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