ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/ev.3
(Generate patch)

Comparing libev/ev.3 (file contents):
Revision 1.11 by root, Sat Nov 24 07:14:26 2007 UTC vs.
Revision 1.44 by root, Sat Dec 8 15:30:26 2007 UTC

127.\} 127.\}
128.rm #[ #] #H #V #F C 128.rm #[ #] #H #V #F C
129.\" ======================================================================== 129.\" ========================================================================
130.\" 130.\"
131.IX Title ""<STANDARD INPUT>" 1" 131.IX Title ""<STANDARD INPUT>" 1"
132.TH "<STANDARD INPUT>" 1 "2007-11-24" "perl v5.8.8" "User Contributed Perl Documentation" 132.TH "<STANDARD INPUT>" 1 "2007-12-08" "perl v5.8.8" "User Contributed Perl Documentation"
133.SH "NAME" 133.SH "NAME"
134libev \- a high performance full\-featured event loop written in C 134libev \- a high performance full\-featured event loop written in C
135.SH "SYNOPSIS" 135.SH "SYNOPSIS"
136.IX Header "SYNOPSIS" 136.IX Header "SYNOPSIS"
137.Vb 1 137.Vb 1
138\& #include <ev.h> 138\& #include <ev.h>
139.Ve 139.Ve
140.SH "EXAMPLE PROGRAM"
141.IX Header "EXAMPLE PROGRAM"
142.Vb 1
143\& #include <ev.h>
144.Ve
145.PP
146.Vb 2
147\& ev_io stdin_watcher;
148\& ev_timer timeout_watcher;
149.Ve
150.PP
151.Vb 8
152\& /* called when data readable on stdin */
153\& static void
154\& stdin_cb (EV_P_ struct ev_io *w, int revents)
155\& {
156\& /* puts ("stdin ready"); */
157\& ev_io_stop (EV_A_ w); /* just a syntax example */
158\& ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */
159\& }
160.Ve
161.PP
162.Vb 6
163\& static void
164\& timeout_cb (EV_P_ struct ev_timer *w, int revents)
165\& {
166\& /* puts ("timeout"); */
167\& ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */
168\& }
169.Ve
170.PP
171.Vb 4
172\& int
173\& main (void)
174\& {
175\& struct ev_loop *loop = ev_default_loop (0);
176.Ve
177.PP
178.Vb 3
179\& /* initialise an io watcher, then start it */
180\& ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
181\& ev_io_start (loop, &stdin_watcher);
182.Ve
183.PP
184.Vb 3
185\& /* simple non-repeating 5.5 second timeout */
186\& ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
187\& ev_timer_start (loop, &timeout_watcher);
188.Ve
189.PP
190.Vb 2
191\& /* loop till timeout or data ready */
192\& ev_loop (loop, 0);
193.Ve
194.PP
195.Vb 2
196\& return 0;
197\& }
198.Ve
140.SH "DESCRIPTION" 199.SH "DESCRIPTION"
141.IX Header "DESCRIPTION" 200.IX Header "DESCRIPTION"
201The newest version of this document is also available as a html-formatted
202web page you might find easier to navigate when reading it for the first
203time: <http://cvs.schmorp.de/libev/ev.html>.
204.PP
142Libev is an event loop: you register interest in certain events (such as a 205Libev is an event loop: you register interest in certain events (such as a
143file descriptor being readable or a timeout occuring), and it will manage 206file descriptor being readable or a timeout occuring), and it will manage
144these event sources and provide your program with events. 207these event sources and provide your program with events.
145.PP 208.PP
146To do this, it must take more or less complete control over your process 209To do this, it must take more or less complete control over your process
151watchers\fR, which are relatively small C structures you initialise with the 214watchers\fR, which are relatively small C structures you initialise with the
152details of the event, and then hand it over to libev by \fIstarting\fR the 215details of the event, and then hand it over to libev by \fIstarting\fR the
153watcher. 216watcher.
154.SH "FEATURES" 217.SH "FEATURES"
155.IX Header "FEATURES" 218.IX Header "FEATURES"
156Libev supports select, poll, the linux-specific epoll and the bsd-specific 219Libev supports \f(CW\*(C`select\*(C'\fR, \f(CW\*(C`poll\*(C'\fR, the Linux-specific \f(CW\*(C`epoll\*(C'\fR, the
157kqueue mechanisms for file descriptor events, relative timers, absolute 220BSD-specific \f(CW\*(C`kqueue\*(C'\fR and the Solaris-specific event port mechanisms
158timers with customised rescheduling, signal events, process status change 221for file descriptor events (\f(CW\*(C`ev_io\*(C'\fR), the Linux \f(CW\*(C`inotify\*(C'\fR interface
159events (related to \s-1SIGCHLD\s0), and event watchers dealing with the event 222(for \f(CW\*(C`ev_stat\*(C'\fR), relative timers (\f(CW\*(C`ev_timer\*(C'\fR), absolute timers
160loop mechanism itself (idle, prepare and check watchers). It also is quite 223with customised rescheduling (\f(CW\*(C`ev_periodic\*(C'\fR), synchronous signals
161fast (see this benchmark comparing 224(\f(CW\*(C`ev_signal\*(C'\fR), process status change events (\f(CW\*(C`ev_child\*(C'\fR), and event
162it to libevent for example). 225watchers dealing with the event loop mechanism itself (\f(CW\*(C`ev_idle\*(C'\fR,
226\&\f(CW\*(C`ev_embed\*(C'\fR, \f(CW\*(C`ev_prepare\*(C'\fR and \f(CW\*(C`ev_check\*(C'\fR watchers) as well as
227file watchers (\f(CW\*(C`ev_stat\*(C'\fR) and even limited support for fork events
228(\f(CW\*(C`ev_fork\*(C'\fR).
229.PP
230It also is quite fast (see this
231benchmark comparing it to libevent
232for example).
163.SH "CONVENTIONS" 233.SH "CONVENTIONS"
164.IX Header "CONVENTIONS" 234.IX Header "CONVENTIONS"
165Libev is very configurable. In this manual the default configuration 235Libev is very configurable. In this manual the default configuration will
166will be described, which supports multiple event loops. For more info 236be described, which supports multiple event loops. For more info about
167about various configuration options please have a look at the file 237various configuration options please have a look at \fB\s-1EMBED\s0\fR section in
168\&\fI\s-1README\s0.embed\fR in the libev distribution. If libev was configured without 238this manual. If libev was configured without support for multiple event
169support for multiple event loops, then all functions taking an initial 239loops, then all functions taking an initial argument of name \f(CW\*(C`loop\*(C'\fR
170argument of name \f(CW\*(C`loop\*(C'\fR (which is always of type \f(CW\*(C`struct ev_loop *\*(C'\fR) 240(which is always of type \f(CW\*(C`struct ev_loop *\*(C'\fR) will not have this argument.
171will not have this argument.
172.SH "TIME REPRESENTATION" 241.SH "TIME REPRESENTATION"
173.IX Header "TIME REPRESENTATION" 242.IX Header "TIME REPRESENTATION"
174Libev represents time as a single floating point number, representing the 243Libev represents time as a single floating point number, representing the
175(fractional) number of seconds since the (\s-1POSIX\s0) epoch (somewhere near 244(fractional) number of seconds since the (\s-1POSIX\s0) epoch (somewhere near
176the beginning of 1970, details are complicated, don't ask). This type is 245the beginning of 1970, details are complicated, don't ask). This type is
201Usually, it's a good idea to terminate if the major versions mismatch, 270Usually, it's a good idea to terminate if the major versions mismatch,
202as this indicates an incompatible change. Minor versions are usually 271as this indicates an incompatible change. Minor versions are usually
203compatible to older versions, so a larger minor version alone is usually 272compatible to older versions, so a larger minor version alone is usually
204not a problem. 273not a problem.
205.Sp 274.Sp
206Example: make sure we haven't accidentally been linked against the wrong 275Example: Make sure we haven't accidentally been linked against the wrong
207version: 276version.
208.Sp 277.Sp
209.Vb 3 278.Vb 3
210\& assert (("libev version mismatch", 279\& assert (("libev version mismatch",
211\& ev_version_major () == EV_VERSION_MAJOR 280\& ev_version_major () == EV_VERSION_MAJOR
212\& && ev_version_minor () >= EV_VERSION_MINOR)); 281\& && ev_version_minor () >= EV_VERSION_MINOR));
242recommended ones. 311recommended ones.
243.Sp 312.Sp
244See the description of \f(CW\*(C`ev_embed\*(C'\fR watchers for more info. 313See the description of \f(CW\*(C`ev_embed\*(C'\fR watchers for more info.
245.IP "ev_set_allocator (void *(*cb)(void *ptr, long size))" 4 314.IP "ev_set_allocator (void *(*cb)(void *ptr, long size))" 4
246.IX Item "ev_set_allocator (void *(*cb)(void *ptr, long size))" 315.IX Item "ev_set_allocator (void *(*cb)(void *ptr, long size))"
247Sets the allocation function to use (the prototype is similar to the 316Sets the allocation function to use (the prototype is similar \- the
248realloc C function, the semantics are identical). It is used to allocate 317semantics is identical \- to the realloc C function). It is used to
249and free memory (no surprises here). If it returns zero when memory 318allocate and free memory (no surprises here). If it returns zero when
250needs to be allocated, the library might abort or take some potentially 319memory needs to be allocated, the library might abort or take some
251destructive action. The default is your system realloc function. 320potentially destructive action. The default is your system realloc
321function.
252.Sp 322.Sp
253You could override this function in high-availability programs to, say, 323You could override this function in high-availability programs to, say,
254free some memory if it cannot allocate memory, to use a special allocator, 324free some memory if it cannot allocate memory, to use a special allocator,
255or even to sleep a while and retry until some memory is available. 325or even to sleep a while and retry until some memory is available.
256.Sp 326.Sp
257Example: replace the libev allocator with one that waits a bit and then 327Example: Replace the libev allocator with one that waits a bit and then
258retries: better than mine). 328retries).
259.Sp 329.Sp
260.Vb 6 330.Vb 6
261\& static void * 331\& static void *
262\& persistent_realloc (void *ptr, long size) 332\& persistent_realloc (void *ptr, size_t size)
263\& { 333\& {
264\& for (;;) 334\& for (;;)
265\& { 335\& {
266\& void *newptr = realloc (ptr, size); 336\& void *newptr = realloc (ptr, size);
267.Ve 337.Ve
289callback is set, then libev will expect it to remedy the sitution, no 359callback is set, then libev will expect it to remedy the sitution, no
290matter what, when it returns. That is, libev will generally retry the 360matter what, when it returns. That is, libev will generally retry the
291requested operation, or, if the condition doesn't go away, do bad stuff 361requested operation, or, if the condition doesn't go away, do bad stuff
292(such as abort). 362(such as abort).
293.Sp 363.Sp
294Example: do the same thing as libev does internally: 364Example: This is basically the same thing that libev does internally, too.
295.Sp 365.Sp
296.Vb 6 366.Vb 6
297\& static void 367\& static void
298\& fatal_error (const char *msg) 368\& fatal_error (const char *msg)
299\& { 369\& {
345or setgid) then libev will \fInot\fR look at the environment variable 415or setgid) then libev will \fInot\fR look at the environment variable
346\&\f(CW\*(C`LIBEV_FLAGS\*(C'\fR. Otherwise (the default), this environment variable will 416\&\f(CW\*(C`LIBEV_FLAGS\*(C'\fR. Otherwise (the default), this environment variable will
347override the flags completely if it is found in the environment. This is 417override the flags completely if it is found in the environment. This is
348useful to try out specific backends to test their performance, or to work 418useful to try out specific backends to test their performance, or to work
349around bugs. 419around bugs.
420.ie n .IP """EVFLAG_FORKCHECK""" 4
421.el .IP "\f(CWEVFLAG_FORKCHECK\fR" 4
422.IX Item "EVFLAG_FORKCHECK"
423Instead of calling \f(CW\*(C`ev_default_fork\*(C'\fR or \f(CW\*(C`ev_loop_fork\*(C'\fR manually after
424a fork, you can also make libev check for a fork in each iteration by
425enabling this flag.
426.Sp
427This works by calling \f(CW\*(C`getpid ()\*(C'\fR on every iteration of the loop,
428and thus this might slow down your event loop if you do a lot of loop
429iterations and little real work, but is usually not noticeable (on my
430Linux system for example, \f(CW\*(C`getpid\*(C'\fR is actually a simple 5\-insn sequence
431without a syscall and thus \fIvery\fR fast, but my Linux system also has
432\&\f(CW\*(C`pthread_atfork\*(C'\fR which is even faster).
433.Sp
434The big advantage of this flag is that you can forget about fork (and
435forget about forgetting to tell libev about forking) when you use this
436flag.
437.Sp
438This flag setting cannot be overriden or specified in the \f(CW\*(C`LIBEV_FLAGS\*(C'\fR
439environment variable.
350.ie n .IP """EVBACKEND_SELECT"" (value 1, portable select backend)" 4 440.ie n .IP """EVBACKEND_SELECT"" (value 1, portable select backend)" 4
351.el .IP "\f(CWEVBACKEND_SELECT\fR (value 1, portable select backend)" 4 441.el .IP "\f(CWEVBACKEND_SELECT\fR (value 1, portable select backend)" 4
352.IX Item "EVBACKEND_SELECT (value 1, portable select backend)" 442.IX Item "EVBACKEND_SELECT (value 1, portable select backend)"
353This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as 443This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as
354libev tries to roll its own fd_set with no limits on the number of fds, 444libev tries to roll its own fd_set with no limits on the number of fds,
448Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is 538Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is
449always distinct from the default loop. Unlike the default loop, it cannot 539always distinct from the default loop. Unlike the default loop, it cannot
450handle signal and child watchers, and attempts to do so will be greeted by 540handle signal and child watchers, and attempts to do so will be greeted by
451undefined behaviour (or a failed assertion if assertions are enabled). 541undefined behaviour (or a failed assertion if assertions are enabled).
452.Sp 542.Sp
453Example: try to create a event loop that uses epoll and nothing else. 543Example: Try to create a event loop that uses epoll and nothing else.
454.Sp 544.Sp
455.Vb 3 545.Vb 3
456\& struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV); 546\& struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
457\& if (!epoller) 547\& if (!epoller)
458\& fatal ("no epoll found here, maybe it hides under your chair"); 548\& fatal ("no epoll found here, maybe it hides under your chair");
459.Ve 549.Ve
460.IP "ev_default_destroy ()" 4 550.IP "ev_default_destroy ()" 4
461.IX Item "ev_default_destroy ()" 551.IX Item "ev_default_destroy ()"
462Destroys the default loop again (frees all memory and kernel state 552Destroys the default loop again (frees all memory and kernel state
463etc.). This stops all registered event watchers (by not touching them in 553etc.). None of the active event watchers will be stopped in the normal
464any way whatsoever, although you cannot rely on this :). 554sense, so e.g. \f(CW\*(C`ev_is_active\*(C'\fR might still return true. It is your
555responsibility to either stop all watchers cleanly yoursef \fIbefore\fR
556calling this function, or cope with the fact afterwards (which is usually
557the easiest thing, youc na just ignore the watchers and/or \f(CW\*(C`free ()\*(C'\fR them
558for example).
465.IP "ev_loop_destroy (loop)" 4 559.IP "ev_loop_destroy (loop)" 4
466.IX Item "ev_loop_destroy (loop)" 560.IX Item "ev_loop_destroy (loop)"
467Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an 561Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an
468earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR. 562earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR.
469.IP "ev_default_fork ()" 4 563.IP "ev_default_fork ()" 4
491.IP "ev_loop_fork (loop)" 4 585.IP "ev_loop_fork (loop)" 4
492.IX Item "ev_loop_fork (loop)" 586.IX Item "ev_loop_fork (loop)"
493Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by 587Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by
494\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop 588\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop
495after fork, and how you do this is entirely your own problem. 589after fork, and how you do this is entirely your own problem.
590.IP "unsigned int ev_loop_count (loop)" 4
591.IX Item "unsigned int ev_loop_count (loop)"
592Returns the count of loop iterations for the loop, which is identical to
593the number of times libev did poll for new events. It starts at \f(CW0\fR and
594happily wraps around with enough iterations.
595.Sp
596This value can sometimes be useful as a generation counter of sorts (it
597\&\*(L"ticks\*(R" the number of loop iterations), as it roughly corresponds with
598\&\f(CW\*(C`ev_prepare\*(C'\fR and \f(CW\*(C`ev_check\*(C'\fR calls.
496.IP "unsigned int ev_backend (loop)" 4 599.IP "unsigned int ev_backend (loop)" 4
497.IX Item "unsigned int ev_backend (loop)" 600.IX Item "unsigned int ev_backend (loop)"
498Returns one of the \f(CW\*(C`EVBACKEND_*\*(C'\fR flags indicating the event backend in 601Returns one of the \f(CW\*(C`EVBACKEND_*\*(C'\fR flags indicating the event backend in
499use. 602use.
500.IP "ev_tstamp ev_now (loop)" 4 603.IP "ev_tstamp ev_now (loop)" 4
552\& be handled here by queueing them when their watcher gets executed. 655\& be handled here by queueing them when their watcher gets executed.
553\& - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK 656\& - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
554\& were used, return, otherwise continue with step *. 657\& were used, return, otherwise continue with step *.
555.Ve 658.Ve
556.Sp 659.Sp
557Example: queue some jobs and then loop until no events are outsanding 660Example: Queue some jobs and then loop until no events are outsanding
558anymore. 661anymore.
559.Sp 662.Sp
560.Vb 4 663.Vb 4
561\& ... queue jobs here, make sure they register event watchers as long 664\& ... queue jobs here, make sure they register event watchers as long
562\& ... as they still have work to do (even an idle watcher will do..) 665\& ... as they still have work to do (even an idle watcher will do..)
584visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if 687visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if
585no event watchers registered by it are active. It is also an excellent 688no event watchers registered by it are active. It is also an excellent
586way to do this for generic recurring timers or from within third-party 689way to do this for generic recurring timers or from within third-party
587libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR. 690libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR.
588.Sp 691.Sp
589Example: create a signal watcher, but keep it from keeping \f(CW\*(C`ev_loop\*(C'\fR 692Example: Create a signal watcher, but keep it from keeping \f(CW\*(C`ev_loop\*(C'\fR
590running when nothing else is active. 693running when nothing else is active.
591.Sp 694.Sp
592.Vb 4 695.Vb 4
593\& struct dv_signal exitsig; 696\& struct ev_signal exitsig;
594\& ev_signal_init (&exitsig, sig_cb, SIGINT); 697\& ev_signal_init (&exitsig, sig_cb, SIGINT);
595\& ev_signal_start (myloop, &exitsig); 698\& ev_signal_start (loop, &exitsig);
596\& evf_unref (myloop); 699\& evf_unref (loop);
597.Ve 700.Ve
598.Sp 701.Sp
599Example: for some weird reason, unregister the above signal handler again. 702Example: For some weird reason, unregister the above signal handler again.
600.Sp 703.Sp
601.Vb 2 704.Vb 2
602\& ev_ref (myloop); 705\& ev_ref (loop);
603\& ev_signal_stop (myloop, &exitsig); 706\& ev_signal_stop (loop, &exitsig);
604.Ve 707.Ve
605.SH "ANATOMY OF A WATCHER" 708.SH "ANATOMY OF A WATCHER"
606.IX Header "ANATOMY OF A WATCHER" 709.IX Header "ANATOMY OF A WATCHER"
607A watcher is a structure that you create and register to record your 710A watcher is a structure that you create and register to record your
608interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to 711interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to
680The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread. 783The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread.
681.ie n .IP """EV_CHILD""" 4 784.ie n .IP """EV_CHILD""" 4
682.el .IP "\f(CWEV_CHILD\fR" 4 785.el .IP "\f(CWEV_CHILD\fR" 4
683.IX Item "EV_CHILD" 786.IX Item "EV_CHILD"
684The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change. 787The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change.
788.ie n .IP """EV_STAT""" 4
789.el .IP "\f(CWEV_STAT\fR" 4
790.IX Item "EV_STAT"
791The path specified in the \f(CW\*(C`ev_stat\*(C'\fR watcher changed its attributes somehow.
685.ie n .IP """EV_IDLE""" 4 792.ie n .IP """EV_IDLE""" 4
686.el .IP "\f(CWEV_IDLE\fR" 4 793.el .IP "\f(CWEV_IDLE\fR" 4
687.IX Item "EV_IDLE" 794.IX Item "EV_IDLE"
688The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do. 795The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do.
689.ie n .IP """EV_PREPARE""" 4 796.ie n .IP """EV_PREPARE""" 4
699\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any 806\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any
700received events. Callbacks of both watcher types can start and stop as 807received events. Callbacks of both watcher types can start and stop as
701many watchers as they want, and all of them will be taken into account 808many watchers as they want, and all of them will be taken into account
702(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep 809(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep
703\&\f(CW\*(C`ev_loop\*(C'\fR from blocking). 810\&\f(CW\*(C`ev_loop\*(C'\fR from blocking).
811.ie n .IP """EV_EMBED""" 4
812.el .IP "\f(CWEV_EMBED\fR" 4
813.IX Item "EV_EMBED"
814The embedded event loop specified in the \f(CW\*(C`ev_embed\*(C'\fR watcher needs attention.
815.ie n .IP """EV_FORK""" 4
816.el .IP "\f(CWEV_FORK\fR" 4
817.IX Item "EV_FORK"
818The event loop has been resumed in the child process after fork (see
819\&\f(CW\*(C`ev_fork\*(C'\fR).
704.ie n .IP """EV_ERROR""" 4 820.ie n .IP """EV_ERROR""" 4
705.el .IP "\f(CWEV_ERROR\fR" 4 821.el .IP "\f(CWEV_ERROR\fR" 4
706.IX Item "EV_ERROR" 822.IX Item "EV_ERROR"
707An unspecified error has occured, the watcher has been stopped. This might 823An unspecified error has occured, the watcher has been stopped. This might
708happen because the watcher could not be properly started because libev 824happen because the watcher could not be properly started because libev
713Libev will usually signal a few \*(L"dummy\*(R" events together with an error, 829Libev will usually signal a few \*(L"dummy\*(R" events together with an error,
714for example it might indicate that a fd is readable or writable, and if 830for example it might indicate that a fd is readable or writable, and if
715your callbacks is well-written it can just attempt the operation and cope 831your callbacks is well-written it can just attempt the operation and cope
716with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded 832with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded
717programs, though, so beware. 833programs, though, so beware.
718.Sh "\s-1SUMMARY\s0 \s-1OF\s0 \s-1GENERIC\s0 \s-1WATCHER\s0 \s-1FUNCTIONS\s0" 834.Sh "\s-1GENERIC\s0 \s-1WATCHER\s0 \s-1FUNCTIONS\s0"
719.IX Subsection "SUMMARY OF GENERIC WATCHER FUNCTIONS" 835.IX Subsection "GENERIC WATCHER FUNCTIONS"
720In the following description, \f(CW\*(C`TYPE\*(C'\fR stands for the watcher type, 836In the following description, \f(CW\*(C`TYPE\*(C'\fR stands for the watcher type,
721e.g. \f(CW\*(C`timer\*(C'\fR for \f(CW\*(C`ev_timer\*(C'\fR watchers and \f(CW\*(C`io\*(C'\fR for \f(CW\*(C`ev_io\*(C'\fR watchers. 837e.g. \f(CW\*(C`timer\*(C'\fR for \f(CW\*(C`ev_timer\*(C'\fR watchers and \f(CW\*(C`io\*(C'\fR for \f(CW\*(C`ev_io\*(C'\fR watchers.
722.ie n .IP """ev_init"" (ev_TYPE *watcher, callback)" 4 838.ie n .IP """ev_init"" (ev_TYPE *watcher, callback)" 4
723.el .IP "\f(CWev_init\fR (ev_TYPE *watcher, callback)" 4 839.el .IP "\f(CWev_init\fR (ev_TYPE *watcher, callback)" 4
724.IX Item "ev_init (ev_TYPE *watcher, callback)" 840.IX Item "ev_init (ev_TYPE *watcher, callback)"
730which rolls both calls into one. 846which rolls both calls into one.
731.Sp 847.Sp
732You can reinitialise a watcher at any time as long as it has been stopped 848You can reinitialise a watcher at any time as long as it has been stopped
733(or never started) and there are no pending events outstanding. 849(or never started) and there are no pending events outstanding.
734.Sp 850.Sp
735The callbakc is always of type \f(CW\*(C`void (*)(ev_loop *loop, ev_TYPE *watcher, 851The callback is always of type \f(CW\*(C`void (*)(ev_loop *loop, ev_TYPE *watcher,
736int revents)\*(C'\fR. 852int revents)\*(C'\fR.
737.ie n .IP """ev_TYPE_set"" (ev_TYPE *, [args])" 4 853.ie n .IP """ev_TYPE_set"" (ev_TYPE *, [args])" 4
738.el .IP "\f(CWev_TYPE_set\fR (ev_TYPE *, [args])" 4 854.el .IP "\f(CWev_TYPE_set\fR (ev_TYPE *, [args])" 4
739.IX Item "ev_TYPE_set (ev_TYPE *, [args])" 855.IX Item "ev_TYPE_set (ev_TYPE *, [args])"
740This macro initialises the type-specific parts of a watcher. You need to 856This macro initialises the type-specific parts of a watcher. You need to
773.IP "bool ev_is_pending (ev_TYPE *watcher)" 4 889.IP "bool ev_is_pending (ev_TYPE *watcher)" 4
774.IX Item "bool ev_is_pending (ev_TYPE *watcher)" 890.IX Item "bool ev_is_pending (ev_TYPE *watcher)"
775Returns a true value iff the watcher is pending, (i.e. it has outstanding 891Returns a true value iff the watcher is pending, (i.e. it has outstanding
776events but its callback has not yet been invoked). As long as a watcher 892events but its callback has not yet been invoked). As long as a watcher
777is pending (but not active) you must not call an init function on it (but 893is pending (but not active) you must not call an init function on it (but
778\&\f(CW\*(C`ev_TYPE_set\*(C'\fR is safe) and you must make sure the watcher is available to 894\&\f(CW\*(C`ev_TYPE_set\*(C'\fR is safe), you must not change its priority, and you must
779libev (e.g. you cnanot \f(CW\*(C`free ()\*(C'\fR it). 895make sure the watcher is available to libev (e.g. you cannot \f(CW\*(C`free ()\*(C'\fR
896it).
780.IP "callback = ev_cb (ev_TYPE *watcher)" 4 897.IP "callback ev_cb (ev_TYPE *watcher)" 4
781.IX Item "callback = ev_cb (ev_TYPE *watcher)" 898.IX Item "callback ev_cb (ev_TYPE *watcher)"
782Returns the callback currently set on the watcher. 899Returns the callback currently set on the watcher.
783.IP "ev_cb_set (ev_TYPE *watcher, callback)" 4 900.IP "ev_cb_set (ev_TYPE *watcher, callback)" 4
784.IX Item "ev_cb_set (ev_TYPE *watcher, callback)" 901.IX Item "ev_cb_set (ev_TYPE *watcher, callback)"
785Change the callback. You can change the callback at virtually any time 902Change the callback. You can change the callback at virtually any time
786(modulo threads). 903(modulo threads).
904.IP "ev_set_priority (ev_TYPE *watcher, priority)" 4
905.IX Item "ev_set_priority (ev_TYPE *watcher, priority)"
906.PD 0
907.IP "int ev_priority (ev_TYPE *watcher)" 4
908.IX Item "int ev_priority (ev_TYPE *watcher)"
909.PD
910Set and query the priority of the watcher. The priority is a small
911integer between \f(CW\*(C`EV_MAXPRI\*(C'\fR (default: \f(CW2\fR) and \f(CW\*(C`EV_MINPRI\*(C'\fR
912(default: \f(CW\*(C`\-2\*(C'\fR). Pending watchers with higher priority will be invoked
913before watchers with lower priority, but priority will not keep watchers
914from being executed (except for \f(CW\*(C`ev_idle\*(C'\fR watchers).
915.Sp
916This means that priorities are \fIonly\fR used for ordering callback
917invocation after new events have been received. This is useful, for
918example, to reduce latency after idling, or more often, to bind two
919watchers on the same event and make sure one is called first.
920.Sp
921If you need to suppress invocation when higher priority events are pending
922you need to look at \f(CW\*(C`ev_idle\*(C'\fR watchers, which provide this functionality.
923.Sp
924You \fImust not\fR change the priority of a watcher as long as it is active or
925pending.
926.Sp
927The default priority used by watchers when no priority has been set is
928always \f(CW0\fR, which is supposed to not be too high and not be too low :).
929.Sp
930Setting a priority outside the range of \f(CW\*(C`EV_MINPRI\*(C'\fR to \f(CW\*(C`EV_MAXPRI\*(C'\fR is
931fine, as long as you do not mind that the priority value you query might
932or might not have been adjusted to be within valid range.
933.IP "ev_invoke (loop, ev_TYPE *watcher, int revents)" 4
934.IX Item "ev_invoke (loop, ev_TYPE *watcher, int revents)"
935Invoke the \f(CW\*(C`watcher\*(C'\fR with the given \f(CW\*(C`loop\*(C'\fR and \f(CW\*(C`revents\*(C'\fR. Neither
936\&\f(CW\*(C`loop\*(C'\fR nor \f(CW\*(C`revents\*(C'\fR need to be valid as long as the watcher callback
937can deal with that fact.
938.IP "int ev_clear_pending (loop, ev_TYPE *watcher)" 4
939.IX Item "int ev_clear_pending (loop, ev_TYPE *watcher)"
940If the watcher is pending, this function returns clears its pending status
941and returns its \f(CW\*(C`revents\*(C'\fR bitset (as if its callback was invoked). If the
942watcher isn't pending it does nothing and returns \f(CW0\fR.
787.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0" 943.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0"
788.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER" 944.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER"
789Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change 945Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change
790and read at any time, libev will completely ignore it. This can be used 946and read at any time, libev will completely ignore it. This can be used
791to associate arbitrary data with your watcher. If you need more data and 947to associate arbitrary data with your watcher. If you need more data and
812\& struct my_io *w = (struct my_io *)w_; 968\& struct my_io *w = (struct my_io *)w_;
813\& ... 969\& ...
814\& } 970\& }
815.Ve 971.Ve
816.PP 972.PP
817More interesting and less C\-conformant ways of catsing your callback type 973More interesting and less C\-conformant ways of casting your callback type
818have been omitted.... 974instead have been omitted.
975.PP
976Another common scenario is having some data structure with multiple
977watchers:
978.PP
979.Vb 6
980\& struct my_biggy
981\& {
982\& int some_data;
983\& ev_timer t1;
984\& ev_timer t2;
985\& }
986.Ve
987.PP
988In this case getting the pointer to \f(CW\*(C`my_biggy\*(C'\fR is a bit more complicated,
989you need to use \f(CW\*(C`offsetof\*(C'\fR:
990.PP
991.Vb 1
992\& #include <stddef.h>
993.Ve
994.PP
995.Vb 6
996\& static void
997\& t1_cb (EV_P_ struct ev_timer *w, int revents)
998\& {
999\& struct my_biggy big = (struct my_biggy *
1000\& (((char *)w) - offsetof (struct my_biggy, t1));
1001\& }
1002.Ve
1003.PP
1004.Vb 6
1005\& static void
1006\& t2_cb (EV_P_ struct ev_timer *w, int revents)
1007\& {
1008\& struct my_biggy big = (struct my_biggy *
1009\& (((char *)w) - offsetof (struct my_biggy, t2));
1010\& }
1011.Ve
819.SH "WATCHER TYPES" 1012.SH "WATCHER TYPES"
820.IX Header "WATCHER TYPES" 1013.IX Header "WATCHER TYPES"
821This section describes each watcher in detail, but will not repeat 1014This section describes each watcher in detail, but will not repeat
822information given in the last section. 1015information given in the last section. Any initialisation/set macros,
1016functions and members specific to the watcher type are explained.
1017.PP
1018Members are additionally marked with either \fI[read\-only]\fR, meaning that,
1019while the watcher is active, you can look at the member and expect some
1020sensible content, but you must not modify it (you can modify it while the
1021watcher is stopped to your hearts content), or \fI[read\-write]\fR, which
1022means you can expect it to have some sensible content while the watcher
1023is active, but you can also modify it. Modifying it may not do something
1024sensible or take immediate effect (or do anything at all), but libev will
1025not crash or malfunction in any way.
823.ie n .Sh """ev_io"" \- is this file descriptor readable or writable" 1026.ie n .Sh """ev_io"" \- is this file descriptor readable or writable?"
824.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable" 1027.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable?"
825.IX Subsection "ev_io - is this file descriptor readable or writable" 1028.IX Subsection "ev_io - is this file descriptor readable or writable?"
826I/O watchers check whether a file descriptor is readable or writable 1029I/O watchers check whether a file descriptor is readable or writable
827in each iteration of the event loop (This behaviour is called 1030in each iteration of the event loop, or, more precisely, when reading
828level-triggering because you keep receiving events as long as the 1031would not block the process and writing would at least be able to write
829condition persists. Remember you can stop the watcher if you don't want to 1032some data. This behaviour is called level-triggering because you keep
830act on the event and neither want to receive future events). 1033receiving events as long as the condition persists. Remember you can stop
1034the watcher if you don't want to act on the event and neither want to
1035receive future events.
831.PP 1036.PP
832In general you can register as many read and/or write event watchers per 1037In general you can register as many read and/or write event watchers per
833fd as you want (as long as you don't confuse yourself). Setting all file 1038fd as you want (as long as you don't confuse yourself). Setting all file
834descriptors to non-blocking mode is also usually a good idea (but not 1039descriptors to non-blocking mode is also usually a good idea (but not
835required if you know what you are doing). 1040required if you know what you are doing).
836.PP 1041.PP
837You have to be careful with dup'ed file descriptors, though. Some backends 1042You have to be careful with dup'ed file descriptors, though. Some backends
838(the linux epoll backend is a notable example) cannot handle dup'ed file 1043(the linux epoll backend is a notable example) cannot handle dup'ed file
839descriptors correctly if you register interest in two or more fds pointing 1044descriptors correctly if you register interest in two or more fds pointing
840to the same underlying file/socket etc. description (that is, they share 1045to the same underlying file/socket/etc. description (that is, they share
841the same underlying \*(L"file open\*(R"). 1046the same underlying \*(L"file open\*(R").
842.PP 1047.PP
843If you must do this, then force the use of a known-to-be-good backend 1048If you must do this, then force the use of a known-to-be-good backend
844(at the time of this writing, this includes only \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and 1049(at the time of this writing, this includes only \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and
845\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR). 1050\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR).
1051.PP
1052Another thing you have to watch out for is that it is quite easy to
1053receive \*(L"spurious\*(R" readyness notifications, that is your callback might
1054be called with \f(CW\*(C`EV_READ\*(C'\fR but a subsequent \f(CW\*(C`read\*(C'\fR(2) will actually block
1055because there is no data. Not only are some backends known to create a
1056lot of those (for example solaris ports), it is very easy to get into
1057this situation even with a relatively standard program structure. Thus
1058it is best to always use non-blocking I/O: An extra \f(CW\*(C`read\*(C'\fR(2) returning
1059\&\f(CW\*(C`EAGAIN\*(C'\fR is far preferable to a program hanging until some data arrives.
1060.PP
1061If you cannot run the fd in non-blocking mode (for example you should not
1062play around with an Xlib connection), then you have to seperately re-test
1063whether a file descriptor is really ready with a known-to-be good interface
1064such as poll (fortunately in our Xlib example, Xlib already does this on
1065its own, so its quite safe to use).
846.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4 1066.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4
847.IX Item "ev_io_init (ev_io *, callback, int fd, int events)" 1067.IX Item "ev_io_init (ev_io *, callback, int fd, int events)"
848.PD 0 1068.PD 0
849.IP "ev_io_set (ev_io *, int fd, int events)" 4 1069.IP "ev_io_set (ev_io *, int fd, int events)" 4
850.IX Item "ev_io_set (ev_io *, int fd, int events)" 1070.IX Item "ev_io_set (ev_io *, int fd, int events)"
851.PD 1071.PD
852Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The fd is the file descriptor to rceeive 1072Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The \f(CW\*(C`fd\*(C'\fR is the file descriptor to
853events for and events is either \f(CW\*(C`EV_READ\*(C'\fR, \f(CW\*(C`EV_WRITE\*(C'\fR or \f(CW\*(C`EV_READ | 1073rceeive events for and events is either \f(CW\*(C`EV_READ\*(C'\fR, \f(CW\*(C`EV_WRITE\*(C'\fR or
854EV_WRITE\*(C'\fR to receive the given events. 1074\&\f(CW\*(C`EV_READ | EV_WRITE\*(C'\fR to receive the given events.
855.Sp 1075.IP "int fd [read\-only]" 4
856Please note that most of the more scalable backend mechanisms (for example 1076.IX Item "int fd [read-only]"
857epoll and solaris ports) can result in spurious readyness notifications 1077The file descriptor being watched.
858for file descriptors, so you practically need to use non-blocking I/O (and 1078.IP "int events [read\-only]" 4
859treat callback invocation as hint only), or retest separately with a safe 1079.IX Item "int events [read-only]"
860interface before doing I/O (XLib can do this), or force the use of either 1080The events being watched.
861\&\f(CW\*(C`EVBACKEND_SELECT\*(C'\fR or \f(CW\*(C`EVBACKEND_POLL\*(C'\fR, which don't suffer from this
862problem. Also note that it is quite easy to have your callback invoked
863when the readyness condition is no longer valid even when employing
864typical ways of handling events, so its a good idea to use non-blocking
865I/O unconditionally.
866.PP 1081.PP
867Example: call \f(CW\*(C`stdin_readable_cb\*(C'\fR when \s-1STDIN_FILENO\s0 has become, well 1082Example: Call \f(CW\*(C`stdin_readable_cb\*(C'\fR when \s-1STDIN_FILENO\s0 has become, well
868readable, but only once. Since it is likely line\-buffered, you could 1083readable, but only once. Since it is likely line\-buffered, you could
869attempt to read a whole line in the callback: 1084attempt to read a whole line in the callback.
870.PP 1085.PP
871.Vb 6 1086.Vb 6
872\& static void 1087\& static void
873\& stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents) 1088\& stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
874\& { 1089\& {
883\& struct ev_io stdin_readable; 1098\& struct ev_io stdin_readable;
884\& ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ); 1099\& ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
885\& ev_io_start (loop, &stdin_readable); 1100\& ev_io_start (loop, &stdin_readable);
886\& ev_loop (loop, 0); 1101\& ev_loop (loop, 0);
887.Ve 1102.Ve
888.ie n .Sh """ev_timer"" \- relative and optionally recurring timeouts" 1103.ie n .Sh """ev_timer"" \- relative and optionally repeating timeouts"
889.el .Sh "\f(CWev_timer\fP \- relative and optionally recurring timeouts" 1104.el .Sh "\f(CWev_timer\fP \- relative and optionally repeating timeouts"
890.IX Subsection "ev_timer - relative and optionally recurring timeouts" 1105.IX Subsection "ev_timer - relative and optionally repeating timeouts"
891Timer watchers are simple relative timers that generate an event after a 1106Timer watchers are simple relative timers that generate an event after a
892given time, and optionally repeating in regular intervals after that. 1107given time, and optionally repeating in regular intervals after that.
893.PP 1108.PP
894The timers are based on real time, that is, if you register an event that 1109The timers are based on real time, that is, if you register an event that
895times out after an hour and you reset your system clock to last years 1110times out after an hour and you reset your system clock to last years
929.IP "ev_timer_again (loop)" 4 1144.IP "ev_timer_again (loop)" 4
930.IX Item "ev_timer_again (loop)" 1145.IX Item "ev_timer_again (loop)"
931This will act as if the timer timed out and restart it again if it is 1146This will act as if the timer timed out and restart it again if it is
932repeating. The exact semantics are: 1147repeating. The exact semantics are:
933.Sp 1148.Sp
1149If the timer is pending, its pending status is cleared.
1150.Sp
934If the timer is started but nonrepeating, stop it. 1151If the timer is started but nonrepeating, stop it (as if it timed out).
935.Sp 1152.Sp
936If the timer is repeating, either start it if necessary (with the repeat 1153If the timer is repeating, either start it if necessary (with the
937value), or reset the running timer to the repeat value. 1154\&\f(CW\*(C`repeat\*(C'\fR value), or reset the running timer to the \f(CW\*(C`repeat\*(C'\fR value.
938.Sp 1155.Sp
939This sounds a bit complicated, but here is a useful and typical 1156This sounds a bit complicated, but here is a useful and typical
940example: Imagine you have a tcp connection and you want a so-called idle 1157example: Imagine you have a tcp connection and you want a so-called idle
941timeout, that is, you want to be called when there have been, say, 60 1158timeout, that is, you want to be called when there have been, say, 60
942seconds of inactivity on the socket. The easiest way to do this is to 1159seconds of inactivity on the socket. The easiest way to do this is to
943configure an \f(CW\*(C`ev_timer\*(C'\fR with after=repeat=60 and calling ev_timer_again each 1160configure an \f(CW\*(C`ev_timer\*(C'\fR with a \f(CW\*(C`repeat\*(C'\fR value of \f(CW60\fR and then call
944time you successfully read or write some data. If you go into an idle 1161\&\f(CW\*(C`ev_timer_again\*(C'\fR each time you successfully read or write some data. If
945state where you do not expect data to travel on the socket, you can stop 1162you go into an idle state where you do not expect data to travel on the
946the timer, and again will automatically restart it if need be. 1163socket, you can \f(CW\*(C`ev_timer_stop\*(C'\fR the timer, and \f(CW\*(C`ev_timer_again\*(C'\fR will
1164automatically restart it if need be.
1165.Sp
1166That means you can ignore the \f(CW\*(C`after\*(C'\fR value and \f(CW\*(C`ev_timer_start\*(C'\fR
1167altogether and only ever use the \f(CW\*(C`repeat\*(C'\fR value and \f(CW\*(C`ev_timer_again\*(C'\fR:
1168.Sp
1169.Vb 8
1170\& ev_timer_init (timer, callback, 0., 5.);
1171\& ev_timer_again (loop, timer);
1172\& ...
1173\& timer->again = 17.;
1174\& ev_timer_again (loop, timer);
1175\& ...
1176\& timer->again = 10.;
1177\& ev_timer_again (loop, timer);
1178.Ve
1179.Sp
1180This is more slightly efficient then stopping/starting the timer each time
1181you want to modify its timeout value.
1182.IP "ev_tstamp repeat [read\-write]" 4
1183.IX Item "ev_tstamp repeat [read-write]"
1184The current \f(CW\*(C`repeat\*(C'\fR value. Will be used each time the watcher times out
1185or \f(CW\*(C`ev_timer_again\*(C'\fR is called and determines the next timeout (if any),
1186which is also when any modifications are taken into account.
947.PP 1187.PP
948Example: create a timer that fires after 60 seconds. 1188Example: Create a timer that fires after 60 seconds.
949.PP 1189.PP
950.Vb 5 1190.Vb 5
951\& static void 1191\& static void
952\& one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents) 1192\& one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
953\& { 1193\& {
959\& struct ev_timer mytimer; 1199\& struct ev_timer mytimer;
960\& ev_timer_init (&mytimer, one_minute_cb, 60., 0.); 1200\& ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
961\& ev_timer_start (loop, &mytimer); 1201\& ev_timer_start (loop, &mytimer);
962.Ve 1202.Ve
963.PP 1203.PP
964Example: create a timeout timer that times out after 10 seconds of 1204Example: Create a timeout timer that times out after 10 seconds of
965inactivity. 1205inactivity.
966.PP 1206.PP
967.Vb 5 1207.Vb 5
968\& static void 1208\& static void
969\& timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents) 1209\& timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
982.Vb 3 1222.Vb 3
983\& // and in some piece of code that gets executed on any "activity": 1223\& // and in some piece of code that gets executed on any "activity":
984\& // reset the timeout to start ticking again at 10 seconds 1224\& // reset the timeout to start ticking again at 10 seconds
985\& ev_timer_again (&mytimer); 1225\& ev_timer_again (&mytimer);
986.Ve 1226.Ve
987.ie n .Sh """ev_periodic"" \- to cron or not to cron" 1227.ie n .Sh """ev_periodic"" \- to cron or not to cron?"
988.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron" 1228.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron?"
989.IX Subsection "ev_periodic - to cron or not to cron" 1229.IX Subsection "ev_periodic - to cron or not to cron?"
990Periodic watchers are also timers of a kind, but they are very versatile 1230Periodic watchers are also timers of a kind, but they are very versatile
991(and unfortunately a bit complex). 1231(and unfortunately a bit complex).
992.PP 1232.PP
993Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time) 1233Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time)
994but on wallclock time (absolute time). You can tell a periodic watcher 1234but on wallclock time (absolute time). You can tell a periodic watcher
995to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a 1235to trigger \*(L"at\*(R" 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 () 1236periodic watcher to trigger in 10 seconds (by specifiying e.g. \f(CW\*(C`ev_now ()
997+ 10.>) and then reset your system clock to the last year, then it will 1237+ 10.\*(C'\fR) and then reset your system clock to the last year, then it will
998take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger 1238take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger
999roughly 10 seconds later and of course not if you reset your system time 1239roughly 10 seconds later and of course not if you reset your system time
1000again). 1240again).
1001.PP 1241.PP
1002They can also be used to implement vastly more complex timers, such as 1242They can also be used to implement vastly more complex timers, such as
1083.IX Item "ev_periodic_again (loop, ev_periodic *)" 1323.IX Item "ev_periodic_again (loop, ev_periodic *)"
1084Simply stops and restarts the periodic watcher again. This is only useful 1324Simply stops and restarts the periodic watcher again. This is only useful
1085when you changed some parameters or the reschedule callback would return 1325when you changed some parameters or the reschedule callback would return
1086a different time than the last time it was called (e.g. in a crond like 1326a different time than the last time it was called (e.g. in a crond like
1087program when the crontabs have changed). 1327program when the crontabs have changed).
1328.IP "ev_tstamp interval [read\-write]" 4
1329.IX Item "ev_tstamp interval [read-write]"
1330The current interval value. Can be modified any time, but changes only
1331take effect when the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being
1332called.
1333.IP "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read\-write]" 4
1334.IX Item "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]"
1335The current reschedule callback, or \f(CW0\fR, if this functionality is
1336switched off. Can be changed any time, but changes only take effect when
1337the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being called.
1088.PP 1338.PP
1089Example: call a callback every hour, or, more precisely, whenever the 1339Example: Call a callback every hour, or, more precisely, whenever the
1090system clock is divisible by 3600. The callback invocation times have 1340system clock is divisible by 3600. The callback invocation times have
1091potentially a lot of jittering, but good long-term stability. 1341potentially a lot of jittering, but good long-term stability.
1092.PP 1342.PP
1093.Vb 5 1343.Vb 5
1094\& static void 1344\& static void
1102\& struct ev_periodic hourly_tick; 1352\& struct ev_periodic hourly_tick;
1103\& ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0); 1353\& ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1104\& ev_periodic_start (loop, &hourly_tick); 1354\& ev_periodic_start (loop, &hourly_tick);
1105.Ve 1355.Ve
1106.PP 1356.PP
1107Example: the same as above, but use a reschedule callback to do it: 1357Example: The same as above, but use a reschedule callback to do it:
1108.PP 1358.PP
1109.Vb 1 1359.Vb 1
1110\& #include <math.h> 1360\& #include <math.h>
1111.Ve 1361.Ve
1112.PP 1362.PP
1120.PP 1370.PP
1121.Vb 1 1371.Vb 1
1122\& ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb); 1372\& ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1123.Ve 1373.Ve
1124.PP 1374.PP
1125Example: call a callback every hour, starting now: 1375Example: Call a callback every hour, starting now:
1126.PP 1376.PP
1127.Vb 4 1377.Vb 4
1128\& struct ev_periodic hourly_tick; 1378\& struct ev_periodic hourly_tick;
1129\& ev_periodic_init (&hourly_tick, clock_cb, 1379\& ev_periodic_init (&hourly_tick, clock_cb,
1130\& fmod (ev_now (loop), 3600.), 3600., 0); 1380\& fmod (ev_now (loop), 3600.), 3600., 0);
1131\& ev_periodic_start (loop, &hourly_tick); 1381\& ev_periodic_start (loop, &hourly_tick);
1132.Ve 1382.Ve
1133.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled" 1383.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled!"
1134.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled" 1384.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled!"
1135.IX Subsection "ev_signal - signal me when a signal gets signalled" 1385.IX Subsection "ev_signal - signal me when a signal gets signalled!"
1136Signal watchers will trigger an event when the process receives a specific 1386Signal watchers will trigger an event when the process receives a specific
1137signal one or more times. Even though signals are very asynchronous, libev 1387signal one or more times. Even though signals are very asynchronous, libev
1138will try it's best to deliver signals synchronously, i.e. as part of the 1388will try it's best to deliver signals synchronously, i.e. as part of the
1139normal event processing, like any other event. 1389normal event processing, like any other event.
1140.PP 1390.PP
1150.IP "ev_signal_set (ev_signal *, int signum)" 4 1400.IP "ev_signal_set (ev_signal *, int signum)" 4
1151.IX Item "ev_signal_set (ev_signal *, int signum)" 1401.IX Item "ev_signal_set (ev_signal *, int signum)"
1152.PD 1402.PD
1153Configures the watcher to trigger on the given signal number (usually one 1403Configures the watcher to trigger on the given signal number (usually one
1154of the \f(CW\*(C`SIGxxx\*(C'\fR constants). 1404of the \f(CW\*(C`SIGxxx\*(C'\fR constants).
1405.IP "int signum [read\-only]" 4
1406.IX Item "int signum [read-only]"
1407The signal the watcher watches out for.
1155.ie n .Sh """ev_child"" \- wait for pid status changes" 1408.ie n .Sh """ev_child"" \- watch out for process status changes"
1156.el .Sh "\f(CWev_child\fP \- wait for pid status changes" 1409.el .Sh "\f(CWev_child\fP \- watch out for process status changes"
1157.IX Subsection "ev_child - wait for pid status changes" 1410.IX Subsection "ev_child - watch out for process status changes"
1158Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to 1411Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to
1159some child status changes (most typically when a child of yours dies). 1412some child status changes (most typically when a child of yours dies).
1160.IP "ev_child_init (ev_child *, callback, int pid)" 4 1413.IP "ev_child_init (ev_child *, callback, int pid)" 4
1161.IX Item "ev_child_init (ev_child *, callback, int pid)" 1414.IX Item "ev_child_init (ev_child *, callback, int pid)"
1162.PD 0 1415.PD 0
1167\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look 1420\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look
1168at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see 1421at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see
1169the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems 1422the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems
1170\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the 1423\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the
1171process causing the status change. 1424process causing the status change.
1425.IP "int pid [read\-only]" 4
1426.IX Item "int pid [read-only]"
1427The process id this watcher watches out for, or \f(CW0\fR, meaning any process id.
1428.IP "int rpid [read\-write]" 4
1429.IX Item "int rpid [read-write]"
1430The process id that detected a status change.
1431.IP "int rstatus [read\-write]" 4
1432.IX Item "int rstatus [read-write]"
1433The process exit/trace status caused by \f(CW\*(C`rpid\*(C'\fR (see your systems
1434\&\f(CW\*(C`waitpid\*(C'\fR and \f(CW\*(C`sys/wait.h\*(C'\fR documentation for details).
1172.PP 1435.PP
1173Example: try to exit cleanly on \s-1SIGINT\s0 and \s-1SIGTERM\s0. 1436Example: Try to exit cleanly on \s-1SIGINT\s0 and \s-1SIGTERM\s0.
1174.PP 1437.PP
1175.Vb 5 1438.Vb 5
1176\& static void 1439\& static void
1177\& sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents) 1440\& sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1178\& { 1441\& {
1183.Vb 3 1446.Vb 3
1184\& struct ev_signal signal_watcher; 1447\& struct ev_signal signal_watcher;
1185\& ev_signal_init (&signal_watcher, sigint_cb, SIGINT); 1448\& ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1186\& ev_signal_start (loop, &sigint_cb); 1449\& ev_signal_start (loop, &sigint_cb);
1187.Ve 1450.Ve
1451.ie n .Sh """ev_stat"" \- did the file attributes just change?"
1452.el .Sh "\f(CWev_stat\fP \- did the file attributes just change?"
1453.IX Subsection "ev_stat - did the file attributes just change?"
1454This watches a filesystem path for attribute changes. That is, it calls
1455\&\f(CW\*(C`stat\*(C'\fR regularly (or when the \s-1OS\s0 says it changed) and sees if it changed
1456compared to the last time, invoking the callback if it did.
1457.PP
1458The path does not need to exist: changing from \*(L"path exists\*(R" to \*(L"path does
1459not exist\*(R" is a status change like any other. The condition \*(L"path does
1460not exist\*(R" is signified by the \f(CW\*(C`st_nlink\*(C'\fR field being zero (which is
1461otherwise always forced to be at least one) and all the other fields of
1462the stat buffer having unspecified contents.
1463.PP
1464The path \fIshould\fR be absolute and \fImust not\fR end in a slash. If it is
1465relative and your working directory changes, the behaviour is undefined.
1466.PP
1467Since there is no standard to do this, the portable implementation simply
1468calls \f(CW\*(C`stat (2)\*(C'\fR regularly on the path to see if it changed somehow. You
1469can specify a recommended polling interval for this case. If you specify
1470a polling interval of \f(CW0\fR (highly recommended!) then a \fIsuitable,
1471unspecified default\fR value will be used (which you can expect to be around
1472five seconds, although this might change dynamically). Libev will also
1473impose a minimum interval which is currently around \f(CW0.1\fR, but thats
1474usually overkill.
1475.PP
1476This watcher type is not meant for massive numbers of stat watchers,
1477as even with OS-supported change notifications, this can be
1478resource\-intensive.
1479.PP
1480At the time of this writing, only the Linux inotify interface is
1481implemented (implementing kqueue support is left as an exercise for the
1482reader). Inotify will be used to give hints only and should not change the
1483semantics of \f(CW\*(C`ev_stat\*(C'\fR watchers, which means that libev sometimes needs
1484to fall back to regular polling again even with inotify, but changes are
1485usually detected immediately, and if the file exists there will be no
1486polling.
1487.IP "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)" 4
1488.IX Item "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)"
1489.PD 0
1490.IP "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)" 4
1491.IX Item "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)"
1492.PD
1493Configures the watcher to wait for status changes of the given
1494\&\f(CW\*(C`path\*(C'\fR. The \f(CW\*(C`interval\*(C'\fR is a hint on how quickly a change is expected to
1495be detected and should normally be specified as \f(CW0\fR to let libev choose
1496a suitable value. The memory pointed to by \f(CW\*(C`path\*(C'\fR must point to the same
1497path for as long as the watcher is active.
1498.Sp
1499The callback will be receive \f(CW\*(C`EV_STAT\*(C'\fR when a change was detected,
1500relative to the attributes at the time the watcher was started (or the
1501last change was detected).
1502.IP "ev_stat_stat (ev_stat *)" 4
1503.IX Item "ev_stat_stat (ev_stat *)"
1504Updates the stat buffer immediately with new values. If you change the
1505watched path in your callback, you could call this fucntion to avoid
1506detecting this change (while introducing a race condition). Can also be
1507useful simply to find out the new values.
1508.IP "ev_statdata attr [read\-only]" 4
1509.IX Item "ev_statdata attr [read-only]"
1510The most-recently detected attributes of the file. Although the type is of
1511\&\f(CW\*(C`ev_statdata\*(C'\fR, this is usually the (or one of the) \f(CW\*(C`struct stat\*(C'\fR types
1512suitable for your system. If the \f(CW\*(C`st_nlink\*(C'\fR member is \f(CW0\fR, then there
1513was some error while \f(CW\*(C`stat\*(C'\fRing the file.
1514.IP "ev_statdata prev [read\-only]" 4
1515.IX Item "ev_statdata prev [read-only]"
1516The previous attributes of the file. The callback gets invoked whenever
1517\&\f(CW\*(C`prev\*(C'\fR != \f(CW\*(C`attr\*(C'\fR.
1518.IP "ev_tstamp interval [read\-only]" 4
1519.IX Item "ev_tstamp interval [read-only]"
1520The specified interval.
1521.IP "const char *path [read\-only]" 4
1522.IX Item "const char *path [read-only]"
1523The filesystem path that is being watched.
1524.PP
1525Example: Watch \f(CW\*(C`/etc/passwd\*(C'\fR for attribute changes.
1526.PP
1527.Vb 15
1528\& static void
1529\& passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1530\& {
1531\& /* /etc/passwd changed in some way */
1532\& if (w->attr.st_nlink)
1533\& {
1534\& printf ("passwd current size %ld\en", (long)w->attr.st_size);
1535\& printf ("passwd current atime %ld\en", (long)w->attr.st_mtime);
1536\& printf ("passwd current mtime %ld\en", (long)w->attr.st_mtime);
1537\& }
1538\& else
1539\& /* you shalt not abuse printf for puts */
1540\& puts ("wow, /etc/passwd is not there, expect problems. "
1541\& "if this is windows, they already arrived\en");
1542\& }
1543.Ve
1544.PP
1545.Vb 2
1546\& ...
1547\& ev_stat passwd;
1548.Ve
1549.PP
1550.Vb 2
1551\& ev_stat_init (&passwd, passwd_cb, "/etc/passwd");
1552\& ev_stat_start (loop, &passwd);
1553.Ve
1188.ie n .Sh """ev_idle"" \- when you've got nothing better to do" 1554.ie n .Sh """ev_idle"" \- when you've got nothing better to do..."
1189.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do" 1555.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do..."
1190.IX Subsection "ev_idle - when you've got nothing better to do" 1556.IX Subsection "ev_idle - when you've got nothing better to do..."
1191Idle watchers trigger events when there are no other events are pending 1557Idle watchers trigger events when no other events of the same or higher
1192(prepare, check and other idle watchers do not count). That is, as long 1558priority are pending (prepare, check and other idle watchers do not
1193as your process is busy handling sockets or timeouts (or even signals, 1559count).
1194imagine) it will not be triggered. But when your process is idle all idle 1560.PP
1195watchers are being called again and again, once per event loop iteration \- 1561That is, as long as your process is busy handling sockets or timeouts
1562(or even signals, imagine) of the same or higher priority it will not be
1563triggered. But when your process is idle (or only lower-priority watchers
1564are pending), the idle watchers are being called once per event loop
1196until stopped, that is, or your process receives more events and becomes 1565iteration \- until stopped, that is, or your process receives more events
1197busy. 1566and becomes busy again with higher priority stuff.
1198.PP 1567.PP
1199The most noteworthy effect is that as long as any idle watchers are 1568The most noteworthy effect is that as long as any idle watchers are
1200active, the process will not block when waiting for new events. 1569active, the process will not block when waiting for new events.
1201.PP 1570.PP
1202Apart from keeping your process non-blocking (which is a useful 1571Apart from keeping your process non-blocking (which is a useful
1207.IX Item "ev_idle_init (ev_signal *, callback)" 1576.IX Item "ev_idle_init (ev_signal *, callback)"
1208Initialises and configures the idle watcher \- it has no parameters of any 1577Initialises and configures the idle watcher \- it has no parameters of any
1209kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless, 1578kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless,
1210believe me. 1579believe me.
1211.PP 1580.PP
1212Example: dynamically allocate an \f(CW\*(C`ev_idle\*(C'\fR, start it, and in the 1581Example: Dynamically allocate an \f(CW\*(C`ev_idle\*(C'\fR watcher, start it, and in the
1213callback, free it. Alos, use no error checking, as usual. 1582callback, free it. Also, use no error checking, as usual.
1214.PP 1583.PP
1215.Vb 7 1584.Vb 7
1216\& static void 1585\& static void
1217\& idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents) 1586\& idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1218\& { 1587\& {
1225.Vb 3 1594.Vb 3
1226\& struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle)); 1595\& struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1227\& ev_idle_init (idle_watcher, idle_cb); 1596\& ev_idle_init (idle_watcher, idle_cb);
1228\& ev_idle_start (loop, idle_cb); 1597\& ev_idle_start (loop, idle_cb);
1229.Ve 1598.Ve
1230.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop" 1599.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop!"
1231.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop" 1600.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop!"
1232.IX Subsection "ev_prepare and ev_check - customise your event loop" 1601.IX Subsection "ev_prepare and ev_check - customise your event loop!"
1233Prepare and check watchers are usually (but not always) used in tandem: 1602Prepare and check watchers are usually (but not always) used in tandem:
1234prepare watchers get invoked before the process blocks and check watchers 1603prepare watchers get invoked before the process blocks and check watchers
1235afterwards. 1604afterwards.
1236.PP 1605.PP
1606You \fImust not\fR call \f(CW\*(C`ev_loop\*(C'\fR or similar functions that enter
1607the current event loop from either \f(CW\*(C`ev_prepare\*(C'\fR or \f(CW\*(C`ev_check\*(C'\fR
1608watchers. Other loops than the current one are fine, however. The
1609rationale behind this is that you do not need to check for recursion in
1610those watchers, i.e. the sequence will always be \f(CW\*(C`ev_prepare\*(C'\fR, blocking,
1611\&\f(CW\*(C`ev_check\*(C'\fR so if you have one watcher of each kind they will always be
1612called in pairs bracketing the blocking call.
1613.PP
1237Their main purpose is to integrate other event mechanisms into libev and 1614Their main purpose is to integrate other event mechanisms into libev and
1238their use is somewhat advanced. This could be used, for example, to track 1615their use is somewhat advanced. This could be used, for example, to track
1239variable changes, implement your own watchers, integrate net-snmp or a 1616variable changes, implement your own watchers, integrate net-snmp or a
1240coroutine library and lots more. 1617coroutine library and lots more. They are also occasionally useful if
1618you cache some data and want to flush it before blocking (for example,
1619in X programs you might want to do an \f(CW\*(C`XFlush ()\*(C'\fR in an \f(CW\*(C`ev_prepare\*(C'\fR
1620watcher).
1241.PP 1621.PP
1242This is done by examining in each prepare call which file descriptors need 1622This is done by examining in each prepare call which file descriptors need
1243to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for 1623to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for
1244them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries 1624them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries
1245provide just this functionality). Then, in the check watcher you check for 1625provide just this functionality). Then, in the check watcher you check for
1264.PD 1644.PD
1265Initialises and configures the prepare or check watcher \- they have no 1645Initialises and configures the prepare or check watcher \- they have no
1266parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR 1646parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR
1267macros, but using them is utterly, utterly and completely pointless. 1647macros, but using them is utterly, utterly and completely pointless.
1268.PP 1648.PP
1269Example: *TODO*. 1649There are a number of principal ways to embed other event loops or modules
1650into libev. Here are some ideas on how to include libadns into libev
1651(there is a Perl module named \f(CW\*(C`EV::ADNS\*(C'\fR that does this, which you could
1652use for an actually working example. Another Perl module named \f(CW\*(C`EV::Glib\*(C'\fR
1653embeds a Glib main context into libev, and finally, \f(CW\*(C`Glib::EV\*(C'\fR embeds \s-1EV\s0
1654into the Glib event loop).
1655.PP
1656Method 1: Add \s-1IO\s0 watchers and a timeout watcher in a prepare handler,
1657and in a check watcher, destroy them and call into libadns. What follows
1658is pseudo-code only of course. This requires you to either use a low
1659priority for the check watcher or use \f(CW\*(C`ev_clear_pending\*(C'\fR explicitly, as
1660the callbacks for the IO/timeout watchers might not have been called yet.
1661.PP
1662.Vb 2
1663\& static ev_io iow [nfd];
1664\& static ev_timer tw;
1665.Ve
1666.PP
1667.Vb 4
1668\& static void
1669\& io_cb (ev_loop *loop, ev_io *w, int revents)
1670\& {
1671\& }
1672.Ve
1673.PP
1674.Vb 8
1675\& // create io watchers for each fd and a timer before blocking
1676\& static void
1677\& adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1678\& {
1679\& int timeout = 3600000;
1680\& struct pollfd fds [nfd];
1681\& // actual code will need to loop here and realloc etc.
1682\& adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1683.Ve
1684.PP
1685.Vb 3
1686\& /* the callback is illegal, but won't be called as we stop during check */
1687\& ev_timer_init (&tw, 0, timeout * 1e-3);
1688\& ev_timer_start (loop, &tw);
1689.Ve
1690.PP
1691.Vb 6
1692\& // create one ev_io per pollfd
1693\& for (int i = 0; i < nfd; ++i)
1694\& {
1695\& ev_io_init (iow + i, io_cb, fds [i].fd,
1696\& ((fds [i].events & POLLIN ? EV_READ : 0)
1697\& | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1698.Ve
1699.PP
1700.Vb 4
1701\& fds [i].revents = 0;
1702\& ev_io_start (loop, iow + i);
1703\& }
1704\& }
1705.Ve
1706.PP
1707.Vb 5
1708\& // stop all watchers after blocking
1709\& static void
1710\& adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1711\& {
1712\& ev_timer_stop (loop, &tw);
1713.Ve
1714.PP
1715.Vb 8
1716\& for (int i = 0; i < nfd; ++i)
1717\& {
1718\& // set the relevant poll flags
1719\& // could also call adns_processreadable etc. here
1720\& struct pollfd *fd = fds + i;
1721\& int revents = ev_clear_pending (iow + i);
1722\& if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1723\& if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1724.Ve
1725.PP
1726.Vb 3
1727\& // now stop the watcher
1728\& ev_io_stop (loop, iow + i);
1729\& }
1730.Ve
1731.PP
1732.Vb 2
1733\& adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1734\& }
1735.Ve
1736.PP
1737Method 2: This would be just like method 1, but you run \f(CW\*(C`adns_afterpoll\*(C'\fR
1738in the prepare watcher and would dispose of the check watcher.
1739.PP
1740Method 3: If the module to be embedded supports explicit event
1741notification (adns does), you can also make use of the actual watcher
1742callbacks, and only destroy/create the watchers in the prepare watcher.
1743.PP
1744.Vb 5
1745\& static void
1746\& timer_cb (EV_P_ ev_timer *w, int revents)
1747\& {
1748\& adns_state ads = (adns_state)w->data;
1749\& update_now (EV_A);
1750.Ve
1751.PP
1752.Vb 2
1753\& adns_processtimeouts (ads, &tv_now);
1754\& }
1755.Ve
1756.PP
1757.Vb 5
1758\& static void
1759\& io_cb (EV_P_ ev_io *w, int revents)
1760\& {
1761\& adns_state ads = (adns_state)w->data;
1762\& update_now (EV_A);
1763.Ve
1764.PP
1765.Vb 3
1766\& if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
1767\& if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
1768\& }
1769.Ve
1770.PP
1771.Vb 1
1772\& // do not ever call adns_afterpoll
1773.Ve
1774.PP
1775Method 4: Do not use a prepare or check watcher because the module you
1776want to embed is too inflexible to support it. Instead, youc na override
1777their poll function. The drawback with this solution is that the main
1778loop is now no longer controllable by \s-1EV\s0. The \f(CW\*(C`Glib::EV\*(C'\fR module does
1779this.
1780.PP
1781.Vb 4
1782\& static gint
1783\& event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1784\& {
1785\& int got_events = 0;
1786.Ve
1787.PP
1788.Vb 2
1789\& for (n = 0; n < nfds; ++n)
1790\& // create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1791.Ve
1792.PP
1793.Vb 2
1794\& if (timeout >= 0)
1795\& // create/start timer
1796.Ve
1797.PP
1798.Vb 2
1799\& // poll
1800\& ev_loop (EV_A_ 0);
1801.Ve
1802.PP
1803.Vb 3
1804\& // stop timer again
1805\& if (timeout >= 0)
1806\& ev_timer_stop (EV_A_ &to);
1807.Ve
1808.PP
1809.Vb 3
1810\& // stop io watchers again - their callbacks should have set
1811\& for (n = 0; n < nfds; ++n)
1812\& ev_io_stop (EV_A_ iow [n]);
1813.Ve
1814.PP
1815.Vb 2
1816\& return got_events;
1817\& }
1818.Ve
1270.ie n .Sh """ev_embed"" \- when one backend isn't enough" 1819.ie n .Sh """ev_embed"" \- when one backend isn't enough..."
1271.el .Sh "\f(CWev_embed\fP \- when one backend isn't enough" 1820.el .Sh "\f(CWev_embed\fP \- when one backend isn't enough..."
1272.IX Subsection "ev_embed - when one backend isn't enough" 1821.IX Subsection "ev_embed - when one backend isn't enough..."
1273This is a rather advanced watcher type that lets you embed one event loop 1822This is a rather advanced watcher type that lets you embed one event loop
1274into another (currently only \f(CW\*(C`ev_io\*(C'\fR events are supported in the embedded 1823into another (currently only \f(CW\*(C`ev_io\*(C'\fR events are supported in the embedded
1275loop, other types of watchers might be handled in a delayed or incorrect 1824loop, other types of watchers might be handled in a delayed or incorrect
1276fashion and must not be used). 1825fashion and must not be used).
1277.PP 1826.PP
1357.IP "ev_embed_sweep (loop, ev_embed *)" 4 1906.IP "ev_embed_sweep (loop, ev_embed *)" 4
1358.IX Item "ev_embed_sweep (loop, ev_embed *)" 1907.IX Item "ev_embed_sweep (loop, ev_embed *)"
1359Make a single, non-blocking sweep over the embedded loop. This works 1908Make a single, non-blocking sweep over the embedded loop. This works
1360similarly to \f(CW\*(C`ev_loop (embedded_loop, EVLOOP_NONBLOCK)\*(C'\fR, but in the most 1909similarly to \f(CW\*(C`ev_loop (embedded_loop, EVLOOP_NONBLOCK)\*(C'\fR, but in the most
1361apropriate way for embedded loops. 1910apropriate way for embedded loops.
1911.IP "struct ev_loop *loop [read\-only]" 4
1912.IX Item "struct ev_loop *loop [read-only]"
1913The embedded event loop.
1914.ie n .Sh """ev_fork"" \- the audacity to resume the event loop after a fork"
1915.el .Sh "\f(CWev_fork\fP \- the audacity to resume the event loop after a fork"
1916.IX Subsection "ev_fork - the audacity to resume the event loop after a fork"
1917Fork watchers are called when a \f(CW\*(C`fork ()\*(C'\fR was detected (usually because
1918whoever is a good citizen cared to tell libev about it by calling
1919\&\f(CW\*(C`ev_default_fork\*(C'\fR or \f(CW\*(C`ev_loop_fork\*(C'\fR). The invocation is done before the
1920event loop blocks next and before \f(CW\*(C`ev_check\*(C'\fR watchers are being called,
1921and only in the child after the fork. If whoever good citizen calling
1922\&\f(CW\*(C`ev_default_fork\*(C'\fR cheats and calls it in the wrong process, the fork
1923handlers will be invoked, too, of course.
1924.IP "ev_fork_init (ev_signal *, callback)" 4
1925.IX Item "ev_fork_init (ev_signal *, callback)"
1926Initialises and configures the fork watcher \- it has no parameters of any
1927kind. There is a \f(CW\*(C`ev_fork_set\*(C'\fR macro, but using it is utterly pointless,
1928believe me.
1362.SH "OTHER FUNCTIONS" 1929.SH "OTHER FUNCTIONS"
1363.IX Header "OTHER FUNCTIONS" 1930.IX Header "OTHER FUNCTIONS"
1364There are some other functions of possible interest. Described. Here. Now. 1931There are some other functions of possible interest. Described. Here. Now.
1365.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4 1932.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4
1366.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 1933.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)"
1428.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4 1995.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4
1429.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library." 1996.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library."
1430.PD 1997.PD
1431.SH "\*(C+ SUPPORT" 1998.SH "\*(C+ SUPPORT"
1432.IX Header " SUPPORT" 1999.IX Header " SUPPORT"
1433\&\s-1TBD\s0. 2000Libev comes with some simplistic wrapper classes for \*(C+ that mainly allow
2001you to use some convinience methods to start/stop watchers and also change
2002the callback model to a model using method callbacks on objects.
2003.PP
2004To use it,
2005.PP
2006.Vb 1
2007\& #include <ev++.h>
2008.Ve
2009.PP
2010This automatically includes \fIev.h\fR and puts all of its definitions (many
2011of them macros) into the global namespace. All \*(C+ specific things are
2012put into the \f(CW\*(C`ev\*(C'\fR namespace. It should support all the same embedding
2013options as \fIev.h\fR, most notably \f(CW\*(C`EV_MULTIPLICITY\*(C'\fR.
2014.PP
2015Care has been taken to keep the overhead low. The only data member the \*(C+
2016classes add (compared to plain C\-style watchers) is the event loop pointer
2017that the watcher is associated with (or no additional members at all if
2018you disable \f(CW\*(C`EV_MULTIPLICITY\*(C'\fR when embedding libev).
2019.PP
2020Currently, functions, and static and non-static member functions can be
2021used as callbacks. Other types should be easy to add as long as they only
2022need one additional pointer for context. If you need support for other
2023types of functors please contact the author (preferably after implementing
2024it).
2025.PP
2026Here is a list of things available in the \f(CW\*(C`ev\*(C'\fR namespace:
2027.ie n .IP """ev::READ""\fR, \f(CW""ev::WRITE"" etc." 4
2028.el .IP "\f(CWev::READ\fR, \f(CWev::WRITE\fR etc." 4
2029.IX Item "ev::READ, ev::WRITE etc."
2030These are just enum values with the same values as the \f(CW\*(C`EV_READ\*(C'\fR etc.
2031macros from \fIev.h\fR.
2032.ie n .IP """ev::tstamp""\fR, \f(CW""ev::now""" 4
2033.el .IP "\f(CWev::tstamp\fR, \f(CWev::now\fR" 4
2034.IX Item "ev::tstamp, ev::now"
2035Aliases to the same types/functions as with the \f(CW\*(C`ev_\*(C'\fR prefix.
2036.ie n .IP """ev::io""\fR, \f(CW""ev::timer""\fR, \f(CW""ev::periodic""\fR, \f(CW""ev::idle""\fR, \f(CW""ev::sig"" etc." 4
2037.el .IP "\f(CWev::io\fR, \f(CWev::timer\fR, \f(CWev::periodic\fR, \f(CWev::idle\fR, \f(CWev::sig\fR etc." 4
2038.IX Item "ev::io, ev::timer, ev::periodic, ev::idle, ev::sig etc."
2039For each \f(CW\*(C`ev_TYPE\*(C'\fR watcher in \fIev.h\fR there is a corresponding class of
2040the same name in the \f(CW\*(C`ev\*(C'\fR namespace, with the exception of \f(CW\*(C`ev_signal\*(C'\fR
2041which is called \f(CW\*(C`ev::sig\*(C'\fR to avoid clashes with the \f(CW\*(C`signal\*(C'\fR macro
2042defines by many implementations.
2043.Sp
2044All of those classes have these methods:
2045.RS 4
2046.IP "ev::TYPE::TYPE ()" 4
2047.IX Item "ev::TYPE::TYPE ()"
2048.PD 0
2049.IP "ev::TYPE::TYPE (struct ev_loop *)" 4
2050.IX Item "ev::TYPE::TYPE (struct ev_loop *)"
2051.IP "ev::TYPE::~TYPE" 4
2052.IX Item "ev::TYPE::~TYPE"
2053.PD
2054The constructor (optionally) takes an event loop to associate the watcher
2055with. If it is omitted, it will use \f(CW\*(C`EV_DEFAULT\*(C'\fR.
2056.Sp
2057The constructor calls \f(CW\*(C`ev_init\*(C'\fR for you, which means you have to call the
2058\&\f(CW\*(C`set\*(C'\fR method before starting it.
2059.Sp
2060It will not set a callback, however: You have to call the templated \f(CW\*(C`set\*(C'\fR
2061method to set a callback before you can start the watcher.
2062.Sp
2063(The reason why you have to use a method is a limitation in \*(C+ which does
2064not allow explicit template arguments for constructors).
2065.Sp
2066The destructor automatically stops the watcher if it is active.
2067.IP "w\->set<class, &class::method> (object *)" 4
2068.IX Item "w->set<class, &class::method> (object *)"
2069This method sets the callback method to call. The method has to have a
2070signature of \f(CW\*(C`void (*)(ev_TYPE &, int)\*(C'\fR, it receives the watcher as
2071first argument and the \f(CW\*(C`revents\*(C'\fR as second. The object must be given as
2072parameter and is stored in the \f(CW\*(C`data\*(C'\fR member of the watcher.
2073.Sp
2074This method synthesizes efficient thunking code to call your method from
2075the C callback that libev requires. If your compiler can inline your
2076callback (i.e. it is visible to it at the place of the \f(CW\*(C`set\*(C'\fR call and
2077your compiler is good :), then the method will be fully inlined into the
2078thunking function, making it as fast as a direct C callback.
2079.Sp
2080Example: simple class declaration and watcher initialisation
2081.Sp
2082.Vb 4
2083\& struct myclass
2084\& {
2085\& void io_cb (ev::io &w, int revents) { }
2086\& }
2087.Ve
2088.Sp
2089.Vb 3
2090\& myclass obj;
2091\& ev::io iow;
2092\& iow.set <myclass, &myclass::io_cb> (&obj);
2093.Ve
2094.IP "w\->set<function> (void *data = 0)" 4
2095.IX Item "w->set<function> (void *data = 0)"
2096Also sets a callback, but uses a static method or plain function as
2097callback. The optional \f(CW\*(C`data\*(C'\fR argument will be stored in the watcher's
2098\&\f(CW\*(C`data\*(C'\fR member and is free for you to use.
2099.Sp
2100The prototype of the \f(CW\*(C`function\*(C'\fR must be \f(CW\*(C`void (*)(ev::TYPE &w, int)\*(C'\fR.
2101.Sp
2102See the method\-\f(CW\*(C`set\*(C'\fR above for more details.
2103.Sp
2104Example:
2105.Sp
2106.Vb 2
2107\& static void io_cb (ev::io &w, int revents) { }
2108\& iow.set <io_cb> ();
2109.Ve
2110.IP "w\->set (struct ev_loop *)" 4
2111.IX Item "w->set (struct ev_loop *)"
2112Associates a different \f(CW\*(C`struct ev_loop\*(C'\fR with this watcher. You can only
2113do this when the watcher is inactive (and not pending either).
2114.IP "w\->set ([args])" 4
2115.IX Item "w->set ([args])"
2116Basically the same as \f(CW\*(C`ev_TYPE_set\*(C'\fR, with the same args. Must be
2117called at least once. Unlike the C counterpart, an active watcher gets
2118automatically stopped and restarted when reconfiguring it with this
2119method.
2120.IP "w\->start ()" 4
2121.IX Item "w->start ()"
2122Starts the watcher. Note that there is no \f(CW\*(C`loop\*(C'\fR argument, as the
2123constructor already stores the event loop.
2124.IP "w\->stop ()" 4
2125.IX Item "w->stop ()"
2126Stops the watcher if it is active. Again, no \f(CW\*(C`loop\*(C'\fR argument.
2127.ie n .IP "w\->again () ""ev::timer""\fR, \f(CW""ev::periodic"" only" 4
2128.el .IP "w\->again () \f(CWev::timer\fR, \f(CWev::periodic\fR only" 4
2129.IX Item "w->again () ev::timer, ev::periodic only"
2130For \f(CW\*(C`ev::timer\*(C'\fR and \f(CW\*(C`ev::periodic\*(C'\fR, this invokes the corresponding
2131\&\f(CW\*(C`ev_TYPE_again\*(C'\fR function.
2132.ie n .IP "w\->sweep () ""ev::embed"" only" 4
2133.el .IP "w\->sweep () \f(CWev::embed\fR only" 4
2134.IX Item "w->sweep () ev::embed only"
2135Invokes \f(CW\*(C`ev_embed_sweep\*(C'\fR.
2136.ie n .IP "w\->update () ""ev::stat"" only" 4
2137.el .IP "w\->update () \f(CWev::stat\fR only" 4
2138.IX Item "w->update () ev::stat only"
2139Invokes \f(CW\*(C`ev_stat_stat\*(C'\fR.
2140.RE
2141.RS 4
2142.RE
2143.PP
2144Example: Define a class with an \s-1IO\s0 and idle watcher, start one of them in
2145the constructor.
2146.PP
2147.Vb 4
2148\& class myclass
2149\& {
2150\& ev_io io; void io_cb (ev::io &w, int revents);
2151\& ev_idle idle void idle_cb (ev::idle &w, int revents);
2152.Ve
2153.PP
2154.Vb 2
2155\& myclass ();
2156\& }
2157.Ve
2158.PP
2159.Vb 4
2160\& myclass::myclass (int fd)
2161\& {
2162\& io .set <myclass, &myclass::io_cb > (this);
2163\& idle.set <myclass, &myclass::idle_cb> (this);
2164.Ve
2165.PP
2166.Vb 2
2167\& io.start (fd, ev::READ);
2168\& }
2169.Ve
2170.SH "MACRO MAGIC"
2171.IX Header "MACRO MAGIC"
2172Libev can be compiled with a variety of options, the most fundemantal is
2173\&\f(CW\*(C`EV_MULTIPLICITY\*(C'\fR. This option determines whether (most) functions and
2174callbacks have an initial \f(CW\*(C`struct ev_loop *\*(C'\fR argument.
2175.PP
2176To make it easier to write programs that cope with either variant, the
2177following macros are defined:
2178.ie n .IP """EV_A""\fR, \f(CW""EV_A_""" 4
2179.el .IP "\f(CWEV_A\fR, \f(CWEV_A_\fR" 4
2180.IX Item "EV_A, EV_A_"
2181This provides the loop \fIargument\fR for functions, if one is required (\*(L"ev
2182loop argument\*(R"). The \f(CW\*(C`EV_A\*(C'\fR form is used when this is the sole argument,
2183\&\f(CW\*(C`EV_A_\*(C'\fR is used when other arguments are following. Example:
2184.Sp
2185.Vb 3
2186\& ev_unref (EV_A);
2187\& ev_timer_add (EV_A_ watcher);
2188\& ev_loop (EV_A_ 0);
2189.Ve
2190.Sp
2191It assumes the variable \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR is in scope,
2192which is often provided by the following macro.
2193.ie n .IP """EV_P""\fR, \f(CW""EV_P_""" 4
2194.el .IP "\f(CWEV_P\fR, \f(CWEV_P_\fR" 4
2195.IX Item "EV_P, EV_P_"
2196This provides the loop \fIparameter\fR for functions, if one is required (\*(L"ev
2197loop parameter\*(R"). The \f(CW\*(C`EV_P\*(C'\fR form is used when this is the sole parameter,
2198\&\f(CW\*(C`EV_P_\*(C'\fR is used when other parameters are following. Example:
2199.Sp
2200.Vb 2
2201\& // this is how ev_unref is being declared
2202\& static void ev_unref (EV_P);
2203.Ve
2204.Sp
2205.Vb 2
2206\& // this is how you can declare your typical callback
2207\& static void cb (EV_P_ ev_timer *w, int revents)
2208.Ve
2209.Sp
2210It declares a parameter \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR, quite
2211suitable for use with \f(CW\*(C`EV_A\*(C'\fR.
2212.ie n .IP """EV_DEFAULT""\fR, \f(CW""EV_DEFAULT_""" 4
2213.el .IP "\f(CWEV_DEFAULT\fR, \f(CWEV_DEFAULT_\fR" 4
2214.IX Item "EV_DEFAULT, EV_DEFAULT_"
2215Similar to the other two macros, this gives you the value of the default
2216loop, if multiple loops are supported (\*(L"ev loop default\*(R").
2217.PP
2218Example: Declare and initialise a check watcher, utilising the above
2219macros so it will work regardless of whether multiple loops are supported
2220or not.
2221.PP
2222.Vb 5
2223\& static void
2224\& check_cb (EV_P_ ev_timer *w, int revents)
2225\& {
2226\& ev_check_stop (EV_A_ w);
2227\& }
2228.Ve
2229.PP
2230.Vb 4
2231\& ev_check check;
2232\& ev_check_init (&check, check_cb);
2233\& ev_check_start (EV_DEFAULT_ &check);
2234\& ev_loop (EV_DEFAULT_ 0);
2235.Ve
2236.SH "EMBEDDING"
2237.IX Header "EMBEDDING"
2238Libev can (and often is) directly embedded into host
2239applications. Examples of applications that embed it include the Deliantra
2240Game Server, the \s-1EV\s0 perl module, the \s-1GNU\s0 Virtual Private Ethernet (gvpe)
2241and rxvt\-unicode.
2242.PP
2243The goal is to enable you to just copy the neecssary files into your
2244source directory without having to change even a single line in them, so
2245you can easily upgrade by simply copying (or having a checked-out copy of
2246libev somewhere in your source tree).
2247.Sh "\s-1FILESETS\s0"
2248.IX Subsection "FILESETS"
2249Depending on what features you need you need to include one or more sets of files
2250in your app.
2251.PP
2252\fI\s-1CORE\s0 \s-1EVENT\s0 \s-1LOOP\s0\fR
2253.IX Subsection "CORE EVENT LOOP"
2254.PP
2255To include only the libev core (all the \f(CW\*(C`ev_*\*(C'\fR functions), with manual
2256configuration (no autoconf):
2257.PP
2258.Vb 2
2259\& #define EV_STANDALONE 1
2260\& #include "ev.c"
2261.Ve
2262.PP
2263This will automatically include \fIev.h\fR, too, and should be done in a
2264single C source file only to provide the function implementations. To use
2265it, do the same for \fIev.h\fR in all files wishing to use this \s-1API\s0 (best
2266done by writing a wrapper around \fIev.h\fR that you can include instead and
2267where you can put other configuration options):
2268.PP
2269.Vb 2
2270\& #define EV_STANDALONE 1
2271\& #include "ev.h"
2272.Ve
2273.PP
2274Both header files and implementation files can be compiled with a \*(C+
2275compiler (at least, thats a stated goal, and breakage will be treated
2276as a bug).
2277.PP
2278You need the following files in your source tree, or in a directory
2279in your include path (e.g. in libev/ when using \-Ilibev):
2280.PP
2281.Vb 4
2282\& ev.h
2283\& ev.c
2284\& ev_vars.h
2285\& ev_wrap.h
2286.Ve
2287.PP
2288.Vb 1
2289\& ev_win32.c required on win32 platforms only
2290.Ve
2291.PP
2292.Vb 5
2293\& ev_select.c only when select backend is enabled (which is enabled by default)
2294\& ev_poll.c only when poll backend is enabled (disabled by default)
2295\& ev_epoll.c only when the epoll backend is enabled (disabled by default)
2296\& ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2297\& ev_port.c only when the solaris port backend is enabled (disabled by default)
2298.Ve
2299.PP
2300\&\fIev.c\fR includes the backend files directly when enabled, so you only need
2301to compile this single file.
2302.PP
2303\fI\s-1LIBEVENT\s0 \s-1COMPATIBILITY\s0 \s-1API\s0\fR
2304.IX Subsection "LIBEVENT COMPATIBILITY API"
2305.PP
2306To include the libevent compatibility \s-1API\s0, also include:
2307.PP
2308.Vb 1
2309\& #include "event.c"
2310.Ve
2311.PP
2312in the file including \fIev.c\fR, and:
2313.PP
2314.Vb 1
2315\& #include "event.h"
2316.Ve
2317.PP
2318in the files that want to use the libevent \s-1API\s0. This also includes \fIev.h\fR.
2319.PP
2320You need the following additional files for this:
2321.PP
2322.Vb 2
2323\& event.h
2324\& event.c
2325.Ve
2326.PP
2327\fI\s-1AUTOCONF\s0 \s-1SUPPORT\s0\fR
2328.IX Subsection "AUTOCONF SUPPORT"
2329.PP
2330Instead of using \f(CW\*(C`EV_STANDALONE=1\*(C'\fR and providing your config in
2331whatever way you want, you can also \f(CW\*(C`m4_include([libev.m4])\*(C'\fR in your
2332\&\fIconfigure.ac\fR and leave \f(CW\*(C`EV_STANDALONE\*(C'\fR undefined. \fIev.c\fR will then
2333include \fIconfig.h\fR and configure itself accordingly.
2334.PP
2335For this of course you need the m4 file:
2336.PP
2337.Vb 1
2338\& libev.m4
2339.Ve
2340.Sh "\s-1PREPROCESSOR\s0 \s-1SYMBOLS/MACROS\s0"
2341.IX Subsection "PREPROCESSOR SYMBOLS/MACROS"
2342Libev can be configured via a variety of preprocessor symbols you have to define
2343before including any of its files. The default is not to build for multiplicity
2344and only include the select backend.
2345.IP "\s-1EV_STANDALONE\s0" 4
2346.IX Item "EV_STANDALONE"
2347Must always be \f(CW1\fR if you do not use autoconf configuration, which
2348keeps libev from including \fIconfig.h\fR, and it also defines dummy
2349implementations for some libevent functions (such as logging, which is not
2350supported). It will also not define any of the structs usually found in
2351\&\fIevent.h\fR that are not directly supported by the libev core alone.
2352.IP "\s-1EV_USE_MONOTONIC\s0" 4
2353.IX Item "EV_USE_MONOTONIC"
2354If defined to be \f(CW1\fR, libev will try to detect the availability of the
2355monotonic clock option at both compiletime and runtime. Otherwise no use
2356of the monotonic clock option will be attempted. If you enable this, you
2357usually have to link against librt or something similar. Enabling it when
2358the functionality isn't available is safe, though, althoguh you have
2359to make sure you link against any libraries where the \f(CW\*(C`clock_gettime\*(C'\fR
2360function is hiding in (often \fI\-lrt\fR).
2361.IP "\s-1EV_USE_REALTIME\s0" 4
2362.IX Item "EV_USE_REALTIME"
2363If defined to be \f(CW1\fR, libev will try to detect the availability of the
2364realtime clock option at compiletime (and assume its availability at
2365runtime if successful). Otherwise no use of the realtime clock option will
2366be attempted. This effectively replaces \f(CW\*(C`gettimeofday\*(C'\fR by \f(CW\*(C`clock_get
2367(CLOCK_REALTIME, ...)\*(C'\fR and will not normally affect correctness. See tzhe note about libraries
2368in the description of \f(CW\*(C`EV_USE_MONOTONIC\*(C'\fR, though.
2369.IP "\s-1EV_USE_SELECT\s0" 4
2370.IX Item "EV_USE_SELECT"
2371If undefined or defined to be \f(CW1\fR, libev will compile in support for the
2372\&\f(CW\*(C`select\*(C'\fR(2) backend. No attempt at autodetection will be done: if no
2373other method takes over, select will be it. Otherwise the select backend
2374will not be compiled in.
2375.IP "\s-1EV_SELECT_USE_FD_SET\s0" 4
2376.IX Item "EV_SELECT_USE_FD_SET"
2377If defined to \f(CW1\fR, then the select backend will use the system \f(CW\*(C`fd_set\*(C'\fR
2378structure. This is useful if libev doesn't compile due to a missing
2379\&\f(CW\*(C`NFDBITS\*(C'\fR or \f(CW\*(C`fd_mask\*(C'\fR definition or it misguesses the bitset layout on
2380exotic systems. This usually limits the range of file descriptors to some
2381low limit such as 1024 or might have other limitations (winsocket only
2382allows 64 sockets). The \f(CW\*(C`FD_SETSIZE\*(C'\fR macro, set before compilation, might
2383influence the size of the \f(CW\*(C`fd_set\*(C'\fR used.
2384.IP "\s-1EV_SELECT_IS_WINSOCKET\s0" 4
2385.IX Item "EV_SELECT_IS_WINSOCKET"
2386When defined to \f(CW1\fR, the select backend will assume that
2387select/socket/connect etc. don't understand file descriptors but
2388wants osf handles on win32 (this is the case when the select to
2389be used is the winsock select). This means that it will call
2390\&\f(CW\*(C`_get_osfhandle\*(C'\fR on the fd to convert it to an \s-1OS\s0 handle. Otherwise,
2391it is assumed that all these functions actually work on fds, even
2392on win32. Should not be defined on non\-win32 platforms.
2393.IP "\s-1EV_USE_POLL\s0" 4
2394.IX Item "EV_USE_POLL"
2395If defined to be \f(CW1\fR, libev will compile in support for the \f(CW\*(C`poll\*(C'\fR(2)
2396backend. Otherwise it will be enabled on non\-win32 platforms. It
2397takes precedence over select.
2398.IP "\s-1EV_USE_EPOLL\s0" 4
2399.IX Item "EV_USE_EPOLL"
2400If defined to be \f(CW1\fR, libev will compile in support for the Linux
2401\&\f(CW\*(C`epoll\*(C'\fR(7) backend. Its availability will be detected at runtime,
2402otherwise another method will be used as fallback. This is the
2403preferred backend for GNU/Linux systems.
2404.IP "\s-1EV_USE_KQUEUE\s0" 4
2405.IX Item "EV_USE_KQUEUE"
2406If defined to be \f(CW1\fR, libev will compile in support for the \s-1BSD\s0 style
2407\&\f(CW\*(C`kqueue\*(C'\fR(2) backend. Its actual availability will be detected at runtime,
2408otherwise another method will be used as fallback. This is the preferred
2409backend for \s-1BSD\s0 and BSD-like systems, although on most BSDs kqueue only
2410supports some types of fds correctly (the only platform we found that
2411supports ptys for example was NetBSD), so kqueue might be compiled in, but
2412not be used unless explicitly requested. The best way to use it is to find
2413out whether kqueue supports your type of fd properly and use an embedded
2414kqueue loop.
2415.IP "\s-1EV_USE_PORT\s0" 4
2416.IX Item "EV_USE_PORT"
2417If defined to be \f(CW1\fR, libev will compile in support for the Solaris
241810 port style backend. Its availability will be detected at runtime,
2419otherwise another method will be used as fallback. This is the preferred
2420backend for Solaris 10 systems.
2421.IP "\s-1EV_USE_DEVPOLL\s0" 4
2422.IX Item "EV_USE_DEVPOLL"
2423reserved for future expansion, works like the \s-1USE\s0 symbols above.
2424.IP "\s-1EV_USE_INOTIFY\s0" 4
2425.IX Item "EV_USE_INOTIFY"
2426If defined to be \f(CW1\fR, libev will compile in support for the Linux inotify
2427interface to speed up \f(CW\*(C`ev_stat\*(C'\fR watchers. Its actual availability will
2428be detected at runtime.
2429.IP "\s-1EV_H\s0" 4
2430.IX Item "EV_H"
2431The name of the \fIev.h\fR header file used to include it. The default if
2432undefined is \f(CW\*(C`<ev.h>\*(C'\fR in \fIevent.h\fR and \f(CW"ev.h"\fR in \fIev.c\fR. This
2433can be used to virtually rename the \fIev.h\fR header file in case of conflicts.
2434.IP "\s-1EV_CONFIG_H\s0" 4
2435.IX Item "EV_CONFIG_H"
2436If \f(CW\*(C`EV_STANDALONE\*(C'\fR isn't \f(CW1\fR, this variable can be used to override
2437\&\fIev.c\fR's idea of where to find the \fIconfig.h\fR file, similarly to
2438\&\f(CW\*(C`EV_H\*(C'\fR, above.
2439.IP "\s-1EV_EVENT_H\s0" 4
2440.IX Item "EV_EVENT_H"
2441Similarly to \f(CW\*(C`EV_H\*(C'\fR, this macro can be used to override \fIevent.c\fR's idea
2442of how the \fIevent.h\fR header can be found.
2443.IP "\s-1EV_PROTOTYPES\s0" 4
2444.IX Item "EV_PROTOTYPES"
2445If defined to be \f(CW0\fR, then \fIev.h\fR will not define any function
2446prototypes, but still define all the structs and other symbols. This is
2447occasionally useful if you want to provide your own wrapper functions
2448around libev functions.
2449.IP "\s-1EV_MULTIPLICITY\s0" 4
2450.IX Item "EV_MULTIPLICITY"
2451If undefined or defined to \f(CW1\fR, then all event-loop-specific functions
2452will have the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument, and you can create
2453additional independent event loops. Otherwise there will be no support
2454for multiple event loops and there is no first event loop pointer
2455argument. Instead, all functions act on the single default loop.
2456.IP "\s-1EV_MINPRI\s0" 4
2457.IX Item "EV_MINPRI"
2458.PD 0
2459.IP "\s-1EV_MAXPRI\s0" 4
2460.IX Item "EV_MAXPRI"
2461.PD
2462The range of allowed priorities. \f(CW\*(C`EV_MINPRI\*(C'\fR must be smaller or equal to
2463\&\f(CW\*(C`EV_MAXPRI\*(C'\fR, but otherwise there are no non-obvious limitations. You can
2464provide for more priorities by overriding those symbols (usually defined
2465to be \f(CW\*(C`\-2\*(C'\fR and \f(CW2\fR, respectively).
2466.Sp
2467When doing priority-based operations, libev usually has to linearly search
2468all the priorities, so having many of them (hundreds) uses a lot of space
2469and time, so using the defaults of five priorities (\-2 .. +2) is usually
2470fine.
2471.Sp
2472If your embedding app does not need any priorities, defining these both to
2473\&\f(CW0\fR will save some memory and cpu.
2474.IP "\s-1EV_PERIODIC_ENABLE\s0" 4
2475.IX Item "EV_PERIODIC_ENABLE"
2476If undefined or defined to be \f(CW1\fR, then periodic timers are supported. If
2477defined to be \f(CW0\fR, then they are not. Disabling them saves a few kB of
2478code.
2479.IP "\s-1EV_IDLE_ENABLE\s0" 4
2480.IX Item "EV_IDLE_ENABLE"
2481If undefined or defined to be \f(CW1\fR, then idle watchers are supported. If
2482defined to be \f(CW0\fR, then they are not. Disabling them saves a few kB of
2483code.
2484.IP "\s-1EV_EMBED_ENABLE\s0" 4
2485.IX Item "EV_EMBED_ENABLE"
2486If undefined or defined to be \f(CW1\fR, then embed watchers are supported. If
2487defined to be \f(CW0\fR, then they are not.
2488.IP "\s-1EV_STAT_ENABLE\s0" 4
2489.IX Item "EV_STAT_ENABLE"
2490If undefined or defined to be \f(CW1\fR, then stat watchers are supported. If
2491defined to be \f(CW0\fR, then they are not.
2492.IP "\s-1EV_FORK_ENABLE\s0" 4
2493.IX Item "EV_FORK_ENABLE"
2494If undefined or defined to be \f(CW1\fR, then fork watchers are supported. If
2495defined to be \f(CW0\fR, then they are not.
2496.IP "\s-1EV_MINIMAL\s0" 4
2497.IX Item "EV_MINIMAL"
2498If you need to shave off some kilobytes of code at the expense of some
2499speed, define this symbol to \f(CW1\fR. Currently only used for gcc to override
2500some inlining decisions, saves roughly 30% codesize of amd64.
2501.IP "\s-1EV_PID_HASHSIZE\s0" 4
2502.IX Item "EV_PID_HASHSIZE"
2503\&\f(CW\*(C`ev_child\*(C'\fR watchers use a small hash table to distribute workload by
2504pid. The default size is \f(CW16\fR (or \f(CW1\fR with \f(CW\*(C`EV_MINIMAL\*(C'\fR), usually more
2505than enough. If you need to manage thousands of children you might want to
2506increase this value (\fImust\fR be a power of two).
2507.IP "\s-1EV_INOTIFY_HASHSIZE\s0" 4
2508.IX Item "EV_INOTIFY_HASHSIZE"
2509\&\f(CW\*(C`ev_staz\*(C'\fR watchers use a small hash table to distribute workload by
2510inotify watch id. The default size is \f(CW16\fR (or \f(CW1\fR with \f(CW\*(C`EV_MINIMAL\*(C'\fR),
2511usually more than enough. If you need to manage thousands of \f(CW\*(C`ev_stat\*(C'\fR
2512watchers you might want to increase this value (\fImust\fR be a power of
2513two).
2514.IP "\s-1EV_COMMON\s0" 4
2515.IX Item "EV_COMMON"
2516By default, all watchers have a \f(CW\*(C`void *data\*(C'\fR member. By redefining
2517this macro to a something else you can include more and other types of
2518members. You have to define it each time you include one of the files,
2519though, and it must be identical each time.
2520.Sp
2521For example, the perl \s-1EV\s0 module uses something like this:
2522.Sp
2523.Vb 3
2524\& #define EV_COMMON \e
2525\& SV *self; /* contains this struct */ \e
2526\& SV *cb_sv, *fh /* note no trailing ";" */
2527.Ve
2528.IP "\s-1EV_CB_DECLARE\s0 (type)" 4
2529.IX Item "EV_CB_DECLARE (type)"
2530.PD 0
2531.IP "\s-1EV_CB_INVOKE\s0 (watcher, revents)" 4
2532.IX Item "EV_CB_INVOKE (watcher, revents)"
2533.IP "ev_set_cb (ev, cb)" 4
2534.IX Item "ev_set_cb (ev, cb)"
2535.PD
2536Can be used to change the callback member declaration in each watcher,
2537and the way callbacks are invoked and set. Must expand to a struct member
2538definition and a statement, respectively. See the \fIev.v\fR header file for
2539their default definitions. One possible use for overriding these is to
2540avoid the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument in all cases, or to use
2541method calls instead of plain function calls in \*(C+.
2542.Sh "\s-1EXAMPLES\s0"
2543.IX Subsection "EXAMPLES"
2544For a real-world example of a program the includes libev
2545verbatim, you can have a look at the \s-1EV\s0 perl module
2546(<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
2547the \fIlibev/\fR subdirectory and includes them in the \fI\s-1EV/EVAPI\s0.h\fR (public
2548interface) and \fI\s-1EV\s0.xs\fR (implementation) files. Only the \fI\s-1EV\s0.xs\fR file
2549will be compiled. It is pretty complex because it provides its own header
2550file.
2551.Sp
2552The usage in rxvt-unicode is simpler. It has a \fIev_cpp.h\fR header file
2553that everybody includes and which overrides some configure choices:
2554.Sp
2555.Vb 9
2556\& #define EV_MINIMAL 1
2557\& #define EV_USE_POLL 0
2558\& #define EV_MULTIPLICITY 0
2559\& #define EV_PERIODIC_ENABLE 0
2560\& #define EV_STAT_ENABLE 0
2561\& #define EV_FORK_ENABLE 0
2562\& #define EV_CONFIG_H <config.h>
2563\& #define EV_MINPRI 0
2564\& #define EV_MAXPRI 0
2565.Ve
2566.Sp
2567.Vb 1
2568\& #include "ev++.h"
2569.Ve
2570.Sp
2571And a \fIev_cpp.C\fR implementation file that contains libev proper and is compiled:
2572.Sp
2573.Vb 2
2574\& #include "ev_cpp.h"
2575\& #include "ev.c"
2576.Ve
2577.SH "COMPLEXITIES"
2578.IX Header "COMPLEXITIES"
2579In this section the complexities of (many of) the algorithms used inside
2580libev will be explained. For complexity discussions about backends see the
2581documentation for \f(CW\*(C`ev_default_init\*(C'\fR.
2582.Sp
2583All of the following are about amortised time: If an array needs to be
2584extended, libev needs to realloc and move the whole array, but this
2585happens asymptotically never with higher number of elements, so O(1) might
2586mean it might do a lengthy realloc operation in rare cases, but on average
2587it is much faster and asymptotically approaches constant time.
2588.RS 4
2589.IP "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)" 4
2590.IX Item "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)"
2591This means that, when you have a watcher that triggers in one hour and
2592there are 100 watchers that would trigger before that then inserting will
2593have to skip those 100 watchers.
2594.IP "Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)" 4
2595.IX Item "Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)"
2596That means that for changing a timer costs less than removing/adding them
2597as only the relative motion in the event queue has to be paid for.
2598.IP "Starting io/check/prepare/idle/signal/child watchers: O(1)" 4
2599.IX Item "Starting io/check/prepare/idle/signal/child watchers: O(1)"
2600These just add the watcher into an array or at the head of a list.
2601=item Stopping check/prepare/idle watchers: O(1)
2602.IP "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % \s-1EV_PID_HASHSIZE\s0))" 4
2603.IX Item "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))"
2604These watchers are stored in lists then need to be walked to find the
2605correct watcher to remove. The lists are usually short (you don't usually
2606have many watchers waiting for the same fd or signal).
2607.IP "Finding the next timer per loop iteration: O(1)" 4
2608.IX Item "Finding the next timer per loop iteration: O(1)"
2609.PD 0
2610.IP "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)" 4
2611.IX Item "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)"
2612.PD
2613A change means an I/O watcher gets started or stopped, which requires
2614libev to recalculate its status (and possibly tell the kernel).
2615.IP "Activating one watcher: O(1)" 4
2616.IX Item "Activating one watcher: O(1)"
2617.PD 0
2618.IP "Priority handling: O(number_of_priorities)" 4
2619.IX Item "Priority handling: O(number_of_priorities)"
2620.PD
2621Priorities are implemented by allocating some space for each
2622priority. When doing priority-based operations, libev usually has to
2623linearly search all the priorities.
2624.RE
2625.RS 4
1434.SH "AUTHOR" 2626.SH "AUTHOR"
1435.IX Header "AUTHOR" 2627.IX Header "AUTHOR"
1436Marc Lehmann <libev@schmorp.de>. 2628Marc Lehmann <libev@schmorp.de>.

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines