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

Comparing libev/ev.pod (file contents):
Revision 1.334 by root, Mon Oct 25 10:30:23 2010 UTC vs.
Revision 1.369 by root, Mon May 30 18:34:28 2011 UTC

58 ev_timer_start (loop, &timeout_watcher); 58 ev_timer_start (loop, &timeout_watcher);
59 59
60 // now wait for events to arrive 60 // now wait for events to arrive
61 ev_run (loop, 0); 61 ev_run (loop, 0);
62 62
63 // unloop was called, so exit 63 // break was called, so exit
64 return 0; 64 return 0;
65 } 65 }
66 66
67=head1 ABOUT THIS DOCUMENT 67=head1 ABOUT THIS DOCUMENT
68 68
241the current system, you would need to look at C<ev_embeddable_backends () 241the current system, you would need to look at C<ev_embeddable_backends ()
242& ev_supported_backends ()>, likewise for recommended ones. 242& ev_supported_backends ()>, likewise for recommended ones.
243 243
244See the description of C<ev_embed> watchers for more info. 244See the description of C<ev_embed> watchers for more info.
245 245
246=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT] 246=item ev_set_allocator (void *(*cb)(void *ptr, long size))
247 247
248Sets the allocation function to use (the prototype is similar - the 248Sets the allocation function to use (the prototype is similar - the
249semantics are identical to the C<realloc> C89/SuS/POSIX function). It is 249semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
250used to allocate and free memory (no surprises here). If it returns zero 250used to allocate and free memory (no surprises here). If it returns zero
251when memory needs to be allocated (C<size != 0>), the library might abort 251when memory needs to be allocated (C<size != 0>), the library might abort
277 } 277 }
278 278
279 ... 279 ...
280 ev_set_allocator (persistent_realloc); 280 ev_set_allocator (persistent_realloc);
281 281
282=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT] 282=item ev_set_syserr_cb (void (*cb)(const char *msg))
283 283
284Set the callback function to call on a retryable system call error (such 284Set the callback function to call on a retryable system call error (such
285as failed select, poll, epoll_wait). The message is a printable string 285as failed select, poll, epoll_wait). The message is a printable string
286indicating the system call or subsystem causing the problem. If this 286indicating the system call or subsystem causing the problem. If this
287callback is set, then libev will expect it to remedy the situation, no 287callback is set, then libev will expect it to remedy the situation, no
299 } 299 }
300 300
301 ... 301 ...
302 ev_set_syserr_cb (fatal_error); 302 ev_set_syserr_cb (fatal_error);
303 303
304=item ev_feed_signal (int signum)
305
306This function can be used to "simulate" a signal receive. It is completely
307safe to call this function at any time, from any context, including signal
308handlers or random threads.
309
310Its main use is to customise signal handling in your process, especially
311in the presence of threads. For example, you could block signals
312by default in all threads (and specifying C<EVFLAG_NOSIGMASK> when
313creating any loops), and in one thread, use C<sigwait> or any other
314mechanism to wait for signals, then "deliver" them to libev by calling
315C<ev_feed_signal>.
316
304=back 317=back
305 318
306=head1 FUNCTIONS CONTROLLING EVENT LOOPS 319=head1 FUNCTIONS CONTROLLING EVENT LOOPS
307 320
308An event loop is described by a C<struct ev_loop *> (the C<struct> is 321An event loop is described by a C<struct ev_loop *> (the C<struct> is
355=item struct ev_loop *ev_loop_new (unsigned int flags) 368=item struct ev_loop *ev_loop_new (unsigned int flags)
356 369
357This will create and initialise a new event loop object. If the loop 370This will create and initialise a new event loop object. If the loop
358could not be initialised, returns false. 371could not be initialised, returns false.
359 372
360Note that this function I<is> thread-safe, and one common way to use 373This function is thread-safe, and one common way to use libev with
361libev with threads is indeed to create one loop per thread, and using the 374threads is indeed to create one loop per thread, and using the default
362default loop in the "main" or "initial" thread. 375loop in the "main" or "initial" thread.
363 376
364The flags argument can be used to specify special behaviour or specific 377The flags argument can be used to specify special behaviour or specific
365backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>). 378backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
366 379
367The following flags are supported: 380The following flags are supported:
402environment variable. 415environment variable.
403 416
404=item C<EVFLAG_NOINOTIFY> 417=item C<EVFLAG_NOINOTIFY>
405 418
406When this flag is specified, then libev will not attempt to use the 419When this flag is specified, then libev will not attempt to use the
407I<inotify> API for it's C<ev_stat> watchers. Apart from debugging and 420I<inotify> API for its C<ev_stat> watchers. Apart from debugging and
408testing, this flag can be useful to conserve inotify file descriptors, as 421testing, this flag can be useful to conserve inotify file descriptors, as
409otherwise each loop using C<ev_stat> watchers consumes one inotify handle. 422otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
410 423
411=item C<EVFLAG_SIGNALFD> 424=item C<EVFLAG_SIGNALFD>
412 425
413When this flag is specified, then libev will attempt to use the 426When this flag is specified, then libev will attempt to use the
414I<signalfd> API for it's C<ev_signal> (and C<ev_child>) watchers. This API 427I<signalfd> API for its C<ev_signal> (and C<ev_child>) watchers. This API
415delivers signals synchronously, which makes it both faster and might make 428delivers signals synchronously, which makes it both faster and might make
416it possible to get the queued signal data. It can also simplify signal 429it possible to get the queued signal data. It can also simplify signal
417handling with threads, as long as you properly block signals in your 430handling with threads, as long as you properly block signals in your
418threads that are not interested in handling them. 431threads that are not interested in handling them.
419 432
420Signalfd will not be used by default as this changes your signal mask, and 433Signalfd will not be used by default as this changes your signal mask, and
421there are a lot of shoddy libraries and programs (glib's threadpool for 434there are a lot of shoddy libraries and programs (glib's threadpool for
422example) that can't properly initialise their signal masks. 435example) that can't properly initialise their signal masks.
436
437=item C<EVFLAG_NOSIGMASK>
438
439When this flag is specified, then libev will avoid to modify the signal
440mask. Specifically, this means you ahve to make sure signals are unblocked
441when you want to receive them.
442
443This behaviour is useful when you want to do your own signal handling, or
444want to handle signals only in specific threads and want to avoid libev
445unblocking the signals.
446
447It's also required by POSIX in a threaded program, as libev calls
448C<sigprocmask>, whose behaviour is officially unspecified.
449
450This flag's behaviour will become the default in future versions of libev.
423 451
424=item C<EVBACKEND_SELECT> (value 1, portable select backend) 452=item C<EVBACKEND_SELECT> (value 1, portable select backend)
425 453
426This is your standard select(2) backend. Not I<completely> standard, as 454This is your standard select(2) backend. Not I<completely> standard, as
427libev tries to roll its own fd_set with no limits on the number of fds, 455libev tries to roll its own fd_set with no limits on the number of fds,
455=item C<EVBACKEND_EPOLL> (value 4, Linux) 483=item C<EVBACKEND_EPOLL> (value 4, Linux)
456 484
457Use the linux-specific epoll(7) interface (for both pre- and post-2.6.9 485Use the linux-specific epoll(7) interface (for both pre- and post-2.6.9
458kernels). 486kernels).
459 487
460For few fds, this backend is a bit little slower than poll and select, 488For few fds, this backend is a bit little slower than poll and select, but
461but it scales phenomenally better. While poll and select usually scale 489it scales phenomenally better. While poll and select usually scale like
462like O(total_fds) where n is the total number of fds (or the highest fd), 490O(total_fds) where total_fds is the total number of fds (or the highest
463epoll scales either O(1) or O(active_fds). 491fd), epoll scales either O(1) or O(active_fds).
464 492
465The epoll mechanism deserves honorable mention as the most misdesigned 493The epoll mechanism deserves honorable mention as the most misdesigned
466of the more advanced event mechanisms: mere annoyances include silently 494of the more advanced event mechanisms: mere annoyances include silently
467dropping file descriptors, requiring a system call per change per file 495dropping file descriptors, requiring a system call per change per file
468descriptor (and unnecessary guessing of parameters), problems with dup and 496descriptor (and unnecessary guessing of parameters), problems with dup,
497returning before the timeout value, resulting in additional iterations
498(and only giving 5ms accuracy while select on the same platform gives
469so on. The biggest issue is fork races, however - if a program forks then 4990.1ms) and so on. The biggest issue is fork races, however - if a program
470I<both> parent and child process have to recreate the epoll set, which can 500forks then I<both> parent and child process have to recreate the epoll
471take considerable time (one syscall per file descriptor) and is of course 501set, which can take considerable time (one syscall per file descriptor)
472hard to detect. 502and is of course hard to detect.
473 503
474Epoll is also notoriously buggy - embedding epoll fds I<should> work, but 504Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
475of course I<doesn't>, and epoll just loves to report events for totally 505of course I<doesn't>, and epoll just loves to report events for totally
476I<different> file descriptors (even already closed ones, so one cannot 506I<different> file descriptors (even already closed ones, so one cannot
477even remove them from the set) than registered in the set (especially 507even remove them from the set) than registered in the set (especially
479employing an additional generation counter and comparing that against the 509employing an additional generation counter and comparing that against the
480events to filter out spurious ones, recreating the set when required. Last 510events to filter out spurious ones, recreating the set when required. Last
481not least, it also refuses to work with some file descriptors which work 511not least, it also refuses to work with some file descriptors which work
482perfectly fine with C<select> (files, many character devices...). 512perfectly fine with C<select> (files, many character devices...).
483 513
514Epoll is truly the train wreck analog among event poll mechanisms,
515a frankenpoll, cobbled together in a hurry, no thought to design or
516interaction with others.
517
484While stopping, setting and starting an I/O watcher in the same iteration 518While stopping, setting and starting an I/O watcher in the same iteration
485will result in some caching, there is still a system call per such 519will result in some caching, there is still a system call per such
486incident (because the same I<file descriptor> could point to a different 520incident (because the same I<file descriptor> could point to a different
487I<file description> now), so its best to avoid that. Also, C<dup ()>'ed 521I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
488file descriptors might not work very well if you register events for both 522file descriptors might not work very well if you register events for both
553=item C<EVBACKEND_PORT> (value 32, Solaris 10) 587=item C<EVBACKEND_PORT> (value 32, Solaris 10)
554 588
555This uses the Solaris 10 event port mechanism. As with everything on Solaris, 589This uses the Solaris 10 event port mechanism. As with everything on Solaris,
556it's really slow, but it still scales very well (O(active_fds)). 590it's really slow, but it still scales very well (O(active_fds)).
557 591
558Please note that Solaris event ports can deliver a lot of spurious
559notifications, so you need to use non-blocking I/O or other means to avoid
560blocking when no data (or space) is available.
561
562While this backend scales well, it requires one system call per active 592While this backend scales well, it requires one system call per active
563file descriptor per loop iteration. For small and medium numbers of file 593file descriptor per loop iteration. For small and medium numbers of file
564descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend 594descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
565might perform better. 595might perform better.
566 596
567On the positive side, with the exception of the spurious readiness 597On the positive side, this backend actually performed fully to
568notifications, this backend actually performed fully to specification
569in all tests and is fully embeddable, which is a rare feat among the 598specification in all tests and is fully embeddable, which is a rare feat
570OS-specific backends (I vastly prefer correctness over speed hacks). 599among the OS-specific backends (I vastly prefer correctness over speed
600hacks).
601
602On the negative side, the interface is I<bizarre> - so bizarre that
603even sun itself gets it wrong in their code examples: The event polling
604function sometimes returning events to the caller even though an error
605occurred, but with no indication whether it has done so or not (yes, it's
606even documented that way) - deadly for edge-triggered interfaces where
607you absolutely have to know whether an event occurred or not because you
608have to re-arm the watcher.
609
610Fortunately libev seems to be able to work around these idiocies.
571 611
572This backend maps C<EV_READ> and C<EV_WRITE> in the same way as 612This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
573C<EVBACKEND_POLL>. 613C<EVBACKEND_POLL>.
574 614
575=item C<EVBACKEND_ALL> 615=item C<EVBACKEND_ALL>
576 616
577Try all backends (even potentially broken ones that wouldn't be tried 617Try all backends (even potentially broken ones that wouldn't be tried
578with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as 618with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
579C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>. 619C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
580 620
581It is definitely not recommended to use this flag. 621It is definitely not recommended to use this flag, use whatever
622C<ev_recommended_backends ()> returns, or simply do not specify a backend
623at all.
624
625=item C<EVBACKEND_MASK>
626
627Not a backend at all, but a mask to select all backend bits from a
628C<flags> value, in case you want to mask out any backends from a flags
629value (e.g. when modifying the C<LIBEV_FLAGS> environment variable).
582 630
583=back 631=back
584 632
585If one or more of the backend flags are or'ed into the flags value, 633If one or more of the backend flags are or'ed into the flags value,
586then only these backends will be tried (in the reverse order as listed 634then only these backends will be tried (in the reverse order as listed
615This function is normally used on loop objects allocated by 663This function is normally used on loop objects allocated by
616C<ev_loop_new>, but it can also be used on the default loop returned by 664C<ev_loop_new>, but it can also be used on the default loop returned by
617C<ev_default_loop>, in which case it is not thread-safe. 665C<ev_default_loop>, in which case it is not thread-safe.
618 666
619Note that it is not advisable to call this function on the default loop 667Note that it is not advisable to call this function on the default loop
620except in the rare occasion where you really need to free it's resources. 668except in the rare occasion where you really need to free its resources.
621If you need dynamically allocated loops it is better to use C<ev_loop_new> 669If you need dynamically allocated loops it is better to use C<ev_loop_new>
622and C<ev_loop_destroy>. 670and C<ev_loop_destroy>.
623 671
624=item ev_loop_fork (loop) 672=item ev_loop_fork (loop)
625 673
673prepare and check phases. 721prepare and check phases.
674 722
675=item unsigned int ev_depth (loop) 723=item unsigned int ev_depth (loop)
676 724
677Returns the number of times C<ev_run> was entered minus the number of 725Returns the number of times C<ev_run> was entered minus the number of
678times C<ev_run> was exited, in other words, the recursion depth. 726times C<ev_run> was exited normally, in other words, the recursion depth.
679 727
680Outside C<ev_run>, this number is zero. In a callback, this number is 728Outside C<ev_run>, this number is zero. In a callback, this number is
681C<1>, unless C<ev_run> was invoked recursively (or from another thread), 729C<1>, unless C<ev_run> was invoked recursively (or from another thread),
682in which case it is higher. 730in which case it is higher.
683 731
684Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread 732Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread,
685etc.), doesn't count as "exit" - consider this as a hint to avoid such 733throwing an exception etc.), doesn't count as "exit" - consider this
686ungentleman-like behaviour unless it's really convenient. 734as a hint to avoid such ungentleman-like behaviour unless it's really
735convenient, in which case it is fully supported.
687 736
688=item unsigned int ev_backend (loop) 737=item unsigned int ev_backend (loop)
689 738
690Returns one of the C<EVBACKEND_*> flags indicating the event backend in 739Returns one of the C<EVBACKEND_*> flags indicating the event backend in
691use. 740use.
753finished (especially in interactive programs), but having a program 802finished (especially in interactive programs), but having a program
754that automatically loops as long as it has to and no longer by virtue 803that automatically loops as long as it has to and no longer by virtue
755of relying on its watchers stopping correctly, that is truly a thing of 804of relying on its watchers stopping correctly, that is truly a thing of
756beauty. 805beauty.
757 806
807This function is also I<mostly> exception-safe - you can break out of
808a C<ev_run> call by calling C<longjmp> in a callback, throwing a C++
809exception and so on. This does not decrement the C<ev_depth> value, nor
810will it clear any outstanding C<EVBREAK_ONE> breaks.
811
758A flags value of C<EVRUN_NOWAIT> will look for new events, will handle 812A flags value of C<EVRUN_NOWAIT> will look for new events, will handle
759those events and any already outstanding ones, but will not wait and 813those events and any already outstanding ones, but will not wait and
760block your process in case there are no events and will return after one 814block your process in case there are no events and will return after one
761iteration of the loop. This is sometimes useful to poll and handle new 815iteration of the loop. This is sometimes useful to poll and handle new
762events while doing lengthy calculations, to keep the program responsive. 816events while doing lengthy calculations, to keep the program responsive.
771This is useful if you are waiting for some external event in conjunction 825This is useful if you are waiting for some external event in conjunction
772with something not expressible using other libev watchers (i.e. "roll your 826with something not expressible using other libev watchers (i.e. "roll your
773own C<ev_run>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is 827own C<ev_run>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
774usually a better approach for this kind of thing. 828usually a better approach for this kind of thing.
775 829
776Here are the gory details of what C<ev_run> does: 830Here are the gory details of what C<ev_run> does (this is for your
831understanding, not a guarantee that things will work exactly like this in
832future versions):
777 833
778 - Increment loop depth. 834 - Increment loop depth.
779 - Reset the ev_break status. 835 - Reset the ev_break status.
780 - Before the first iteration, call any pending watchers. 836 - Before the first iteration, call any pending watchers.
781 LOOP: 837 LOOP:
814anymore. 870anymore.
815 871
816 ... queue jobs here, make sure they register event watchers as long 872 ... queue jobs here, make sure they register event watchers as long
817 ... as they still have work to do (even an idle watcher will do..) 873 ... as they still have work to do (even an idle watcher will do..)
818 ev_run (my_loop, 0); 874 ev_run (my_loop, 0);
819 ... jobs done or somebody called unloop. yeah! 875 ... jobs done or somebody called break. yeah!
820 876
821=item ev_break (loop, how) 877=item ev_break (loop, how)
822 878
823Can be used to make a call to C<ev_run> return early (but only after it 879Can be used to make a call to C<ev_run> return early (but only after it
824has processed all outstanding events). The C<how> argument must be either 880has processed all outstanding events). The C<how> argument must be either
825C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or 881C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or
826C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return. 882C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return.
827 883
828This "unloop state" will be cleared when entering C<ev_run> again. 884This "break state" will be cleared on the next call to C<ev_run>.
829 885
830It is safe to call C<ev_break> from outside any C<ev_run> calls. ##TODO## 886It is safe to call C<ev_break> from outside any C<ev_run> calls, too, in
887which case it will have no effect.
831 888
832=item ev_ref (loop) 889=item ev_ref (loop)
833 890
834=item ev_unref (loop) 891=item ev_unref (loop)
835 892
856running when nothing else is active. 913running when nothing else is active.
857 914
858 ev_signal exitsig; 915 ev_signal exitsig;
859 ev_signal_init (&exitsig, sig_cb, SIGINT); 916 ev_signal_init (&exitsig, sig_cb, SIGINT);
860 ev_signal_start (loop, &exitsig); 917 ev_signal_start (loop, &exitsig);
861 evf_unref (loop); 918 ev_unref (loop);
862 919
863Example: For some weird reason, unregister the above signal handler again. 920Example: For some weird reason, unregister the above signal handler again.
864 921
865 ev_ref (loop); 922 ev_ref (loop);
866 ev_signal_stop (loop, &exitsig); 923 ev_signal_stop (loop, &exitsig);
978See also the locking example in the C<THREADS> section later in this 1035See also the locking example in the C<THREADS> section later in this
979document. 1036document.
980 1037
981=item ev_set_userdata (loop, void *data) 1038=item ev_set_userdata (loop, void *data)
982 1039
983=item ev_userdata (loop) 1040=item void *ev_userdata (loop)
984 1041
985Set and retrieve a single C<void *> associated with a loop. When 1042Set and retrieve a single C<void *> associated with a loop. When
986C<ev_set_userdata> has never been called, then C<ev_userdata> returns 1043C<ev_set_userdata> has never been called, then C<ev_userdata> returns
987C<0.> 1044C<0>.
988 1045
989These two functions can be used to associate arbitrary data with a loop, 1046These two functions can be used to associate arbitrary data with a loop,
990and are intended solely for the C<invoke_pending_cb>, C<release> and 1047and are intended solely for the C<invoke_pending_cb>, C<release> and
991C<acquire> callbacks described above, but of course can be (ab-)used for 1048C<acquire> callbacks described above, but of course can be (ab-)used for
992any other purpose as well. 1049any other purpose as well.
1154programs, though, as the fd could already be closed and reused for another 1211programs, though, as the fd could already be closed and reused for another
1155thing, so beware. 1212thing, so beware.
1156 1213
1157=back 1214=back
1158 1215
1216=head2 GENERIC WATCHER FUNCTIONS
1217
1218=over 4
1219
1220=item C<ev_init> (ev_TYPE *watcher, callback)
1221
1222This macro initialises the generic portion of a watcher. The contents
1223of the watcher object can be arbitrary (so C<malloc> will do). Only
1224the generic parts of the watcher are initialised, you I<need> to call
1225the type-specific C<ev_TYPE_set> macro afterwards to initialise the
1226type-specific parts. For each type there is also a C<ev_TYPE_init> macro
1227which rolls both calls into one.
1228
1229You can reinitialise a watcher at any time as long as it has been stopped
1230(or never started) and there are no pending events outstanding.
1231
1232The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
1233int revents)>.
1234
1235Example: Initialise an C<ev_io> watcher in two steps.
1236
1237 ev_io w;
1238 ev_init (&w, my_cb);
1239 ev_io_set (&w, STDIN_FILENO, EV_READ);
1240
1241=item C<ev_TYPE_set> (ev_TYPE *watcher, [args])
1242
1243This macro initialises the type-specific parts of a watcher. You need to
1244call C<ev_init> at least once before you call this macro, but you can
1245call C<ev_TYPE_set> any number of times. You must not, however, call this
1246macro on a watcher that is active (it can be pending, however, which is a
1247difference to the C<ev_init> macro).
1248
1249Although some watcher types do not have type-specific arguments
1250(e.g. C<ev_prepare>) you still need to call its C<set> macro.
1251
1252See C<ev_init>, above, for an example.
1253
1254=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
1255
1256This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
1257calls into a single call. This is the most convenient method to initialise
1258a watcher. The same limitations apply, of course.
1259
1260Example: Initialise and set an C<ev_io> watcher in one step.
1261
1262 ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
1263
1264=item C<ev_TYPE_start> (loop, ev_TYPE *watcher)
1265
1266Starts (activates) the given watcher. Only active watchers will receive
1267events. If the watcher is already active nothing will happen.
1268
1269Example: Start the C<ev_io> watcher that is being abused as example in this
1270whole section.
1271
1272 ev_io_start (EV_DEFAULT_UC, &w);
1273
1274=item C<ev_TYPE_stop> (loop, ev_TYPE *watcher)
1275
1276Stops the given watcher if active, and clears the pending status (whether
1277the watcher was active or not).
1278
1279It is possible that stopped watchers are pending - for example,
1280non-repeating timers are being stopped when they become pending - but
1281calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
1282pending. If you want to free or reuse the memory used by the watcher it is
1283therefore a good idea to always call its C<ev_TYPE_stop> function.
1284
1285=item bool ev_is_active (ev_TYPE *watcher)
1286
1287Returns a true value iff the watcher is active (i.e. it has been started
1288and not yet been stopped). As long as a watcher is active you must not modify
1289it.
1290
1291=item bool ev_is_pending (ev_TYPE *watcher)
1292
1293Returns a true value iff the watcher is pending, (i.e. it has outstanding
1294events but its callback has not yet been invoked). As long as a watcher
1295is pending (but not active) you must not call an init function on it (but
1296C<ev_TYPE_set> is safe), you must not change its priority, and you must
1297make sure the watcher is available to libev (e.g. you cannot C<free ()>
1298it).
1299
1300=item callback ev_cb (ev_TYPE *watcher)
1301
1302Returns the callback currently set on the watcher.
1303
1304=item ev_cb_set (ev_TYPE *watcher, callback)
1305
1306Change the callback. You can change the callback at virtually any time
1307(modulo threads).
1308
1309=item ev_set_priority (ev_TYPE *watcher, int priority)
1310
1311=item int ev_priority (ev_TYPE *watcher)
1312
1313Set and query the priority of the watcher. The priority is a small
1314integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
1315(default: C<-2>). Pending watchers with higher priority will be invoked
1316before watchers with lower priority, but priority will not keep watchers
1317from being executed (except for C<ev_idle> watchers).
1318
1319If you need to suppress invocation when higher priority events are pending
1320you need to look at C<ev_idle> watchers, which provide this functionality.
1321
1322You I<must not> change the priority of a watcher as long as it is active or
1323pending.
1324
1325Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
1326fine, as long as you do not mind that the priority value you query might
1327or might not have been clamped to the valid range.
1328
1329The default priority used by watchers when no priority has been set is
1330always C<0>, which is supposed to not be too high and not be too low :).
1331
1332See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
1333priorities.
1334
1335=item ev_invoke (loop, ev_TYPE *watcher, int revents)
1336
1337Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
1338C<loop> nor C<revents> need to be valid as long as the watcher callback
1339can deal with that fact, as both are simply passed through to the
1340callback.
1341
1342=item int ev_clear_pending (loop, ev_TYPE *watcher)
1343
1344If the watcher is pending, this function clears its pending status and
1345returns its C<revents> bitset (as if its callback was invoked). If the
1346watcher isn't pending it does nothing and returns C<0>.
1347
1348Sometimes it can be useful to "poll" a watcher instead of waiting for its
1349callback to be invoked, which can be accomplished with this function.
1350
1351=item ev_feed_event (loop, ev_TYPE *watcher, int revents)
1352
1353Feeds the given event set into the event loop, as if the specified event
1354had happened for the specified watcher (which must be a pointer to an
1355initialised but not necessarily started event watcher). Obviously you must
1356not free the watcher as long as it has pending events.
1357
1358Stopping the watcher, letting libev invoke it, or calling
1359C<ev_clear_pending> will clear the pending event, even if the watcher was
1360not started in the first place.
1361
1362See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
1363functions that do not need a watcher.
1364
1365=back
1366
1367See also the L<ASSOCIATING CUSTOM DATA WITH A WATCHER> and L<BUILDING YOUR
1368OWN COMPOSITE WATCHERS> idioms.
1369
1159=head2 WATCHER STATES 1370=head2 WATCHER STATES
1160 1371
1161There are various watcher states mentioned throughout this manual - 1372There are various watcher states mentioned throughout this manual -
1162active, pending and so on. In this section these states and the rules to 1373active, pending and so on. In this section these states and the rules to
1163transition between them will be described in more detail - and while these 1374transition between them will be described in more detail - and while these
1169 1380
1170Before a watcher can be registered with the event looop it has to be 1381Before a watcher can be registered with the event looop it has to be
1171initialised. This can be done with a call to C<ev_TYPE_init>, or calls to 1382initialised. This can be done with a call to C<ev_TYPE_init>, or calls to
1172C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function. 1383C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function.
1173 1384
1174In this state it is simply some block of memory that is suitable for use 1385In this state it is simply some block of memory that is suitable for
1175in an event loop. It can be moved around, freed, reused etc. at will. 1386use in an event loop. It can be moved around, freed, reused etc. at
1387will - as long as you either keep the memory contents intact, or call
1388C<ev_TYPE_init> again.
1176 1389
1177=item started/running/active 1390=item started/running/active
1178 1391
1179Once a watcher has been started with a call to C<ev_TYPE_start> it becomes 1392Once a watcher has been started with a call to C<ev_TYPE_start> it becomes
1180property of the event loop, and is actively waiting for events. While in 1393property of the event loop, and is actively waiting for events. While in
1208latter will clear any pending state the watcher might be in, regardless 1421latter will clear any pending state the watcher might be in, regardless
1209of whether it was active or not, so stopping a watcher explicitly before 1422of whether it was active or not, so stopping a watcher explicitly before
1210freeing it is often a good idea. 1423freeing it is often a good idea.
1211 1424
1212While stopped (and not pending) the watcher is essentially in the 1425While stopped (and not pending) the watcher is essentially in the
1213initialised state, that is it can be reused, moved, modified in any way 1426initialised state, that is, it can be reused, moved, modified in any way
1214you wish. 1427you wish (but when you trash the memory block, you need to C<ev_TYPE_init>
1428it again).
1215 1429
1216=back 1430=back
1217
1218=head2 GENERIC WATCHER FUNCTIONS
1219
1220=over 4
1221
1222=item C<ev_init> (ev_TYPE *watcher, callback)
1223
1224This macro initialises the generic portion of a watcher. The contents
1225of the watcher object can be arbitrary (so C<malloc> will do). Only
1226the generic parts of the watcher are initialised, you I<need> to call
1227the type-specific C<ev_TYPE_set> macro afterwards to initialise the
1228type-specific parts. For each type there is also a C<ev_TYPE_init> macro
1229which rolls both calls into one.
1230
1231You can reinitialise a watcher at any time as long as it has been stopped
1232(or never started) and there are no pending events outstanding.
1233
1234The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
1235int revents)>.
1236
1237Example: Initialise an C<ev_io> watcher in two steps.
1238
1239 ev_io w;
1240 ev_init (&w, my_cb);
1241 ev_io_set (&w, STDIN_FILENO, EV_READ);
1242
1243=item C<ev_TYPE_set> (ev_TYPE *watcher, [args])
1244
1245This macro initialises the type-specific parts of a watcher. You need to
1246call C<ev_init> at least once before you call this macro, but you can
1247call C<ev_TYPE_set> any number of times. You must not, however, call this
1248macro on a watcher that is active (it can be pending, however, which is a
1249difference to the C<ev_init> macro).
1250
1251Although some watcher types do not have type-specific arguments
1252(e.g. C<ev_prepare>) you still need to call its C<set> macro.
1253
1254See C<ev_init>, above, for an example.
1255
1256=item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
1257
1258This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
1259calls into a single call. This is the most convenient method to initialise
1260a watcher. The same limitations apply, of course.
1261
1262Example: Initialise and set an C<ev_io> watcher in one step.
1263
1264 ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
1265
1266=item C<ev_TYPE_start> (loop, ev_TYPE *watcher)
1267
1268Starts (activates) the given watcher. Only active watchers will receive
1269events. If the watcher is already active nothing will happen.
1270
1271Example: Start the C<ev_io> watcher that is being abused as example in this
1272whole section.
1273
1274 ev_io_start (EV_DEFAULT_UC, &w);
1275
1276=item C<ev_TYPE_stop> (loop, ev_TYPE *watcher)
1277
1278Stops the given watcher if active, and clears the pending status (whether
1279the watcher was active or not).
1280
1281It is possible that stopped watchers are pending - for example,
1282non-repeating timers are being stopped when they become pending - but
1283calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
1284pending. If you want to free or reuse the memory used by the watcher it is
1285therefore a good idea to always call its C<ev_TYPE_stop> function.
1286
1287=item bool ev_is_active (ev_TYPE *watcher)
1288
1289Returns a true value iff the watcher is active (i.e. it has been started
1290and not yet been stopped). As long as a watcher is active you must not modify
1291it.
1292
1293=item bool ev_is_pending (ev_TYPE *watcher)
1294
1295Returns a true value iff the watcher is pending, (i.e. it has outstanding
1296events but its callback has not yet been invoked). As long as a watcher
1297is pending (but not active) you must not call an init function on it (but
1298C<ev_TYPE_set> is safe), you must not change its priority, and you must
1299make sure the watcher is available to libev (e.g. you cannot C<free ()>
1300it).
1301
1302=item callback ev_cb (ev_TYPE *watcher)
1303
1304Returns the callback currently set on the watcher.
1305
1306=item ev_cb_set (ev_TYPE *watcher, callback)
1307
1308Change the callback. You can change the callback at virtually any time
1309(modulo threads).
1310
1311=item ev_set_priority (ev_TYPE *watcher, int priority)
1312
1313=item int ev_priority (ev_TYPE *watcher)
1314
1315Set and query the priority of the watcher. The priority is a small
1316integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
1317(default: C<-2>). Pending watchers with higher priority will be invoked
1318before watchers with lower priority, but priority will not keep watchers
1319from being executed (except for C<ev_idle> watchers).
1320
1321If you need to suppress invocation when higher priority events are pending
1322you need to look at C<ev_idle> watchers, which provide this functionality.
1323
1324You I<must not> change the priority of a watcher as long as it is active or
1325pending.
1326
1327Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
1328fine, as long as you do not mind that the priority value you query might
1329or might not have been clamped to the valid range.
1330
1331The default priority used by watchers when no priority has been set is
1332always C<0>, which is supposed to not be too high and not be too low :).
1333
1334See L<WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
1335priorities.
1336
1337=item ev_invoke (loop, ev_TYPE *watcher, int revents)
1338
1339Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
1340C<loop> nor C<revents> need to be valid as long as the watcher callback
1341can deal with that fact, as both are simply passed through to the
1342callback.
1343
1344=item int ev_clear_pending (loop, ev_TYPE *watcher)
1345
1346If the watcher is pending, this function clears its pending status and
1347returns its C<revents> bitset (as if its callback was invoked). If the
1348watcher isn't pending it does nothing and returns C<0>.
1349
1350Sometimes it can be useful to "poll" a watcher instead of waiting for its
1351callback to be invoked, which can be accomplished with this function.
1352
1353=item ev_feed_event (loop, ev_TYPE *watcher, int revents)
1354
1355Feeds the given event set into the event loop, as if the specified event
1356had happened for the specified watcher (which must be a pointer to an
1357initialised but not necessarily started event watcher). Obviously you must
1358not free the watcher as long as it has pending events.
1359
1360Stopping the watcher, letting libev invoke it, or calling
1361C<ev_clear_pending> will clear the pending event, even if the watcher was
1362not started in the first place.
1363
1364See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
1365functions that do not need a watcher.
1366
1367=back
1368
1369
1370=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
1371
1372Each watcher has, by default, a member C<void *data> that you can change
1373and read at any time: libev will completely ignore it. This can be used
1374to associate arbitrary data with your watcher. If you need more data and
1375don't want to allocate memory and store a pointer to it in that data
1376member, you can also "subclass" the watcher type and provide your own
1377data:
1378
1379 struct my_io
1380 {
1381 ev_io io;
1382 int otherfd;
1383 void *somedata;
1384 struct whatever *mostinteresting;
1385 };
1386
1387 ...
1388 struct my_io w;
1389 ev_io_init (&w.io, my_cb, fd, EV_READ);
1390
1391And since your callback will be called with a pointer to the watcher, you
1392can cast it back to your own type:
1393
1394 static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
1395 {
1396 struct my_io *w = (struct my_io *)w_;
1397 ...
1398 }
1399
1400More interesting and less C-conformant ways of casting your callback type
1401instead have been omitted.
1402
1403Another common scenario is to use some data structure with multiple
1404embedded watchers:
1405
1406 struct my_biggy
1407 {
1408 int some_data;
1409 ev_timer t1;
1410 ev_timer t2;
1411 }
1412
1413In this case getting the pointer to C<my_biggy> is a bit more
1414complicated: Either you store the address of your C<my_biggy> struct
1415in the C<data> member of the watcher (for woozies), or you need to use
1416some pointer arithmetic using C<offsetof> inside your watchers (for real
1417programmers):
1418
1419 #include <stddef.h>
1420
1421 static void
1422 t1_cb (EV_P_ ev_timer *w, int revents)
1423 {
1424 struct my_biggy big = (struct my_biggy *)
1425 (((char *)w) - offsetof (struct my_biggy, t1));
1426 }
1427
1428 static void
1429 t2_cb (EV_P_ ev_timer *w, int revents)
1430 {
1431 struct my_biggy big = (struct my_biggy *)
1432 (((char *)w) - offsetof (struct my_biggy, t2));
1433 }
1434 1431
1435=head2 WATCHER PRIORITY MODELS 1432=head2 WATCHER PRIORITY MODELS
1436 1433
1437Many event loops support I<watcher priorities>, which are usually small 1434Many event loops support I<watcher priorities>, which are usually small
1438integers that influence the ordering of event callback invocation 1435integers that influence the ordering of event callback invocation
1565In general you can register as many read and/or write event watchers per 1562In general you can register as many read and/or write event watchers per
1566fd as you want (as long as you don't confuse yourself). Setting all file 1563fd as you want (as long as you don't confuse yourself). Setting all file
1567descriptors to non-blocking mode is also usually a good idea (but not 1564descriptors to non-blocking mode is also usually a good idea (but not
1568required if you know what you are doing). 1565required if you know what you are doing).
1569 1566
1570If you cannot use non-blocking mode, then force the use of a
1571known-to-be-good backend (at the time of this writing, this includes only
1572C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
1573descriptors for which non-blocking operation makes no sense (such as
1574files) - libev doesn't guarantee any specific behaviour in that case.
1575
1576Another thing you have to watch out for is that it is quite easy to 1567Another thing you have to watch out for is that it is quite easy to
1577receive "spurious" readiness notifications, that is your callback might 1568receive "spurious" readiness notifications, that is, your callback might
1578be called with C<EV_READ> but a subsequent C<read>(2) will actually block 1569be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1579because there is no data. Not only are some backends known to create a 1570because there is no data. It is very easy to get into this situation even
1580lot of those (for example Solaris ports), it is very easy to get into 1571with a relatively standard program structure. Thus it is best to always
1581this situation even with a relatively standard program structure. Thus 1572use non-blocking I/O: An extra C<read>(2) returning C<EAGAIN> is far
1582it is best to always use non-blocking I/O: An extra C<read>(2) returning
1583C<EAGAIN> is far preferable to a program hanging until some data arrives. 1573preferable to a program hanging until some data arrives.
1584 1574
1585If you cannot run the fd in non-blocking mode (for example you should 1575If you cannot run the fd in non-blocking mode (for example you should
1586not play around with an Xlib connection), then you have to separately 1576not play around with an Xlib connection), then you have to separately
1587re-test whether a file descriptor is really ready with a known-to-be good 1577re-test whether a file descriptor is really ready with a known-to-be good
1588interface such as poll (fortunately in our Xlib example, Xlib already 1578interface such as poll (fortunately in the case of Xlib, it already does
1589does this on its own, so its quite safe to use). Some people additionally 1579this on its own, so its quite safe to use). Some people additionally
1590use C<SIGALRM> and an interval timer, just to be sure you won't block 1580use C<SIGALRM> and an interval timer, just to be sure you won't block
1591indefinitely. 1581indefinitely.
1592 1582
1593But really, best use non-blocking mode. 1583But really, best use non-blocking mode.
1594 1584
1622 1612
1623There is no workaround possible except not registering events 1613There is no workaround possible except not registering events
1624for potentially C<dup ()>'ed file descriptors, or to resort to 1614for potentially C<dup ()>'ed file descriptors, or to resort to
1625C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>. 1615C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
1626 1616
1617=head3 The special problem of files
1618
1619Many people try to use C<select> (or libev) on file descriptors
1620representing files, and expect it to become ready when their program
1621doesn't block on disk accesses (which can take a long time on their own).
1622
1623However, this cannot ever work in the "expected" way - you get a readiness
1624notification as soon as the kernel knows whether and how much data is
1625there, and in the case of open files, that's always the case, so you
1626always get a readiness notification instantly, and your read (or possibly
1627write) will still block on the disk I/O.
1628
1629Another way to view it is that in the case of sockets, pipes, character
1630devices and so on, there is another party (the sender) that delivers data
1631on its own, but in the case of files, there is no such thing: the disk
1632will not send data on its own, simply because it doesn't know what you
1633wish to read - you would first have to request some data.
1634
1635Since files are typically not-so-well supported by advanced notification
1636mechanism, libev tries hard to emulate POSIX behaviour with respect
1637to files, even though you should not use it. The reason for this is
1638convenience: sometimes you want to watch STDIN or STDOUT, which is
1639usually a tty, often a pipe, but also sometimes files or special devices
1640(for example, C<epoll> on Linux works with F</dev/random> but not with
1641F</dev/urandom>), and even though the file might better be served with
1642asynchronous I/O instead of with non-blocking I/O, it is still useful when
1643it "just works" instead of freezing.
1644
1645So avoid file descriptors pointing to files when you know it (e.g. use
1646libeio), but use them when it is convenient, e.g. for STDIN/STDOUT, or
1647when you rarely read from a file instead of from a socket, and want to
1648reuse the same code path.
1649
1627=head3 The special problem of fork 1650=head3 The special problem of fork
1628 1651
1629Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit 1652Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
1630useless behaviour. Libev fully supports fork, but needs to be told about 1653useless behaviour. Libev fully supports fork, but needs to be told about
1631it in the child. 1654it in the child if you want to continue to use it in the child.
1632 1655
1633To support fork in your programs, you either have to call 1656To support fork in your child processes, you have to call C<ev_loop_fork
1634C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child, 1657()> after a fork in the child, enable C<EVFLAG_FORKCHECK>, or resort to
1635enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or 1658C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
1636C<EVBACKEND_POLL>.
1637 1659
1638=head3 The special problem of SIGPIPE 1660=head3 The special problem of SIGPIPE
1639 1661
1640While not really specific to libev, it is easy to forget about C<SIGPIPE>: 1662While not really specific to libev, it is easy to forget about C<SIGPIPE>:
1641when writing to a pipe whose other end has been closed, your program gets 1663when writing to a pipe whose other end has been closed, your program gets
2131 2153
2132Another way to think about it (for the mathematically inclined) is that 2154Another way to think about it (for the mathematically inclined) is that
2133C<ev_periodic> will try to run the callback in this mode at the next possible 2155C<ev_periodic> will try to run the callback in this mode at the next possible
2134time where C<time = offset (mod interval)>, regardless of any time jumps. 2156time where C<time = offset (mod interval)>, regardless of any time jumps.
2135 2157
2136For numerical stability it is preferable that the C<offset> value is near 2158The C<interval> I<MUST> be positive, and for numerical stability, the
2137C<ev_now ()> (the current time), but there is no range requirement for 2159interval value should be higher than C<1/8192> (which is around 100
2138this value, and in fact is often specified as zero. 2160microseconds) and C<offset> should be higher than C<0> and should have
2161at most a similar magnitude as the current time (say, within a factor of
2162ten). Typical values for offset are, in fact, C<0> or something between
2163C<0> and C<interval>, which is also the recommended range.
2139 2164
2140Note also that there is an upper limit to how often a timer can fire (CPU 2165Note also that there is an upper limit to how often a timer can fire (CPU
2141speed for example), so if C<interval> is very small then timing stability 2166speed for example), so if C<interval> is very small then timing stability
2142will of course deteriorate. Libev itself tries to be exact to be about one 2167will of course deteriorate. Libev itself tries to be exact to be about one
2143millisecond (if the OS supports it and the machine is fast enough). 2168millisecond (if the OS supports it and the machine is fast enough).
2257 2282
2258=head2 C<ev_signal> - signal me when a signal gets signalled! 2283=head2 C<ev_signal> - signal me when a signal gets signalled!
2259 2284
2260Signal watchers will trigger an event when the process receives a specific 2285Signal watchers will trigger an event when the process receives a specific
2261signal one or more times. Even though signals are very asynchronous, libev 2286signal one or more times. Even though signals are very asynchronous, libev
2262will try it's best to deliver signals synchronously, i.e. as part of the 2287will try its best to deliver signals synchronously, i.e. as part of the
2263normal event processing, like any other event. 2288normal event processing, like any other event.
2264 2289
2265If you want signals to be delivered truly asynchronously, just use 2290If you want signals to be delivered truly asynchronously, just use
2266C<sigaction> as you would do without libev and forget about sharing 2291C<sigaction> as you would do without libev and forget about sharing
2267the signal. You can even use C<ev_async> from a signal handler to 2292the signal. You can even use C<ev_async> from a signal handler to
2286=head3 The special problem of inheritance over fork/execve/pthread_create 2311=head3 The special problem of inheritance over fork/execve/pthread_create
2287 2312
2288Both the signal mask (C<sigprocmask>) and the signal disposition 2313Both the signal mask (C<sigprocmask>) and the signal disposition
2289(C<sigaction>) are unspecified after starting a signal watcher (and after 2314(C<sigaction>) are unspecified after starting a signal watcher (and after
2290stopping it again), that is, libev might or might not block the signal, 2315stopping it again), that is, libev might or might not block the signal,
2291and might or might not set or restore the installed signal handler. 2316and might or might not set or restore the installed signal handler (but
2317see C<EVFLAG_NOSIGMASK>).
2292 2318
2293While this does not matter for the signal disposition (libev never 2319While this does not matter for the signal disposition (libev never
2294sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on 2320sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
2295C<execve>), this matters for the signal mask: many programs do not expect 2321C<execve>), this matters for the signal mask: many programs do not expect
2296certain signals to be blocked. 2322certain signals to be blocked.
2309I<has> to modify the signal mask, at least temporarily. 2335I<has> to modify the signal mask, at least temporarily.
2310 2336
2311So I can't stress this enough: I<If you do not reset your signal mask when 2337So I can't stress this enough: I<If you do not reset your signal mask when
2312you expect it to be empty, you have a race condition in your code>. This 2338you expect it to be empty, you have a race condition in your code>. This
2313is not a libev-specific thing, this is true for most event libraries. 2339is not a libev-specific thing, this is true for most event libraries.
2340
2341=head3 The special problem of threads signal handling
2342
2343POSIX threads has problematic signal handling semantics, specifically,
2344a lot of functionality (sigfd, sigwait etc.) only really works if all
2345threads in a process block signals, which is hard to achieve.
2346
2347When you want to use sigwait (or mix libev signal handling with your own
2348for the same signals), you can tackle this problem by globally blocking
2349all signals before creating any threads (or creating them with a fully set
2350sigprocmask) and also specifying the C<EVFLAG_NOSIGMASK> when creating
2351loops. Then designate one thread as "signal receiver thread" which handles
2352these signals. You can pass on any signals that libev might be interested
2353in by calling C<ev_feed_signal>.
2314 2354
2315=head3 Watcher-Specific Functions and Data Members 2355=head3 Watcher-Specific Functions and Data Members
2316 2356
2317=over 4 2357=over 4
2318 2358
3153 atexit (program_exits); 3193 atexit (program_exits);
3154 3194
3155 3195
3156=head2 C<ev_async> - how to wake up an event loop 3196=head2 C<ev_async> - how to wake up an event loop
3157 3197
3158In general, you cannot use an C<ev_run> from multiple threads or other 3198In general, you cannot use an C<ev_loop> from multiple threads or other
3159asynchronous sources such as signal handlers (as opposed to multiple event 3199asynchronous sources such as signal handlers (as opposed to multiple event
3160loops - those are of course safe to use in different threads). 3200loops - those are of course safe to use in different threads).
3161 3201
3162Sometimes, however, you need to wake up an event loop you do not control, 3202Sometimes, however, you need to wake up an event loop you do not control,
3163for example because it belongs to another thread. This is what C<ev_async> 3203for example because it belongs to another thread. This is what C<ev_async>
3165it by calling C<ev_async_send>, which is thread- and signal safe. 3205it by calling C<ev_async_send>, which is thread- and signal safe.
3166 3206
3167This functionality is very similar to C<ev_signal> watchers, as signals, 3207This functionality is very similar to C<ev_signal> watchers, as signals,
3168too, are asynchronous in nature, and signals, too, will be compressed 3208too, are asynchronous in nature, and signals, too, will be compressed
3169(i.e. the number of callback invocations may be less than the number of 3209(i.e. the number of callback invocations may be less than the number of
3170C<ev_async_sent> calls). 3210C<ev_async_sent> calls). In fact, you could use signal watchers as a kind
3211of "global async watchers" by using a watcher on an otherwise unused
3212signal, and C<ev_feed_signal> to signal this watcher from another thread,
3213even without knowing which loop owns the signal.
3171 3214
3172Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not 3215Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
3173just the default loop. 3216just the default loop.
3174 3217
3175=head3 Queueing 3218=head3 Queueing
3270trust me. 3313trust me.
3271 3314
3272=item ev_async_send (loop, ev_async *) 3315=item ev_async_send (loop, ev_async *)
3273 3316
3274Sends/signals/activates the given C<ev_async> watcher, that is, feeds 3317Sends/signals/activates the given C<ev_async> watcher, that is, feeds
3275an C<EV_ASYNC> event on the watcher into the event loop. Unlike 3318an C<EV_ASYNC> event on the watcher into the event loop, and instantly
3319returns.
3320
3276C<ev_feed_event>, this call is safe to do from other threads, signal or 3321Unlike C<ev_feed_event>, this call is safe to do from other threads,
3277similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding 3322signal or similar contexts (see the discussion of C<EV_ATOMIC_T> in the
3278section below on what exactly this means). 3323embedding section below on what exactly this means).
3279 3324
3280Note that, as with other watchers in libev, multiple events might get 3325Note that, as with other watchers in libev, multiple events might get
3281compressed into a single callback invocation (another way to look at this 3326compressed into a single callback invocation (another way to look at this
3282is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>, 3327is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
3283reset when the event loop detects that). 3328reset when the event loop detects that).
3351Feed an event on the given fd, as if a file descriptor backend detected 3396Feed an event on the given fd, as if a file descriptor backend detected
3352the given events it. 3397the given events it.
3353 3398
3354=item ev_feed_signal_event (loop, int signum) 3399=item ev_feed_signal_event (loop, int signum)
3355 3400
3356Feed an event as if the given signal occurred (C<loop> must be the default 3401Feed an event as if the given signal occurred. See also C<ev_feed_signal>,
3357loop!). 3402which is async-safe.
3358 3403
3359=back 3404=back
3405
3406
3407=head1 COMMON OR USEFUL IDIOMS (OR BOTH)
3408
3409This section explains some common idioms that are not immediately
3410obvious. Note that examples are sprinkled over the whole manual, and this
3411section only contains stuff that wouldn't fit anywhere else.
3412
3413=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
3414
3415Each watcher has, by default, a C<void *data> member that you can read
3416or modify at any time: libev will completely ignore it. This can be used
3417to associate arbitrary data with your watcher. If you need more data and
3418don't want to allocate memory separately and store a pointer to it in that
3419data member, you can also "subclass" the watcher type and provide your own
3420data:
3421
3422 struct my_io
3423 {
3424 ev_io io;
3425 int otherfd;
3426 void *somedata;
3427 struct whatever *mostinteresting;
3428 };
3429
3430 ...
3431 struct my_io w;
3432 ev_io_init (&w.io, my_cb, fd, EV_READ);
3433
3434And since your callback will be called with a pointer to the watcher, you
3435can cast it back to your own type:
3436
3437 static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
3438 {
3439 struct my_io *w = (struct my_io *)w_;
3440 ...
3441 }
3442
3443More interesting and less C-conformant ways of casting your callback
3444function type instead have been omitted.
3445
3446=head2 BUILDING YOUR OWN COMPOSITE WATCHERS
3447
3448Another common scenario is to use some data structure with multiple
3449embedded watchers, in effect creating your own watcher that combines
3450multiple libev event sources into one "super-watcher":
3451
3452 struct my_biggy
3453 {
3454 int some_data;
3455 ev_timer t1;
3456 ev_timer t2;
3457 }
3458
3459In this case getting the pointer to C<my_biggy> is a bit more
3460complicated: Either you store the address of your C<my_biggy> struct in
3461the C<data> member of the watcher (for woozies or C++ coders), or you need
3462to use some pointer arithmetic using C<offsetof> inside your watchers (for
3463real programmers):
3464
3465 #include <stddef.h>
3466
3467 static void
3468 t1_cb (EV_P_ ev_timer *w, int revents)
3469 {
3470 struct my_biggy big = (struct my_biggy *)
3471 (((char *)w) - offsetof (struct my_biggy, t1));
3472 }
3473
3474 static void
3475 t2_cb (EV_P_ ev_timer *w, int revents)
3476 {
3477 struct my_biggy big = (struct my_biggy *)
3478 (((char *)w) - offsetof (struct my_biggy, t2));
3479 }
3480
3481=head2 MODEL/NESTED EVENT LOOP INVOCATIONS AND EXIT CONDITIONS
3482
3483Often (especially in GUI toolkits) there are places where you have
3484I<modal> interaction, which is most easily implemented by recursively
3485invoking C<ev_run>.
3486
3487This brings the problem of exiting - a callback might want to finish the
3488main C<ev_run> call, but not the nested one (e.g. user clicked "Quit", but
3489a modal "Are you sure?" dialog is still waiting), or just the nested one
3490and not the main one (e.g. user clocked "Ok" in a modal dialog), or some
3491other combination: In these cases, C<ev_break> will not work alone.
3492
3493The solution is to maintain "break this loop" variable for each C<ev_run>
3494invocation, and use a loop around C<ev_run> until the condition is
3495triggered, using C<EVRUN_ONCE>:
3496
3497 // main loop
3498 int exit_main_loop = 0;
3499
3500 while (!exit_main_loop)
3501 ev_run (EV_DEFAULT_ EVRUN_ONCE);
3502
3503 // in a model watcher
3504 int exit_nested_loop = 0;
3505
3506 while (!exit_nested_loop)
3507 ev_run (EV_A_ EVRUN_ONCE);
3508
3509To exit from any of these loops, just set the corresponding exit variable:
3510
3511 // exit modal loop
3512 exit_nested_loop = 1;
3513
3514 // exit main program, after modal loop is finished
3515 exit_main_loop = 1;
3516
3517 // exit both
3518 exit_main_loop = exit_nested_loop = 1;
3519
3520=head2 THREAD LOCKING EXAMPLE
3521
3522Here is a fictitious example of how to run an event loop in a different
3523thread from where callbacks are being invoked and watchers are
3524created/added/removed.
3525
3526For a real-world example, see the C<EV::Loop::Async> perl module,
3527which uses exactly this technique (which is suited for many high-level
3528languages).
3529
3530The example uses a pthread mutex to protect the loop data, a condition
3531variable to wait for callback invocations, an async watcher to notify the
3532event loop thread and an unspecified mechanism to wake up the main thread.
3533
3534First, you need to associate some data with the event loop:
3535
3536 typedef struct {
3537 mutex_t lock; /* global loop lock */
3538 ev_async async_w;
3539 thread_t tid;
3540 cond_t invoke_cv;
3541 } userdata;
3542
3543 void prepare_loop (EV_P)
3544 {
3545 // for simplicity, we use a static userdata struct.
3546 static userdata u;
3547
3548 ev_async_init (&u->async_w, async_cb);
3549 ev_async_start (EV_A_ &u->async_w);
3550
3551 pthread_mutex_init (&u->lock, 0);
3552 pthread_cond_init (&u->invoke_cv, 0);
3553
3554 // now associate this with the loop
3555 ev_set_userdata (EV_A_ u);
3556 ev_set_invoke_pending_cb (EV_A_ l_invoke);
3557 ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
3558
3559 // then create the thread running ev_run
3560 pthread_create (&u->tid, 0, l_run, EV_A);
3561 }
3562
3563The callback for the C<ev_async> watcher does nothing: the watcher is used
3564solely to wake up the event loop so it takes notice of any new watchers
3565that might have been added:
3566
3567 static void
3568 async_cb (EV_P_ ev_async *w, int revents)
3569 {
3570 // just used for the side effects
3571 }
3572
3573The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
3574protecting the loop data, respectively.
3575
3576 static void
3577 l_release (EV_P)
3578 {
3579 userdata *u = ev_userdata (EV_A);
3580 pthread_mutex_unlock (&u->lock);
3581 }
3582
3583 static void
3584 l_acquire (EV_P)
3585 {
3586 userdata *u = ev_userdata (EV_A);
3587 pthread_mutex_lock (&u->lock);
3588 }
3589
3590The event loop thread first acquires the mutex, and then jumps straight
3591into C<ev_run>:
3592
3593 void *
3594 l_run (void *thr_arg)
3595 {
3596 struct ev_loop *loop = (struct ev_loop *)thr_arg;
3597
3598 l_acquire (EV_A);
3599 pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
3600 ev_run (EV_A_ 0);
3601 l_release (EV_A);
3602
3603 return 0;
3604 }
3605
3606Instead of invoking all pending watchers, the C<l_invoke> callback will
3607signal the main thread via some unspecified mechanism (signals? pipe
3608writes? C<Async::Interrupt>?) and then waits until all pending watchers
3609have been called (in a while loop because a) spurious wakeups are possible
3610and b) skipping inter-thread-communication when there are no pending
3611watchers is very beneficial):
3612
3613 static void
3614 l_invoke (EV_P)
3615 {
3616 userdata *u = ev_userdata (EV_A);
3617
3618 while (ev_pending_count (EV_A))
3619 {
3620 wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
3621 pthread_cond_wait (&u->invoke_cv, &u->lock);
3622 }
3623 }
3624
3625Now, whenever the main thread gets told to invoke pending watchers, it
3626will grab the lock, call C<ev_invoke_pending> and then signal the loop
3627thread to continue:
3628
3629 static void
3630 real_invoke_pending (EV_P)
3631 {
3632 userdata *u = ev_userdata (EV_A);
3633
3634 pthread_mutex_lock (&u->lock);
3635 ev_invoke_pending (EV_A);
3636 pthread_cond_signal (&u->invoke_cv);
3637 pthread_mutex_unlock (&u->lock);
3638 }
3639
3640Whenever you want to start/stop a watcher or do other modifications to an
3641event loop, you will now have to lock:
3642
3643 ev_timer timeout_watcher;
3644 userdata *u = ev_userdata (EV_A);
3645
3646 ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
3647
3648 pthread_mutex_lock (&u->lock);
3649 ev_timer_start (EV_A_ &timeout_watcher);
3650 ev_async_send (EV_A_ &u->async_w);
3651 pthread_mutex_unlock (&u->lock);
3652
3653Note that sending the C<ev_async> watcher is required because otherwise
3654an event loop currently blocking in the kernel will have no knowledge
3655about the newly added timer. By waking up the loop it will pick up any new
3656watchers in the next event loop iteration.
3657
3658=head2 THREADS, COROUTINES, CONTINUATIONS, QUEUES... INSTEAD OF CALLBACKS
3659
3660While the overhead of a callback that e.g. schedules a thread is small, it
3661is still an overhead. If you embed libev, and your main usage is with some
3662kind of threads or coroutines, you might want to customise libev so that
3663doesn't need callbacks anymore.
3664
3665Imagine you have coroutines that you can switch to using a function
3666C<switch_to (coro)>, that libev runs in a coroutine called C<libev_coro>
3667and that due to some magic, the currently active coroutine is stored in a
3668global called C<current_coro>. Then you can build your own "wait for libev
3669event" primitive by changing C<EV_CB_DECLARE> and C<EV_CB_INVOKE> (note
3670the differing C<;> conventions):
3671
3672 #define EV_CB_DECLARE(type) struct my_coro *cb;
3673 #define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb)
3674
3675That means instead of having a C callback function, you store the
3676coroutine to switch to in each watcher, and instead of having libev call
3677your callback, you instead have it switch to that coroutine.
3678
3679A coroutine might now wait for an event with a function called
3680C<wait_for_event>. (the watcher needs to be started, as always, but it doesn't
3681matter when, or whether the watcher is active or not when this function is
3682called):
3683
3684 void
3685 wait_for_event (ev_watcher *w)
3686 {
3687 ev_cb_set (w) = current_coro;
3688 switch_to (libev_coro);
3689 }
3690
3691That basically suspends the coroutine inside C<wait_for_event> and
3692continues the libev coroutine, which, when appropriate, switches back to
3693this or any other coroutine. I am sure if you sue this your own :)
3694
3695You can do similar tricks if you have, say, threads with an event queue -
3696instead of storing a coroutine, you store the queue object and instead of
3697switching to a coroutine, you push the watcher onto the queue and notify
3698any waiters.
3699
3700To embed libev, see L<EMBEDDING>, but in short, it's easiest to create two
3701files, F<my_ev.h> and F<my_ev.c> that include the respective libev files:
3702
3703 // my_ev.h
3704 #define EV_CB_DECLARE(type) struct my_coro *cb;
3705 #define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb);
3706 #include "../libev/ev.h"
3707
3708 // my_ev.c
3709 #define EV_H "my_ev.h"
3710 #include "../libev/ev.c"
3711
3712And then use F<my_ev.h> when you would normally use F<ev.h>, and compile
3713F<my_ev.c> into your project. When properly specifying include paths, you
3714can even use F<ev.h> as header file name directly.
3360 3715
3361 3716
3362=head1 LIBEVENT EMULATION 3717=head1 LIBEVENT EMULATION
3363 3718
3364Libev offers a compatibility emulation layer for libevent. It cannot 3719Libev offers a compatibility emulation layer for libevent. It cannot
3365emulate the internals of libevent, so here are some usage hints: 3720emulate the internals of libevent, so here are some usage hints:
3366 3721
3367=over 4 3722=over 4
3723
3724=item * Only the libevent-1.4.1-beta API is being emulated.
3725
3726This was the newest libevent version available when libev was implemented,
3727and is still mostly unchanged in 2010.
3368 3728
3369=item * Use it by including <event.h>, as usual. 3729=item * Use it by including <event.h>, as usual.
3370 3730
3371=item * The following members are fully supported: ev_base, ev_callback, 3731=item * The following members are fully supported: ev_base, ev_callback,
3372ev_arg, ev_fd, ev_res, ev_events. 3732ev_arg, ev_fd, ev_res, ev_events.
3378=item * Priorities are not currently supported. Initialising priorities 3738=item * Priorities are not currently supported. Initialising priorities
3379will fail and all watchers will have the same priority, even though there 3739will fail and all watchers will have the same priority, even though there
3380is an ev_pri field. 3740is an ev_pri field.
3381 3741
3382=item * In libevent, the last base created gets the signals, in libev, the 3742=item * In libevent, the last base created gets the signals, in libev, the
3383first base created (== the default loop) gets the signals. 3743base that registered the signal gets the signals.
3384 3744
3385=item * Other members are not supported. 3745=item * Other members are not supported.
3386 3746
3387=item * The libev emulation is I<not> ABI compatible to libevent, you need 3747=item * The libev emulation is I<not> ABI compatible to libevent, you need
3388to use the libev header file and library. 3748to use the libev header file and library.
3407Care has been taken to keep the overhead low. The only data member the C++ 3767Care has been taken to keep the overhead low. The only data member the C++
3408classes add (compared to plain C-style watchers) is the event loop pointer 3768classes add (compared to plain C-style watchers) is the event loop pointer
3409that the watcher is associated with (or no additional members at all if 3769that the watcher is associated with (or no additional members at all if
3410you disable C<EV_MULTIPLICITY> when embedding libev). 3770you disable C<EV_MULTIPLICITY> when embedding libev).
3411 3771
3412Currently, functions, and static and non-static member functions can be 3772Currently, functions, static and non-static member functions and classes
3413used as callbacks. Other types should be easy to add as long as they only 3773with C<operator ()> can be used as callbacks. Other types should be easy
3414need one additional pointer for context. If you need support for other 3774to add as long as they only need one additional pointer for context. If
3415types of functors please contact the author (preferably after implementing 3775you need support for other types of functors please contact the author
3416it). 3776(preferably after implementing it).
3417 3777
3418Here is a list of things available in the C<ev> namespace: 3778Here is a list of things available in the C<ev> namespace:
3419 3779
3420=over 4 3780=over 4
3421 3781
3849F<event.h> that are not directly supported by the libev core alone. 4209F<event.h> that are not directly supported by the libev core alone.
3850 4210
3851In standalone mode, libev will still try to automatically deduce the 4211In standalone mode, libev will still try to automatically deduce the
3852configuration, but has to be more conservative. 4212configuration, but has to be more conservative.
3853 4213
4214=item EV_USE_FLOOR
4215
4216If defined to be C<1>, libev will use the C<floor ()> function for its
4217periodic reschedule calculations, otherwise libev will fall back on a
4218portable (slower) implementation. If you enable this, you usually have to
4219link against libm or something equivalent. Enabling this when the C<floor>
4220function is not available will fail, so the safe default is to not enable
4221this.
4222
3854=item EV_USE_MONOTONIC 4223=item EV_USE_MONOTONIC
3855 4224
3856If defined to be C<1>, libev will try to detect the availability of the 4225If defined to be C<1>, libev will try to detect the availability of the
3857monotonic clock option at both compile time and runtime. Otherwise no 4226monotonic clock option at both compile time and runtime. Otherwise no
3858use of the monotonic clock option will be attempted. If you enable this, 4227use of the monotonic clock option will be attempted. If you enable this,
4289And a F<ev_cpp.C> implementation file that contains libev proper and is compiled: 4658And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
4290 4659
4291 #include "ev_cpp.h" 4660 #include "ev_cpp.h"
4292 #include "ev.c" 4661 #include "ev.c"
4293 4662
4294=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES 4663=head1 INTERACTION WITH OTHER PROGRAMS, LIBRARIES OR THE ENVIRONMENT
4295 4664
4296=head2 THREADS AND COROUTINES 4665=head2 THREADS AND COROUTINES
4297 4666
4298=head3 THREADS 4667=head3 THREADS
4299 4668
4350default loop and triggering an C<ev_async> watcher from the default loop 4719default loop and triggering an C<ev_async> watcher from the default loop
4351watcher callback into the event loop interested in the signal. 4720watcher callback into the event loop interested in the signal.
4352 4721
4353=back 4722=back
4354 4723
4355=head4 THREAD LOCKING EXAMPLE 4724See also L<THREAD LOCKING EXAMPLE>.
4356
4357Here is a fictitious example of how to run an event loop in a different
4358thread than where callbacks are being invoked and watchers are
4359created/added/removed.
4360
4361For a real-world example, see the C<EV::Loop::Async> perl module,
4362which uses exactly this technique (which is suited for many high-level
4363languages).
4364
4365The example uses a pthread mutex to protect the loop data, a condition
4366variable to wait for callback invocations, an async watcher to notify the
4367event loop thread and an unspecified mechanism to wake up the main thread.
4368
4369First, you need to associate some data with the event loop:
4370
4371 typedef struct {
4372 mutex_t lock; /* global loop lock */
4373 ev_async async_w;
4374 thread_t tid;
4375 cond_t invoke_cv;
4376 } userdata;
4377
4378 void prepare_loop (EV_P)
4379 {
4380 // for simplicity, we use a static userdata struct.
4381 static userdata u;
4382
4383 ev_async_init (&u->async_w, async_cb);
4384 ev_async_start (EV_A_ &u->async_w);
4385
4386 pthread_mutex_init (&u->lock, 0);
4387 pthread_cond_init (&u->invoke_cv, 0);
4388
4389 // now associate this with the loop
4390 ev_set_userdata (EV_A_ u);
4391 ev_set_invoke_pending_cb (EV_A_ l_invoke);
4392 ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
4393
4394 // then create the thread running ev_loop
4395 pthread_create (&u->tid, 0, l_run, EV_A);
4396 }
4397
4398The callback for the C<ev_async> watcher does nothing: the watcher is used
4399solely to wake up the event loop so it takes notice of any new watchers
4400that might have been added:
4401
4402 static void
4403 async_cb (EV_P_ ev_async *w, int revents)
4404 {
4405 // just used for the side effects
4406 }
4407
4408The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
4409protecting the loop data, respectively.
4410
4411 static void
4412 l_release (EV_P)
4413 {
4414 userdata *u = ev_userdata (EV_A);
4415 pthread_mutex_unlock (&u->lock);
4416 }
4417
4418 static void
4419 l_acquire (EV_P)
4420 {
4421 userdata *u = ev_userdata (EV_A);
4422 pthread_mutex_lock (&u->lock);
4423 }
4424
4425The event loop thread first acquires the mutex, and then jumps straight
4426into C<ev_run>:
4427
4428 void *
4429 l_run (void *thr_arg)
4430 {
4431 struct ev_loop *loop = (struct ev_loop *)thr_arg;
4432
4433 l_acquire (EV_A);
4434 pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
4435 ev_run (EV_A_ 0);
4436 l_release (EV_A);
4437
4438 return 0;
4439 }
4440
4441Instead of invoking all pending watchers, the C<l_invoke> callback will
4442signal the main thread via some unspecified mechanism (signals? pipe
4443writes? C<Async::Interrupt>?) and then waits until all pending watchers
4444have been called (in a while loop because a) spurious wakeups are possible
4445and b) skipping inter-thread-communication when there are no pending
4446watchers is very beneficial):
4447
4448 static void
4449 l_invoke (EV_P)
4450 {
4451 userdata *u = ev_userdata (EV_A);
4452
4453 while (ev_pending_count (EV_A))
4454 {
4455 wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
4456 pthread_cond_wait (&u->invoke_cv, &u->lock);
4457 }
4458 }
4459
4460Now, whenever the main thread gets told to invoke pending watchers, it
4461will grab the lock, call C<ev_invoke_pending> and then signal the loop
4462thread to continue:
4463
4464 static void
4465 real_invoke_pending (EV_P)
4466 {
4467 userdata *u = ev_userdata (EV_A);
4468
4469 pthread_mutex_lock (&u->lock);
4470 ev_invoke_pending (EV_A);
4471 pthread_cond_signal (&u->invoke_cv);
4472 pthread_mutex_unlock (&u->lock);
4473 }
4474
4475Whenever you want to start/stop a watcher or do other modifications to an
4476event loop, you will now have to lock:
4477
4478 ev_timer timeout_watcher;
4479 userdata *u = ev_userdata (EV_A);
4480
4481 ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
4482
4483 pthread_mutex_lock (&u->lock);
4484 ev_timer_start (EV_A_ &timeout_watcher);
4485 ev_async_send (EV_A_ &u->async_w);
4486 pthread_mutex_unlock (&u->lock);
4487
4488Note that sending the C<ev_async> watcher is required because otherwise
4489an event loop currently blocking in the kernel will have no knowledge
4490about the newly added timer. By waking up the loop it will pick up any new
4491watchers in the next event loop iteration.
4492 4725
4493=head3 COROUTINES 4726=head3 COROUTINES
4494 4727
4495Libev is very accommodating to coroutines ("cooperative threads"): 4728Libev is very accommodating to coroutines ("cooperative threads"):
4496libev fully supports nesting calls to its functions from different 4729libev fully supports nesting calls to its functions from different
5005The physical time that is observed. It is apparently strictly monotonic :) 5238The physical time that is observed. It is apparently strictly monotonic :)
5006 5239
5007=item wall-clock time 5240=item wall-clock time
5008 5241
5009The time and date as shown on clocks. Unlike real time, it can actually 5242The time and date as shown on clocks. Unlike real time, it can actually
5010be wrong and jump forwards and backwards, e.g. when the you adjust your 5243be wrong and jump forwards and backwards, e.g. when you adjust your
5011clock. 5244clock.
5012 5245
5013=item watcher 5246=item watcher
5014 5247
5015A data structure that describes interest in certain events. Watchers need 5248A data structure that describes interest in certain events. Watchers need
5018=back 5251=back
5019 5252
5020=head1 AUTHOR 5253=head1 AUTHOR
5021 5254
5022Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael 5255Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael
5023Magnusson and Emanuele Giaquinta. 5256Magnusson and Emanuele Giaquinta, and minor corrections by many others.
5024 5257

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines