[ViewVC] Diff of: cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.17 by root, Mon Nov 12 08:57:03 2007 UTC vs.
Revision 1.30 by root, Fri Nov 23 04:36:03 2007 UTC

…		…
49	called C<ev_tstamp>, which is what you should use too. It usually aliases	49	called C<ev_tstamp>, which is what you should use too. It usually aliases
50	to the double type in C.	50	to the double type in C.
51		51
52	=head1 GLOBAL FUNCTIONS	52	=head1 GLOBAL FUNCTIONS
53		53
		54	These functions can be called anytime, even before initialising the
		55	library in any way.
		56
54	=over 4	57	=over 4
55		58
56	=item ev_tstamp ev_time ()	59	=item ev_tstamp ev_time ()
57		60
58	Returns the current time as libev would use it.	61	Returns the current time as libev would use it. Please note that the
		62	C<ev_now> function is usually faster and also often returns the timestamp
		63	you actually want to know.
59		64
60	=item int ev_version_major ()	65	=item int ev_version_major ()
61		66
62	=item int ev_version_minor ()	67	=item int ev_version_minor ()
63		68
…		…
140	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will	145	C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
141	override the flags completely if it is found in the environment. This is	146	override the flags completely if it is found in the environment. This is
142	useful to try out specific backends to test their performance, or to work	147	useful to try out specific backends to test their performance, or to work
143	around bugs.	148	around bugs.
144		149
145	=item C<EVMETHOD_SELECT> (portable select backend)	150	=item C<EVMETHOD_SELECT> (value 1, portable select backend)
146		151
		152	This is your standard select(2) backend. Not I<completely> standard, as
		153	libev tries to roll its own fd_set with no limits on the number of fds,
		154	but if that fails, expect a fairly low limit on the number of fds when
		155	using this backend. It doesn't scale too well (O(highest_fd)), but its usually
		156	the fastest backend for a low number of fds.
		157
147	=item C<EVMETHOD_POLL> (poll backend, available everywhere except on windows)	158	=item C<EVMETHOD_POLL> (value 2, poll backend, available everywhere except on windows)
148		159
		160	And this is your standard poll(2) backend. It's more complicated than
		161	select, but handles sparse fds better and has no artificial limit on the
		162	number of fds you can use (except it will slow down considerably with a
		163	lot of inactive fds). It scales similarly to select, i.e. O(total_fds).
		164
149	=item C<EVMETHOD_EPOLL> (linux only)	165	=item C<EVMETHOD_EPOLL> (value 4, Linux)
150		166
151	=item C<EVMETHOD_KQUEUE> (some bsds only)	167	For few fds, this backend is a bit little slower than poll and select,
		168	but it scales phenomenally better. While poll and select usually scale like
		169	O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
		170	either O(1) or O(active_fds).
152		171
		172	While stopping and starting an I/O watcher in the same iteration will
		173	result in some caching, there is still a syscall per such incident
		174	(because the fd could point to a different file description now), so its
		175	best to avoid that. Also, dup()ed file descriptors might not work very
		176	well if you register events for both fds.
		177
		178	=item C<EVMETHOD_KQUEUE> (value 8, most BSD clones)
		179
		180	Kqueue deserves special mention, as at the time of this writing, it
		181	was broken on all BSDs except NetBSD (usually it doesn't work with
		182	anything but sockets and pipes, except on Darwin, where of course its
		183	completely useless). For this reason its not being "autodetected" unless
		184	you explicitly specify the flags (i.e. you don't use EVFLAG_AUTO).
		185
		186	It scales in the same way as the epoll backend, but the interface to the
		187	kernel is more efficient (which says nothing about its actual speed, of
		188	course). While starting and stopping an I/O watcher does not cause an
		189	extra syscall as with epoll, it still adds up to four event changes per
		190	incident, so its best to avoid that.
		191
153	=item C<EVMETHOD_DEVPOLL> (solaris 8 only)	192	=item C<EVMETHOD_DEVPOLL> (value 16, Solaris 8)
154		193
		194	This is not implemented yet (and might never be).
		195
155	=item C<EVMETHOD_PORT> (solaris 10 only)	196	=item C<EVMETHOD_PORT> (value 32, Solaris 10)
		197
		198	This uses the Solaris 10 port mechanism. As with everything on Solaris,
		199	it's really slow, but it still scales very well (O(active_fds)).
		200
		201	=item C<EVMETHOD_ALL>
		202
		203	Try all backends (even potentially broken ones that wouldn't be tried
		204	with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
		205	C<EVMETHOD_ALL & ~EVMETHOD_KQUEUE>.
		206
		207	=back
156		208
157	If one or more of these are ored into the flags value, then only these	209	If one or more of these are ored into the flags value, then only these
158	backends will be tried (in the reverse order as given here). If one are	210	backends will be tried (in the reverse order as given here). If none are
159	specified, any backend will do.	211	specified, most compiled-in backend will be tried, usually in reverse
160		212	order of their flag values :)
161	=back
162		213
163	=item struct ev_loop *ev_loop_new (unsigned int flags)	214	=item struct ev_loop *ev_loop_new (unsigned int flags)
164		215
165	Similar to C<ev_default_loop>, but always creates a new event loop that is	216	Similar to C<ev_default_loop>, but always creates a new event loop that is
166	always distinct from the default loop. Unlike the default loop, it cannot	217	always distinct from the default loop. Unlike the default loop, it cannot
…		…
183	This function reinitialises the kernel state for backends that have	234	This function reinitialises the kernel state for backends that have
184	one. Despite the name, you can call it anytime, but it makes most sense	235	one. Despite the name, you can call it anytime, but it makes most sense
185	after forking, in either the parent or child process (or both, but that	236	after forking, in either the parent or child process (or both, but that
186	again makes little sense).	237	again makes little sense).
187		238
188	You I<must> call this function after forking if and only if you want to	239	You I<must> call this function in the child process after forking if and
189	use the event library in both processes. If you just fork+exec, you don't	240	only if you want to use the event library in both processes. If you just
190	have to call it.	241	fork+exec, you don't have to call it.
191		242
192	The function itself is quite fast and it's usually not a problem to call	243	The function itself is quite fast and it's usually not a problem to call
193	it just in case after a fork. To make this easy, the function will fit in	244	it just in case after a fork. To make this easy, the function will fit in
194	quite nicely into a call to C<pthread_atfork>:	245	quite nicely into a call to C<pthread_atfork>:
195		246
…		…
234		285
235	This flags value could be used to implement alternative looping	286	This flags value could be used to implement alternative looping
236	constructs, but the C<prepare> and C<check> watchers provide a better and	287	constructs, but the C<prepare> and C<check> watchers provide a better and
237	more generic mechanism.	288	more generic mechanism.
238		289
		290	Here are the gory details of what ev_loop does:
		291
		292	1. If there are no active watchers (reference count is zero), return.
		293	2. Queue and immediately call all prepare watchers.
		294	3. If we have been forked, recreate the kernel state.
		295	4. Update the kernel state with all outstanding changes.
		296	5. Update the "event loop time".
		297	6. Calculate for how long to block.
		298	7. Block the process, waiting for events.
		299	8. Update the "event loop time" and do time jump handling.
		300	9. Queue all outstanding timers.
		301	10. Queue all outstanding periodics.
		302	11. If no events are pending now, queue all idle watchers.
		303	12. Queue all check watchers.
		304	13. Call all queued watchers in reverse order (i.e. check watchers first).
		305	14. If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
		306	was used, return, otherwise continue with step #1.
		307
239	=item ev_unloop (loop, how)	308	=item ev_unloop (loop, how)
240		309
241	Can be used to make a call to C<ev_loop> return early (but only after it	310	Can be used to make a call to C<ev_loop> return early (but only after it
242	has processed all outstanding events). The C<how> argument must be either	311	has processed all outstanding events). The C<how> argument must be either
243	C<EVUNLOOP_ONCE>, which will make the innermost C<ev_loop> call return, or	312	C<EVUNLOOP_ONE>, which will make the innermost C<ev_loop> call return, or
244	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.	313	C<EVUNLOOP_ALL>, which will make all nested C<ev_loop> calls return.
245		314
246	=item ev_ref (loop)	315	=item ev_ref (loop)
247		316
248	=item ev_unref (loop)	317	=item ev_unref (loop)
…		…
414	in each iteration of the event loop (This behaviour is called	483	in each iteration of the event loop (This behaviour is called
415	level-triggering because you keep receiving events as long as the	484	level-triggering because you keep receiving events as long as the
416	condition persists. Remember you can stop the watcher if you don't want to	485	condition persists. Remember you can stop the watcher if you don't want to
417	act on the event and neither want to receive future events).	486	act on the event and neither want to receive future events).
418		487
419	In general you can register as many read and/or write event watchers oer	488	In general you can register as many read and/or write event watchers per
420	fd as you want (as long as you don't confuse yourself). Setting all file	489	fd as you want (as long as you don't confuse yourself). Setting all file
421	descriptors to non-blocking mode is also usually a good idea (but not	490	descriptors to non-blocking mode is also usually a good idea (but not
422	required if you know what you are doing).	491	required if you know what you are doing).
423		492
424	You have to be careful with dup'ed file descriptors, though. Some backends	493	You have to be careful with dup'ed file descriptors, though. Some backends
425	(the linux epoll backend is a notable example) cannot handle dup'ed file	494	(the linux epoll backend is a notable example) cannot handle dup'ed file
426	descriptors correctly if you register interest in two or more fds pointing	495	descriptors correctly if you register interest in two or more fds pointing
427	to the same file/socket etc. description.	496	to the same underlying file/socket etc. description (that is, they share
		497	the same underlying "file open").
428		498
429	If you must do this, then force the use of a known-to-be-good backend	499	If you must do this, then force the use of a known-to-be-good backend
430	(at the time of this writing, this includes only EVMETHOD_SELECT and	500	(at the time of this writing, this includes only EVMETHOD_SELECT and
431	EVMETHOD_POLL).	501	EVMETHOD_POLL).
432		502
…		…
446		516
447	Timer watchers are simple relative timers that generate an event after a	517	Timer watchers are simple relative timers that generate an event after a
448	given time, and optionally repeating in regular intervals after that.	518	given time, and optionally repeating in regular intervals after that.
449		519
450	The timers are based on real time, that is, if you register an event that	520	The timers are based on real time, that is, if you register an event that
451	times out after an hour and youreset your system clock to last years	521	times out after an hour and you reset your system clock to last years
452	time, it will still time out after (roughly) and hour. "Roughly" because	522	time, it will still time out after (roughly) and hour. "Roughly" because
453	detecting time jumps is hard, and soem inaccuracies are unavoidable (the	523	detecting time jumps is hard, and some inaccuracies are unavoidable (the
454	monotonic clock option helps a lot here).	524	monotonic clock option helps a lot here).
455		525
456	The relative timeouts are calculated relative to the C<ev_now ()>	526	The relative timeouts are calculated relative to the C<ev_now ()>
457	time. This is usually the right thing as this timestamp refers to the time	527	time. This is usually the right thing as this timestamp refers to the time
458	of the event triggering whatever timeout you are modifying/starting. If	528	of the event triggering whatever timeout you are modifying/starting. If
459	you suspect event processing to be delayed and you need to base the timeout	529	you suspect event processing to be delayed and you I<need> to base the timeout
460	ion the current time, use something like this to adjust for this:	530	on the current time, use something like this to adjust for this:
461		531
462	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);	532	ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
		533
		534	The callback is guarenteed to be invoked only when its timeout has passed,
		535	but if multiple timers become ready during the same loop iteration then
		536	order of execution is undefined.
463		537
464	=over 4	538	=over 4
465		539
466	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)	540	=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
467		541
…		…
473	later, again, and again, until stopped manually.	547	later, again, and again, until stopped manually.
474		548
475	The timer itself will do a best-effort at avoiding drift, that is, if you	549	The timer itself will do a best-effort at avoiding drift, that is, if you
476	configure a timer to trigger every 10 seconds, then it will trigger at	550	configure a timer to trigger every 10 seconds, then it will trigger at
477	exactly 10 second intervals. If, however, your program cannot keep up with	551	exactly 10 second intervals. If, however, your program cannot keep up with
478	the timer (ecause it takes longer than those 10 seconds to do stuff) the	552	the timer (because it takes longer than those 10 seconds to do stuff) the
479	timer will not fire more than once per event loop iteration.	553	timer will not fire more than once per event loop iteration.
480		554
481	=item ev_timer_again (loop)	555	=item ev_timer_again (loop)
482		556
483	This will act as if the timer timed out and restart it again if it is	557	This will act as if the timer timed out and restart it again if it is
…		…
514	again).	588	again).
515		589
516	They can also be used to implement vastly more complex timers, such as	590	They can also be used to implement vastly more complex timers, such as
517	triggering an event on eahc midnight, local time.	591	triggering an event on eahc midnight, local time.
518		592
		593	As with timers, the callback is guarenteed to be invoked only when the
		594	time (C<at>) has been passed, but if multiple periodic timers become ready
		595	during the same loop iteration then order of execution is undefined.
		596
519	=over 4	597	=over 4
520		598
521	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)	599	=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
522		600
523	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)	601	=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
524		602
525	Lots of arguments, lets sort it out... There are basically three modes of	603	Lots of arguments, lets sort it out... There are basically three modes of
526	operation, and we will explain them from simplest to complex:	604	operation, and we will explain them from simplest to complex:
527
528		605
529	=over 4	606	=over 4
530		607
531	=item * absolute timer (interval = reschedule_cb = 0)	608	=item * absolute timer (interval = reschedule_cb = 0)
532		609
…		…
560	In this mode the values for C<interval> and C<at> are both being	637	In this mode the values for C<interval> and C<at> are both being
561	ignored. Instead, each time the periodic watcher gets scheduled, the	638	ignored. Instead, each time the periodic watcher gets scheduled, the
562	reschedule callback will be called with the watcher as first, and the	639	reschedule callback will be called with the watcher as first, and the
563	current time as second argument.	640	current time as second argument.
564		641
565	NOTE: I<This callback MUST NOT stop or destroy the periodic or any other	642	NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
566	periodic watcher, ever, or make any event loop modifications>. If you need	643	ever, or make any event loop modifications>. If you need to stop it,
567	to stop it, return C<now + 1e30> (or so, fudge fudge) and stop it afterwards.	644	return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
568		645	starting a prepare watcher).
569	Also, I<< this callback must always return a time that is later than the
570	passed C<now> value >>. Not even C<now> itself will be ok.
571		646
572	Its prototype is C<ev_tstamp (reschedule_cb)(struct ev_periodic w,	647	Its prototype is C<ev_tstamp (reschedule_cb)(struct ev_periodic w,
573	ev_tstamp now)>, e.g.:	648	ev_tstamp now)>, e.g.:
574		649
575	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)	650	static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
…		…
580	It must return the next time to trigger, based on the passed time value	655	It must return the next time to trigger, based on the passed time value
581	(that is, the lowest time value larger than to the second argument). It	656	(that is, the lowest time value larger than to the second argument). It
582	will usually be called just before the callback will be triggered, but	657	will usually be called just before the callback will be triggered, but
583	might be called at other times, too.	658	might be called at other times, too.
584		659
		660	NOTE: I<< This callback must always return a time that is later than the
		661	passed C<now> value >>. Not even C<now> itself will do, it I<must> be larger.
		662
585	This can be used to create very complex timers, such as a timer that	663	This can be used to create very complex timers, such as a timer that
586	triggers on each midnight, local time. To do this, you would calculate the	664	triggers on each midnight, local time. To do this, you would calculate the
587	next midnight after C<now> and return the timestamp value for this. How you do this	665	next midnight after C<now> and return the timestamp value for this. How
588	is, again, up to you (but it is not trivial).	666	you do this is, again, up to you (but it is not trivial, which is the main
		667	reason I omitted it as an example).
589		668
590	=back	669	=back
591		670
592	=item ev_periodic_again (loop, ev_periodic *)	671	=item ev_periodic_again (loop, ev_periodic *)
593		672
…		…
672	=back	751	=back
673		752
674	=head2 C<ev_prepare> and C<ev_check> - customise your event loop	753	=head2 C<ev_prepare> and C<ev_check> - customise your event loop
675		754
676	Prepare and check watchers are usually (but not always) used in tandem:	755	Prepare and check watchers are usually (but not always) used in tandem:
677	Prepare watchers get invoked before the process blocks and check watchers	756	prepare watchers get invoked before the process blocks and check watchers
678	afterwards.	757	afterwards.
679		758
680	Their main purpose is to integrate other event mechanisms into libev. This	759	Their main purpose is to integrate other event mechanisms into libev. This
681	could be used, for example, to track variable changes, implement your own	760	could be used, for example, to track variable changes, implement your own
682	watchers, integrate net-snmp or a coroutine library and lots more.	761	watchers, integrate net-snmp or a coroutine library and lots more.
…		…
685	to be watched by the other library, registering C<ev_io> watchers for	764	to be watched by the other library, registering C<ev_io> watchers for
686	them and starting an C<ev_timer> watcher for any timeouts (many libraries	765	them and starting an C<ev_timer> watcher for any timeouts (many libraries
687	provide just this functionality). Then, in the check watcher you check for	766	provide just this functionality). Then, in the check watcher you check for
688	any events that occured (by checking the pending status of all watchers	767	any events that occured (by checking the pending status of all watchers
689	and stopping them) and call back into the library. The I/O and timer	768	and stopping them) and call back into the library. The I/O and timer
690	callbacks will never actually be called (but must be valid neverthelles,	769	callbacks will never actually be called (but must be valid nevertheless,
691	because you never know, you know?).	770	because you never know, you know?).
692		771
693	As another example, the Perl Coro module uses these hooks to integrate	772	As another example, the Perl Coro module uses these hooks to integrate
694	coroutines into libev programs, by yielding to other active coroutines	773	coroutines into libev programs, by yielding to other active coroutines
695	during each prepare and only letting the process block if no coroutines	774	during each prepare and only letting the process block if no coroutines
696	are ready to run (its actually more complicated, it only runs coroutines	775	are ready to run (it's actually more complicated: it only runs coroutines
697	with priority higher than the event loop and one lower priority once,	776	with priority higher than or equal to the event loop and one coroutine
698	using idle watchers to keep the event loop from blocking if lower-priority	777	of lower priority, but only once, using idle watchers to keep the event
699	coroutines exist, thus mapping low-priority coroutines to idle/background	778	loop from blocking if lower-priority coroutines are active, thus mapping
700	tasks).	779	low-priority coroutines to idle/background tasks).
701		780
702	=over 4	781	=over 4
703		782
704	=item ev_prepare_init (ev_prepare *, callback)	783	=item ev_prepare_init (ev_prepare *, callback)
705		784
…		…
720	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)	799	=item ev_once (loop, int fd, int events, ev_tstamp timeout, callback)
721		800
722	This function combines a simple timer and an I/O watcher, calls your	801	This function combines a simple timer and an I/O watcher, calls your
723	callback on whichever event happens first and automatically stop both	802	callback on whichever event happens first and automatically stop both
724	watchers. This is useful if you want to wait for a single event on an fd	803	watchers. This is useful if you want to wait for a single event on an fd
725	or timeout without havign to allocate/configure/start/stop/free one or	804	or timeout without having to allocate/configure/start/stop/free one or
726	more watchers yourself.	805	more watchers yourself.
727		806
728	If C<fd> is less than 0, then no I/O watcher will be started and events	807	If C<fd> is less than 0, then no I/O watcher will be started and events
729	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and	808	is being ignored. Otherwise, an C<ev_io> watcher for the given C<fd> and
730	C<events> set will be craeted and started.	809	C<events> set will be craeted and started.
…		…
733	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and	812	started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
734	repeat = 0) will be started. While C<0> is a valid timeout, it is of	813	repeat = 0) will be started. While C<0> is a valid timeout, it is of
735	dubious value.	814	dubious value.
736		815
737	The callback has the type C<void (cb)(int revents, void arg)> and gets	816	The callback has the type C<void (cb)(int revents, void arg)> and gets
738	passed an events set like normal event callbacks (with a combination of	817	passed an C<revents> set like normal event callbacks (a combination of
739	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>	818	C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMEOUT>) and the C<arg>
740	value passed to C<ev_once>:	819	value passed to C<ev_once>:
741		820
742	static void stdin_ready (int revents, void *arg)	821	static void stdin_ready (int revents, void *arg)
743	{	822	{
…		…
764		843
765	Feed an event as if the given signal occured (loop must be the default loop!).	844	Feed an event as if the given signal occured (loop must be the default loop!).
766		845
767	=back	846	=back
768		847
		848	=head1 LIBEVENT EMULATION
		849
		850	Libev offers a compatibility emulation layer for libevent. It cannot
		851	emulate the internals of libevent, so here are some usage hints:
		852
		853	=over 4
		854
		855	=item * Use it by including <event.h>, as usual.
		856
		857	=item * The following members are fully supported: ev_base, ev_callback,
		858	ev_arg, ev_fd, ev_res, ev_events.
		859
		860	=item * Avoid using ev_flags and the EVLIST_*-macros, while it is
		861	maintained by libev, it does not work exactly the same way as in libevent (consider
		862	it a private API).
		863
		864	=item * Priorities are not currently supported. Initialising priorities
		865	will fail and all watchers will have the same priority, even though there
		866	is an ev_pri field.
		867
		868	=item * Other members are not supported.
		869
		870	=item * The libev emulation is I<not> ABI compatible to libevent, you need
		871	to use the libev header file and library.
		872
		873	=back
		874
		875	=head1 C++ SUPPORT
		876
		877	TBD.
		878
769	=head1 AUTHOR	879	=head1 AUTHOR
770		880
771	Marc Lehmann <libev@schmorp.de>.	881	Marc Lehmann <libev@schmorp.de>.
772		882

Diff Legend

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

Comparing libev/ev.pod (file contents): Revision 1.17 by root, Mon Nov 12 08:57:03 2007 UTC vs. Revision 1.30 by root, Fri Nov 23 04:36:03 2007 UTC

Diff Legend

Comparing libev/ev.pod (file contents):
Revision 1.17 by root, Mon Nov 12 08:57:03 2007 UTC vs.
Revision 1.30 by root, Fri Nov 23 04:36:03 2007 UTC