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

Comparing EV/README (file contents):
Revision 1.4 by root, Mon Oct 29 19:53:21 2007 UTC vs.
Revision 1.20 by root, Sat Dec 22 12:05:42 2007 UTC

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

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines