $w = EV::io $fileno_or_fh, $eventmask, $callback 385=item $w = EV::io $fileno_or_fh, $eventmask, $callback
140 386
141=item my $w = EV::io_ns $fileno_or_fh, $eventmask, $callback 387=item $w = EV::io_ns $fileno_or_fh, $eventmask, $callback
388
389=item $w = $loop->io ($fileno_or_fh, $eventmask, $callback)
390
391=item $w = $loop->io_ns ($fileno_or_fh, $eventmask, $callback)
142 392
143As long as the returned watcher object is alive, call the C<$callback> 393As long as the returned watcher object is alive, call the C<$callback>
144when the events specified in C<$eventmask> happen. Initially, the timeout 394when at least one of events specified in C<$eventmask> occurs.
145is disabled.
146 395
147You can additionall set a timeout to occur on the watcher, but note that
148this timeout will not be reset when you get an I/O event in the EV::PERSIST
149case, and reaching a timeout will always stop the watcher even in the
150EV::PERSIST case.
151
152If you want a timeout to occur only after a specific time of inactivity, set
153a repeating timeout and do NOT use EV::PERSIST.
154
155Eventmask can be one or more of these constants ORed together: 396The $eventmask can be one or more of these constants ORed together:
156 397
157 EV::READ wait until read() wouldn't block anymore 398 EV::READ wait until read() wouldn't block anymore
158 EV::WRITE wait until write() wouldn't block anymore 399 EV::WRITE wait until write() wouldn't block anymore
159 EV::PERSIST stay active after a (non-timeout) event occured
160 400
161The C<io_ns> variant doesn't add/start the newly created watcher. 401The C<io_ns> variant doesn't start (activate) the newly created watcher.
162 402
163=item my $w = EV::timed_io $fileno_or_fh, $eventmask, $timeout, $callback 403=item $w->set ($fileno_or_fh, $eventmask)
164 404
165=item my $w = EV::timed_io_ns $fileno_or_fh, $eventmask, $timeout, $callback 405Reconfigures the watcher, see the constructor above for details. Can be
406called at any time.
166 407
167Same as C<io> and C<io_ns>, but also specifies a timeout (as if there was 408=item $current_fh = $w->fh
168a call to C<< $w->timeout ($timout, 1) >>. The persist flag is not allowed
169and will automatically be cleared. The watcher will be restarted after each event.
170 409
171If the timeout is zero or undef, no timeout will be set, and a normal 410=item $old_fh = $w->fh ($new_fh)
172watcher (with the persist flag set!) will be created.
173 411
174This has the effect of timing out after the specified period of inactivity 412Returns the previously set filehandle and optionally set a new one.
175has happened.
176 413
177Due to the design of libevent, this is also relatively inefficient, having 414=item $current_eventmask = $w->events
178one or two io watchers and a separate timeout watcher that you reset on
179activity (by calling its C<start> method) is usually more efficient.
180 415
416=item $old_eventmask = $w->events ($new_eventmask)
417
418Returns the previously set event mask and optionally set a new one.
419
420=back
421
422
423=head3 TIMER WATCHERS - relative and optionally repeating timeouts
424
425=over 4
426
181=item my $w = EV::timer $after, $repeat, $callback 427=item $w = EV::timer $after, $repeat, $callback
182 428
183=item my $w = EV::timer_ns $after, $repeat, $callback 429=item $w = EV::timer_ns $after, $repeat, $callback
184 430
185Calls the callback after C<$after> seconds. If C<$repeat> is true, the 431=item $w = $loop->timer ($after, $repeat, $callback)
186timer will be restarted after the callback returns. This means that the
187callback would be called roughly every C<$after> seconds, prolonged by the
188time the callback takes.
189 432
433=item $w = $loop->timer_ns ($after, $repeat, $callback)
434
435Calls the callback after C<$after> seconds (which may be fractional). If
436C<$repeat> is non-zero, the timer will be restarted (with the $repeat
437value as $after) after the callback returns.
438
439This means that the callback would be called roughly after C<$after>
440seconds, and then every C<$repeat> seconds. The timer does his best not
441to drift, but it will not invoke the timer more often then once per event
442loop iteration, and might drift in other cases. If that isn't acceptable,
443look at EV::periodic, which can provide long-term stable timers.
444
445The timer is based on a monotonic clock, that is, if somebody is sitting
446in front of the machine while the timer is running and changes the system
447clock, the timer will nevertheless run (roughly) the same time.
448
190The C<timer_ns> variant doesn't add/start the newly created watcher. 449The C<timer_ns> variant doesn't start (activate) the newly created watcher.
191 450
192=item my $w = EV::timer_abs $at, $interval, $callback 451=item $w->set ($after, $repeat)
193 452
194=item my $w = EV::timer_abs_ns $at, $interval, $callback 453Reconfigures the watcher, see the constructor above for details. Can be called at
454any time.
195 455
196Similar to EV::timer, but the time is given as an absolute point in time 456=item $w->again
197(C<$at>), plus an optional C<$interval>.
198 457
199If the C<$interval> is zero, then the callback will be called at the time 458Similar to the C<start> method, but has special semantics for repeating timers:
200C<$at> if that is in the future, or as soon as possible if its in the
201past. It will not automatically repeat.
202 459
203If the C<$interval> is nonzero, then the watcher will always be scheduled 460If the timer is active and non-repeating, it will be stopped.
204to time out at the next C<$at + integer * $interval> time.
205 461
206This can be used to schedule a callback to run at very regular intervals, 462If the timer is active and repeating, reset the timeout to occur
207as long as the processing time is less then the interval (otherwise 463C<$repeat> seconds after now.
208obviously events will be skipped). 464
465If the timer is inactive and repeating, start it using the repeat value.
466
467Otherwise do nothing.
468
469This behaviour is useful when you have a timeout for some IO
470operation. You create a timer object with the same value for C<$after> and
471C<$repeat>, and then, in the read/write watcher, run the C<again> method
472on the timeout.
473
474=back
475
476
477=head3 PERIODIC WATCHERS - to cron or not to cron?
478
479=over 4
480
481=item $w = EV::periodic $at, $interval, $reschedule_cb, $callback
482
483=item $w = EV::periodic_ns $at, $interval, $reschedule_cb, $callback
484
485=item $w = $loop->periodic ($at, $interval, $reschedule_cb, $callback)
486
487=item $w = $loop->periodic_ns ($at, $interval, $reschedule_cb, $callback)
488
489Similar to EV::timer, but is not based on relative timeouts but on
490absolute times. Apart from creating "simple" timers that trigger "at" the
491specified time, it can also be used for non-drifting absolute timers and
492more complex, cron-like, setups that are not adversely affected by time
493jumps (i.e. when the system clock is changed by explicit date -s or other
494means such as ntpd). It is also the most complex watcher type in EV.
495
496It has three distinct "modes":
497
498=over 4
499
500=item * absolute timer ($interval = $reschedule_cb = 0)
501
502This time simply fires at the wallclock time C<$at> and doesn't repeat. It
503will not adjust when a time jump occurs, that is, if it is to be run
504at January 1st 2011 then it will run when the system time reaches or
505surpasses this time.
506
507=item * non-repeating interval timer ($interval > 0, $reschedule_cb = 0)
508
509In this mode the watcher will always be scheduled to time out at the
510next C<$at + N * $interval> time (for some integer N) and then repeat,
511regardless of any time jumps.
512
513This can be used to create timers that do not drift with respect to system
514time:
515
516 my $hourly = EV::periodic 0, 3600, 0, sub { print "once/hour\n" };
517
518That doesn't mean there will always be 3600 seconds in between triggers,
519but only that the the clalback will be called when the system time shows a
520full hour (UTC).
209 521
210Another way to think about it (for the mathematically inclined) is that 522Another way to think about it (for the mathematically inclined) is that
211C<timer_abs> will try to tun the callback at the next possible time where 523EV::periodic will try to run the callback in this mode at the next
212C<$time = $at (mod $interval)>, regardless of any time jumps. 524possible time where C<$time = $at (mod $interval)>, regardless of any time
525jumps.
213 526
527=item * manual reschedule mode ($reschedule_cb = coderef)
528
529In this mode $interval and $at are both being ignored. Instead, each
530time the periodic watcher gets scheduled, the reschedule callback
531($reschedule_cb) will be called with the watcher as first, and the current
532time as second argument.
533
534I<This callback MUST NOT stop or destroy this or any other periodic
535watcher, ever>. If you need to stop it, return 1e30 and stop it
536afterwards.
537
538It must return the next time to trigger, based on the passed time value
539(that is, the lowest time value larger than to the second argument). It
540will usually be called just before the callback will be triggered, but
541might be called at other times, too.
542
543This can be used to create very complex timers, such as a timer that
544triggers on each midnight, local time (actually 24 hours after the last
545midnight, to keep the example simple. If you know a way to do it correctly
546in about the same space (without requiring elaborate modules), drop me a
547note :):
548
549 my $daily = EV::periodic 0, 0, sub {
550 my ($w, $now) = @_;
551
552 use Time::Local ();
553 my (undef, undef, undef, $d, $m, $y) = localtime $now;
554 86400 + Time::Local::timelocal 0, 0, 0, $d, $m, $y
555 }, sub {
556 print "it's midnight or likely shortly after, now\n";
557 };
558
559=back
560
214The C<timer_abs_ns> variant doesn't add/start the newly created watcher. 561The C<periodic_ns> variant doesn't start (activate) the newly created watcher.
215 562
563=item $w->set ($at, $interval, $reschedule_cb)
564
565Reconfigures the watcher, see the constructor above for details. Can be called at
566any time.
567
568=item $w->again
569
570Simply stops and starts the watcher again.
571
572=item $time = $w->at
573
574Return the time that the watcher is expected to trigger next.
575
576=back
577
578
579=head3 SIGNAL WATCHERS - signal me when a signal gets signalled!
580
581=over 4
582
216=item my $w = EV::signal $signal, $callback 583=item $w = EV::signal $signal, $callback
217 584
218=item my $w = EV::signal_ns $signal, $callback 585=item $w = EV::signal_ns $signal, $callback
219 586
220Call the callback when $signal is received (the signal can be specified 587Call the callback when $signal is received (the signal can be specified by
221by number or by name, just as with kill or %SIG). Signal watchers are 588number or by name, just as with C<kill> or C<%SIG>).
222persistent no natter what.
223 589
224EV will grab the signal for the process (the kernel only allows one 590EV will grab the signal for the process (the kernel only allows one
225component to receive signals) when you start a signal watcher, and 591component to receive a signal at a time) when you start a signal watcher,
226removes it again when you stop it. Pelr does the same when you add/remove 592and removes it again when you stop it. Perl does the same when you
227callbacks to %SIG, so watch out. 593add/remove callbacks to C<%SIG>, so watch out.
228 594
229Unfortunately, only one handler can be registered per signal. Screw 595You can have as many signal watchers per signal as you want.
230libevent.
231 596
232The C<signal_ns> variant doesn't add/start the newly created watcher. 597The C<signal_ns> variant doesn't start (activate) the newly created watcher.
233 598
234=back 599=item $w->set ($signal)
235 600
236=head1 THE EV::Event CLASS 601Reconfigures the watcher, see the constructor above for details. Can be
602called at any time.
237 603
238All EV functions creating an event watcher (designated by C<my $w =>
239above) support the following methods on the returned watcher object:
240
241=over 4
242
243=item $w->add ($timeout)
244
245Stops and (re-)starts the event watcher, setting the optional timeout to
246the given value, or clearing the timeout if none is given.
247
248=item $w->start
249
250Stops and (re-)starts the event watcher without touching the timeout.
251
252=item $w->del
253
254=item $w->stop
255
256Stop the event watcher if it was started.
257
258=item $current_callback = $w->cb
259
260=item $old_callback = $w->cb ($new_callback)
261
262Return the previously set callback and optionally set a new one.
263
264=item $current_fh = $w->fh
265
266=item $old_fh = $w->fh ($new_fh)
267
268Returns the previously set filehandle and optionally set a new one (also
269clears the EV::SIGNAL flag when setting a filehandle).
270
271=item $current_signal = $w->signal 604=item $current_signum = $w->signal
272 605
273=item $old_signal = $w->signal ($new_signal) 606=item $old_signum = $w->signal ($new_signal)
274 607
608Returns the previously set signal (always as a number not name) and
609optionally set a new one.
610
611=back
612
613
614=head3 CHILD WATCHERS - watch out for process status changes
615
616=over 4
617
618=item $w = EV::child $pid, $callback
619
620=item $w = EV::child_ns $pid, $callback
621
622=item $w = $loop->child ($pid, $callback)
623
624=item $w = $loop->child_ns ($pid, $callback)
625
626Call the callback when a status change for pid C<$pid> (or any pid if
627C<$pid> is 0) has been received. More precisely: when the process receives
628a C<SIGCHLD>, EV will fetch the outstanding exit/wait status for all
629changed/zombie children and call the callback.
630
631It is valid (and fully supported) to install a child watcher after a child
632has exited but before the event loop has started its next iteration (for
633example, first you C<fork>, then the new child process might exit, and
634only then do you install a child watcher in the parent for the new pid).
635
636You can access both exit (or tracing) status and pid by using the
637C<rstatus> and C<rpid> methods on the watcher object.
638
639You can have as many pid watchers per pid as you want, they will all be
640called.
641
642The C<child_ns> variant doesn't start (activate) the newly created watcher.
643
644=item $w->set ($pid)
645
646Reconfigures the watcher, see the constructor above for details. Can be called at
647any time.
648
649=item $current_pid = $w->pid
650
651=item $old_pid = $w->pid ($new_pid)
652
275Returns the previously set signal number and optionally set a new one (also sets 653Returns the previously set process id and optionally set a new one.
276the EV::SIGNAL flag when setting a signal).
277 654
278=item $current_eventmask = $w->events 655=item $exit_status = $w->rstatus
279 656
280=item $old_eventmask = $w->events ($new_eventmask) 657Return the exit/wait status (as returned by waitpid, see the waitpid entry
658in perlfunc).
281 659
660=item $pid = $w->rpid
661
662Return the pid of the awaited child (useful when you have installed a
663watcher for all pids).
664
665=back
666
667
668=head3 STAT WATCHERS - did the file attributes just change?
669
670=over 4
671
672=item $w = EV::stat $path, $interval, $callback
673
674=item $w = EV::stat_ns $path, $interval, $callback
675
676=item $w = $loop->stat ($path, $interval, $callback)
677
678=item $w = $loop->stat_ns ($path, $interval, $callback)
679
680Call the callback when a file status change has been detected on
681C<$path>. The C<$path> does not need to exist, changing from "path exists"
682to "path does not exist" is a status change like any other.
683
684The C<$interval> is a recommended polling interval for systems where
685OS-supported change notifications don't exist or are not supported. If
686you use C<0> then an unspecified default is used (which is highly
687recommended!), which is to be expected to be around five seconds usually.
688
689This watcher type is not meant for massive numbers of stat watchers,
690as even with OS-supported change notifications, this can be
691resource-intensive.
692
693The C<stat_ns> variant doesn't start (activate) the newly created watcher.
694
695=item ... = $w->stat
696
697This call is very similar to the perl C<stat> built-in: It stats (using
698C<lstat>) the path specified in the watcher and sets perls stat cache (as
699well as EV's idea of the current stat values) to the values found.
700
701In scalar context, a boolean is return indicating success or failure of
702the stat. In list context, the same 13-value list as with stat is returned
703(except that the blksize and blocks fields are not reliable).
704
705In the case of an error, errno is set to C<ENOENT> (regardless of the
706actual error value) and the C<nlink> value is forced to zero (if the stat
707was successful then nlink is guaranteed to be non-zero).
708
709See also the next two entries for more info.
710
711=item ... = $w->attr
712
713Just like C<< $w->stat >>, but without the initial stat'ing: this returns
714the values most recently detected by EV. See the next entry for more info.
715
716=item ... = $w->prev
717
718Just like C<< $w->stat >>, but without the initial stat'ing: this returns
719the previous set of values, before the change.
720
721That is, when the watcher callback is invoked, C<< $w->prev >> will be set
722to the values found I<before> a change was detected, while C<< $w->attr >>
723returns the values found leading to the change detection. The difference (if any)
724between C<prev> and C<attr> is what triggered the callback.
725
726If you did something to the filesystem object and do not want to trigger
727yet another change, you can call C<stat> to update EV's idea of what the
728current attributes are.
729
730=item $w->set ($path, $interval)
731
732Reconfigures the watcher, see the constructor above for details. Can be
733called at any time.
734
735=item $current_path = $w->path
736
737=item $old_path = $w->path ($new_path)
738
282Returns the previously set event mask and optionally set a new one. 739Returns the previously set path and optionally set a new one.
283 740
284=item $w->timeout ($after, $repeat) 741=item $current_interval = $w->interval
285 742
286Resets the timeout (see C<EV::timer> for details). 743=item $old_interval = $w->interval ($new_interval)
287 744
288=item $w->timeout_abs ($at, $interval) 745Returns the previously set interval and optionally set a new one. Can be
746used to query the actual interval used.
289 747
290Resets the timeout (see C<EV::timer_abs> for details).
291
292=item $w->priority_set ($priority)
293
294Set the priority of the watcher to C<$priority> (0 <= $priority < $EV::NPRI).
295
296=back 748=back
749
750
751=head3 IDLE WATCHERS - when you've got nothing better to do...
752
753=over 4
754
755=item $w = EV::idle $callback
756
757=item $w = EV::idle_ns $callback
758
759=item $w = $loop->idle ($callback)
760
761=item $w = $loop->idle_ns ($callback)
762
763Call the callback when there are no other pending watchers of the same or
764higher priority (excluding check, prepare and other idle watchers of the
765same or lower priority, of course). They are called idle watchers because
766when the watcher is the highest priority pending event in the process, the
767process is considered to be idle at that priority.
768
769If you want a watcher that is only ever called when I<no> other events are
770outstanding you have to set the priority to C<EV::MINPRI>.
771
772The process will not block as long as any idle watchers are active, and
773they will be called repeatedly until stopped.
774
775For example, if you have idle watchers at priority C<0> and C<1>, and
776an I/O watcher at priority C<0>, then the idle watcher at priority C<1>
777and the I/O watcher will always run when ready. Only when the idle watcher
778at priority C<1> is stopped and the I/O watcher at priority C<0> is not
779pending with the C<0>-priority idle watcher be invoked.
780
781The C<idle_ns> variant doesn't start (activate) the newly created watcher.
782
783=back
784
785
786=head3 PREPARE WATCHERS - customise your event loop!
787
788=over 4
789
790=item $w = EV::prepare $callback
791
792=item $w = EV::prepare_ns $callback
793
794=item $w = $loop->prepare ($callback)
795
796=item $w = $loop->prepare_ns ($callback)
797
798Call the callback just before the process would block. You can still
799create/modify any watchers at this point.
800
801See the EV::check watcher, below, for explanations and an example.
802
803The C<prepare_ns> variant doesn't start (activate) the newly created watcher.
804
805=back
806
807
808=head3 CHECK WATCHERS - customise your event loop even more!
809
810=over 4
811
812=item $w = EV::check $callback
813
814=item $w = EV::check_ns $callback
815
816=item $w = $loop->check ($callback)
817
818=item $w = $loop->check_ns ($callback)
819
820Call the callback just after the process wakes up again (after it has
821gathered events), but before any other callbacks have been invoked.
822
823This is used to integrate other event-based software into the EV
824mainloop: You register a prepare callback and in there, you create io and
825timer watchers as required by the other software. Here is a real-world
826example of integrating Net::SNMP (with some details left out):
827
828 our @snmp_watcher;
829
830 our $snmp_prepare = EV::prepare sub {
831 # do nothing unless active
832 $dispatcher->{_event_queue_h}
833 or return;
834
835 # make the dispatcher handle any outstanding stuff
836 ... not shown
837
838 # create an I/O watcher for each and every socket
839 @snmp_watcher = (
840 (map { EV::io $_, EV::READ, sub { } }
841 keys %{ $dispatcher->{_descriptors} }),
842
843 EV::timer +($event->[Net::SNMP::Dispatcher::_ACTIVE]
844 ? $event->[Net::SNMP::Dispatcher::_TIME] - EV::now : 0),
845 0, sub { },
846 );
847 };
848
849The callbacks are irrelevant (and are not even being called), the
850only purpose of those watchers is to wake up the process as soon as
851one of those events occurs (socket readable, or timer timed out). The
852corresponding EV::check watcher will then clean up:
853
854 our $snmp_check = EV::check sub {
855 # destroy all watchers
856 @snmp_watcher = ();
857
858 # make the dispatcher handle any new stuff
859 ... not shown
860 };
861
862The callbacks of the created watchers will not be called as the watchers
863are destroyed before this cna happen (remember EV::check gets called
864first).
865
866The C<check_ns> variant doesn't start (activate) the newly created watcher.
867
868=back
869
870
871=head3 FORK WATCHERS - the audacity to resume the event loop after a fork
872
873Fork watchers are called when a C<fork ()> was detected. The invocation
874is done before the event loop blocks next and before C<check> watchers
875are being called, and only in the child after the fork.
876
877=over 4
878
879=item $w = EV::fork $callback
880
881=item $w = EV::fork_ns $callback
882
883=item $w = $loop->fork ($callback)
884
885=item $w = $loop->fork_ns ($callback)
886
887Call the callback before the event loop is resumed in the child process
888after a fork.
889
890The C<fork_ns> variant doesn't start (activate) the newly created watcher.
891
892=back
893
894
895=head1 PERL SIGNALS
896
897While Perl signal handling (C<%SIG>) is not affected by EV, the behaviour
898with EV is as the same as any other C library: Perl-signals will only be
899handled when Perl runs, which means your signal handler might be invoked
900only the next time an event callback is invoked.
901
902The solution is to use EV signal watchers (see C<EV::signal>), which will
903ensure proper operations with regards to other event watchers.
904
905If you cannot do this for whatever reason, you can also force a watcher
906to be called on every event loop iteration by installing a C<EV::check>
907watcher:
908
909 my $async_check = EV::check sub { };
910
911This ensures that perl gets into control for a short time to handle any
912pending signals, and also ensures (slightly) slower overall operation.
297 913
298=head1 THREADS 914=head1 THREADS
299 915
300Threads are not supported by this in any way. Perl pseudo-threads is evil 916Threads are not supported by this module in any way. Perl pseudo-threads
301and must die. 917is evil stuff and must die. As soon as Perl gains real threads I will work
918on thread support for it.
919
920=head1 FORK
921
922Most of the "improved" event delivering mechanisms of modern operating
923systems have quite a few problems with fork(2) (to put it bluntly: it is
924not supported and usually destructive). Libev makes it possible to work
925around this by having a function that recreates the kernel state after
926fork in the child.
927
928On non-win32 platforms, this module requires the pthread_atfork
929functionality to do this automatically for you. This function is quite
930buggy on most BSDs, though, so YMMV. The overhead for this is quite
931negligible, because everything the function currently does is set a flag
932that is checked only when the event loop gets used the next time, so when
933you do fork but not use EV, the overhead is minimal.
934
935On win32, there is no notion of fork so all this doesn't apply, of course.
302 936
303=cut 937=cut
304 938
305our $DIED = sub { 939our $DIED = sub {
306 warn "EV: error in callback (ignoring): $@"; 940 warn "EV: error in callback (ignoring): $@";
307}; 941};
308 942
309init; 943default_loop
310 944 or die 'EV: cannot initialise libev backend. bad $ENV{LIBEV_FLAGS}?';
311push @AnyEvent::REGISTRY, [EV => "EV::AnyEvent"];
312 945
3131; 9461;
314 947
315=head1 SEE ALSO 948=head1 SEE ALSO
316 949
317 L<EV::DNS>, L<event(3)>, L<event.h>, L<evdns.h>. 950L<EV::ADNS> (asynchronous DNS), L<Glib::EV> (makes Glib/Gtk2 use EV as
318 L<EV::AnyEvent>. 951event loop), L<EV::Glib> (embed Glib into EV), L<Coro::EV> (efficient
952coroutines with EV), L<Net::SNMP::EV> (asynchronous SNMP).
319 953
320=head1 AUTHOR 954=head1 AUTHOR
321 955
322 Marc Lehmann <schmorp@schmorp.de> 956 Marc Lehmann <schmorp@schmorp.de>
323 http://home.schmorp.de/ 957 http://home.schmorp.de/

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines