… | |
… | |
256 | file descriptors. To improve latency and performance, you can ask |
256 | file descriptors. To improve latency and performance, you can ask |
257 | C<EV::Loop::Async> to loop an additional number of times in the foreground |
257 | C<EV::Loop::Async> to loop an additional number of times in the foreground |
258 | after invoking the callbacks, effectively doing the polling in the |
258 | after invoking the callbacks, effectively doing the polling in the |
259 | foreground. |
259 | foreground. |
260 | |
260 | |
261 | The default is C<0>, meaning thatno foreground polling will be done. A |
261 | The default is C<0>, meaning that no foreground polling will be done. A |
262 | value of C<1> means that, after handling the pending events, it will call |
262 | value of C<1> means that, after handling the pending events, it will call |
263 | C<< $loop->loop (EV::LOOP_NONBLOCK) >> and handle the resulting events, if |
263 | C<< $loop->loop (EV::LOOP_NONBLOCK) >> and handle the resulting events, if |
264 | any. A value of C<2> means that this will be iterated twice. |
264 | any. A value of C<2> means that this will be iterated twice. |
265 | |
265 | |
266 | When a foreground event poll does not yield any new events, then no |
266 | When a foreground event poll does not yield any new events, then no |
267 | further iterations will be made, so this is only a I<maximum> value of |
267 | further iterations will be made, so this is only a I<maximum> value of |
268 | additional loop runs. |
268 | additional loop runs. |
269 | |
269 | |
|
|
270 | Take also note of the standard EV C<set_io_collect_interval> |
|
|
271 | functionality, which can achieve a similar, but different, effect - YMMV. |
|
|
272 | |
270 | =back |
273 | =back |
271 | |
274 | |
272 | =head1 SEE ALSO |
275 | =head1 SEE ALSO |
273 | |
276 | |
274 | L<EV>, L<Async::Interrupt>. |
277 | L<EV>, L<Async::Interrupt>. |