ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/ev.pod
(Generate patch)

Comparing libev/ev.pod (file contents):
Revision 1.56 by root, Wed Nov 28 11:15:55 2007 UTC vs.
Revision 1.112 by root, Wed Dec 26 08:06:09 2007 UTC

4 4
5=head1 SYNOPSIS 5=head1 SYNOPSIS
6 6
7 #include <ev.h> 7 #include <ev.h>
8 8
9=head1 EXAMPLE PROGRAM 9=head2 EXAMPLE PROGRAM
10 10
11 #include <ev.h> 11 #include <ev.h>
12 12
13 ev_io stdin_watcher; 13 ev_io stdin_watcher;
14 ev_timer timeout_watcher; 14 ev_timer timeout_watcher;
48 return 0; 48 return 0;
49 } 49 }
50 50
51=head1 DESCRIPTION 51=head1 DESCRIPTION
52 52
53The newest version of this document is also available as a html-formatted
54web page you might find easier to navigate when reading it for the first
55time: L<http://cvs.schmorp.de/libev/ev.html>.
56
53Libev is an event loop: you register interest in certain events (such as a 57Libev is an event loop: you register interest in certain events (such as a
54file descriptor being readable or a timeout occuring), and it will manage 58file descriptor being readable or a timeout occurring), and it will manage
55these event sources and provide your program with events. 59these event sources and provide your program with events.
56 60
57To do this, it must take more or less complete control over your process 61To do this, it must take more or less complete control over your process
58(or thread) by executing the I<event loop> handler, and will then 62(or thread) by executing the I<event loop> handler, and will then
59communicate events via a callback mechanism. 63communicate events via a callback mechanism.
61You register interest in certain events by registering so-called I<event 65You register interest in certain events by registering so-called I<event
62watchers>, which are relatively small C structures you initialise with the 66watchers>, which are relatively small C structures you initialise with the
63details of the event, and then hand it over to libev by I<starting> the 67details of the event, and then hand it over to libev by I<starting> the
64watcher. 68watcher.
65 69
66=head1 FEATURES 70=head2 FEATURES
67 71
68Libev supports C<select>, C<poll>, the linux-specific C<epoll>, the 72Libev supports C<select>, C<poll>, the Linux-specific C<epoll>, the
69bsd-specific C<kqueue> and the solaris-specific event port mechanisms 73BSD-specific C<kqueue> and the Solaris-specific event port mechanisms
70for file descriptor events (C<ev_io>), relative timers (C<ev_timer>), 74for file descriptor events (C<ev_io>), the Linux C<inotify> interface
75(for C<ev_stat>), relative timers (C<ev_timer>), absolute timers
71absolute timers with customised rescheduling (C<ev_periodic>), synchronous 76with customised rescheduling (C<ev_periodic>), synchronous signals
72signals (C<ev_signal>), process status change events (C<ev_child>), and 77(C<ev_signal>), process status change events (C<ev_child>), and event
73event watchers dealing with the event loop mechanism itself (C<ev_idle>, 78watchers dealing with the event loop mechanism itself (C<ev_idle>,
74C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as 79C<ev_embed>, C<ev_prepare> and C<ev_check> watchers) as well as
75file watchers (C<ev_stat>) and even limited support for fork events 80file watchers (C<ev_stat>) and even limited support for fork events
76(C<ev_fork>). 81(C<ev_fork>).
77 82
78It also is quite fast (see this 83It also is quite fast (see this
79L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent 84L<benchmark|http://libev.schmorp.de/bench.html> comparing it to libevent
80for example). 85for example).
81 86
82=head1 CONVENTIONS 87=head2 CONVENTIONS
83 88
84Libev is very configurable. In this manual the default configuration will 89Libev is very configurable. In this manual the default configuration will
85be described, which supports multiple event loops. For more info about 90be described, which supports multiple event loops. For more info about
86various configuration options please have a look at B<EMBED> section in 91various configuration options please have a look at B<EMBED> section in
87this manual. If libev was configured without support for multiple event 92this manual. If libev was configured without support for multiple event
88loops, then all functions taking an initial argument of name C<loop> 93loops, then all functions taking an initial argument of name C<loop>
89(which is always of type C<struct ev_loop *>) will not have this argument. 94(which is always of type C<struct ev_loop *>) will not have this argument.
90 95
91=head1 TIME REPRESENTATION 96=head2 TIME REPRESENTATION
92 97
93Libev represents time as a single floating point number, representing the 98Libev represents time as a single floating point number, representing the
94(fractional) number of seconds since the (POSIX) epoch (somewhere near 99(fractional) number of seconds since the (POSIX) epoch (somewhere near
95the beginning of 1970, details are complicated, don't ask). This type is 100the beginning of 1970, details are complicated, don't ask). This type is
96called C<ev_tstamp>, which is what you should use too. It usually aliases 101called C<ev_tstamp>, which is what you should use too. It usually aliases
97to the C<double> type in C, and when you need to do any calculations on 102to the C<double> type in C, and when you need to do any calculations on
98it, you should treat it as such. 103it, you should treat it as some floatingpoint value. Unlike the name
104component C<stamp> might indicate, it is also used for time differences
105throughout libev.
99 106
100=head1 GLOBAL FUNCTIONS 107=head1 GLOBAL FUNCTIONS
101 108
102These functions can be called anytime, even before initialising the 109These functions can be called anytime, even before initialising the
103library in any way. 110library in any way.
108 115
109Returns the current time as libev would use it. Please note that the 116Returns the current time as libev would use it. Please note that the
110C<ev_now> function is usually faster and also often returns the timestamp 117C<ev_now> function is usually faster and also often returns the timestamp
111you actually want to know. 118you actually want to know.
112 119
120=item ev_sleep (ev_tstamp interval)
121
122Sleep for the given interval: The current thread will be blocked until
123either it is interrupted or the given time interval has passed. Basically
124this is a subsecond-resolution C<sleep ()>.
125
113=item int ev_version_major () 126=item int ev_version_major ()
114 127
115=item int ev_version_minor () 128=item int ev_version_minor ()
116 129
117You can find out the major and minor version numbers of the library 130You can find out the major and minor ABI version numbers of the library
118you linked against by calling the functions C<ev_version_major> and 131you linked against by calling the functions C<ev_version_major> and
119C<ev_version_minor>. If you want, you can compare against the global 132C<ev_version_minor>. If you want, you can compare against the global
120symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the 133symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
121version of the library your program was compiled against. 134version of the library your program was compiled against.
122 135
136These version numbers refer to the ABI version of the library, not the
137release version.
138
123Usually, it's a good idea to terminate if the major versions mismatch, 139Usually, it's a good idea to terminate if the major versions mismatch,
124as this indicates an incompatible change. Minor versions are usually 140as this indicates an incompatible change. Minor versions are usually
125compatible to older versions, so a larger minor version alone is usually 141compatible to older versions, so a larger minor version alone is usually
126not a problem. 142not a problem.
127 143
128Example: Make sure we haven't accidentally been linked against the wrong 144Example: Make sure we haven't accidentally been linked against the wrong
129version. 145version.
162C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for 178C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for
163recommended ones. 179recommended ones.
164 180
165See the description of C<ev_embed> watchers for more info. 181See the description of C<ev_embed> watchers for more info.
166 182
167=item ev_set_allocator (void *(*cb)(void *ptr, size_t size)) 183=item ev_set_allocator (void *(*cb)(void *ptr, long size))
168 184
169Sets the allocation function to use (the prototype and semantics are 185Sets the allocation function to use (the prototype is similar - the
170identical to the realloc C function). It is used to allocate and free 186semantics is identical - to the realloc C function). It is used to
171memory (no surprises here). If it returns zero when memory needs to be 187allocate and free memory (no surprises here). If it returns zero when
172allocated, the library might abort or take some potentially destructive 188memory needs to be allocated, the library might abort or take some
173action. The default is your system realloc function. 189potentially destructive action. The default is your system realloc
190function.
174 191
175You could override this function in high-availability programs to, say, 192You could override this function in high-availability programs to, say,
176free some memory if it cannot allocate memory, to use a special allocator, 193free some memory if it cannot allocate memory, to use a special allocator,
177or even to sleep a while and retry until some memory is available. 194or even to sleep a while and retry until some memory is available.
178 195
264C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will 281C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
265override the flags completely if it is found in the environment. This is 282override the flags completely if it is found in the environment. This is
266useful to try out specific backends to test their performance, or to work 283useful to try out specific backends to test their performance, or to work
267around bugs. 284around bugs.
268 285
286=item C<EVFLAG_FORKCHECK>
287
288Instead of calling C<ev_default_fork> or C<ev_loop_fork> manually after
289a fork, you can also make libev check for a fork in each iteration by
290enabling this flag.
291
292This works by calling C<getpid ()> on every iteration of the loop,
293and thus this might slow down your event loop if you do a lot of loop
294iterations and little real work, but is usually not noticeable (on my
295Linux system for example, C<getpid> is actually a simple 5-insn sequence
296without a syscall and thus I<very> fast, but my Linux system also has
297C<pthread_atfork> which is even faster).
298
299The big advantage of this flag is that you can forget about fork (and
300forget about forgetting to tell libev about forking) when you use this
301flag.
302
303This flag setting cannot be overriden or specified in the C<LIBEV_FLAGS>
304environment variable.
305
269=item C<EVBACKEND_SELECT> (value 1, portable select backend) 306=item C<EVBACKEND_SELECT> (value 1, portable select backend)
270 307
271This is your standard select(2) backend. Not I<completely> standard, as 308This is your standard select(2) backend. Not I<completely> standard, as
272libev tries to roll its own fd_set with no limits on the number of fds, 309libev tries to roll its own fd_set with no limits on the number of fds,
273but if that fails, expect a fairly low limit on the number of fds when 310but if that fails, expect a fairly low limit on the number of fds when
274using this backend. It doesn't scale too well (O(highest_fd)), but its usually 311using this backend. It doesn't scale too well (O(highest_fd)), but its
275the fastest backend for a low number of fds. 312usually the fastest backend for a low number of (low-numbered :) fds.
313
314To get good performance out of this backend you need a high amount of
315parallelity (most of the file descriptors should be busy). If you are
316writing a server, you should C<accept ()> in a loop to accept as many
317connections as possible during one iteration. You might also want to have
318a look at C<ev_set_io_collect_interval ()> to increase the amount of
319readyness notifications you get per iteration.
276 320
277=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows) 321=item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
278 322
279And this is your standard poll(2) backend. It's more complicated than 323And this is your standard poll(2) backend. It's more complicated
280select, but handles sparse fds better and has no artificial limit on the 324than select, but handles sparse fds better and has no artificial
281number of fds you can use (except it will slow down considerably with a 325limit on the number of fds you can use (except it will slow down
282lot of inactive fds). It scales similarly to select, i.e. O(total_fds). 326considerably with a lot of inactive fds). It scales similarly to select,
327i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
328performance tips.
283 329
284=item C<EVBACKEND_EPOLL> (value 4, Linux) 330=item C<EVBACKEND_EPOLL> (value 4, Linux)
285 331
286For few fds, this backend is a bit little slower than poll and select, 332For few fds, this backend is a bit little slower than poll and select,
287but it scales phenomenally better. While poll and select usually scale like 333but it scales phenomenally better. While poll and select usually scale
288O(total_fds) where n is the total number of fds (or the highest fd), epoll scales 334like O(total_fds) where n is the total number of fds (or the highest fd),
289either O(1) or O(active_fds). 335epoll scales either O(1) or O(active_fds). The epoll design has a number
336of shortcomings, such as silently dropping events in some hard-to-detect
337cases and rewiring a syscall per fd change, no fork support and bad
338support for dup.
290 339
291While stopping and starting an I/O watcher in the same iteration will 340While stopping, setting and starting an I/O watcher in the same iteration
292result in some caching, there is still a syscall per such incident 341will result in some caching, there is still a syscall per such incident
293(because the fd could point to a different file description now), so its 342(because the fd could point to a different file description now), so its
294best to avoid that. Also, dup()ed file descriptors might not work very 343best to avoid that. Also, C<dup ()>'ed file descriptors might not work
295well if you register events for both fds. 344very well if you register events for both fds.
296 345
297Please note that epoll sometimes generates spurious notifications, so you 346Please note that epoll sometimes generates spurious notifications, so you
298need to use non-blocking I/O or other means to avoid blocking when no data 347need to use non-blocking I/O or other means to avoid blocking when no data
299(or space) is available. 348(or space) is available.
300 349
350Best performance from this backend is achieved by not unregistering all
351watchers for a file descriptor until it has been closed, if possible, i.e.
352keep at least one watcher active per fd at all times.
353
354While nominally embeddeble in other event loops, this feature is broken in
355all kernel versions tested so far.
356
301=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones) 357=item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
302 358
303Kqueue deserves special mention, as at the time of this writing, it 359Kqueue deserves special mention, as at the time of this writing, it
304was broken on all BSDs except NetBSD (usually it doesn't work with 360was broken on all BSDs except NetBSD (usually it doesn't work reliably
305anything but sockets and pipes, except on Darwin, where of course its 361with anything but sockets and pipes, except on Darwin, where of course
306completely useless). For this reason its not being "autodetected" 362it's completely useless). For this reason it's not being "autodetected"
307unless you explicitly specify it explicitly in the flags (i.e. using 363unless you explicitly specify it explicitly in the flags (i.e. using
308C<EVBACKEND_KQUEUE>). 364C<EVBACKEND_KQUEUE>) or libev was compiled on a known-to-be-good (-enough)
365system like NetBSD.
366
367You still can embed kqueue into a normal poll or select backend and use it
368only for sockets (after having made sure that sockets work with kqueue on
369the target platform). See C<ev_embed> watchers for more info.
309 370
310It scales in the same way as the epoll backend, but the interface to the 371It scales in the same way as the epoll backend, but the interface to the
311kernel is more efficient (which says nothing about its actual speed, of 372kernel is more efficient (which says nothing about its actual speed, of
312course). While starting and stopping an I/O watcher does not cause an 373course). While stopping, setting and starting an I/O watcher does never
313extra syscall as with epoll, it still adds up to four event changes per 374cause an extra syscall as with C<EVBACKEND_EPOLL>, it still adds up to
314incident, so its best to avoid that. 375two event changes per incident, support for C<fork ()> is very bad and it
376drops fds silently in similarly hard-to-detect cases.
377
378This backend usually performs well under most conditions.
379
380While nominally embeddable in other event loops, this doesn't work
381everywhere, so you might need to test for this. And since it is broken
382almost everywhere, you should only use it when you have a lot of sockets
383(for which it usually works), by embedding it into another event loop
384(e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>) and using it only for
385sockets.
315 386
316=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8) 387=item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
317 388
318This is not implemented yet (and might never be). 389This is not implemented yet (and might never be, unless you send me an
390implementation). According to reports, C</dev/poll> only supports sockets
391and is not embeddable, which would limit the usefulness of this backend
392immensely.
319 393
320=item C<EVBACKEND_PORT> (value 32, Solaris 10) 394=item C<EVBACKEND_PORT> (value 32, Solaris 10)
321 395
322This uses the Solaris 10 port mechanism. As with everything on Solaris, 396This uses the Solaris 10 event port mechanism. As with everything on Solaris,
323it's really slow, but it still scales very well (O(active_fds)). 397it's really slow, but it still scales very well (O(active_fds)).
324 398
325Please note that solaris ports can result in a lot of spurious 399Please note that solaris event ports can deliver a lot of spurious
326notifications, so you need to use non-blocking I/O or other means to avoid 400notifications, so you need to use non-blocking I/O or other means to avoid
327blocking when no data (or space) is available. 401blocking when no data (or space) is available.
402
403While this backend scales well, it requires one system call per active
404file descriptor per loop iteration. For small and medium numbers of file
405descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
406might perform better.
328 407
329=item C<EVBACKEND_ALL> 408=item C<EVBACKEND_ALL>
330 409
331Try all backends (even potentially broken ones that wouldn't be tried 410Try all backends (even potentially broken ones that wouldn't be tried
332with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as 411with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
333C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>. 412C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
413
414It is definitely not recommended to use this flag.
334 415
335=back 416=back
336 417
337If one or more of these are ored into the flags value, then only these 418If one or more of these are ored into the flags value, then only these
338backends will be tried (in the reverse order as given here). If none are 419backends will be tried (in the reverse order as given here). If none are
373Destroys the default loop again (frees all memory and kernel state 454Destroys the default loop again (frees all memory and kernel state
374etc.). None of the active event watchers will be stopped in the normal 455etc.). None of the active event watchers will be stopped in the normal
375sense, so e.g. C<ev_is_active> might still return true. It is your 456sense, so e.g. C<ev_is_active> might still return true. It is your
376responsibility to either stop all watchers cleanly yoursef I<before> 457responsibility to either stop all watchers cleanly yoursef I<before>
377calling this function, or cope with the fact afterwards (which is usually 458calling this function, or cope with the fact afterwards (which is usually
378the easiest thing, youc na just ignore the watchers and/or C<free ()> them 459the easiest thing, you can just ignore the watchers and/or C<free ()> them
379for example). 460for example).
461
462Note that certain global state, such as signal state, will not be freed by
463this function, and related watchers (such as signal and child watchers)
464would need to be stopped manually.
465
466In general it is not advisable to call this function except in the
467rare occasion where you really need to free e.g. the signal handling
468pipe fds. If you need dynamically allocated loops it is better to use
469C<ev_loop_new> and C<ev_loop_destroy>).
380 470
381=item ev_loop_destroy (loop) 471=item ev_loop_destroy (loop)
382 472
383Like C<ev_default_destroy>, but destroys an event loop created by an 473Like C<ev_default_destroy>, but destroys an event loop created by an
384earlier call to C<ev_loop_new>. 474earlier call to C<ev_loop_new>.
408 498
409Like C<ev_default_fork>, but acts on an event loop created by 499Like C<ev_default_fork>, but acts on an event loop created by
410C<ev_loop_new>. Yes, you have to call this on every allocated event loop 500C<ev_loop_new>. Yes, you have to call this on every allocated event loop
411after fork, and how you do this is entirely your own problem. 501after fork, and how you do this is entirely your own problem.
412 502
503=item unsigned int ev_loop_count (loop)
504
505Returns the count of loop iterations for the loop, which is identical to
506the number of times libev did poll for new events. It starts at C<0> and
507happily wraps around with enough iterations.
508
509This value can sometimes be useful as a generation counter of sorts (it
510"ticks" the number of loop iterations), as it roughly corresponds with
511C<ev_prepare> and C<ev_check> calls.
512
413=item unsigned int ev_backend (loop) 513=item unsigned int ev_backend (loop)
414 514
415Returns one of the C<EVBACKEND_*> flags indicating the event backend in 515Returns one of the C<EVBACKEND_*> flags indicating the event backend in
416use. 516use.
417 517
419 519
420Returns the current "event loop time", which is the time the event loop 520Returns the current "event loop time", which is the time the event loop
421received events and started processing them. This timestamp does not 521received events and started processing them. This timestamp does not
422change as long as callbacks are being processed, and this is also the base 522change as long as callbacks are being processed, and this is also the base
423time used for relative timers. You can treat it as the timestamp of the 523time used for relative timers. You can treat it as the timestamp of the
424event occuring (or more correctly, libev finding out about it). 524event occurring (or more correctly, libev finding out about it).
425 525
426=item ev_loop (loop, int flags) 526=item ev_loop (loop, int flags)
427 527
428Finally, this is it, the event handler. This function usually is called 528Finally, this is it, the event handler. This function usually is called
429after you initialised all your watchers and you want to start handling 529after you initialised all your watchers and you want to start handling
450libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is 550libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is
451usually a better approach for this kind of thing. 551usually a better approach for this kind of thing.
452 552
453Here are the gory details of what C<ev_loop> does: 553Here are the gory details of what C<ev_loop> does:
454 554
555 - Before the first iteration, call any pending watchers.
455 * If there are no active watchers (reference count is zero), return. 556 * If there are no active watchers (reference count is zero), return.
456 - Queue prepare watchers and then call all outstanding watchers. 557 - Queue all prepare watchers and then call all outstanding watchers.
457 - If we have been forked, recreate the kernel state. 558 - If we have been forked, recreate the kernel state.
458 - Update the kernel state with all outstanding changes. 559 - Update the kernel state with all outstanding changes.
459 - Update the "event loop time". 560 - Update the "event loop time".
460 - Calculate for how long to block. 561 - Calculate for how long to block.
461 - Block the process, waiting for any events. 562 - Block the process, waiting for any events.
512Example: For some weird reason, unregister the above signal handler again. 613Example: For some weird reason, unregister the above signal handler again.
513 614
514 ev_ref (loop); 615 ev_ref (loop);
515 ev_signal_stop (loop, &exitsig); 616 ev_signal_stop (loop, &exitsig);
516 617
618=item ev_set_io_collect_interval (loop, ev_tstamp interval)
619
620=item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
621
622These advanced functions influence the time that libev will spend waiting
623for events. Both are by default C<0>, meaning that libev will try to
624invoke timer/periodic callbacks and I/O callbacks with minimum latency.
625
626Setting these to a higher value (the C<interval> I<must> be >= C<0>)
627allows libev to delay invocation of I/O and timer/periodic callbacks to
628increase efficiency of loop iterations.
629
630The background is that sometimes your program runs just fast enough to
631handle one (or very few) event(s) per loop iteration. While this makes
632the program responsive, it also wastes a lot of CPU time to poll for new
633events, especially with backends like C<select ()> which have a high
634overhead for the actual polling but can deliver many events at once.
635
636By setting a higher I<io collect interval> you allow libev to spend more
637time collecting I/O events, so you can handle more events per iteration,
638at the cost of increasing latency. Timeouts (both C<ev_periodic> and
639C<ev_timer>) will be not affected. Setting this to a non-null value will
640introduce an additional C<ev_sleep ()> call into most loop iterations.
641
642Likewise, by setting a higher I<timeout collect interval> you allow libev
643to spend more time collecting timeouts, at the expense of increased
644latency (the watcher callback will be called later). C<ev_io> watchers
645will not be affected. Setting this to a non-null value will not introduce
646any overhead in libev.
647
648Many (busy) programs can usually benefit by setting the io collect
649interval to a value near C<0.1> or so, which is often enough for
650interactive servers (of course not for games), likewise for timeouts. It
651usually doesn't make much sense to set it to a lower value than C<0.01>,
652as this approsaches the timing granularity of most systems.
653
517=back 654=back
518 655
519 656
520=head1 ANATOMY OF A WATCHER 657=head1 ANATOMY OF A WATCHER
521 658
700=item bool ev_is_pending (ev_TYPE *watcher) 837=item bool ev_is_pending (ev_TYPE *watcher)
701 838
702Returns a true value iff the watcher is pending, (i.e. it has outstanding 839Returns a true value iff the watcher is pending, (i.e. it has outstanding
703events but its callback has not yet been invoked). As long as a watcher 840events but its callback has not yet been invoked). As long as a watcher
704is pending (but not active) you must not call an init function on it (but 841is pending (but not active) you must not call an init function on it (but
705C<ev_TYPE_set> is safe) and you must make sure the watcher is available to 842C<ev_TYPE_set> is safe), you must not change its priority, and you must
706libev (e.g. you cnanot C<free ()> it). 843make sure the watcher is available to libev (e.g. you cannot C<free ()>
844it).
707 845
708=item callback ev_cb (ev_TYPE *watcher) 846=item callback ev_cb (ev_TYPE *watcher)
709 847
710Returns the callback currently set on the watcher. 848Returns the callback currently set on the watcher.
711 849
712=item ev_cb_set (ev_TYPE *watcher, callback) 850=item ev_cb_set (ev_TYPE *watcher, callback)
713 851
714Change the callback. You can change the callback at virtually any time 852Change the callback. You can change the callback at virtually any time
715(modulo threads). 853(modulo threads).
854
855=item ev_set_priority (ev_TYPE *watcher, priority)
856
857=item int ev_priority (ev_TYPE *watcher)
858
859Set and query the priority of the watcher. The priority is a small
860integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
861(default: C<-2>). Pending watchers with higher priority will be invoked
862before watchers with lower priority, but priority will not keep watchers
863from being executed (except for C<ev_idle> watchers).
864
865This means that priorities are I<only> used for ordering callback
866invocation after new events have been received. This is useful, for
867example, to reduce latency after idling, or more often, to bind two
868watchers on the same event and make sure one is called first.
869
870If you need to suppress invocation when higher priority events are pending
871you need to look at C<ev_idle> watchers, which provide this functionality.
872
873You I<must not> change the priority of a watcher as long as it is active or
874pending.
875
876The default priority used by watchers when no priority has been set is
877always C<0>, which is supposed to not be too high and not be too low :).
878
879Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
880fine, as long as you do not mind that the priority value you query might
881or might not have been adjusted to be within valid range.
882
883=item ev_invoke (loop, ev_TYPE *watcher, int revents)
884
885Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
886C<loop> nor C<revents> need to be valid as long as the watcher callback
887can deal with that fact.
888
889=item int ev_clear_pending (loop, ev_TYPE *watcher)
890
891If the watcher is pending, this function returns clears its pending status
892and returns its C<revents> bitset (as if its callback was invoked). If the
893watcher isn't pending it does nothing and returns C<0>.
716 894
717=back 895=back
718 896
719 897
720=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER 898=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
805In general you can register as many read and/or write event watchers per 983In general you can register as many read and/or write event watchers per
806fd as you want (as long as you don't confuse yourself). Setting all file 984fd as you want (as long as you don't confuse yourself). Setting all file
807descriptors to non-blocking mode is also usually a good idea (but not 985descriptors to non-blocking mode is also usually a good idea (but not
808required if you know what you are doing). 986required if you know what you are doing).
809 987
810You have to be careful with dup'ed file descriptors, though. Some backends
811(the linux epoll backend is a notable example) cannot handle dup'ed file
812descriptors correctly if you register interest in two or more fds pointing
813to the same underlying file/socket/etc. description (that is, they share
814the same underlying "file open").
815
816If you must do this, then force the use of a known-to-be-good backend 988If you must do this, then force the use of a known-to-be-good backend
817(at the time of this writing, this includes only C<EVBACKEND_SELECT> and 989(at the time of this writing, this includes only C<EVBACKEND_SELECT> and
818C<EVBACKEND_POLL>). 990C<EVBACKEND_POLL>).
819 991
820Another thing you have to watch out for is that it is quite easy to 992Another thing you have to watch out for is that it is quite easy to
826it is best to always use non-blocking I/O: An extra C<read>(2) returning 998it is best to always use non-blocking I/O: An extra C<read>(2) returning
827C<EAGAIN> is far preferable to a program hanging until some data arrives. 999C<EAGAIN> is far preferable to a program hanging until some data arrives.
828 1000
829If you cannot run the fd in non-blocking mode (for example you should not 1001If you cannot run the fd in non-blocking mode (for example you should not
830play around with an Xlib connection), then you have to seperately re-test 1002play around with an Xlib connection), then you have to seperately re-test
831wether a file descriptor is really ready with a known-to-be good interface 1003whether a file descriptor is really ready with a known-to-be good interface
832such as poll (fortunately in our Xlib example, Xlib already does this on 1004such as poll (fortunately in our Xlib example, Xlib already does this on
833its own, so its quite safe to use). 1005its own, so its quite safe to use).
1006
1007=head3 The special problem of disappearing file descriptors
1008
1009Some backends (e.g. kqueue, epoll) need to be told about closing a file
1010descriptor (either by calling C<close> explicitly or by any other means,
1011such as C<dup>). The reason is that you register interest in some file
1012descriptor, but when it goes away, the operating system will silently drop
1013this interest. If another file descriptor with the same number then is
1014registered with libev, there is no efficient way to see that this is, in
1015fact, a different file descriptor.
1016
1017To avoid having to explicitly tell libev about such cases, libev follows
1018the following policy: Each time C<ev_io_set> is being called, libev
1019will assume that this is potentially a new file descriptor, otherwise
1020it is assumed that the file descriptor stays the same. That means that
1021you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
1022descriptor even if the file descriptor number itself did not change.
1023
1024This is how one would do it normally anyway, the important point is that
1025the libev application should not optimise around libev but should leave
1026optimisations to libev.
1027
1028=head3 The special problem of dup'ed file descriptors
1029
1030Some backends (e.g. epoll), cannot register events for file descriptors,
1031but only events for the underlying file descriptions. That means when you
1032have C<dup ()>'ed file descriptors or weirder constellations, and register
1033events for them, only one file descriptor might actually receive events.
1034
1035There is no workaround possible except not registering events
1036for potentially C<dup ()>'ed file descriptors, or to resort to
1037C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
1038
1039=head3 The special problem of fork
1040
1041Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
1042useless behaviour. Libev fully supports fork, but needs to be told about
1043it in the child.
1044
1045To support fork in your programs, you either have to call
1046C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child,
1047enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or
1048C<EVBACKEND_POLL>.
1049
1050
1051=head3 Watcher-Specific Functions
834 1052
835=over 4 1053=over 4
836 1054
837=item ev_io_init (ev_io *, callback, int fd, int events) 1055=item ev_io_init (ev_io *, callback, int fd, int events)
838 1056
849=item int events [read-only] 1067=item int events [read-only]
850 1068
851The events being watched. 1069The events being watched.
852 1070
853=back 1071=back
1072
1073=head3 Examples
854 1074
855Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well 1075Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
856readable, but only once. Since it is likely line-buffered, you could 1076readable, but only once. Since it is likely line-buffered, you could
857attempt to read a whole line in the callback. 1077attempt to read a whole line in the callback.
858 1078
892 1112
893The callback is guarenteed to be invoked only when its timeout has passed, 1113The callback is guarenteed to be invoked only when its timeout has passed,
894but if multiple timers become ready during the same loop iteration then 1114but if multiple timers become ready during the same loop iteration then
895order of execution is undefined. 1115order of execution is undefined.
896 1116
1117=head3 Watcher-Specific Functions and Data Members
1118
897=over 4 1119=over 4
898 1120
899=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat) 1121=item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
900 1122
901=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat) 1123=item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
914=item ev_timer_again (loop) 1136=item ev_timer_again (loop)
915 1137
916This will act as if the timer timed out and restart it again if it is 1138This will act as if the timer timed out and restart it again if it is
917repeating. The exact semantics are: 1139repeating. The exact semantics are:
918 1140
1141If the timer is pending, its pending status is cleared.
1142
919If the timer is started but nonrepeating, stop it. 1143If the timer is started but nonrepeating, stop it (as if it timed out).
920 1144
921If the timer is repeating, either start it if necessary (with the repeat 1145If the timer is repeating, either start it if necessary (with the
922value), or reset the running timer to the repeat value. 1146C<repeat> value), or reset the running timer to the C<repeat> value.
923 1147
924This sounds a bit complicated, but here is a useful and typical 1148This sounds a bit complicated, but here is a useful and typical
925example: Imagine you have a tcp connection and you want a so-called 1149example: Imagine you have a tcp connection and you want a so-called idle
926idle timeout, that is, you want to be called when there have been, 1150timeout, that is, you want to be called when there have been, say, 60
927say, 60 seconds of inactivity on the socket. The easiest way to do 1151seconds of inactivity on the socket. The easiest way to do this is to
928this is to configure an C<ev_timer> with C<after>=C<repeat>=C<60> and calling 1152configure an C<ev_timer> with a C<repeat> value of C<60> and then call
929C<ev_timer_again> each time you successfully read or write some data. If 1153C<ev_timer_again> each time you successfully read or write some data. If
930you go into an idle state where you do not expect data to travel on the 1154you go into an idle state where you do not expect data to travel on the
931socket, you can stop the timer, and again will automatically restart it if 1155socket, you can C<ev_timer_stop> the timer, and C<ev_timer_again> will
932need be. 1156automatically restart it if need be.
933 1157
934You can also ignore the C<after> value and C<ev_timer_start> altogether 1158That means you can ignore the C<after> value and C<ev_timer_start>
935and only ever use the C<repeat> value: 1159altogether and only ever use the C<repeat> value and C<ev_timer_again>:
936 1160
937 ev_timer_init (timer, callback, 0., 5.); 1161 ev_timer_init (timer, callback, 0., 5.);
938 ev_timer_again (loop, timer); 1162 ev_timer_again (loop, timer);
939 ... 1163 ...
940 timer->again = 17.; 1164 timer->again = 17.;
941 ev_timer_again (loop, timer); 1165 ev_timer_again (loop, timer);
942 ... 1166 ...
943 timer->again = 10.; 1167 timer->again = 10.;
944 ev_timer_again (loop, timer); 1168 ev_timer_again (loop, timer);
945 1169
946This is more efficient then stopping/starting the timer eahc time you want 1170This is more slightly efficient then stopping/starting the timer each time
947to modify its timeout value. 1171you want to modify its timeout value.
948 1172
949=item ev_tstamp repeat [read-write] 1173=item ev_tstamp repeat [read-write]
950 1174
951The current C<repeat> value. Will be used each time the watcher times out 1175The current C<repeat> value. Will be used each time the watcher times out
952or C<ev_timer_again> is called and determines the next timeout (if any), 1176or C<ev_timer_again> is called and determines the next timeout (if any),
953which is also when any modifications are taken into account. 1177which is also when any modifications are taken into account.
954 1178
955=back 1179=back
1180
1181=head3 Examples
956 1182
957Example: Create a timer that fires after 60 seconds. 1183Example: Create a timer that fires after 60 seconds.
958 1184
959 static void 1185 static void
960 one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents) 1186 one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
994but on wallclock time (absolute time). You can tell a periodic watcher 1220but on wallclock time (absolute time). You can tell a periodic watcher
995to trigger "at" some specific point in time. For example, if you tell a 1221to trigger "at" some specific point in time. For example, if you tell a
996periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now () 1222periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now ()
997+ 10.>) and then reset your system clock to the last year, then it will 1223+ 10.>) and then reset your system clock to the last year, then it will
998take a year to trigger the event (unlike an C<ev_timer>, which would trigger 1224take a year to trigger the event (unlike an C<ev_timer>, which would trigger
999roughly 10 seconds later and of course not if you reset your system time 1225roughly 10 seconds later).
1000again).
1001 1226
1002They can also be used to implement vastly more complex timers, such as 1227They can also be used to implement vastly more complex timers, such as
1003triggering an event on eahc midnight, local time. 1228triggering an event on each midnight, local time or other, complicated,
1229rules.
1004 1230
1005As with timers, the callback is guarenteed to be invoked only when the 1231As with timers, the callback is guarenteed to be invoked only when the
1006time (C<at>) has been passed, but if multiple periodic timers become ready 1232time (C<at>) has been passed, but if multiple periodic timers become ready
1007during the same loop iteration then order of execution is undefined. 1233during the same loop iteration then order of execution is undefined.
1008 1234
1235=head3 Watcher-Specific Functions and Data Members
1236
1009=over 4 1237=over 4
1010 1238
1011=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb) 1239=item ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)
1012 1240
1013=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb) 1241=item ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)
1015Lots of arguments, lets sort it out... There are basically three modes of 1243Lots of arguments, lets sort it out... There are basically three modes of
1016operation, and we will explain them from simplest to complex: 1244operation, and we will explain them from simplest to complex:
1017 1245
1018=over 4 1246=over 4
1019 1247
1020=item * absolute timer (interval = reschedule_cb = 0) 1248=item * absolute timer (at = time, interval = reschedule_cb = 0)
1021 1249
1022In this configuration the watcher triggers an event at the wallclock time 1250In this configuration the watcher triggers an event at the wallclock time
1023C<at> and doesn't repeat. It will not adjust when a time jump occurs, 1251C<at> and doesn't repeat. It will not adjust when a time jump occurs,
1024that is, if it is to be run at January 1st 2011 then it will run when the 1252that is, if it is to be run at January 1st 2011 then it will run when the
1025system time reaches or surpasses this time. 1253system time reaches or surpasses this time.
1026 1254
1027=item * non-repeating interval timer (interval > 0, reschedule_cb = 0) 1255=item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)
1028 1256
1029In this mode the watcher will always be scheduled to time out at the next 1257In this mode the watcher will always be scheduled to time out at the next
1030C<at + N * interval> time (for some integer N) and then repeat, regardless 1258C<at + N * interval> time (for some integer N, which can also be negative)
1031of any time jumps. 1259and then repeat, regardless of any time jumps.
1032 1260
1033This can be used to create timers that do not drift with respect to system 1261This can be used to create timers that do not drift with respect to system
1034time: 1262time:
1035 1263
1036 ev_periodic_set (&periodic, 0., 3600., 0); 1264 ev_periodic_set (&periodic, 0., 3600., 0);
1042 1270
1043Another way to think about it (for the mathematically inclined) is that 1271Another way to think about it (for the mathematically inclined) is that
1044C<ev_periodic> will try to run the callback in this mode at the next possible 1272C<ev_periodic> will try to run the callback in this mode at the next possible
1045time where C<time = at (mod interval)>, regardless of any time jumps. 1273time where C<time = at (mod interval)>, regardless of any time jumps.
1046 1274
1275For numerical stability it is preferable that the C<at> value is near
1276C<ev_now ()> (the current time), but there is no range requirement for
1277this value.
1278
1047=item * manual reschedule mode (reschedule_cb = callback) 1279=item * manual reschedule mode (at and interval ignored, reschedule_cb = callback)
1048 1280
1049In this mode the values for C<interval> and C<at> are both being 1281In this mode the values for C<interval> and C<at> are both being
1050ignored. Instead, each time the periodic watcher gets scheduled, the 1282ignored. Instead, each time the periodic watcher gets scheduled, the
1051reschedule callback will be called with the watcher as first, and the 1283reschedule callback will be called with the watcher as first, and the
1052current time as second argument. 1284current time as second argument.
1053 1285
1054NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, 1286NOTE: I<This callback MUST NOT stop or destroy any periodic watcher,
1055ever, or make any event loop modifications>. If you need to stop it, 1287ever, or make any event loop modifications>. If you need to stop it,
1056return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by 1288return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by
1057starting a prepare watcher). 1289starting an C<ev_prepare> watcher, which is legal).
1058 1290
1059Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w, 1291Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
1060ev_tstamp now)>, e.g.: 1292ev_tstamp now)>, e.g.:
1061 1293
1062 static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now) 1294 static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now)
1085Simply stops and restarts the periodic watcher again. This is only useful 1317Simply stops and restarts the periodic watcher again. This is only useful
1086when you changed some parameters or the reschedule callback would return 1318when you changed some parameters or the reschedule callback would return
1087a different time than the last time it was called (e.g. in a crond like 1319a different time than the last time it was called (e.g. in a crond like
1088program when the crontabs have changed). 1320program when the crontabs have changed).
1089 1321
1322=item ev_tstamp offset [read-write]
1323
1324When repeating, this contains the offset value, otherwise this is the
1325absolute point in time (the C<at> value passed to C<ev_periodic_set>).
1326
1327Can be modified any time, but changes only take effect when the periodic
1328timer fires or C<ev_periodic_again> is being called.
1329
1090=item ev_tstamp interval [read-write] 1330=item ev_tstamp interval [read-write]
1091 1331
1092The current interval value. Can be modified any time, but changes only 1332The current interval value. Can be modified any time, but changes only
1093take effect when the periodic timer fires or C<ev_periodic_again> is being 1333take effect when the periodic timer fires or C<ev_periodic_again> is being
1094called. 1334called.
1097 1337
1098The current reschedule callback, or C<0>, if this functionality is 1338The current reschedule callback, or C<0>, if this functionality is
1099switched off. Can be changed any time, but changes only take effect when 1339switched off. Can be changed any time, but changes only take effect when
1100the periodic timer fires or C<ev_periodic_again> is being called. 1340the periodic timer fires or C<ev_periodic_again> is being called.
1101 1341
1342=item ev_tstamp at [read-only]
1343
1344When active, contains the absolute time that the watcher is supposed to
1345trigger next.
1346
1102=back 1347=back
1348
1349=head3 Examples
1103 1350
1104Example: Call a callback every hour, or, more precisely, whenever the 1351Example: Call a callback every hour, or, more precisely, whenever the
1105system clock is divisible by 3600. The callback invocation times have 1352system clock is divisible by 3600. The callback invocation times have
1106potentially a lot of jittering, but good long-term stability. 1353potentially a lot of jittering, but good long-term stability.
1107 1354
1147with the kernel (thus it coexists with your own signal handlers as long 1394with the kernel (thus it coexists with your own signal handlers as long
1148as you don't register any with libev). Similarly, when the last signal 1395as you don't register any with libev). Similarly, when the last signal
1149watcher for a signal is stopped libev will reset the signal handler to 1396watcher for a signal is stopped libev will reset the signal handler to
1150SIG_DFL (regardless of what it was set to before). 1397SIG_DFL (regardless of what it was set to before).
1151 1398
1399=head3 Watcher-Specific Functions and Data Members
1400
1152=over 4 1401=over 4
1153 1402
1154=item ev_signal_init (ev_signal *, callback, int signum) 1403=item ev_signal_init (ev_signal *, callback, int signum)
1155 1404
1156=item ev_signal_set (ev_signal *, int signum) 1405=item ev_signal_set (ev_signal *, int signum)
1167 1416
1168=head2 C<ev_child> - watch out for process status changes 1417=head2 C<ev_child> - watch out for process status changes
1169 1418
1170Child watchers trigger when your process receives a SIGCHLD in response to 1419Child watchers trigger when your process receives a SIGCHLD in response to
1171some child status changes (most typically when a child of yours dies). 1420some child status changes (most typically when a child of yours dies).
1421
1422=head3 Watcher-Specific Functions and Data Members
1172 1423
1173=over 4 1424=over 4
1174 1425
1175=item ev_child_init (ev_child *, callback, int pid) 1426=item ev_child_init (ev_child *, callback, int pid)
1176 1427
1196The process exit/trace status caused by C<rpid> (see your systems 1447The process exit/trace status caused by C<rpid> (see your systems
1197C<waitpid> and C<sys/wait.h> documentation for details). 1448C<waitpid> and C<sys/wait.h> documentation for details).
1198 1449
1199=back 1450=back
1200 1451
1452=head3 Examples
1453
1201Example: Try to exit cleanly on SIGINT and SIGTERM. 1454Example: Try to exit cleanly on SIGINT and SIGTERM.
1202 1455
1203 static void 1456 static void
1204 sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents) 1457 sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1205 { 1458 {
1221not exist" is a status change like any other. The condition "path does 1474not exist" is a status change like any other. The condition "path does
1222not exist" is signified by the C<st_nlink> field being zero (which is 1475not exist" is signified by the C<st_nlink> field being zero (which is
1223otherwise always forced to be at least one) and all the other fields of 1476otherwise always forced to be at least one) and all the other fields of
1224the stat buffer having unspecified contents. 1477the stat buffer having unspecified contents.
1225 1478
1479The path I<should> be absolute and I<must not> end in a slash. If it is
1480relative and your working directory changes, the behaviour is undefined.
1481
1226Since there is no standard to do this, the portable implementation simply 1482Since there is no standard to do this, the portable implementation simply
1227calls C<stat (2)> regulalry on the path to see if it changed somehow. You 1483calls C<stat (2)> regularly on the path to see if it changed somehow. You
1228can specify a recommended polling interval for this case. If you specify 1484can specify a recommended polling interval for this case. If you specify
1229a polling interval of C<0> (highly recommended!) then a I<suitable, 1485a polling interval of C<0> (highly recommended!) then a I<suitable,
1230unspecified default> value will be used (which you can expect to be around 1486unspecified default> value will be used (which you can expect to be around
1231five seconds, although this might change dynamically). Libev will also 1487five seconds, although this might change dynamically). Libev will also
1232impose a minimum interval which is currently around C<0.1>, but thats 1488impose a minimum interval which is currently around C<0.1>, but thats
1234 1490
1235This watcher type is not meant for massive numbers of stat watchers, 1491This watcher type is not meant for massive numbers of stat watchers,
1236as even with OS-supported change notifications, this can be 1492as even with OS-supported change notifications, this can be
1237resource-intensive. 1493resource-intensive.
1238 1494
1239At the time of this writing, no specific OS backends are implemented, but 1495At the time of this writing, only the Linux inotify interface is
1240if demand increases, at least a kqueue and inotify backend will be added. 1496implemented (implementing kqueue support is left as an exercise for the
1497reader). Inotify will be used to give hints only and should not change the
1498semantics of C<ev_stat> watchers, which means that libev sometimes needs
1499to fall back to regular polling again even with inotify, but changes are
1500usually detected immediately, and if the file exists there will be no
1501polling.
1502
1503=head3 Inotify
1504
1505When C<inotify (7)> support has been compiled into libev (generally only
1506available on Linux) and present at runtime, it will be used to speed up
1507change detection where possible. The inotify descriptor will be created lazily
1508when the first C<ev_stat> watcher is being started.
1509
1510Inotify presense does not change the semantics of C<ev_stat> watchers
1511except that changes might be detected earlier, and in some cases, to avoid
1512making regular C<stat> calls. Even in the presense of inotify support
1513there are many cases where libev has to resort to regular C<stat> polling.
1514
1515(There is no support for kqueue, as apparently it cannot be used to
1516implement this functionality, due to the requirement of having a file
1517descriptor open on the object at all times).
1518
1519=head3 The special problem of stat time resolution
1520
1521The C<stat ()> syscall only supports full-second resolution portably, and
1522even on systems where the resolution is higher, many filesystems still
1523only support whole seconds.
1524
1525That means that, if the time is the only thing that changes, you might
1526miss updates: on the first update, C<ev_stat> detects a change and calls
1527your callback, which does something. When there is another update within
1528the same second, C<ev_stat> will be unable to detect it.
1529
1530The solution to this is to delay acting on a change for a second (or till
1531the next second boundary), using a roughly one-second delay C<ev_timer>
1532(C<ev_timer_set (w, 0., 1.01); ev_timer_again (loop, w)>). The C<.01>
1533is added to work around small timing inconsistencies of some operating
1534systems.
1535
1536=head3 Watcher-Specific Functions and Data Members
1241 1537
1242=over 4 1538=over 4
1243 1539
1244=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval) 1540=item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
1245 1541
1281=item const char *path [read-only] 1577=item const char *path [read-only]
1282 1578
1283The filesystem path that is being watched. 1579The filesystem path that is being watched.
1284 1580
1285=back 1581=back
1582
1583=head3 Examples
1286 1584
1287Example: Watch C</etc/passwd> for attribute changes. 1585Example: Watch C</etc/passwd> for attribute changes.
1288 1586
1289 static void 1587 static void
1290 passwd_cb (struct ev_loop *loop, ev_stat *w, int revents) 1588 passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1303 } 1601 }
1304 1602
1305 ... 1603 ...
1306 ev_stat passwd; 1604 ev_stat passwd;
1307 1605
1308 ev_stat_init (&passwd, passwd_cb, "/etc/passwd"); 1606 ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1309 ev_stat_start (loop, &passwd); 1607 ev_stat_start (loop, &passwd);
1310 1608
1609Example: Like above, but additionally use a one-second delay so we do not
1610miss updates (however, frequent updates will delay processing, too, so
1611one might do the work both on C<ev_stat> callback invocation I<and> on
1612C<ev_timer> callback invocation).
1613
1614 static ev_stat passwd;
1615 static ev_timer timer;
1616
1617 static void
1618 timer_cb (EV_P_ ev_timer *w, int revents)
1619 {
1620 ev_timer_stop (EV_A_ w);
1621
1622 /* now it's one second after the most recent passwd change */
1623 }
1624
1625 static void
1626 stat_cb (EV_P_ ev_stat *w, int revents)
1627 {
1628 /* reset the one-second timer */
1629 ev_timer_again (EV_A_ &timer);
1630 }
1631
1632 ...
1633 ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
1634 ev_stat_start (loop, &passwd);
1635 ev_timer_init (&timer, timer_cb, 0., 1.01);
1636
1311 1637
1312=head2 C<ev_idle> - when you've got nothing better to do... 1638=head2 C<ev_idle> - when you've got nothing better to do...
1313 1639
1314Idle watchers trigger events when there are no other events are pending 1640Idle watchers trigger events when no other events of the same or higher
1315(prepare, check and other idle watchers do not count). That is, as long 1641priority are pending (prepare, check and other idle watchers do not
1316as your process is busy handling sockets or timeouts (or even signals, 1642count).
1317imagine) it will not be triggered. But when your process is idle all idle 1643
1318watchers are being called again and again, once per event loop iteration - 1644That is, as long as your process is busy handling sockets or timeouts
1645(or even signals, imagine) of the same or higher priority it will not be
1646triggered. But when your process is idle (or only lower-priority watchers
1647are pending), the idle watchers are being called once per event loop
1319until stopped, that is, or your process receives more events and becomes 1648iteration - until stopped, that is, or your process receives more events
1320busy. 1649and becomes busy again with higher priority stuff.
1321 1650
1322The most noteworthy effect is that as long as any idle watchers are 1651The most noteworthy effect is that as long as any idle watchers are
1323active, the process will not block when waiting for new events. 1652active, the process will not block when waiting for new events.
1324 1653
1325Apart from keeping your process non-blocking (which is a useful 1654Apart from keeping your process non-blocking (which is a useful
1326effect on its own sometimes), idle watchers are a good place to do 1655effect on its own sometimes), idle watchers are a good place to do
1327"pseudo-background processing", or delay processing stuff to after the 1656"pseudo-background processing", or delay processing stuff to after the
1328event loop has handled all outstanding events. 1657event loop has handled all outstanding events.
1329 1658
1659=head3 Watcher-Specific Functions and Data Members
1660
1330=over 4 1661=over 4
1331 1662
1332=item ev_idle_init (ev_signal *, callback) 1663=item ev_idle_init (ev_signal *, callback)
1333 1664
1334Initialises and configures the idle watcher - it has no parameters of any 1665Initialises and configures the idle watcher - it has no parameters of any
1335kind. There is a C<ev_idle_set> macro, but using it is utterly pointless, 1666kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
1336believe me. 1667believe me.
1337 1668
1338=back 1669=back
1670
1671=head3 Examples
1339 1672
1340Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the 1673Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
1341callback, free it. Also, use no error checking, as usual. 1674callback, free it. Also, use no error checking, as usual.
1342 1675
1343 static void 1676 static void
1391with priority higher than or equal to the event loop and one coroutine 1724with priority higher than or equal to the event loop and one coroutine
1392of lower priority, but only once, using idle watchers to keep the event 1725of lower priority, but only once, using idle watchers to keep the event
1393loop from blocking if lower-priority coroutines are active, thus mapping 1726loop from blocking if lower-priority coroutines are active, thus mapping
1394low-priority coroutines to idle/background tasks). 1727low-priority coroutines to idle/background tasks).
1395 1728
1729It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>)
1730priority, to ensure that they are being run before any other watchers
1731after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers,
1732too) should not activate ("feed") events into libev. While libev fully
1733supports this, they will be called before other C<ev_check> watchers
1734did their job. As C<ev_check> watchers are often used to embed other
1735(non-libev) event loops those other event loops might be in an unusable
1736state until their C<ev_check> watcher ran (always remind yourself to
1737coexist peacefully with others).
1738
1739=head3 Watcher-Specific Functions and Data Members
1740
1396=over 4 1741=over 4
1397 1742
1398=item ev_prepare_init (ev_prepare *, callback) 1743=item ev_prepare_init (ev_prepare *, callback)
1399 1744
1400=item ev_check_init (ev_check *, callback) 1745=item ev_check_init (ev_check *, callback)
1403parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set> 1748parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
1404macros, but using them is utterly, utterly and completely pointless. 1749macros, but using them is utterly, utterly and completely pointless.
1405 1750
1406=back 1751=back
1407 1752
1408Example: To include a library such as adns, you would add IO watchers 1753=head3 Examples
1409and a timeout watcher in a prepare handler, as required by libadns, and 1754
1755There are a number of principal ways to embed other event loops or modules
1756into libev. Here are some ideas on how to include libadns into libev
1757(there is a Perl module named C<EV::ADNS> that does this, which you could
1758use for an actually working example. Another Perl module named C<EV::Glib>
1759embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV
1760into the Glib event loop).
1761
1762Method 1: Add IO watchers and a timeout watcher in a prepare handler,
1410in a check watcher, destroy them and call into libadns. What follows is 1763and in a check watcher, destroy them and call into libadns. What follows
1411pseudo-code only of course: 1764is pseudo-code only of course. This requires you to either use a low
1765priority for the check watcher or use C<ev_clear_pending> explicitly, as
1766the callbacks for the IO/timeout watchers might not have been called yet.
1412 1767
1413 static ev_io iow [nfd]; 1768 static ev_io iow [nfd];
1414 static ev_timer tw; 1769 static ev_timer tw;
1415 1770
1416 static void 1771 static void
1417 io_cb (ev_loop *loop, ev_io *w, int revents) 1772 io_cb (ev_loop *loop, ev_io *w, int revents)
1418 { 1773 {
1419 // set the relevant poll flags
1420 // could also call adns_processreadable etc. here
1421 struct pollfd *fd = (struct pollfd *)w->data;
1422 if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1423 if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1424 } 1774 }
1425 1775
1426 // create io watchers for each fd and a timer before blocking 1776 // create io watchers for each fd and a timer before blocking
1427 static void 1777 static void
1428 adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents) 1778 adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1429 { 1779 {
1430 int timeout = 3600000;truct pollfd fds [nfd]; 1780 int timeout = 3600000;
1781 struct pollfd fds [nfd];
1431 // actual code will need to loop here and realloc etc. 1782 // actual code will need to loop here and realloc etc.
1432 adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ())); 1783 adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1433 1784
1434 /* the callback is illegal, but won't be called as we stop during check */ 1785 /* the callback is illegal, but won't be called as we stop during check */
1435 ev_timer_init (&tw, 0, timeout * 1e-3); 1786 ev_timer_init (&tw, 0, timeout * 1e-3);
1436 ev_timer_start (loop, &tw); 1787 ev_timer_start (loop, &tw);
1437 1788
1438 // create on ev_io per pollfd 1789 // create one ev_io per pollfd
1439 for (int i = 0; i < nfd; ++i) 1790 for (int i = 0; i < nfd; ++i)
1440 { 1791 {
1441 ev_io_init (iow + i, io_cb, fds [i].fd, 1792 ev_io_init (iow + i, io_cb, fds [i].fd,
1442 ((fds [i].events & POLLIN ? EV_READ : 0) 1793 ((fds [i].events & POLLIN ? EV_READ : 0)
1443 | (fds [i].events & POLLOUT ? EV_WRITE : 0))); 1794 | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1444 1795
1445 fds [i].revents = 0; 1796 fds [i].revents = 0;
1446 iow [i].data = fds + i;
1447 ev_io_start (loop, iow + i); 1797 ev_io_start (loop, iow + i);
1448 } 1798 }
1449 } 1799 }
1450 1800
1451 // stop all watchers after blocking 1801 // stop all watchers after blocking
1453 adns_check_cb (ev_loop *loop, ev_check *w, int revents) 1803 adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1454 { 1804 {
1455 ev_timer_stop (loop, &tw); 1805 ev_timer_stop (loop, &tw);
1456 1806
1457 for (int i = 0; i < nfd; ++i) 1807 for (int i = 0; i < nfd; ++i)
1808 {
1809 // set the relevant poll flags
1810 // could also call adns_processreadable etc. here
1811 struct pollfd *fd = fds + i;
1812 int revents = ev_clear_pending (iow + i);
1813 if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1814 if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1815
1816 // now stop the watcher
1458 ev_io_stop (loop, iow + i); 1817 ev_io_stop (loop, iow + i);
1818 }
1459 1819
1460 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop)); 1820 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1821 }
1822
1823Method 2: This would be just like method 1, but you run C<adns_afterpoll>
1824in the prepare watcher and would dispose of the check watcher.
1825
1826Method 3: If the module to be embedded supports explicit event
1827notification (adns does), you can also make use of the actual watcher
1828callbacks, and only destroy/create the watchers in the prepare watcher.
1829
1830 static void
1831 timer_cb (EV_P_ ev_timer *w, int revents)
1832 {
1833 adns_state ads = (adns_state)w->data;
1834 update_now (EV_A);
1835
1836 adns_processtimeouts (ads, &tv_now);
1837 }
1838
1839 static void
1840 io_cb (EV_P_ ev_io *w, int revents)
1841 {
1842 adns_state ads = (adns_state)w->data;
1843 update_now (EV_A);
1844
1845 if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
1846 if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
1847 }
1848
1849 // do not ever call adns_afterpoll
1850
1851Method 4: Do not use a prepare or check watcher because the module you
1852want to embed is too inflexible to support it. Instead, youc na override
1853their poll function. The drawback with this solution is that the main
1854loop is now no longer controllable by EV. The C<Glib::EV> module does
1855this.
1856
1857 static gint
1858 event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1859 {
1860 int got_events = 0;
1861
1862 for (n = 0; n < nfds; ++n)
1863 // create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1864
1865 if (timeout >= 0)
1866 // create/start timer
1867
1868 // poll
1869 ev_loop (EV_A_ 0);
1870
1871 // stop timer again
1872 if (timeout >= 0)
1873 ev_timer_stop (EV_A_ &to);
1874
1875 // stop io watchers again - their callbacks should have set
1876 for (n = 0; n < nfds; ++n)
1877 ev_io_stop (EV_A_ iow [n]);
1878
1879 return got_events;
1461 } 1880 }
1462 1881
1463 1882
1464=head2 C<ev_embed> - when one backend isn't enough... 1883=head2 C<ev_embed> - when one backend isn't enough...
1465 1884
1508portable one. 1927portable one.
1509 1928
1510So when you want to use this feature you will always have to be prepared 1929So when you want to use this feature you will always have to be prepared
1511that you cannot get an embeddable loop. The recommended way to get around 1930that you cannot get an embeddable loop. The recommended way to get around
1512this is to have a separate variables for your embeddable loop, try to 1931this is to have a separate variables for your embeddable loop, try to
1513create it, and if that fails, use the normal loop for everything: 1932create it, and if that fails, use the normal loop for everything.
1933
1934=head3 Watcher-Specific Functions and Data Members
1935
1936=over 4
1937
1938=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
1939
1940=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)
1941
1942Configures the watcher to embed the given loop, which must be
1943embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1944invoked automatically, otherwise it is the responsibility of the callback
1945to invoke it (it will continue to be called until the sweep has been done,
1946if you do not want thta, you need to temporarily stop the embed watcher).
1947
1948=item ev_embed_sweep (loop, ev_embed *)
1949
1950Make a single, non-blocking sweep over the embedded loop. This works
1951similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1952apropriate way for embedded loops.
1953
1954=item struct ev_loop *other [read-only]
1955
1956The embedded event loop.
1957
1958=back
1959
1960=head3 Examples
1961
1962Example: Try to get an embeddable event loop and embed it into the default
1963event loop. If that is not possible, use the default loop. The default
1964loop is stored in C<loop_hi>, while the mebeddable loop is stored in
1965C<loop_lo> (which is C<loop_hi> in the acse no embeddable loop can be
1966used).
1514 1967
1515 struct ev_loop *loop_hi = ev_default_init (0); 1968 struct ev_loop *loop_hi = ev_default_init (0);
1516 struct ev_loop *loop_lo = 0; 1969 struct ev_loop *loop_lo = 0;
1517 struct ev_embed embed; 1970 struct ev_embed embed;
1518 1971
1529 ev_embed_start (loop_hi, &embed); 1982 ev_embed_start (loop_hi, &embed);
1530 } 1983 }
1531 else 1984 else
1532 loop_lo = loop_hi; 1985 loop_lo = loop_hi;
1533 1986
1534=over 4 1987Example: Check if kqueue is available but not recommended and create
1988a kqueue backend for use with sockets (which usually work with any
1989kqueue implementation). Store the kqueue/socket-only event loop in
1990C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
1535 1991
1536=item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop) 1992 struct ev_loop *loop = ev_default_init (0);
1993 struct ev_loop *loop_socket = 0;
1994 struct ev_embed embed;
1995
1996 if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
1997 if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
1998 {
1999 ev_embed_init (&embed, 0, loop_socket);
2000 ev_embed_start (loop, &embed);
2001 }
1537 2002
1538=item ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop) 2003 if (!loop_socket)
2004 loop_socket = loop;
1539 2005
1540Configures the watcher to embed the given loop, which must be 2006 // now use loop_socket for all sockets, and loop for everything else
1541embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
1542invoked automatically, otherwise it is the responsibility of the callback
1543to invoke it (it will continue to be called until the sweep has been done,
1544if you do not want thta, you need to temporarily stop the embed watcher).
1545
1546=item ev_embed_sweep (loop, ev_embed *)
1547
1548Make a single, non-blocking sweep over the embedded loop. This works
1549similarly to C<ev_loop (embedded_loop, EVLOOP_NONBLOCK)>, but in the most
1550apropriate way for embedded loops.
1551
1552=item struct ev_loop *loop [read-only]
1553
1554The embedded event loop.
1555
1556=back
1557 2007
1558 2008
1559=head2 C<ev_fork> - the audacity to resume the event loop after a fork 2009=head2 C<ev_fork> - the audacity to resume the event loop after a fork
1560 2010
1561Fork watchers are called when a C<fork ()> was detected (usually because 2011Fork watchers are called when a C<fork ()> was detected (usually because
1564event loop blocks next and before C<ev_check> watchers are being called, 2014event loop blocks next and before C<ev_check> watchers are being called,
1565and only in the child after the fork. If whoever good citizen calling 2015and only in the child after the fork. If whoever good citizen calling
1566C<ev_default_fork> cheats and calls it in the wrong process, the fork 2016C<ev_default_fork> cheats and calls it in the wrong process, the fork
1567handlers will be invoked, too, of course. 2017handlers will be invoked, too, of course.
1568 2018
2019=head3 Watcher-Specific Functions and Data Members
2020
1569=over 4 2021=over 4
1570 2022
1571=item ev_fork_init (ev_signal *, callback) 2023=item ev_fork_init (ev_signal *, callback)
1572 2024
1573Initialises and configures the fork watcher - it has no parameters of any 2025Initialises and configures the fork watcher - it has no parameters of any
1669 2121
1670To use it, 2122To use it,
1671 2123
1672 #include <ev++.h> 2124 #include <ev++.h>
1673 2125
1674(it is not installed by default). This automatically includes F<ev.h> 2126This automatically includes F<ev.h> and puts all of its definitions (many
1675and puts all of its definitions (many of them macros) into the global 2127of them macros) into the global namespace. All C++ specific things are
1676namespace. All C++ specific things are put into the C<ev> namespace. 2128put into the C<ev> namespace. It should support all the same embedding
2129options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
1677 2130
1678It should support all the same embedding options as F<ev.h>, most notably 2131Care has been taken to keep the overhead low. The only data member the C++
1679C<EV_MULTIPLICITY>. 2132classes add (compared to plain C-style watchers) is the event loop pointer
2133that the watcher is associated with (or no additional members at all if
2134you disable C<EV_MULTIPLICITY> when embedding libev).
2135
2136Currently, functions, and static and non-static member functions can be
2137used as callbacks. Other types should be easy to add as long as they only
2138need one additional pointer for context. If you need support for other
2139types of functors please contact the author (preferably after implementing
2140it).
1680 2141
1681Here is a list of things available in the C<ev> namespace: 2142Here is a list of things available in the C<ev> namespace:
1682 2143
1683=over 4 2144=over 4
1684 2145
1700 2161
1701All of those classes have these methods: 2162All of those classes have these methods:
1702 2163
1703=over 4 2164=over 4
1704 2165
1705=item ev::TYPE::TYPE (object *, object::method *) 2166=item ev::TYPE::TYPE ()
1706 2167
1707=item ev::TYPE::TYPE (object *, object::method *, struct ev_loop *) 2168=item ev::TYPE::TYPE (struct ev_loop *)
1708 2169
1709=item ev::TYPE::~TYPE 2170=item ev::TYPE::~TYPE
1710 2171
1711The constructor takes a pointer to an object and a method pointer to 2172The constructor (optionally) takes an event loop to associate the watcher
1712the event handler callback to call in this class. The constructor calls 2173with. If it is omitted, it will use C<EV_DEFAULT>.
1713C<ev_init> for you, which means you have to call the C<set> method 2174
1714before starting it. If you do not specify a loop then the constructor 2175The constructor calls C<ev_init> for you, which means you have to call the
1715automatically associates the default loop with this watcher. 2176C<set> method before starting it.
2177
2178It will not set a callback, however: You have to call the templated C<set>
2179method to set a callback before you can start the watcher.
2180
2181(The reason why you have to use a method is a limitation in C++ which does
2182not allow explicit template arguments for constructors).
1716 2183
1717The destructor automatically stops the watcher if it is active. 2184The destructor automatically stops the watcher if it is active.
2185
2186=item w->set<class, &class::method> (object *)
2187
2188This method sets the callback method to call. The method has to have a
2189signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
2190first argument and the C<revents> as second. The object must be given as
2191parameter and is stored in the C<data> member of the watcher.
2192
2193This method synthesizes efficient thunking code to call your method from
2194the C callback that libev requires. If your compiler can inline your
2195callback (i.e. it is visible to it at the place of the C<set> call and
2196your compiler is good :), then the method will be fully inlined into the
2197thunking function, making it as fast as a direct C callback.
2198
2199Example: simple class declaration and watcher initialisation
2200
2201 struct myclass
2202 {
2203 void io_cb (ev::io &w, int revents) { }
2204 }
2205
2206 myclass obj;
2207 ev::io iow;
2208 iow.set <myclass, &myclass::io_cb> (&obj);
2209
2210=item w->set<function> (void *data = 0)
2211
2212Also sets a callback, but uses a static method or plain function as
2213callback. The optional C<data> argument will be stored in the watcher's
2214C<data> member and is free for you to use.
2215
2216The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
2217
2218See the method-C<set> above for more details.
2219
2220Example:
2221
2222 static void io_cb (ev::io &w, int revents) { }
2223 iow.set <io_cb> ();
1718 2224
1719=item w->set (struct ev_loop *) 2225=item w->set (struct ev_loop *)
1720 2226
1721Associates a different C<struct ev_loop> with this watcher. You can only 2227Associates a different C<struct ev_loop> with this watcher. You can only
1722do this when the watcher is inactive (and not pending either). 2228do this when the watcher is inactive (and not pending either).
1723 2229
1724=item w->set ([args]) 2230=item w->set ([args])
1725 2231
1726Basically the same as C<ev_TYPE_set>, with the same args. Must be 2232Basically the same as C<ev_TYPE_set>, with the same args. Must be
1727called at least once. Unlike the C counterpart, an active watcher gets 2233called at least once. Unlike the C counterpart, an active watcher gets
1728automatically stopped and restarted. 2234automatically stopped and restarted when reconfiguring it with this
2235method.
1729 2236
1730=item w->start () 2237=item w->start ()
1731 2238
1732Starts the watcher. Note that there is no C<loop> argument as the 2239Starts the watcher. Note that there is no C<loop> argument, as the
1733constructor already takes the loop. 2240constructor already stores the event loop.
1734 2241
1735=item w->stop () 2242=item w->stop ()
1736 2243
1737Stops the watcher if it is active. Again, no C<loop> argument. 2244Stops the watcher if it is active. Again, no C<loop> argument.
1738 2245
1739=item w->again () C<ev::timer>, C<ev::periodic> only 2246=item w->again () (C<ev::timer>, C<ev::periodic> only)
1740 2247
1741For C<ev::timer> and C<ev::periodic>, this invokes the corresponding 2248For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
1742C<ev_TYPE_again> function. 2249C<ev_TYPE_again> function.
1743 2250
1744=item w->sweep () C<ev::embed> only 2251=item w->sweep () (C<ev::embed> only)
1745 2252
1746Invokes C<ev_embed_sweep>. 2253Invokes C<ev_embed_sweep>.
1747 2254
1748=item w->update () C<ev::stat> only 2255=item w->update () (C<ev::stat> only)
1749 2256
1750Invokes C<ev_stat_stat>. 2257Invokes C<ev_stat_stat>.
1751 2258
1752=back 2259=back
1753 2260
1763 2270
1764 myclass (); 2271 myclass ();
1765 } 2272 }
1766 2273
1767 myclass::myclass (int fd) 2274 myclass::myclass (int fd)
1768 : io (this, &myclass::io_cb),
1769 idle (this, &myclass::idle_cb)
1770 { 2275 {
2276 io .set <myclass, &myclass::io_cb > (this);
2277 idle.set <myclass, &myclass::idle_cb> (this);
2278
1771 io.start (fd, ev::READ); 2279 io.start (fd, ev::READ);
1772 } 2280 }
1773 2281
1774 2282
1775=head1 MACRO MAGIC 2283=head1 MACRO MAGIC
1776 2284
1777Libev can be compiled with a variety of options, the most fundemantal is 2285Libev can be compiled with a variety of options, the most fundamantal
1778C<EV_MULTIPLICITY>. This option determines wether (most) functions and 2286of which is C<EV_MULTIPLICITY>. This option determines whether (most)
1779callbacks have an initial C<struct ev_loop *> argument. 2287functions and callbacks have an initial C<struct ev_loop *> argument.
1780 2288
1781To make it easier to write programs that cope with either variant, the 2289To make it easier to write programs that cope with either variant, the
1782following macros are defined: 2290following macros are defined:
1783 2291
1784=over 4 2292=over 4
1816Similar to the other two macros, this gives you the value of the default 2324Similar to the other two macros, this gives you the value of the default
1817loop, if multiple loops are supported ("ev loop default"). 2325loop, if multiple loops are supported ("ev loop default").
1818 2326
1819=back 2327=back
1820 2328
1821Example: Declare and initialise a check watcher, working regardless of 2329Example: Declare and initialise a check watcher, utilising the above
1822wether multiple loops are supported or not. 2330macros so it will work regardless of whether multiple loops are supported
2331or not.
1823 2332
1824 static void 2333 static void
1825 check_cb (EV_P_ ev_timer *w, int revents) 2334 check_cb (EV_P_ ev_timer *w, int revents)
1826 { 2335 {
1827 ev_check_stop (EV_A_ w); 2336 ev_check_stop (EV_A_ w);
1830 ev_check check; 2339 ev_check check;
1831 ev_check_init (&check, check_cb); 2340 ev_check_init (&check, check_cb);
1832 ev_check_start (EV_DEFAULT_ &check); 2341 ev_check_start (EV_DEFAULT_ &check);
1833 ev_loop (EV_DEFAULT_ 0); 2342 ev_loop (EV_DEFAULT_ 0);
1834 2343
1835
1836=head1 EMBEDDING 2344=head1 EMBEDDING
1837 2345
1838Libev can (and often is) directly embedded into host 2346Libev can (and often is) directly embedded into host
1839applications. Examples of applications that embed it include the Deliantra 2347applications. Examples of applications that embed it include the Deliantra
1840Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe) 2348Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
1841and rxvt-unicode. 2349and rxvt-unicode.
1842 2350
1843The goal is to enable you to just copy the neecssary files into your 2351The goal is to enable you to just copy the necessary files into your
1844source directory without having to change even a single line in them, so 2352source directory without having to change even a single line in them, so
1845you can easily upgrade by simply copying (or having a checked-out copy of 2353you can easily upgrade by simply copying (or having a checked-out copy of
1846libev somewhere in your source tree). 2354libev somewhere in your source tree).
1847 2355
1848=head2 FILESETS 2356=head2 FILESETS
1879 ev_vars.h 2387 ev_vars.h
1880 ev_wrap.h 2388 ev_wrap.h
1881 2389
1882 ev_win32.c required on win32 platforms only 2390 ev_win32.c required on win32 platforms only
1883 2391
1884 ev_select.c only when select backend is enabled (which is by default) 2392 ev_select.c only when select backend is enabled (which is enabled by default)
1885 ev_poll.c only when poll backend is enabled (disabled by default) 2393 ev_poll.c only when poll backend is enabled (disabled by default)
1886 ev_epoll.c only when the epoll backend is enabled (disabled by default) 2394 ev_epoll.c only when the epoll backend is enabled (disabled by default)
1887 ev_kqueue.c only when the kqueue backend is enabled (disabled by default) 2395 ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
1888 ev_port.c only when the solaris port backend is enabled (disabled by default) 2396 ev_port.c only when the solaris port backend is enabled (disabled by default)
1889 2397
1938 2446
1939If defined to be C<1>, libev will try to detect the availability of the 2447If defined to be C<1>, libev will try to detect the availability of the
1940monotonic clock option at both compiletime and runtime. Otherwise no use 2448monotonic clock option at both compiletime and runtime. Otherwise no use
1941of the monotonic clock option will be attempted. If you enable this, you 2449of the monotonic clock option will be attempted. If you enable this, you
1942usually have to link against librt or something similar. Enabling it when 2450usually have to link against librt or something similar. Enabling it when
1943the functionality isn't available is safe, though, althoguh you have 2451the functionality isn't available is safe, though, although you have
1944to make sure you link against any libraries where the C<clock_gettime> 2452to make sure you link against any libraries where the C<clock_gettime>
1945function is hiding in (often F<-lrt>). 2453function is hiding in (often F<-lrt>).
1946 2454
1947=item EV_USE_REALTIME 2455=item EV_USE_REALTIME
1948 2456
1949If defined to be C<1>, libev will try to detect the availability of the 2457If defined to be C<1>, libev will try to detect the availability of the
1950realtime clock option at compiletime (and assume its availability at 2458realtime clock option at compiletime (and assume its availability at
1951runtime if successful). Otherwise no use of the realtime clock option will 2459runtime if successful). Otherwise no use of the realtime clock option will
1952be attempted. This effectively replaces C<gettimeofday> by C<clock_get 2460be attempted. This effectively replaces C<gettimeofday> by C<clock_get
1953(CLOCK_REALTIME, ...)> and will not normally affect correctness. See tzhe note about libraries 2461(CLOCK_REALTIME, ...)> and will not normally affect correctness. See the
1954in the description of C<EV_USE_MONOTONIC>, though. 2462note about libraries in the description of C<EV_USE_MONOTONIC>, though.
2463
2464=item EV_USE_NANOSLEEP
2465
2466If defined to be C<1>, libev will assume that C<nanosleep ()> is available
2467and will use it for delays. Otherwise it will use C<select ()>.
1955 2468
1956=item EV_USE_SELECT 2469=item EV_USE_SELECT
1957 2470
1958If undefined or defined to be C<1>, libev will compile in support for the 2471If undefined or defined to be C<1>, libev will compile in support for the
1959C<select>(2) backend. No attempt at autodetection will be done: if no 2472C<select>(2) backend. No attempt at autodetection will be done: if no
1977wants osf handles on win32 (this is the case when the select to 2490wants osf handles on win32 (this is the case when the select to
1978be used is the winsock select). This means that it will call 2491be used is the winsock select). This means that it will call
1979C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise, 2492C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
1980it is assumed that all these functions actually work on fds, even 2493it is assumed that all these functions actually work on fds, even
1981on win32. Should not be defined on non-win32 platforms. 2494on win32. Should not be defined on non-win32 platforms.
2495
2496=item EV_FD_TO_WIN32_HANDLE
2497
2498If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
2499file descriptors to socket handles. When not defining this symbol (the
2500default), then libev will call C<_get_osfhandle>, which is usually
2501correct. In some cases, programs use their own file descriptor management,
2502in which case they can provide this function to map fds to socket handles.
1982 2503
1983=item EV_USE_POLL 2504=item EV_USE_POLL
1984 2505
1985If defined to be C<1>, libev will compile in support for the C<poll>(2) 2506If defined to be C<1>, libev will compile in support for the C<poll>(2)
1986backend. Otherwise it will be enabled on non-win32 platforms. It 2507backend. Otherwise it will be enabled on non-win32 platforms. It
2023be detected at runtime. 2544be detected at runtime.
2024 2545
2025=item EV_H 2546=item EV_H
2026 2547
2027The name of the F<ev.h> header file used to include it. The default if 2548The name of the F<ev.h> header file used to include it. The default if
2028undefined is C<< <ev.h> >> in F<event.h> and C<"ev.h"> in F<ev.c>. This 2549undefined is C<"ev.h"> in F<event.h> and F<ev.c>. This can be used to
2029can be used to virtually rename the F<ev.h> header file in case of conflicts. 2550virtually rename the F<ev.h> header file in case of conflicts.
2030 2551
2031=item EV_CONFIG_H 2552=item EV_CONFIG_H
2032 2553
2033If C<EV_STANDALONE> isn't C<1>, this variable can be used to override 2554If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
2034F<ev.c>'s idea of where to find the F<config.h> file, similarly to 2555F<ev.c>'s idea of where to find the F<config.h> file, similarly to
2035C<EV_H>, above. 2556C<EV_H>, above.
2036 2557
2037=item EV_EVENT_H 2558=item EV_EVENT_H
2038 2559
2039Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea 2560Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
2040of how the F<event.h> header can be found. 2561of how the F<event.h> header can be found, the dfeault is C<"event.h">.
2041 2562
2042=item EV_PROTOTYPES 2563=item EV_PROTOTYPES
2043 2564
2044If defined to be C<0>, then F<ev.h> will not define any function 2565If defined to be C<0>, then F<ev.h> will not define any function
2045prototypes, but still define all the structs and other symbols. This is 2566prototypes, but still define all the structs and other symbols. This is
2052will have the C<struct ev_loop *> as first argument, and you can create 2573will have the C<struct ev_loop *> as first argument, and you can create
2053additional independent event loops. Otherwise there will be no support 2574additional independent event loops. Otherwise there will be no support
2054for multiple event loops and there is no first event loop pointer 2575for multiple event loops and there is no first event loop pointer
2055argument. Instead, all functions act on the single default loop. 2576argument. Instead, all functions act on the single default loop.
2056 2577
2578=item EV_MINPRI
2579
2580=item EV_MAXPRI
2581
2582The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
2583C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
2584provide for more priorities by overriding those symbols (usually defined
2585to be C<-2> and C<2>, respectively).
2586
2587When doing priority-based operations, libev usually has to linearly search
2588all the priorities, so having many of them (hundreds) uses a lot of space
2589and time, so using the defaults of five priorities (-2 .. +2) is usually
2590fine.
2591
2592If your embedding app does not need any priorities, defining these both to
2593C<0> will save some memory and cpu.
2594
2057=item EV_PERIODIC_ENABLE 2595=item EV_PERIODIC_ENABLE
2058 2596
2059If undefined or defined to be C<1>, then periodic timers are supported. If 2597If undefined or defined to be C<1>, then periodic timers are supported. If
2598defined to be C<0>, then they are not. Disabling them saves a few kB of
2599code.
2600
2601=item EV_IDLE_ENABLE
2602
2603If undefined or defined to be C<1>, then idle watchers are supported. If
2060defined to be C<0>, then they are not. Disabling them saves a few kB of 2604defined to be C<0>, then they are not. Disabling them saves a few kB of
2061code. 2605code.
2062 2606
2063=item EV_EMBED_ENABLE 2607=item EV_EMBED_ENABLE
2064 2608
2088than enough. If you need to manage thousands of children you might want to 2632than enough. If you need to manage thousands of children you might want to
2089increase this value (I<must> be a power of two). 2633increase this value (I<must> be a power of two).
2090 2634
2091=item EV_INOTIFY_HASHSIZE 2635=item EV_INOTIFY_HASHSIZE
2092 2636
2093C<ev_staz> watchers use a small hash table to distribute workload by 2637C<ev_stat> watchers use a small hash table to distribute workload by
2094inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>), 2638inotify watch id. The default size is C<16> (or C<1> with C<EV_MINIMAL>),
2095usually more than enough. If you need to manage thousands of C<ev_stat> 2639usually more than enough. If you need to manage thousands of C<ev_stat>
2096watchers you might want to increase this value (I<must> be a power of 2640watchers you might want to increase this value (I<must> be a power of
2097two). 2641two).
2098 2642
2115 2659
2116=item ev_set_cb (ev, cb) 2660=item ev_set_cb (ev, cb)
2117 2661
2118Can be used to change the callback member declaration in each watcher, 2662Can be used to change the callback member declaration in each watcher,
2119and the way callbacks are invoked and set. Must expand to a struct member 2663and the way callbacks are invoked and set. Must expand to a struct member
2120definition and a statement, respectively. See the F<ev.v> header file for 2664definition and a statement, respectively. See the F<ev.h> header file for
2121their default definitions. One possible use for overriding these is to 2665their default definitions. One possible use for overriding these is to
2122avoid the C<struct ev_loop *> as first argument in all cases, or to use 2666avoid the C<struct ev_loop *> as first argument in all cases, or to use
2123method calls instead of plain function calls in C++. 2667method calls instead of plain function calls in C++.
2668
2669=head2 EXPORTED API SYMBOLS
2670
2671If you need to re-export the API (e.g. via a dll) and you need a list of
2672exported symbols, you can use the provided F<Symbol.*> files which list
2673all public symbols, one per line:
2674
2675 Symbols.ev for libev proper
2676 Symbols.event for the libevent emulation
2677
2678This can also be used to rename all public symbols to avoid clashes with
2679multiple versions of libev linked together (which is obviously bad in
2680itself, but sometimes it is inconvinient to avoid this).
2681
2682A sed command like this will create wrapper C<#define>'s that you need to
2683include before including F<ev.h>:
2684
2685 <Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
2686
2687This would create a file F<wrap.h> which essentially looks like this:
2688
2689 #define ev_backend myprefix_ev_backend
2690 #define ev_check_start myprefix_ev_check_start
2691 #define ev_check_stop myprefix_ev_check_stop
2692 ...
2124 2693
2125=head2 EXAMPLES 2694=head2 EXAMPLES
2126 2695
2127For a real-world example of a program the includes libev 2696For a real-world example of a program the includes libev
2128verbatim, you can have a look at the EV perl module 2697verbatim, you can have a look at the EV perl module
2131interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file 2700interface) and F<EV.xs> (implementation) files. Only the F<EV.xs> file
2132will be compiled. It is pretty complex because it provides its own header 2701will be compiled. It is pretty complex because it provides its own header
2133file. 2702file.
2134 2703
2135The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file 2704The usage in rxvt-unicode is simpler. It has a F<ev_cpp.h> header file
2136that everybody includes and which overrides some autoconf choices: 2705that everybody includes and which overrides some configure choices:
2137 2706
2707 #define EV_MINIMAL 1
2138 #define EV_USE_POLL 0 2708 #define EV_USE_POLL 0
2139 #define EV_MULTIPLICITY 0 2709 #define EV_MULTIPLICITY 0
2140 #define EV_PERIODICS 0 2710 #define EV_PERIODIC_ENABLE 0
2711 #define EV_STAT_ENABLE 0
2712 #define EV_FORK_ENABLE 0
2141 #define EV_CONFIG_H <config.h> 2713 #define EV_CONFIG_H <config.h>
2714 #define EV_MINPRI 0
2715 #define EV_MAXPRI 0
2142 2716
2143 #include "ev++.h" 2717 #include "ev++.h"
2144 2718
2145And a F<ev_cpp.C> implementation file that contains libev proper and is compiled: 2719And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
2146 2720
2152 2726
2153In this section the complexities of (many of) the algorithms used inside 2727In this section the complexities of (many of) the algorithms used inside
2154libev will be explained. For complexity discussions about backends see the 2728libev will be explained. For complexity discussions about backends see the
2155documentation for C<ev_default_init>. 2729documentation for C<ev_default_init>.
2156 2730
2731All of the following are about amortised time: If an array needs to be
2732extended, libev needs to realloc and move the whole array, but this
2733happens asymptotically never with higher number of elements, so O(1) might
2734mean it might do a lengthy realloc operation in rare cases, but on average
2735it is much faster and asymptotically approaches constant time.
2736
2157=over 4 2737=over 4
2158 2738
2159=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers) 2739=item Starting and stopping timer/periodic watchers: O(log skipped_other_timers)
2160 2740
2741This means that, when you have a watcher that triggers in one hour and
2742there are 100 watchers that would trigger before that then inserting will
2743have to skip roughly seven (C<ld 100>) of these watchers.
2744
2161=item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers) 2745=item Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)
2746
2747That means that changing a timer costs less than removing/adding them
2748as only the relative motion in the event queue has to be paid for.
2162 2749
2163=item Starting io/check/prepare/idle/signal/child watchers: O(1) 2750=item Starting io/check/prepare/idle/signal/child watchers: O(1)
2164 2751
2752These just add the watcher into an array or at the head of a list.
2753
2165=item Stopping check/prepare/idle watchers: O(1) 2754=item Stopping check/prepare/idle watchers: O(1)
2166 2755
2167=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE)) 2756=item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))
2168 2757
2758These watchers are stored in lists then need to be walked to find the
2759correct watcher to remove. The lists are usually short (you don't usually
2760have many watchers waiting for the same fd or signal).
2761
2169=item Finding the next timer per loop iteration: O(1) 2762=item Finding the next timer in each loop iteration: O(1)
2763
2764By virtue of using a binary heap, the next timer is always found at the
2765beginning of the storage array.
2170 2766
2171=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd) 2767=item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)
2172 2768
2173=item Activating one watcher: O(1) 2769A change means an I/O watcher gets started or stopped, which requires
2770libev to recalculate its status (and possibly tell the kernel, depending
2771on backend and wether C<ev_io_set> was used).
2772
2773=item Activating one watcher (putting it into the pending state): O(1)
2774
2775=item Priority handling: O(number_of_priorities)
2776
2777Priorities are implemented by allocating some space for each
2778priority. When doing priority-based operations, libev usually has to
2779linearly search all the priorities, but starting/stopping and activating
2780watchers becomes O(1) w.r.t. prioritiy handling.
2174 2781
2175=back 2782=back
2176 2783
2177 2784
2785=head1 Win32 platform limitations and workarounds
2786
2787Win32 doesn't support any of the standards (e.g. POSIX) that libev
2788requires, and its I/O model is fundamentally incompatible with the POSIX
2789model. Libev still offers limited functionality on this platform in
2790the form of the C<EVBACKEND_SELECT> backend, and only supports socket
2791descriptors. This only applies when using Win32 natively, not when using
2792e.g. cygwin.
2793
2794There is no supported compilation method available on windows except
2795embedding it into other applications.
2796
2797Due to the many, low, and arbitrary limits on the win32 platform and the
2798abysmal performance of winsockets, using a large number of sockets is not
2799recommended (and not reasonable). If your program needs to use more than
2800a hundred or so sockets, then likely it needs to use a totally different
2801implementation for windows, as libev offers the POSIX model, which cannot
2802be implemented efficiently on windows (microsoft monopoly games).
2803
2804=over 4
2805
2806=item The winsocket select function
2807
2808The winsocket C<select> function doesn't follow POSIX in that it requires
2809socket I<handles> and not socket I<file descriptors>. This makes select
2810very inefficient, and also requires a mapping from file descriptors
2811to socket handles. See the discussion of the C<EV_SELECT_USE_FD_SET>,
2812C<EV_SELECT_IS_WINSOCKET> and C<EV_FD_TO_WIN32_HANDLE> preprocessor
2813symbols for more info.
2814
2815The configuration for a "naked" win32 using the microsoft runtime
2816libraries and raw winsocket select is:
2817
2818 #define EV_USE_SELECT 1
2819 #define EV_SELECT_IS_WINSOCKET 1 /* forces EV_SELECT_USE_FD_SET, too */
2820
2821Note that winsockets handling of fd sets is O(n), so you can easily get a
2822complexity in the O(n²) range when using win32.
2823
2824=item Limited number of file descriptors
2825
2826Windows has numerous arbitrary (and low) limits on things. Early versions
2827of winsocket's select only supported waiting for a max. of C<64> handles
2828(probably owning to the fact that all windows kernels can only wait for
2829C<64> things at the same time internally; microsoft recommends spawning a
2830chain of threads and wait for 63 handles and the previous thread in each).
2831
2832Newer versions support more handles, but you need to define C<FD_SETSIZE>
2833to some high number (e.g. C<2048>) before compiling the winsocket select
2834call (which might be in libev or elsewhere, for example, perl does its own
2835select emulation on windows).
2836
2837Another limit is the number of file descriptors in the microsoft runtime
2838libraries, which by default is C<64> (there must be a hidden I<64> fetish
2839or something like this inside microsoft). You can increase this by calling
2840C<_setmaxstdio>, which can increase this limit to C<2048> (another
2841arbitrary limit), but is broken in many versions of the microsoft runtime
2842libraries.
2843
2844This might get you to about C<512> or C<2048> sockets (depending on
2845windows version and/or the phase of the moon). To get more, you need to
2846wrap all I/O functions and provide your own fd management, but the cost of
2847calling select (O(n²)) will likely make this unworkable.
2848
2849=back
2850
2851
2178=head1 AUTHOR 2852=head1 AUTHOR
2179 2853
2180Marc Lehmann <libev@schmorp.de>. 2854Marc Lehmann <libev@schmorp.de>.
2181 2855

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines