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

Comparing libev/ev.3 (file contents):
Revision 1.1 by root, Tue Nov 13 03:11:57 2007 UTC vs.
Revision 1.27 by root, Tue Nov 27 20:15:01 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 "2007-11-13" "perl v5.8.8" "User Contributed Perl Documentation" 132.TH "<STANDARD INPUT>" 1 "2007-11-27" "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 2
138\& /* this is the only header you need */
138\& #include <ev.h> 139\& #include <ev.h>
140.Ve
141.PP
142.Vb 3
143\& /* what follows is a fully working example program */
144\& ev_io stdin_watcher;
145\& ev_timer timeout_watcher;
146.Ve
147.PP
148.Vb 8
149\& /* called when data readable on stdin */
150\& static void
151\& stdin_cb (EV_P_ struct ev_io *w, int revents)
152\& {
153\& /* puts ("stdin ready"); */
154\& ev_io_stop (EV_A_ w); /* just a syntax example */
155\& ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */
156\& }
157.Ve
158.PP
159.Vb 6
160\& static void
161\& timeout_cb (EV_P_ struct ev_timer *w, int revents)
162\& {
163\& /* puts ("timeout"); */
164\& ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */
165\& }
166.Ve
167.PP
168.Vb 4
169\& int
170\& main (void)
171\& {
172\& struct ev_loop *loop = ev_default_loop (0);
173.Ve
174.PP
175.Vb 3
176\& /* initialise an io watcher, then start it */
177\& ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
178\& ev_io_start (loop, &stdin_watcher);
179.Ve
180.PP
181.Vb 3
182\& /* simple non-repeating 5.5 second timeout */
183\& ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
184\& ev_timer_start (loop, &timeout_watcher);
185.Ve
186.PP
187.Vb 2
188\& /* loop till timeout or data ready */
189\& ev_loop (loop, 0);
190.Ve
191.PP
192.Vb 2
193\& return 0;
194\& }
139.Ve 195.Ve
140.SH "DESCRIPTION" 196.SH "DESCRIPTION"
141.IX Header "DESCRIPTION" 197.IX Header "DESCRIPTION"
142Libev is an event loop: you register interest in certain events (such as a 198Libev 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 199file descriptor being readable or a timeout occuring), and it will manage
… …
173.IX Header "TIME REPRESENTATION" 229.IX Header "TIME REPRESENTATION"
174Libev represents time as a single floating point number, representing the 230Libev represents time as a single floating point number, representing the
175(fractional) number of seconds since the (\s-1POSIX\s0) epoch (somewhere near 231(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 232the beginning of 1970, details are complicated, don't ask). This type is
177called \f(CW\*(C`ev_tstamp\*(C'\fR, which is what you should use too. It usually aliases 233called \f(CW\*(C`ev_tstamp\*(C'\fR, which is what you should use too. It usually aliases
178to the double type in C. 234to the \f(CW\*(C`double\*(C'\fR type in C, and when you need to do any calculations on
235it, you should treat it as such.
179.SH "GLOBAL FUNCTIONS" 236.SH "GLOBAL FUNCTIONS"
180.IX Header "GLOBAL FUNCTIONS" 237.IX Header "GLOBAL FUNCTIONS"
181These functions can be called anytime, even before initialising the 238These functions can be called anytime, even before initialising the
182library in any way. 239library in any way.
183.IP "ev_tstamp ev_time ()" 4 240.IP "ev_tstamp ev_time ()" 4
184.IX Item "ev_tstamp ev_time ()" 241.IX Item "ev_tstamp ev_time ()"
185Returns the current time as libev would use it. 242Returns the current time as libev would use it. Please note that the
243\&\f(CW\*(C`ev_now\*(C'\fR function is usually faster and also often returns the timestamp
244you actually want to know.
186.IP "int ev_version_major ()" 4 245.IP "int ev_version_major ()" 4
187.IX Item "int ev_version_major ()" 246.IX Item "int ev_version_major ()"
188.PD 0 247.PD 0
189.IP "int ev_version_minor ()" 4 248.IP "int ev_version_minor ()" 4
190.IX Item "int ev_version_minor ()" 249.IX Item "int ev_version_minor ()"
… …
197.Sp 256.Sp
198Usually, it's a good idea to terminate if the major versions mismatch, 257Usually, it's a good idea to terminate if the major versions mismatch,
199as this indicates an incompatible change. Minor versions are usually 258as this indicates an incompatible change. Minor versions are usually
200compatible to older versions, so a larger minor version alone is usually 259compatible to older versions, so a larger minor version alone is usually
201not a problem. 260not a problem.
261.Sp
262Example: make sure we haven't accidentally been linked against the wrong
263version:
264.Sp
265.Vb 3
266\& assert (("libev version mismatch",
267\& ev_version_major () == EV_VERSION_MAJOR
268\& && ev_version_minor () >= EV_VERSION_MINOR));
269.Ve
270.IP "unsigned int ev_supported_backends ()" 4
271.IX Item "unsigned int ev_supported_backends ()"
272Return the set of all backends (i.e. their corresponding \f(CW\*(C`EV_BACKEND_*\*(C'\fR
273value) compiled into this binary of libev (independent of their
274availability on the system you are running on). See \f(CW\*(C`ev_default_loop\*(C'\fR for
275a description of the set values.
276.Sp
277Example: make sure we have the epoll method, because yeah this is cool and
278a must have and can we have a torrent of it please!!!11
279.Sp
280.Vb 2
281\& assert (("sorry, no epoll, no sex",
282\& ev_supported_backends () & EVBACKEND_EPOLL));
283.Ve
284.IP "unsigned int ev_recommended_backends ()" 4
285.IX Item "unsigned int ev_recommended_backends ()"
286Return the set of all backends compiled into this binary of libev and also
287recommended for this platform. This set is often smaller than the one
288returned by \f(CW\*(C`ev_supported_backends\*(C'\fR, as for example kqueue is broken on
289most BSDs and will not be autodetected unless you explicitly request it
290(assuming you know what you are doing). This is the set of backends that
291libev will probe for if you specify no backends explicitly.
292.IP "unsigned int ev_embeddable_backends ()" 4
293.IX Item "unsigned int ev_embeddable_backends ()"
294Returns the set of backends that are embeddable in other event loops. This
295is the theoretical, all\-platform, value. To find which backends
296might be supported on the current system, you would need to look at
297\&\f(CW\*(C`ev_embeddable_backends () & ev_supported_backends ()\*(C'\fR, likewise for
298recommended ones.
299.Sp
300See the description of \f(CW\*(C`ev_embed\*(C'\fR watchers for more info.
202.IP "ev_set_allocator (void *(*cb)(void *ptr, long size))" 4 301.IP "ev_set_allocator (void *(*cb)(void *ptr, size_t size))" 4
203.IX Item "ev_set_allocator (void *(*cb)(void *ptr, long size))" 302.IX Item "ev_set_allocator (void *(*cb)(void *ptr, size_t size))"
204Sets the allocation function to use (the prototype is similar to the 303Sets the allocation function to use (the prototype and semantics are
205realloc C function, the semantics are identical). It is used to allocate 304identical to the realloc C function). It is used to allocate and free
206and free memory (no surprises here). If it returns zero when memory 305memory (no surprises here). If it returns zero when memory needs to be
207needs to be allocated, the library might abort or take some potentially 306allocated, the library might abort or take some potentially destructive
208destructive action. The default is your system realloc function. 307action. The default is your system realloc function.
209.Sp 308.Sp
210You could override this function in high-availability programs to, say, 309You could override this function in high-availability programs to, say,
211free some memory if it cannot allocate memory, to use a special allocator, 310free some memory if it cannot allocate memory, to use a special allocator,
212or even to sleep a while and retry until some memory is available. 311or even to sleep a while and retry until some memory is available.
312.Sp
313Example: replace the libev allocator with one that waits a bit and then
314retries: better than mine).
315.Sp
316.Vb 6
317\& static void *
318\& persistent_realloc (void *ptr, size_t size)
319\& {
320\& for (;;)
321\& {
322\& void *newptr = realloc (ptr, size);
323.Ve
324.Sp
325.Vb 2
326\& if (newptr)
327\& return newptr;
328.Ve
329.Sp
330.Vb 3
331\& sleep (60);
332\& }
333\& }
334.Ve
335.Sp
336.Vb 2
337\& ...
338\& ev_set_allocator (persistent_realloc);
339.Ve
213.IP "ev_set_syserr_cb (void (*cb)(const char *msg));" 4 340.IP "ev_set_syserr_cb (void (*cb)(const char *msg));" 4
214.IX Item "ev_set_syserr_cb (void (*cb)(const char *msg));" 341.IX Item "ev_set_syserr_cb (void (*cb)(const char *msg));"
215Set the callback function to call on a retryable syscall error (such 342Set the callback function to call on a retryable syscall error (such
216as failed select, poll, epoll_wait). The message is a printable string 343as failed select, poll, epoll_wait). The message is a printable string
217indicating the system call or subsystem causing the problem. If this 344indicating the system call or subsystem causing the problem. If this
218callback is set, then libev will expect it to remedy the sitution, no 345callback is set, then libev will expect it to remedy the sitution, no
219matter what, when it returns. That is, libev will generally retry the 346matter what, when it returns. That is, libev will generally retry the
220requested operation, or, if the condition doesn't go away, do bad stuff 347requested operation, or, if the condition doesn't go away, do bad stuff
221(such as abort). 348(such as abort).
349.Sp
350Example: do the same thing as libev does internally:
351.Sp
352.Vb 6
353\& static void
354\& fatal_error (const char *msg)
355\& {
356\& perror (msg);
357\& abort ();
358\& }
359.Ve
360.Sp
361.Vb 2
362\& ...
363\& ev_set_syserr_cb (fatal_error);
364.Ve
222.SH "FUNCTIONS CONTROLLING THE EVENT LOOP" 365.SH "FUNCTIONS CONTROLLING THE EVENT LOOP"
223.IX Header "FUNCTIONS CONTROLLING THE EVENT LOOP" 366.IX Header "FUNCTIONS CONTROLLING THE EVENT LOOP"
224An event loop is described by a \f(CW\*(C`struct ev_loop *\*(C'\fR. The library knows two 367An event loop is described by a \f(CW\*(C`struct ev_loop *\*(C'\fR. The library knows two
225types of such loops, the \fIdefault\fR loop, which supports signals and child 368types of such loops, the \fIdefault\fR loop, which supports signals and child
226events, and dynamically created loops which do not. 369events, and dynamically created loops which do not.
… …
234.IP "struct ev_loop *ev_default_loop (unsigned int flags)" 4 377.IP "struct ev_loop *ev_default_loop (unsigned int flags)" 4
235.IX Item "struct ev_loop *ev_default_loop (unsigned int flags)" 378.IX Item "struct ev_loop *ev_default_loop (unsigned int flags)"
236This will initialise the default event loop if it hasn't been initialised 379This will initialise the default event loop if it hasn't been initialised
237yet and return it. If the default loop could not be initialised, returns 380yet and return it. If the default loop could not be initialised, returns
238false. If it already was initialised it simply returns it (and ignores the 381false. If it already was initialised it simply returns it (and ignores the
239flags). 382flags. If that is troubling you, check \f(CW\*(C`ev_backend ()\*(C'\fR afterwards).
240.Sp 383.Sp
241If you don't know what event loop to use, use the one returned from this 384If you don't know what event loop to use, use the one returned from this
242function. 385function.
243.Sp 386.Sp
244The flags argument can be used to specify special behaviour or specific 387The flags argument can be used to specify special behaviour or specific
245backends to use, and is usually specified as 0 (or \s-1EVFLAG_AUTO\s0). 388backends to use, and is usually specified as \f(CW0\fR (or \f(CW\*(C`EVFLAG_AUTO\*(C'\fR).
246.Sp 389.Sp
247It supports the following flags: 390The following flags are supported:
248.RS 4 391.RS 4
249.ie n .IP """EVFLAG_AUTO""" 4 392.ie n .IP """EVFLAG_AUTO""" 4
250.el .IP "\f(CWEVFLAG_AUTO\fR" 4 393.el .IP "\f(CWEVFLAG_AUTO\fR" 4
251.IX Item "EVFLAG_AUTO" 394.IX Item "EVFLAG_AUTO"
252The default flags value. Use this if you have no clue (it's the right 395The default flags value. Use this if you have no clue (it's the right
… …
258or setgid) then libev will \fInot\fR look at the environment variable 401or setgid) then libev will \fInot\fR look at the environment variable
259\&\f(CW\*(C`LIBEV_FLAGS\*(C'\fR. Otherwise (the default), this environment variable will 402\&\f(CW\*(C`LIBEV_FLAGS\*(C'\fR. Otherwise (the default), this environment variable will
260override the flags completely if it is found in the environment. This is 403override the flags completely if it is found in the environment. This is
261useful to try out specific backends to test their performance, or to work 404useful to try out specific backends to test their performance, or to work
262around bugs. 405around bugs.
263.ie n .IP """EVMETHOD_SELECT"" (portable select backend)" 4 406.ie n .IP """EVBACKEND_SELECT"" (value 1, portable select backend)" 4
264.el .IP "\f(CWEVMETHOD_SELECT\fR (portable select backend)" 4 407.el .IP "\f(CWEVBACKEND_SELECT\fR (value 1, portable select backend)" 4
265.IX Item "EVMETHOD_SELECT (portable select backend)" 408.IX Item "EVBACKEND_SELECT (value 1, portable select backend)"
266.PD 0 409This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as
410libev tries to roll its own fd_set with no limits on the number of fds,
411but if that fails, expect a fairly low limit on the number of fds when
412using this backend. It doesn't scale too well (O(highest_fd)), but its usually
413the fastest backend for a low number of fds.
267.ie n .IP """EVMETHOD_POLL"" (poll backend, available everywhere except on windows)" 4 414.ie n .IP """EVBACKEND_POLL"" (value 2, poll backend, available everywhere except on windows)" 4
268.el .IP "\f(CWEVMETHOD_POLL\fR (poll backend, available everywhere except on windows)" 4 415.el .IP "\f(CWEVBACKEND_POLL\fR (value 2, poll backend, available everywhere except on windows)" 4
269.IX Item "EVMETHOD_POLL (poll backend, available everywhere except on windows)" 416.IX Item "EVBACKEND_POLL (value 2, poll backend, available everywhere except on windows)"
417And this is your standard \fIpoll\fR\|(2) backend. It's more complicated than
418select, but handles sparse fds better and has no artificial limit on the
419number of fds you can use (except it will slow down considerably with a
420lot of inactive fds). It scales similarly to select, i.e. O(total_fds).
270.ie n .IP """EVMETHOD_EPOLL"" (linux only)" 4 421.ie n .IP """EVBACKEND_EPOLL"" (value 4, Linux)" 4
271.el .IP "\f(CWEVMETHOD_EPOLL\fR (linux only)" 4 422.el .IP "\f(CWEVBACKEND_EPOLL\fR (value 4, Linux)" 4
272.IX Item "EVMETHOD_EPOLL (linux only)" 423.IX Item "EVBACKEND_EPOLL (value 4, Linux)"
273.ie n .IP """EVMETHOD_KQUEUE"" (some bsds only)" 4 424For few fds, this backend is a bit little slower than poll and select,
274.el .IP "\f(CWEVMETHOD_KQUEUE\fR (some bsds only)" 4 425but it scales phenomenally better. While poll and select usually scale like
275.IX Item "EVMETHOD_KQUEUE (some bsds only)" 426O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
427either O(1) or O(active_fds).
428.Sp
429While stopping and starting an I/O watcher in the same iteration will
430result in some caching, there is still a syscall per such incident
431(because the fd could point to a different file description now), so its
432best to avoid that. Also, \fIdup()\fRed file descriptors might not work very
433well if you register events for both fds.
434.Sp
435Please note that epoll sometimes generates spurious notifications, so you
436need to use non-blocking I/O or other means to avoid blocking when no data
437(or space) is available.
438.ie n .IP """EVBACKEND_KQUEUE"" (value 8, most \s-1BSD\s0 clones)" 4
439.el .IP "\f(CWEVBACKEND_KQUEUE\fR (value 8, most \s-1BSD\s0 clones)" 4
440.IX Item "EVBACKEND_KQUEUE (value 8, most BSD clones)"
441Kqueue deserves special mention, as at the time of this writing, it
442was broken on all BSDs except NetBSD (usually it doesn't work with
443anything but sockets and pipes, except on Darwin, where of course its
444completely useless). For this reason its not being \*(L"autodetected\*(R"
445unless you explicitly specify it explicitly in the flags (i.e. using
446\&\f(CW\*(C`EVBACKEND_KQUEUE\*(C'\fR).
447.Sp
448It scales in the same way as the epoll backend, but the interface to the
449kernel is more efficient (which says nothing about its actual speed, of
450course). While starting and stopping an I/O watcher does not cause an
451extra syscall as with epoll, it still adds up to four event changes per
452incident, so its best to avoid that.
276.ie n .IP """EVMETHOD_DEVPOLL"" (solaris 8 only)" 4 453.ie n .IP """EVBACKEND_DEVPOLL"" (value 16, Solaris 8)" 4
277.el .IP "\f(CWEVMETHOD_DEVPOLL\fR (solaris 8 only)" 4 454.el .IP "\f(CWEVBACKEND_DEVPOLL\fR (value 16, Solaris 8)" 4
278.IX Item "EVMETHOD_DEVPOLL (solaris 8 only)" 455.IX Item "EVBACKEND_DEVPOLL (value 16, Solaris 8)"
456This is not implemented yet (and might never be).
279.ie n .IP """EVMETHOD_PORT"" (solaris 10 only)" 4 457.ie n .IP """EVBACKEND_PORT"" (value 32, Solaris 10)" 4
280.el .IP "\f(CWEVMETHOD_PORT\fR (solaris 10 only)" 4 458.el .IP "\f(CWEVBACKEND_PORT\fR (value 32, Solaris 10)" 4
281.IX Item "EVMETHOD_PORT (solaris 10 only)" 459.IX Item "EVBACKEND_PORT (value 32, Solaris 10)"
282.PD 460This uses the Solaris 10 port mechanism. As with everything on Solaris,
283If one or more of these are ored into the flags value, then only these 461it's really slow, but it still scales very well (O(active_fds)).
284backends will be tried (in the reverse order as given here). If one are 462.Sp
285specified, any backend will do. 463Please note that solaris ports can result in a lot of spurious
464notifications, so you need to use non-blocking I/O or other means to avoid
465blocking when no data (or space) is available.
466.ie n .IP """EVBACKEND_ALL""" 4
467.el .IP "\f(CWEVBACKEND_ALL\fR" 4
468.IX Item "EVBACKEND_ALL"
469Try all backends (even potentially broken ones that wouldn't be tried
470with \f(CW\*(C`EVFLAG_AUTO\*(C'\fR). Since this is a mask, you can do stuff such as
471\&\f(CW\*(C`EVBACKEND_ALL & ~EVBACKEND_KQUEUE\*(C'\fR.
286.RE 472.RE
287.RS 4 473.RS 4
474.Sp
475If one or more of these are ored into the flags value, then only these
476backends will be tried (in the reverse order as given here). If none are
477specified, most compiled-in backend will be tried, usually in reverse
478order of their flag values :)
479.Sp
480The most typical usage is like this:
481.Sp
482.Vb 2
483\& if (!ev_default_loop (0))
484\& fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
485.Ve
486.Sp
487Restrict libev to the select and poll backends, and do not allow
488environment settings to be taken into account:
489.Sp
490.Vb 1
491\& ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
492.Ve
493.Sp
494Use whatever libev has to offer, but make sure that kqueue is used if
495available (warning, breaks stuff, best use only with your own private
496event loop and only if you know the \s-1OS\s0 supports your types of fds):
497.Sp
498.Vb 1
499\& ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
500.Ve
288.RE 501.RE
289.IP "struct ev_loop *ev_loop_new (unsigned int flags)" 4 502.IP "struct ev_loop *ev_loop_new (unsigned int flags)" 4
290.IX Item "struct ev_loop *ev_loop_new (unsigned int flags)" 503.IX Item "struct ev_loop *ev_loop_new (unsigned int flags)"
291Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is 504Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is
292always distinct from the default loop. Unlike the default loop, it cannot 505always distinct from the default loop. Unlike the default loop, it cannot
293handle signal and child watchers, and attempts to do so will be greeted by 506handle signal and child watchers, and attempts to do so will be greeted by
294undefined behaviour (or a failed assertion if assertions are enabled). 507undefined behaviour (or a failed assertion if assertions are enabled).
508.Sp
509Example: try to create a event loop that uses epoll and nothing else.
510.Sp
511.Vb 3
512\& struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
513\& if (!epoller)
514\& fatal ("no epoll found here, maybe it hides under your chair");
515.Ve
295.IP "ev_default_destroy ()" 4 516.IP "ev_default_destroy ()" 4
296.IX Item "ev_default_destroy ()" 517.IX Item "ev_default_destroy ()"
297Destroys the default loop again (frees all memory and kernel state 518Destroys the default loop again (frees all memory and kernel state
298etc.). This stops all registered event watchers (by not touching them in 519etc.). None of the active event watchers will be stopped in the normal
299any way whatsoever, although you cannot rely on this :). 520sense, so e.g. \f(CW\*(C`ev_is_active\*(C'\fR might still return true. It is your
521responsibility to either stop all watchers cleanly yoursef \fIbefore\fR
522calling this function, or cope with the fact afterwards (which is usually
523the easiest thing, youc na just ignore the watchers and/or \f(CW\*(C`free ()\*(C'\fR them
524for example).
300.IP "ev_loop_destroy (loop)" 4 525.IP "ev_loop_destroy (loop)" 4
301.IX Item "ev_loop_destroy (loop)" 526.IX Item "ev_loop_destroy (loop)"
302Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an 527Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an
303earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR. 528earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR.
304.IP "ev_default_fork ()" 4 529.IP "ev_default_fork ()" 4
… …
306This function reinitialises the kernel state for backends that have 531This function reinitialises the kernel state for backends that have
307one. Despite the name, you can call it anytime, but it makes most sense 532one. Despite the name, you can call it anytime, but it makes most sense
308after forking, in either the parent or child process (or both, but that 533after forking, in either the parent or child process (or both, but that
309again makes little sense). 534again makes little sense).
310.Sp 535.Sp
311You \fImust\fR call this function after forking if and only if you want to 536You \fImust\fR call this function in the child process after forking if and
312use the event library in both processes. If you just fork+exec, you don't 537only if you want to use the event library in both processes. If you just
313have to call it. 538fork+exec, you don't have to call it.
314.Sp 539.Sp
315The function itself is quite fast and it's usually not a problem to call 540The function itself is quite fast and it's usually not a problem to call
316it just in case after a fork. To make this easy, the function will fit in 541it just in case after a fork. To make this easy, the function will fit in
317quite nicely into a call to \f(CW\*(C`pthread_atfork\*(C'\fR: 542quite nicely into a call to \f(CW\*(C`pthread_atfork\*(C'\fR:
318.Sp 543.Sp
319.Vb 1 544.Vb 1
320\& pthread_atfork (0, 0, ev_default_fork); 545\& pthread_atfork (0, 0, ev_default_fork);
321.Ve 546.Ve
547.Sp
548At the moment, \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and \f(CW\*(C`EVBACKEND_POLL\*(C'\fR are safe to use
549without calling this function, so if you force one of those backends you
550do not need to care.
322.IP "ev_loop_fork (loop)" 4 551.IP "ev_loop_fork (loop)" 4
323.IX Item "ev_loop_fork (loop)" 552.IX Item "ev_loop_fork (loop)"
324Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by 553Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by
325\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop 554\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop
326after fork, and how you do this is entirely your own problem. 555after fork, and how you do this is entirely your own problem.
327.IP "unsigned int ev_method (loop)" 4 556.IP "unsigned int ev_backend (loop)" 4
328.IX Item "unsigned int ev_method (loop)" 557.IX Item "unsigned int ev_backend (loop)"
329Returns one of the \f(CW\*(C`EVMETHOD_*\*(C'\fR flags indicating the event backend in 558Returns one of the \f(CW\*(C`EVBACKEND_*\*(C'\fR flags indicating the event backend in
330use. 559use.
331.IP "ev_tstamp ev_now (loop)" 4 560.IP "ev_tstamp ev_now (loop)" 4
332.IX Item "ev_tstamp ev_now (loop)" 561.IX Item "ev_tstamp ev_now (loop)"
333Returns the current \*(L"event loop time\*(R", which is the time the event loop 562Returns the current \*(L"event loop time\*(R", which is the time the event loop
334got events and started processing them. This timestamp does not change 563received events and started processing them. This timestamp does not
335as long as callbacks are being processed, and this is also the base time 564change as long as callbacks are being processed, and this is also the base
336used for relative timers. You can treat it as the timestamp of the event 565time used for relative timers. You can treat it as the timestamp of the
337occuring (or more correctly, the mainloop finding out about it). 566event occuring (or more correctly, libev finding out about it).
338.IP "ev_loop (loop, int flags)" 4 567.IP "ev_loop (loop, int flags)" 4
339.IX Item "ev_loop (loop, int flags)" 568.IX Item "ev_loop (loop, int flags)"
340Finally, this is it, the event handler. This function usually is called 569Finally, this is it, the event handler. This function usually is called
341after you initialised all your watchers and you want to start handling 570after you initialised all your watchers and you want to start handling
342events. 571events.
343.Sp 572.Sp
344If the flags argument is specified as 0, it will not return until either 573If the flags argument is specified as \f(CW0\fR, it will not return until
345no event watchers are active anymore or \f(CW\*(C`ev_unloop\*(C'\fR was called. 574either no event watchers are active anymore or \f(CW\*(C`ev_unloop\*(C'\fR was called.
575.Sp
576Please note that an explicit \f(CW\*(C`ev_unloop\*(C'\fR is usually better than
577relying on all watchers to be stopped when deciding when a program has
578finished (especially in interactive programs), but having a program that
579automatically loops as long as it has to and no longer by virtue of
580relying on its watchers stopping correctly is a thing of beauty.
346.Sp 581.Sp
347A flags value of \f(CW\*(C`EVLOOP_NONBLOCK\*(C'\fR will look for new events, will handle 582A flags value of \f(CW\*(C`EVLOOP_NONBLOCK\*(C'\fR will look for new events, will handle
348those events and any outstanding ones, but will not block your process in 583those events and any outstanding ones, but will not block your process in
349case there are no events and will return after one iteration of the loop. 584case there are no events and will return after one iteration of the loop.
350.Sp 585.Sp
351A flags value of \f(CW\*(C`EVLOOP_ONESHOT\*(C'\fR will look for new events (waiting if 586A flags value of \f(CW\*(C`EVLOOP_ONESHOT\*(C'\fR will look for new events (waiting if
352neccessary) and will handle those and any outstanding ones. It will block 587neccessary) and will handle those and any outstanding ones. It will block
353your process until at least one new event arrives, and will return after 588your process until at least one new event arrives, and will return after
354one iteration of the loop. 589one iteration of the loop. This is useful if you are waiting for some
590external event in conjunction with something not expressible using other
591libev watchers. However, a pair of \f(CW\*(C`ev_prepare\*(C'\fR/\f(CW\*(C`ev_check\*(C'\fR watchers is
592usually a better approach for this kind of thing.
355.Sp 593.Sp
356This flags value could be used to implement alternative looping 594Here are the gory details of what \f(CW\*(C`ev_loop\*(C'\fR does:
357constructs, but the \f(CW\*(C`prepare\*(C'\fR and \f(CW\*(C`check\*(C'\fR watchers provide a better and 595.Sp
358more generic mechanism. 596.Vb 18
597\& * If there are no active watchers (reference count is zero), return.
598\& - Queue prepare watchers and then call all outstanding watchers.
599\& - If we have been forked, recreate the kernel state.
600\& - Update the kernel state with all outstanding changes.
601\& - Update the "event loop time".
602\& - Calculate for how long to block.
603\& - Block the process, waiting for any events.
604\& - Queue all outstanding I/O (fd) events.
605\& - Update the "event loop time" and do time jump handling.
606\& - Queue all outstanding timers.
607\& - Queue all outstanding periodics.
608\& - If no events are pending now, queue all idle watchers.
609\& - Queue all check watchers.
610\& - Call all queued watchers in reverse order (i.e. check watchers first).
611\& Signals and child watchers are implemented as I/O watchers, and will
612\& be handled here by queueing them when their watcher gets executed.
613\& - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
614\& were used, return, otherwise continue with step *.
615.Ve
616.Sp
617Example: queue some jobs and then loop until no events are outsanding
618anymore.
619.Sp
620.Vb 4
621\& ... queue jobs here, make sure they register event watchers as long
622\& ... as they still have work to do (even an idle watcher will do..)
623\& ev_loop (my_loop, 0);
624\& ... jobs done. yeah!
625.Ve
359.IP "ev_unloop (loop, how)" 4 626.IP "ev_unloop (loop, how)" 4
360.IX Item "ev_unloop (loop, how)" 627.IX Item "ev_unloop (loop, how)"
361Can be used to make a call to \f(CW\*(C`ev_loop\*(C'\fR return early (but only after it 628Can be used to make a call to \f(CW\*(C`ev_loop\*(C'\fR return early (but only after it
362has processed all outstanding events). The \f(CW\*(C`how\*(C'\fR argument must be either 629has processed all outstanding events). The \f(CW\*(C`how\*(C'\fR argument must be either
363\&\f(CW\*(C`EVUNLOOP_ONE\*(C'\fR, which will make the innermost \f(CW\*(C`ev_loop\*(C'\fR call return, or 630\&\f(CW\*(C`EVUNLOOP_ONE\*(C'\fR, which will make the innermost \f(CW\*(C`ev_loop\*(C'\fR call return, or
… …
376example, libev itself uses this for its internal signal pipe: It is not 643example, libev itself uses this for its internal signal pipe: It is not
377visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if 644visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if
378no event watchers registered by it are active. It is also an excellent 645no event watchers registered by it are active. It is also an excellent
379way to do this for generic recurring timers or from within third-party 646way to do this for generic recurring timers or from within third-party
380libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR. 647libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR.
648.Sp
649Example: create a signal watcher, but keep it from keeping \f(CW\*(C`ev_loop\*(C'\fR
650running when nothing else is active.
651.Sp
652.Vb 4
653\& struct dv_signal exitsig;
654\& ev_signal_init (&exitsig, sig_cb, SIGINT);
655\& ev_signal_start (myloop, &exitsig);
656\& evf_unref (myloop);
657.Ve
658.Sp
659Example: for some weird reason, unregister the above signal handler again.
660.Sp
661.Vb 2
662\& ev_ref (myloop);
663\& ev_signal_stop (myloop, &exitsig);
664.Ve
381.SH "ANATOMY OF A WATCHER" 665.SH "ANATOMY OF A WATCHER"
382.IX Header "ANATOMY OF A WATCHER" 666.IX Header "ANATOMY OF A WATCHER"
383A watcher is a structure that you create and register to record your 667A watcher is a structure that you create and register to record your
384interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to 668interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to
385become readable, you would create an \f(CW\*(C`ev_io\*(C'\fR watcher for that: 669become readable, you would create an \f(CW\*(C`ev_io\*(C'\fR watcher for that:
… …
421*)\*(C'\fR), and you can stop watching for events at any time by calling the 705*)\*(C'\fR), and you can stop watching for events at any time by calling the
422corresponding stop function (\f(CW\*(C`ev_<type>_stop (loop, watcher *)\*(C'\fR. 706corresponding stop function (\f(CW\*(C`ev_<type>_stop (loop, watcher *)\*(C'\fR.
423.PP 707.PP
424As long as your watcher is active (has been started but not stopped) you 708As long as your watcher is active (has been started but not stopped) you
425must not touch the values stored in it. Most specifically you must never 709must not touch the values stored in it. Most specifically you must never
426reinitialise it or call its set method. 710reinitialise it or call its \f(CW\*(C`set\*(C'\fR macro.
427.PP
428You can check whether an event is active by calling the \f(CW\*(C`ev_is_active
429(watcher *)\*(C'\fR macro. To see whether an event is outstanding (but the
430callback for it has not been called yet) you can use the \f(CW\*(C`ev_is_pending
431(watcher *)\*(C'\fR macro.
432.PP 711.PP
433Each and every callback receives the event loop pointer as first, the 712Each and every callback receives the event loop pointer as first, the
434registered watcher structure as second, and a bitset of received events as 713registered watcher structure as second, and a bitset of received events as
435third argument. 714third argument.
436.PP 715.PP
… …
461The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread. 740The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread.
462.ie n .IP """EV_CHILD""" 4 741.ie n .IP """EV_CHILD""" 4
463.el .IP "\f(CWEV_CHILD\fR" 4 742.el .IP "\f(CWEV_CHILD\fR" 4
464.IX Item "EV_CHILD" 743.IX Item "EV_CHILD"
465The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change. 744The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change.
745.ie n .IP """EV_STAT""" 4
746.el .IP "\f(CWEV_STAT\fR" 4
747.IX Item "EV_STAT"
748The path specified in the \f(CW\*(C`ev_stat\*(C'\fR watcher changed its attributes somehow.
466.ie n .IP """EV_IDLE""" 4 749.ie n .IP """EV_IDLE""" 4
467.el .IP "\f(CWEV_IDLE\fR" 4 750.el .IP "\f(CWEV_IDLE\fR" 4
468.IX Item "EV_IDLE" 751.IX Item "EV_IDLE"
469The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do. 752The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do.
470.ie n .IP """EV_PREPARE""" 4 753.ie n .IP """EV_PREPARE""" 4
… …
480\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any 763\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any
481received events. Callbacks of both watcher types can start and stop as 764received events. Callbacks of both watcher types can start and stop as
482many watchers as they want, and all of them will be taken into account 765many watchers as they want, and all of them will be taken into account
483(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep 766(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep
484\&\f(CW\*(C`ev_loop\*(C'\fR from blocking). 767\&\f(CW\*(C`ev_loop\*(C'\fR from blocking).
768.ie n .IP """EV_EMBED""" 4
769.el .IP "\f(CWEV_EMBED\fR" 4
770.IX Item "EV_EMBED"
771The embedded event loop specified in the \f(CW\*(C`ev_embed\*(C'\fR watcher needs attention.
772.ie n .IP """EV_FORK""" 4
773.el .IP "\f(CWEV_FORK\fR" 4
774.IX Item "EV_FORK"
775The event loop has been resumed in the child process after fork (see
776\&\f(CW\*(C`ev_fork\*(C'\fR).
485.ie n .IP """EV_ERROR""" 4 777.ie n .IP """EV_ERROR""" 4
486.el .IP "\f(CWEV_ERROR\fR" 4 778.el .IP "\f(CWEV_ERROR\fR" 4
487.IX Item "EV_ERROR" 779.IX Item "EV_ERROR"
488An unspecified error has occured, the watcher has been stopped. This might 780An unspecified error has occured, the watcher has been stopped. This might
489happen because the watcher could not be properly started because libev 781happen because the watcher could not be properly started because libev
… …
494Libev will usually signal a few \*(L"dummy\*(R" events together with an error, 786Libev will usually signal a few \*(L"dummy\*(R" events together with an error,
495for example it might indicate that a fd is readable or writable, and if 787for example it might indicate that a fd is readable or writable, and if
496your callbacks is well-written it can just attempt the operation and cope 788your callbacks is well-written it can just attempt the operation and cope
497with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded 789with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded
498programs, though, so beware. 790programs, though, so beware.
791.Sh "\s-1GENERIC\s0 \s-1WATCHER\s0 \s-1FUNCTIONS\s0"
792.IX Subsection "GENERIC WATCHER FUNCTIONS"
793In the following description, \f(CW\*(C`TYPE\*(C'\fR stands for the watcher type,
794e.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.
795.ie n .IP """ev_init"" (ev_TYPE *watcher, callback)" 4
796.el .IP "\f(CWev_init\fR (ev_TYPE *watcher, callback)" 4
797.IX Item "ev_init (ev_TYPE *watcher, callback)"
798This macro initialises the generic portion of a watcher. The contents
799of the watcher object can be arbitrary (so \f(CW\*(C`malloc\*(C'\fR will do). Only
800the generic parts of the watcher are initialised, you \fIneed\fR to call
801the type-specific \f(CW\*(C`ev_TYPE_set\*(C'\fR macro afterwards to initialise the
802type-specific parts. For each type there is also a \f(CW\*(C`ev_TYPE_init\*(C'\fR macro
803which rolls both calls into one.
804.Sp
805You can reinitialise a watcher at any time as long as it has been stopped
806(or never started) and there are no pending events outstanding.
807.Sp
808The callback is always of type \f(CW\*(C`void (*)(ev_loop *loop, ev_TYPE *watcher,
809int revents)\*(C'\fR.
810.ie n .IP """ev_TYPE_set"" (ev_TYPE *, [args])" 4
811.el .IP "\f(CWev_TYPE_set\fR (ev_TYPE *, [args])" 4
812.IX Item "ev_TYPE_set (ev_TYPE *, [args])"
813This macro initialises the type-specific parts of a watcher. You need to
814call \f(CW\*(C`ev_init\*(C'\fR at least once before you call this macro, but you can
815call \f(CW\*(C`ev_TYPE_set\*(C'\fR any number of times. You must not, however, call this
816macro on a watcher that is active (it can be pending, however, which is a
817difference to the \f(CW\*(C`ev_init\*(C'\fR macro).
818.Sp
819Although some watcher types do not have type-specific arguments
820(e.g. \f(CW\*(C`ev_prepare\*(C'\fR) you still need to call its \f(CW\*(C`set\*(C'\fR macro.
821.ie n .IP """ev_TYPE_init"" (ev_TYPE *watcher, callback, [args])" 4
822.el .IP "\f(CWev_TYPE_init\fR (ev_TYPE *watcher, callback, [args])" 4
823.IX Item "ev_TYPE_init (ev_TYPE *watcher, callback, [args])"
824This convinience macro rolls both \f(CW\*(C`ev_init\*(C'\fR and \f(CW\*(C`ev_TYPE_set\*(C'\fR macro
825calls into a single call. This is the most convinient method to initialise
826a watcher. The same limitations apply, of course.
827.ie n .IP """ev_TYPE_start"" (loop *, ev_TYPE *watcher)" 4
828.el .IP "\f(CWev_TYPE_start\fR (loop *, ev_TYPE *watcher)" 4
829.IX Item "ev_TYPE_start (loop *, ev_TYPE *watcher)"
830Starts (activates) the given watcher. Only active watchers will receive
831events. If the watcher is already active nothing will happen.
832.ie n .IP """ev_TYPE_stop"" (loop *, ev_TYPE *watcher)" 4
833.el .IP "\f(CWev_TYPE_stop\fR (loop *, ev_TYPE *watcher)" 4
834.IX Item "ev_TYPE_stop (loop *, ev_TYPE *watcher)"
835Stops the given watcher again (if active) and clears the pending
836status. It is possible that stopped watchers are pending (for example,
837non-repeating timers are being stopped when they become pending), but
838\&\f(CW\*(C`ev_TYPE_stop\*(C'\fR ensures that the watcher is neither active nor pending. If
839you want to free or reuse the memory used by the watcher it is therefore a
840good idea to always call its \f(CW\*(C`ev_TYPE_stop\*(C'\fR function.
841.IP "bool ev_is_active (ev_TYPE *watcher)" 4
842.IX Item "bool ev_is_active (ev_TYPE *watcher)"
843Returns a true value iff the watcher is active (i.e. it has been started
844and not yet been stopped). As long as a watcher is active you must not modify
845it.
846.IP "bool ev_is_pending (ev_TYPE *watcher)" 4
847.IX Item "bool ev_is_pending (ev_TYPE *watcher)"
848Returns a true value iff the watcher is pending, (i.e. it has outstanding
849events but its callback has not yet been invoked). As long as a watcher
850is pending (but not active) you must not call an init function on it (but
851\&\f(CW\*(C`ev_TYPE_set\*(C'\fR is safe) and you must make sure the watcher is available to
852libev (e.g. you cnanot \f(CW\*(C`free ()\*(C'\fR it).
853.IP "callback = ev_cb (ev_TYPE *watcher)" 4
854.IX Item "callback = ev_cb (ev_TYPE *watcher)"
855Returns the callback currently set on the watcher.
856.IP "ev_cb_set (ev_TYPE *watcher, callback)" 4
857.IX Item "ev_cb_set (ev_TYPE *watcher, callback)"
858Change the callback. You can change the callback at virtually any time
859(modulo threads).
499.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0" 860.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0"
500.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER" 861.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER"
501Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change 862Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change
502and read at any time, libev will completely ignore it. This can be used 863and read at any time, libev will completely ignore it. This can be used
503to associate arbitrary data with your watcher. If you need more data and 864to associate arbitrary data with your watcher. If you need more data and
… …
529More interesting and less C\-conformant ways of catsing your callback type 890More interesting and less C\-conformant ways of catsing your callback type
530have been omitted.... 891have been omitted....
531.SH "WATCHER TYPES" 892.SH "WATCHER TYPES"
532.IX Header "WATCHER TYPES" 893.IX Header "WATCHER TYPES"
533This section describes each watcher in detail, but will not repeat 894This section describes each watcher in detail, but will not repeat
534information given in the last section. 895information given in the last section. Any initialisation/set macros,
896functions and members specific to the watcher type are explained.
897.PP
898Members are additionally marked with either \fI[read\-only]\fR, meaning that,
899while the watcher is active, you can look at the member and expect some
900sensible content, but you must not modify it (you can modify it while the
901watcher is stopped to your hearts content), or \fI[read\-write]\fR, which
902means you can expect it to have some sensible content while the watcher
903is active, but you can also modify it. Modifying it may not do something
904sensible or take immediate effect (or do anything at all), but libev will
905not crash or malfunction in any way.
535.ie n .Sh """ev_io"" \- is this file descriptor readable or writable" 906.ie n .Sh """ev_io"" \- is this file descriptor readable or writable?"
536.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable" 907.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable?"
537.IX Subsection "ev_io - is this file descriptor readable or writable" 908.IX Subsection "ev_io - is this file descriptor readable or writable?"
538I/O watchers check whether a file descriptor is readable or writable 909I/O watchers check whether a file descriptor is readable or writable
539in each iteration of the event loop (This behaviour is called 910in each iteration of the event loop, or, more precisely, when reading
540level-triggering because you keep receiving events as long as the 911would not block the process and writing would at least be able to write
541condition persists. Remember you can stop the watcher if you don't want to 912some data. This behaviour is called level-triggering because you keep
542act on the event and neither want to receive future events). 913receiving events as long as the condition persists. Remember you can stop
914the watcher if you don't want to act on the event and neither want to
915receive future events.
543.PP 916.PP
544In general you can register as many read and/or write event watchers per 917In general you can register as many read and/or write event watchers per
545fd as you want (as long as you don't confuse yourself). Setting all file 918fd as you want (as long as you don't confuse yourself). Setting all file
546descriptors to non-blocking mode is also usually a good idea (but not 919descriptors to non-blocking mode is also usually a good idea (but not
547required if you know what you are doing). 920required if you know what you are doing).
548.PP 921.PP
549You have to be careful with dup'ed file descriptors, though. Some backends 922You have to be careful with dup'ed file descriptors, though. Some backends
550(the linux epoll backend is a notable example) cannot handle dup'ed file 923(the linux epoll backend is a notable example) cannot handle dup'ed file
551descriptors correctly if you register interest in two or more fds pointing 924descriptors correctly if you register interest in two or more fds pointing
552to the same underlying file/socket etc. description (that is, they share 925to the same underlying file/socket/etc. description (that is, they share
553the same underlying \*(L"file open\*(R"). 926the same underlying \*(L"file open\*(R").
554.PP 927.PP
555If you must do this, then force the use of a known-to-be-good backend 928If you must do this, then force the use of a known-to-be-good backend
556(at the time of this writing, this includes only \s-1EVMETHOD_SELECT\s0 and 929(at the time of this writing, this includes only \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and
557\&\s-1EVMETHOD_POLL\s0). 930\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR).
931.PP
932Another thing you have to watch out for is that it is quite easy to
933receive \*(L"spurious\*(R" readyness notifications, that is your callback might
934be called with \f(CW\*(C`EV_READ\*(C'\fR but a subsequent \f(CW\*(C`read\*(C'\fR(2) will actually block
935because there is no data. Not only are some backends known to create a
936lot of those (for example solaris ports), it is very easy to get into
937this situation even with a relatively standard program structure. Thus
938it is best to always use non-blocking I/O: An extra \f(CW\*(C`read\*(C'\fR(2) returning
939\&\f(CW\*(C`EAGAIN\*(C'\fR is far preferable to a program hanging until some data arrives.
940.PP
941If you cannot run the fd in non-blocking mode (for example you should not
942play around with an Xlib connection), then you have to seperately re-test
943wether a file descriptor is really ready with a known-to-be good interface
944such as poll (fortunately in our Xlib example, Xlib already does this on
945its own, so its quite safe to use).
558.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4 946.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4
559.IX Item "ev_io_init (ev_io *, callback, int fd, int events)" 947.IX Item "ev_io_init (ev_io *, callback, int fd, int events)"
560.PD 0 948.PD 0
561.IP "ev_io_set (ev_io *, int fd, int events)" 4 949.IP "ev_io_set (ev_io *, int fd, int events)" 4
562.IX Item "ev_io_set (ev_io *, int fd, int events)" 950.IX Item "ev_io_set (ev_io *, int fd, int events)"
563.PD 951.PD
564Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The fd is the file descriptor to rceeive 952Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The \f(CW\*(C`fd\*(C'\fR is the file descriptor to
565events 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 | 953rceeive events for and events is either \f(CW\*(C`EV_READ\*(C'\fR, \f(CW\*(C`EV_WRITE\*(C'\fR or
566EV_WRITE\*(C'\fR to receive the given events. 954\&\f(CW\*(C`EV_READ | EV_WRITE\*(C'\fR to receive the given events.
955.IP "int fd [read\-only]" 4
956.IX Item "int fd [read-only]"
957The file descriptor being watched.
958.IP "int events [read\-only]" 4
959.IX Item "int events [read-only]"
960The events being watched.
961.PP
962Example: call \f(CW\*(C`stdin_readable_cb\*(C'\fR when \s-1STDIN_FILENO\s0 has become, well
963readable, but only once. Since it is likely line\-buffered, you could
964attempt to read a whole line in the callback:
965.PP
966.Vb 6
967\& static void
968\& stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
969\& {
970\& ev_io_stop (loop, w);
971\& .. read from stdin here (or from w->fd) and haqndle any I/O errors
972\& }
973.Ve
974.PP
975.Vb 6
976\& ...
977\& struct ev_loop *loop = ev_default_init (0);
978\& struct ev_io stdin_readable;
979\& ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
980\& ev_io_start (loop, &stdin_readable);
981\& ev_loop (loop, 0);
982.Ve
567.ie n .Sh """ev_timer"" \- relative and optionally recurring timeouts" 983.ie n .Sh """ev_timer"" \- relative and optionally repeating timeouts"
568.el .Sh "\f(CWev_timer\fP \- relative and optionally recurring timeouts" 984.el .Sh "\f(CWev_timer\fP \- relative and optionally repeating timeouts"
569.IX Subsection "ev_timer - relative and optionally recurring timeouts" 985.IX Subsection "ev_timer - relative and optionally repeating timeouts"
570Timer watchers are simple relative timers that generate an event after a 986Timer watchers are simple relative timers that generate an event after a
571given time, and optionally repeating in regular intervals after that. 987given time, and optionally repeating in regular intervals after that.
572.PP 988.PP
573The timers are based on real time, that is, if you register an event that 989The timers are based on real time, that is, if you register an event that
574times out after an hour and you reset your system clock to last years 990times out after an hour and you reset your system clock to last years
575time, it will still time out after (roughly) and hour. \*(L"Roughly\*(R" because 991time, it will still time out after (roughly) and hour. \*(L"Roughly\*(R" because
576detecting time jumps is hard, and soem inaccuracies are unavoidable (the 992detecting time jumps is hard, and some inaccuracies are unavoidable (the
577monotonic clock option helps a lot here). 993monotonic clock option helps a lot here).
578.PP 994.PP
579The relative timeouts are calculated relative to the \f(CW\*(C`ev_now ()\*(C'\fR 995The relative timeouts are calculated relative to the \f(CW\*(C`ev_now ()\*(C'\fR
580time. This is usually the right thing as this timestamp refers to the time 996time. This is usually the right thing as this timestamp refers to the time
581of the event triggering whatever timeout you are modifying/starting. If 997of the event triggering whatever timeout you are modifying/starting. If
582you suspect event processing to be delayed and you *need* to base the timeout 998you suspect event processing to be delayed and you \fIneed\fR to base the timeout
583on the current time, use something like this to adjust for this: 999on the current time, use something like this to adjust for this:
584.PP 1000.PP
585.Vb 1 1001.Vb 1
586\& ev_timer_set (&timer, after + ev_now () - ev_time (), 0.); 1002\& ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
587.Ve 1003.Ve
1004.PP
1005The callback is guarenteed to be invoked only when its timeout has passed,
1006but if multiple timers become ready during the same loop iteration then
1007order of execution is undefined.
588.IP "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)" 4 1008.IP "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)" 4
589.IX Item "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)" 1009.IX Item "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)"
590.PD 0 1010.PD 0
591.IP "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 4 1011.IP "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 4
592.IX Item "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 1012.IX Item "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)"
… …
610.Sp 1030.Sp
611If the timer is repeating, either start it if necessary (with the repeat 1031If the timer is repeating, either start it if necessary (with the repeat
612value), or reset the running timer to the repeat value. 1032value), or reset the running timer to the repeat value.
613.Sp 1033.Sp
614This sounds a bit complicated, but here is a useful and typical 1034This sounds a bit complicated, but here is a useful and typical
615example: Imagine you have a tcp connection and you want a so-called idle 1035example: Imagine you have a tcp connection and you want a so-called
616timeout, that is, you want to be called when there have been, say, 60 1036idle timeout, that is, you want to be called when there have been,
617seconds of inactivity on the socket. The easiest way to do this is to 1037say, 60 seconds of inactivity on the socket. The easiest way to do
618configure an \f(CW\*(C`ev_timer\*(C'\fR with after=repeat=60 and calling ev_timer_again each 1038this is to configure an \f(CW\*(C`ev_timer\*(C'\fR with \f(CW\*(C`after\*(C'\fR=\f(CW\*(C`repeat\*(C'\fR=\f(CW60\fR and calling
619time you successfully read or write some data. If you go into an idle 1039\&\f(CW\*(C`ev_timer_again\*(C'\fR each time you successfully read or write some data. If
620state where you do not expect data to travel on the socket, you can stop 1040you go into an idle state where you do not expect data to travel on the
621the timer, and again will automatically restart it if need be. 1041socket, you can stop the timer, and again will automatically restart it if
1042need be.
1043.Sp
1044You can also ignore the \f(CW\*(C`after\*(C'\fR value and \f(CW\*(C`ev_timer_start\*(C'\fR altogether
1045and only ever use the \f(CW\*(C`repeat\*(C'\fR value:
1046.Sp
1047.Vb 8
1048\& ev_timer_init (timer, callback, 0., 5.);
1049\& ev_timer_again (loop, timer);
1050\& ...
1051\& timer->again = 17.;
1052\& ev_timer_again (loop, timer);
1053\& ...
1054\& timer->again = 10.;
1055\& ev_timer_again (loop, timer);
1056.Ve
1057.Sp
1058This is more efficient then stopping/starting the timer eahc time you want
1059to modify its timeout value.
1060.IP "ev_tstamp repeat [read\-write]" 4
1061.IX Item "ev_tstamp repeat [read-write]"
1062The current \f(CW\*(C`repeat\*(C'\fR value. Will be used each time the watcher times out
1063or \f(CW\*(C`ev_timer_again\*(C'\fR is called and determines the next timeout (if any),
1064which is also when any modifications are taken into account.
1065.PP
1066Example: create a timer that fires after 60 seconds.
1067.PP
1068.Vb 5
1069\& static void
1070\& one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1071\& {
1072\& .. one minute over, w is actually stopped right here
1073\& }
1074.Ve
1075.PP
1076.Vb 3
1077\& struct ev_timer mytimer;
1078\& ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1079\& ev_timer_start (loop, &mytimer);
1080.Ve
1081.PP
1082Example: create a timeout timer that times out after 10 seconds of
1083inactivity.
1084.PP
1085.Vb 5
1086\& static void
1087\& timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1088\& {
1089\& .. ten seconds without any activity
1090\& }
1091.Ve
1092.PP
1093.Vb 4
1094\& struct ev_timer mytimer;
1095\& ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1096\& ev_timer_again (&mytimer); /* start timer */
1097\& ev_loop (loop, 0);
1098.Ve
1099.PP
1100.Vb 3
1101\& // and in some piece of code that gets executed on any "activity":
1102\& // reset the timeout to start ticking again at 10 seconds
1103\& ev_timer_again (&mytimer);
1104.Ve
622.ie n .Sh """ev_periodic"" \- to cron or not to cron" 1105.ie n .Sh """ev_periodic"" \- to cron or not to cron?"
623.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron" 1106.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron?"
624.IX Subsection "ev_periodic - to cron or not to cron" 1107.IX Subsection "ev_periodic - to cron or not to cron?"
625Periodic watchers are also timers of a kind, but they are very versatile 1108Periodic watchers are also timers of a kind, but they are very versatile
626(and unfortunately a bit complex). 1109(and unfortunately a bit complex).
627.PP 1110.PP
628Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time) 1111Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time)
629but on wallclock time (absolute time). You can tell a periodic watcher 1112but on wallclock time (absolute time). You can tell a periodic watcher
630to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a 1113to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a
631periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now () 1114periodic watcher to trigger in 10 seconds (by specifiying e.g. \f(CW\*(C`ev_now ()
632+ 10.>) and then reset your system clock to the last year, then it will 1115+ 10.\*(C'\fR) and then reset your system clock to the last year, then it will
633take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger 1116take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger
634roughly 10 seconds later and of course not if you reset your system time 1117roughly 10 seconds later and of course not if you reset your system time
635again). 1118again).
636.PP 1119.PP
637They can also be used to implement vastly more complex timers, such as 1120They can also be used to implement vastly more complex timers, such as
638triggering an event on eahc midnight, local time. 1121triggering an event on eahc midnight, local time.
1122.PP
1123As with timers, the callback is guarenteed to be invoked only when the
1124time (\f(CW\*(C`at\*(C'\fR) has been passed, but if multiple periodic timers become ready
1125during the same loop iteration then order of execution is undefined.
639.IP "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)" 4 1126.IP "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)" 4
640.IX Item "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)" 1127.IX Item "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)"
641.PD 0 1128.PD 0
642.IP "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)" 4 1129.IP "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)" 4
643.IX Item "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)" 1130.IX Item "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)"
… …
714.IX Item "ev_periodic_again (loop, ev_periodic *)" 1201.IX Item "ev_periodic_again (loop, ev_periodic *)"
715Simply stops and restarts the periodic watcher again. This is only useful 1202Simply stops and restarts the periodic watcher again. This is only useful
716when you changed some parameters or the reschedule callback would return 1203when you changed some parameters or the reschedule callback would return
717a different time than the last time it was called (e.g. in a crond like 1204a different time than the last time it was called (e.g. in a crond like
718program when the crontabs have changed). 1205program when the crontabs have changed).
1206.IP "ev_tstamp interval [read\-write]" 4
1207.IX Item "ev_tstamp interval [read-write]"
1208The current interval value. Can be modified any time, but changes only
1209take effect when the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being
1210called.
1211.IP "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read\-write]" 4
1212.IX Item "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]"
1213The current reschedule callback, or \f(CW0\fR, if this functionality is
1214switched off. Can be changed any time, but changes only take effect when
1215the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being called.
1216.PP
1217Example: call a callback every hour, or, more precisely, whenever the
1218system clock is divisible by 3600. The callback invocation times have
1219potentially a lot of jittering, but good long-term stability.
1220.PP
1221.Vb 5
1222\& static void
1223\& clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1224\& {
1225\& ... its now a full hour (UTC, or TAI or whatever your clock follows)
1226\& }
1227.Ve
1228.PP
1229.Vb 3
1230\& struct ev_periodic hourly_tick;
1231\& ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1232\& ev_periodic_start (loop, &hourly_tick);
1233.Ve
1234.PP
1235Example: the same as above, but use a reschedule callback to do it:
1236.PP
1237.Vb 1
1238\& #include <math.h>
1239.Ve
1240.PP
1241.Vb 5
1242\& static ev_tstamp
1243\& my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1244\& {
1245\& return fmod (now, 3600.) + 3600.;
1246\& }
1247.Ve
1248.PP
1249.Vb 1
1250\& ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1251.Ve
1252.PP
1253Example: call a callback every hour, starting now:
1254.PP
1255.Vb 4
1256\& struct ev_periodic hourly_tick;
1257\& ev_periodic_init (&hourly_tick, clock_cb,
1258\& fmod (ev_now (loop), 3600.), 3600., 0);
1259\& ev_periodic_start (loop, &hourly_tick);
1260.Ve
719.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled" 1261.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled!"
720.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled" 1262.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled!"
721.IX Subsection "ev_signal - signal me when a signal gets signalled" 1263.IX Subsection "ev_signal - signal me when a signal gets signalled!"
722Signal watchers will trigger an event when the process receives a specific 1264Signal watchers will trigger an event when the process receives a specific
723signal one or more times. Even though signals are very asynchronous, libev 1265signal one or more times. Even though signals are very asynchronous, libev
724will try it's best to deliver signals synchronously, i.e. as part of the 1266will try it's best to deliver signals synchronously, i.e. as part of the
725normal event processing, like any other event. 1267normal event processing, like any other event.
726.PP 1268.PP
… …
736.IP "ev_signal_set (ev_signal *, int signum)" 4 1278.IP "ev_signal_set (ev_signal *, int signum)" 4
737.IX Item "ev_signal_set (ev_signal *, int signum)" 1279.IX Item "ev_signal_set (ev_signal *, int signum)"
738.PD 1280.PD
739Configures the watcher to trigger on the given signal number (usually one 1281Configures the watcher to trigger on the given signal number (usually one
740of the \f(CW\*(C`SIGxxx\*(C'\fR constants). 1282of the \f(CW\*(C`SIGxxx\*(C'\fR constants).
1283.IP "int signum [read\-only]" 4
1284.IX Item "int signum [read-only]"
1285The signal the watcher watches out for.
741.ie n .Sh """ev_child"" \- wait for pid status changes" 1286.ie n .Sh """ev_child"" \- watch out for process status changes"
742.el .Sh "\f(CWev_child\fP \- wait for pid status changes" 1287.el .Sh "\f(CWev_child\fP \- watch out for process status changes"
743.IX Subsection "ev_child - wait for pid status changes" 1288.IX Subsection "ev_child - watch out for process status changes"
744Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to 1289Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to
745some child status changes (most typically when a child of yours dies). 1290some child status changes (most typically when a child of yours dies).
746.IP "ev_child_init (ev_child *, callback, int pid)" 4 1291.IP "ev_child_init (ev_child *, callback, int pid)" 4
747.IX Item "ev_child_init (ev_child *, callback, int pid)" 1292.IX Item "ev_child_init (ev_child *, callback, int pid)"
748.PD 0 1293.PD 0
… …
753\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look 1298\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look
754at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see 1299at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see
755the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems 1300the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems
756\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the 1301\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the
757process causing the status change. 1302process causing the status change.
1303.IP "int pid [read\-only]" 4
1304.IX Item "int pid [read-only]"
1305The process id this watcher watches out for, or \f(CW0\fR, meaning any process id.
1306.IP "int rpid [read\-write]" 4
1307.IX Item "int rpid [read-write]"
1308The process id that detected a status change.
1309.IP "int rstatus [read\-write]" 4
1310.IX Item "int rstatus [read-write]"
1311The process exit/trace status caused by \f(CW\*(C`rpid\*(C'\fR (see your systems
1312\&\f(CW\*(C`waitpid\*(C'\fR and \f(CW\*(C`sys/wait.h\*(C'\fR documentation for details).
1313.PP
1314Example: try to exit cleanly on \s-1SIGINT\s0 and \s-1SIGTERM\s0.
1315.PP
1316.Vb 5
1317\& static void
1318\& sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1319\& {
1320\& ev_unloop (loop, EVUNLOOP_ALL);
1321\& }
1322.Ve
1323.PP
1324.Vb 3
1325\& struct ev_signal signal_watcher;
1326\& ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1327\& ev_signal_start (loop, &sigint_cb);
1328.Ve
1329.ie n .Sh """ev_stat"" \- did the file attributes just change?"
1330.el .Sh "\f(CWev_stat\fP \- did the file attributes just change?"
1331.IX Subsection "ev_stat - did the file attributes just change?"
1332This watches a filesystem path for attribute changes. That is, it calls
1333\&\f(CW\*(C`stat\*(C'\fR regularly (or when the \s-1OS\s0 says it changed) and sees if it changed
1334compared to the last time, invoking the callback if it did.
1335.PP
1336The path does not need to exist: changing from \*(L"path exists\*(R" to \*(L"path does
1337not exist\*(R" is a status change like any other. The condition \*(L"path does
1338not exist\*(R" is signified by the \f(CW\*(C`st_nlink\*(C'\fR field being zero (which is
1339otherwise always forced to be at least one) and all the other fields of
1340the stat buffer having unspecified contents.
1341.PP
1342Since there is no standard to do this, the portable implementation simply
1343calls \f(CW\*(C`stat (2)\*(C'\fR regulalry on the path to see if it changed somehow. You
1344can specify a recommended polling interval for this case. If you specify
1345a polling interval of \f(CW0\fR (highly recommended!) then a \fIsuitable,
1346unspecified default\fR value will be used (which you can expect to be around
1347five seconds, although this might change dynamically). Libev will also
1348impose a minimum interval which is currently around \f(CW0.1\fR, but thats
1349usually overkill.
1350.PP
1351This watcher type is not meant for massive numbers of stat watchers,
1352as even with OS-supported change notifications, this can be
1353resource\-intensive.
1354.PP
1355At the time of this writing, no specific \s-1OS\s0 backends are implemented, but
1356if demand increases, at least a kqueue and inotify backend will be added.
1357.IP "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)" 4
1358.IX Item "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)"
1359.PD 0
1360.IP "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)" 4
1361.IX Item "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)"
1362.PD
1363Configures the watcher to wait for status changes of the given
1364\&\f(CW\*(C`path\*(C'\fR. The \f(CW\*(C`interval\*(C'\fR is a hint on how quickly a change is expected to
1365be detected and should normally be specified as \f(CW0\fR to let libev choose
1366a suitable value. The memory pointed to by \f(CW\*(C`path\*(C'\fR must point to the same
1367path for as long as the watcher is active.
1368.Sp
1369The callback will be receive \f(CW\*(C`EV_STAT\*(C'\fR when a change was detected,
1370relative to the attributes at the time the watcher was started (or the
1371last change was detected).
1372.IP "ev_stat_stat (ev_stat *)" 4
1373.IX Item "ev_stat_stat (ev_stat *)"
1374Updates the stat buffer immediately with new values. If you change the
1375watched path in your callback, you could call this fucntion to avoid
1376detecting this change (while introducing a race condition). Can also be
1377useful simply to find out the new values.
1378.IP "ev_statdata attr [read\-only]" 4
1379.IX Item "ev_statdata attr [read-only]"
1380The most-recently detected attributes of the file. Although the type is of
1381\&\f(CW\*(C`ev_statdata\*(C'\fR, this is usually the (or one of the) \f(CW\*(C`struct stat\*(C'\fR types
1382suitable for your system. If the \f(CW\*(C`st_nlink\*(C'\fR member is \f(CW0\fR, then there
1383was some error while \f(CW\*(C`stat\*(C'\fRing the file.
1384.IP "ev_statdata prev [read\-only]" 4
1385.IX Item "ev_statdata prev [read-only]"
1386The previous attributes of the file. The callback gets invoked whenever
1387\&\f(CW\*(C`prev\*(C'\fR != \f(CW\*(C`attr\*(C'\fR.
1388.IP "ev_tstamp interval [read\-only]" 4
1389.IX Item "ev_tstamp interval [read-only]"
1390The specified interval.
1391.IP "const char *path [read\-only]" 4
1392.IX Item "const char *path [read-only]"
1393The filesystem path that is being watched.
1394.PP
1395Example: Watch \f(CW\*(C`/etc/passwd\*(C'\fR for attribute changes.
1396.PP
1397.Vb 15
1398\& static void
1399\& passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1400\& {
1401\& /* /etc/passwd changed in some way */
1402\& if (w->attr.st_nlink)
1403\& {
1404\& printf ("passwd current size %ld\en", (long)w->attr.st_size);
1405\& printf ("passwd current atime %ld\en", (long)w->attr.st_mtime);
1406\& printf ("passwd current mtime %ld\en", (long)w->attr.st_mtime);
1407\& }
1408\& else
1409\& /* you shalt not abuse printf for puts */
1410\& puts ("wow, /etc/passwd is not there, expect problems. "
1411\& "if this is windows, they already arrived\en");
1412\& }
1413.Ve
1414.PP
1415.Vb 2
1416\& ...
1417\& ev_stat passwd;
1418.Ve
1419.PP
1420.Vb 2
1421\& ev_stat_init (&passwd, passwd_cb, "/etc/passwd");
1422\& ev_stat_start (loop, &passwd);
1423.Ve
758.ie n .Sh """ev_idle"" \- when you've got nothing better to do" 1424.ie n .Sh """ev_idle"" \- when you've got nothing better to do..."
759.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do" 1425.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do..."
760.IX Subsection "ev_idle - when you've got nothing better to do" 1426.IX Subsection "ev_idle - when you've got nothing better to do..."
761Idle watchers trigger events when there are no other events are pending 1427Idle watchers trigger events when there are no other events are pending
762(prepare, check and other idle watchers do not count). That is, as long 1428(prepare, check and other idle watchers do not count). That is, as long
763as your process is busy handling sockets or timeouts (or even signals, 1429as your process is busy handling sockets or timeouts (or even signals,
764imagine) it will not be triggered. But when your process is idle all idle 1430imagine) it will not be triggered. But when your process is idle all idle
765watchers are being called again and again, once per event loop iteration \- 1431watchers are being called again and again, once per event loop iteration \-
… …
776.IP "ev_idle_init (ev_signal *, callback)" 4 1442.IP "ev_idle_init (ev_signal *, callback)" 4
777.IX Item "ev_idle_init (ev_signal *, callback)" 1443.IX Item "ev_idle_init (ev_signal *, callback)"
778Initialises and configures the idle watcher \- it has no parameters of any 1444Initialises and configures the idle watcher \- it has no parameters of any
779kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless, 1445kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless,
780believe me. 1446believe me.
1447.PP
1448Example: dynamically allocate an \f(CW\*(C`ev_idle\*(C'\fR, start it, and in the
1449callback, free it. Alos, use no error checking, as usual.
1450.PP
1451.Vb 7
1452\& static void
1453\& idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1454\& {
1455\& free (w);
1456\& // now do something you wanted to do when the program has
1457\& // no longer asnything immediate to do.
1458\& }
1459.Ve
1460.PP
1461.Vb 3
1462\& struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1463\& ev_idle_init (idle_watcher, idle_cb);
1464\& ev_idle_start (loop, idle_cb);
1465.Ve
781.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop" 1466.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop!"
782.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop" 1467.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop!"
783.IX Subsection "ev_prepare and ev_check - customise your event loop" 1468.IX Subsection "ev_prepare and ev_check - customise your event loop!"
784Prepare and check watchers are usually (but not always) used in tandem: 1469Prepare and check watchers are usually (but not always) used in tandem:
785prepare watchers get invoked before the process blocks and check watchers 1470prepare watchers get invoked before the process blocks and check watchers
786afterwards. 1471afterwards.
787.PP 1472.PP
1473You \fImust not\fR call \f(CW\*(C`ev_loop\*(C'\fR or similar functions that enter
1474the current event loop from either \f(CW\*(C`ev_prepare\*(C'\fR or \f(CW\*(C`ev_check\*(C'\fR
1475watchers. Other loops than the current one are fine, however. The
1476rationale behind this is that you do not need to check for recursion in
1477those watchers, i.e. the sequence will always be \f(CW\*(C`ev_prepare\*(C'\fR, blocking,
1478\&\f(CW\*(C`ev_check\*(C'\fR so if you have one watcher of each kind they will always be
1479called in pairs bracketing the blocking call.
1480.PP
788Their main purpose is to integrate other event mechanisms into libev. This 1481Their main purpose is to integrate other event mechanisms into libev and
789could be used, for example, to track variable changes, implement your own 1482their use is somewhat advanced. This could be used, for example, to track
790watchers, integrate net-snmp or a coroutine library and lots more. 1483variable changes, implement your own watchers, integrate net-snmp or a
1484coroutine library and lots more. They are also occasionally useful if
1485you cache some data and want to flush it before blocking (for example,
1486in X programs you might want to do an \f(CW\*(C`XFlush ()\*(C'\fR in an \f(CW\*(C`ev_prepare\*(C'\fR
1487watcher).
791.PP 1488.PP
792This is done by examining in each prepare call which file descriptors need 1489This is done by examining in each prepare call which file descriptors need
793to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for 1490to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for
794them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries 1491them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries
795provide just this functionality). Then, in the check watcher you check for 1492provide just this functionality). Then, in the check watcher you check for
… …
813.IX Item "ev_check_init (ev_check *, callback)" 1510.IX Item "ev_check_init (ev_check *, callback)"
814.PD 1511.PD
815Initialises and configures the prepare or check watcher \- they have no 1512Initialises and configures the prepare or check watcher \- they have no
816parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR 1513parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR
817macros, but using them is utterly, utterly and completely pointless. 1514macros, but using them is utterly, utterly and completely pointless.
1515.PP
1516Example: To include a library such as adns, you would add \s-1IO\s0 watchers
1517and a timeout watcher in a prepare handler, as required by libadns, and
1518in a check watcher, destroy them and call into libadns. What follows is
1519pseudo-code only of course:
1520.PP
1521.Vb 2
1522\& static ev_io iow [nfd];
1523\& static ev_timer tw;
1524.Ve
1525.PP
1526.Vb 9
1527\& static void
1528\& io_cb (ev_loop *loop, ev_io *w, int revents)
1529\& {
1530\& // set the relevant poll flags
1531\& // could also call adns_processreadable etc. here
1532\& struct pollfd *fd = (struct pollfd *)w->data;
1533\& if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1534\& if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1535\& }
1536.Ve
1537.PP
1538.Vb 7
1539\& // create io watchers for each fd and a timer before blocking
1540\& static void
1541\& adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1542\& {
1543\& int timeout = 3600000;truct pollfd fds [nfd];
1544\& // actual code will need to loop here and realloc etc.
1545\& adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1546.Ve
1547.PP
1548.Vb 3
1549\& /* the callback is illegal, but won't be called as we stop during check */
1550\& ev_timer_init (&tw, 0, timeout * 1e-3);
1551\& ev_timer_start (loop, &tw);
1552.Ve
1553.PP
1554.Vb 6
1555\& // create on ev_io per pollfd
1556\& for (int i = 0; i < nfd; ++i)
1557\& {
1558\& ev_io_init (iow + i, io_cb, fds [i].fd,
1559\& ((fds [i].events & POLLIN ? EV_READ : 0)
1560\& | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1561.Ve
1562.PP
1563.Vb 5
1564\& fds [i].revents = 0;
1565\& iow [i].data = fds + i;
1566\& ev_io_start (loop, iow + i);
1567\& }
1568\& }
1569.Ve
1570.PP
1571.Vb 5
1572\& // stop all watchers after blocking
1573\& static void
1574\& adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1575\& {
1576\& ev_timer_stop (loop, &tw);
1577.Ve
1578.PP
1579.Vb 2
1580\& for (int i = 0; i < nfd; ++i)
1581\& ev_io_stop (loop, iow + i);
1582.Ve
1583.PP
1584.Vb 2
1585\& adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1586\& }
1587.Ve
1588.ie n .Sh """ev_embed"" \- when one backend isn't enough..."
1589.el .Sh "\f(CWev_embed\fP \- when one backend isn't enough..."
1590.IX Subsection "ev_embed - when one backend isn't enough..."
1591This is a rather advanced watcher type that lets you embed one event loop
1592into another (currently only \f(CW\*(C`ev_io\*(C'\fR events are supported in the embedded
1593loop, other types of watchers might be handled in a delayed or incorrect
1594fashion and must not be used).
1595.PP
1596There are primarily two reasons you would want that: work around bugs and
1597prioritise I/O.
1598.PP
1599As an example for a bug workaround, the kqueue backend might only support
1600sockets on some platform, so it is unusable as generic backend, but you
1601still want to make use of it because you have many sockets and it scales
1602so nicely. In this case, you would create a kqueue-based loop and embed it
1603into your default loop (which might use e.g. poll). Overall operation will
1604be a bit slower because first libev has to poll and then call kevent, but
1605at least you can use both at what they are best.
1606.PP
1607As for prioritising I/O: rarely you have the case where some fds have
1608to be watched and handled very quickly (with low latency), and even
1609priorities and idle watchers might have too much overhead. In this case
1610you would put all the high priority stuff in one loop and all the rest in
1611a second one, and embed the second one in the first.
1612.PP
1613As long as the watcher is active, the callback will be invoked every time
1614there might be events pending in the embedded loop. The callback must then
1615call \f(CW\*(C`ev_embed_sweep (mainloop, watcher)\*(C'\fR to make a single sweep and invoke
1616their callbacks (you could also start an idle watcher to give the embedded
1617loop strictly lower priority for example). You can also set the callback
1618to \f(CW0\fR, in which case the embed watcher will automatically execute the
1619embedded loop sweep.
1620.PP
1621As long as the watcher is started it will automatically handle events. The
1622callback will be invoked whenever some events have been handled. You can
1623set the callback to \f(CW0\fR to avoid having to specify one if you are not
1624interested in that.
1625.PP
1626Also, there have not currently been made special provisions for forking:
1627when you fork, you not only have to call \f(CW\*(C`ev_loop_fork\*(C'\fR on both loops,
1628but you will also have to stop and restart any \f(CW\*(C`ev_embed\*(C'\fR watchers
1629yourself.
1630.PP
1631Unfortunately, not all backends are embeddable, only the ones returned by
1632\&\f(CW\*(C`ev_embeddable_backends\*(C'\fR are, which, unfortunately, does not include any
1633portable one.
1634.PP
1635So when you want to use this feature you will always have to be prepared
1636that you cannot get an embeddable loop. The recommended way to get around
1637this is to have a separate variables for your embeddable loop, try to
1638create it, and if that fails, use the normal loop for everything:
1639.PP
1640.Vb 3
1641\& struct ev_loop *loop_hi = ev_default_init (0);
1642\& struct ev_loop *loop_lo = 0;
1643\& struct ev_embed embed;
1644.Ve
1645.PP
1646.Vb 5
1647\& // see if there is a chance of getting one that works
1648\& // (remember that a flags value of 0 means autodetection)
1649\& loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1650\& ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1651\& : 0;
1652.Ve
1653.PP
1654.Vb 8
1655\& // if we got one, then embed it, otherwise default to loop_hi
1656\& if (loop_lo)
1657\& {
1658\& ev_embed_init (&embed, 0, loop_lo);
1659\& ev_embed_start (loop_hi, &embed);
1660\& }
1661\& else
1662\& loop_lo = loop_hi;
1663.Ve
1664.IP "ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)" 4
1665.IX Item "ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)"
1666.PD 0
1667.IP "ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)" 4
1668.IX Item "ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)"
1669.PD
1670Configures the watcher to embed the given loop, which must be
1671embeddable. If the callback is \f(CW0\fR, then \f(CW\*(C`ev_embed_sweep\*(C'\fR will be
1672invoked automatically, otherwise it is the responsibility of the callback
1673to invoke it (it will continue to be called until the sweep has been done,
1674if you do not want thta, you need to temporarily stop the embed watcher).
1675.IP "ev_embed_sweep (loop, ev_embed *)" 4
1676.IX Item "ev_embed_sweep (loop, ev_embed *)"
1677Make a single, non-blocking sweep over the embedded loop. This works
1678similarly to \f(CW\*(C`ev_loop (embedded_loop, EVLOOP_NONBLOCK)\*(C'\fR, but in the most
1679apropriate way for embedded loops.
1680.IP "struct ev_loop *loop [read\-only]" 4
1681.IX Item "struct ev_loop *loop [read-only]"
1682The embedded event loop.
1683.ie n .Sh """ev_fork"" \- the audacity to resume the event loop after a fork"
1684.el .Sh "\f(CWev_fork\fP \- the audacity to resume the event loop after a fork"
1685.IX Subsection "ev_fork - the audacity to resume the event loop after a fork"
1686Fork watchers are called when a \f(CW\*(C`fork ()\*(C'\fR was detected (usually because
1687whoever is a good citizen cared to tell libev about it by calling
1688\&\f(CW\*(C`ev_default_fork\*(C'\fR or \f(CW\*(C`ev_loop_fork\*(C'\fR). The invocation is done before the
1689event loop blocks next and before \f(CW\*(C`ev_check\*(C'\fR watchers are being called,
1690and only in the child after the fork. If whoever good citizen calling
1691\&\f(CW\*(C`ev_default_fork\*(C'\fR cheats and calls it in the wrong process, the fork
1692handlers will be invoked, too, of course.
1693.IP "ev_fork_init (ev_signal *, callback)" 4
1694.IX Item "ev_fork_init (ev_signal *, callback)"
1695Initialises and configures the fork watcher \- it has no parameters of any
1696kind. There is a \f(CW\*(C`ev_fork_set\*(C'\fR macro, but using it is utterly pointless,
1697believe me.
818.SH "OTHER FUNCTIONS" 1698.SH "OTHER FUNCTIONS"
819.IX Header "OTHER FUNCTIONS" 1699.IX Header "OTHER FUNCTIONS"
820There are some other functions of possible interest. Described. Here. Now. 1700There are some other functions of possible interest. Described. Here. Now.
821.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4 1701.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4
822.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 1702.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)"
… …
851.Ve 1731.Ve
852.Sp 1732.Sp
853.Vb 1 1733.Vb 1
854\& ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0); 1734\& ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
855.Ve 1735.Ve
856.IP "ev_feed_event (loop, watcher, int events)" 4 1736.IP "ev_feed_event (ev_loop *, watcher *, int revents)" 4
857.IX Item "ev_feed_event (loop, watcher, int events)" 1737.IX Item "ev_feed_event (ev_loop *, watcher *, int revents)"
858Feeds the given event set into the event loop, as if the specified event 1738Feeds the given event set into the event loop, as if the specified event
859had happened for the specified watcher (which must be a pointer to an 1739had happened for the specified watcher (which must be a pointer to an
860initialised but not necessarily started event watcher). 1740initialised but not necessarily started event watcher).
861.IP "ev_feed_fd_event (loop, int fd, int revents)" 4 1741.IP "ev_feed_fd_event (ev_loop *, int fd, int revents)" 4
862.IX Item "ev_feed_fd_event (loop, int fd, int revents)" 1742.IX Item "ev_feed_fd_event (ev_loop *, int fd, int revents)"
863Feed an event on the given fd, as if a file descriptor backend detected 1743Feed an event on the given fd, as if a file descriptor backend detected
864the given events it. 1744the given events it.
865.IP "ev_feed_signal_event (loop, int signum)" 4 1745.IP "ev_feed_signal_event (ev_loop *loop, int signum)" 4
866.IX Item "ev_feed_signal_event (loop, int signum)" 1746.IX Item "ev_feed_signal_event (ev_loop *loop, int signum)"
867Feed an event as if the given signal occured (loop must be the default loop!). 1747Feed an event as if the given signal occured (\f(CW\*(C`loop\*(C'\fR must be the default
1748loop!).
868.SH "LIBEVENT EMULATION" 1749.SH "LIBEVENT EMULATION"
869.IX Header "LIBEVENT EMULATION" 1750.IX Header "LIBEVENT EMULATION"
870Libev offers a compatibility emulation layer for libevent. It cannot 1751Libev offers a compatibility emulation layer for libevent. It cannot
871emulate the internals of libevent, so here are some usage hints: 1752emulate the internals of libevent, so here are some usage hints:
872.IP "* Use it by including <event.h>, as usual." 4 1753.IP "* Use it by including <event.h>, as usual." 4
… …
883.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4 1764.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4
884.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library." 1765.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library."
885.PD 1766.PD
886.SH "\*(C+ SUPPORT" 1767.SH "\*(C+ SUPPORT"
887.IX Header " SUPPORT" 1768.IX Header " SUPPORT"
888\&\s-1TBD\s0. 1769Libev comes with some simplistic wrapper classes for \*(C+ that mainly allow
1770you to use some convinience methods to start/stop watchers and also change
1771the callback model to a model using method callbacks on objects.
1772.PP
1773To use it,
1774.PP
1775.Vb 1
1776\& #include <ev++.h>
1777.Ve
1778.PP
1779(it is not installed by default). This automatically includes \fIev.h\fR
1780and puts all of its definitions (many of them macros) into the global
1781namespace. All \*(C+ specific things are put into the \f(CW\*(C`ev\*(C'\fR namespace.
1782.PP
1783It should support all the same embedding options as \fIev.h\fR, most notably
1784\&\f(CW\*(C`EV_MULTIPLICITY\*(C'\fR.
1785.PP
1786Here is a list of things available in the \f(CW\*(C`ev\*(C'\fR namespace:
1787.ie n .IP """ev::READ""\fR, \f(CW""ev::WRITE"" etc." 4
1788.el .IP "\f(CWev::READ\fR, \f(CWev::WRITE\fR etc." 4
1789.IX Item "ev::READ, ev::WRITE etc."
1790These are just enum values with the same values as the \f(CW\*(C`EV_READ\*(C'\fR etc.
1791macros from \fIev.h\fR.
1792.ie n .IP """ev::tstamp""\fR, \f(CW""ev::now""" 4
1793.el .IP "\f(CWev::tstamp\fR, \f(CWev::now\fR" 4
1794.IX Item "ev::tstamp, ev::now"
1795Aliases to the same types/functions as with the \f(CW\*(C`ev_\*(C'\fR prefix.
1796.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
1797.el .IP "\f(CWev::io\fR, \f(CWev::timer\fR, \f(CWev::periodic\fR, \f(CWev::idle\fR, \f(CWev::sig\fR etc." 4
1798.IX Item "ev::io, ev::timer, ev::periodic, ev::idle, ev::sig etc."
1799For each \f(CW\*(C`ev_TYPE\*(C'\fR watcher in \fIev.h\fR there is a corresponding class of
1800the same name in the \f(CW\*(C`ev\*(C'\fR namespace, with the exception of \f(CW\*(C`ev_signal\*(C'\fR
1801which is called \f(CW\*(C`ev::sig\*(C'\fR to avoid clashes with the \f(CW\*(C`signal\*(C'\fR macro
1802defines by many implementations.
1803.Sp
1804All of those classes have these methods:
1805.RS 4
1806.IP "ev::TYPE::TYPE (object *, object::method *)" 4
1807.IX Item "ev::TYPE::TYPE (object *, object::method *)"
1808.PD 0
1809.IP "ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)" 4
1810.IX Item "ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)"
1811.IP "ev::TYPE::~TYPE" 4
1812.IX Item "ev::TYPE::~TYPE"
1813.PD
1814The constructor takes a pointer to an object and a method pointer to
1815the event handler callback to call in this class. The constructor calls
1816\&\f(CW\*(C`ev_init\*(C'\fR for you, which means you have to call the \f(CW\*(C`set\*(C'\fR method
1817before starting it. If you do not specify a loop then the constructor
1818automatically associates the default loop with this watcher.
1819.Sp
1820The destructor automatically stops the watcher if it is active.
1821.IP "w\->set (struct ev_loop *)" 4
1822.IX Item "w->set (struct ev_loop *)"
1823Associates a different \f(CW\*(C`struct ev_loop\*(C'\fR with this watcher. You can only
1824do this when the watcher is inactive (and not pending either).
1825.IP "w\->set ([args])" 4
1826.IX Item "w->set ([args])"
1827Basically the same as \f(CW\*(C`ev_TYPE_set\*(C'\fR, with the same args. Must be
1828called at least once. Unlike the C counterpart, an active watcher gets
1829automatically stopped and restarted.
1830.IP "w\->start ()" 4
1831.IX Item "w->start ()"
1832Starts the watcher. Note that there is no \f(CW\*(C`loop\*(C'\fR argument as the
1833constructor already takes the loop.
1834.IP "w\->stop ()" 4
1835.IX Item "w->stop ()"
1836Stops the watcher if it is active. Again, no \f(CW\*(C`loop\*(C'\fR argument.
1837.ie n .IP "w\->again () ""ev::timer""\fR, \f(CW""ev::periodic"" only" 4
1838.el .IP "w\->again () \f(CWev::timer\fR, \f(CWev::periodic\fR only" 4
1839.IX Item "w->again () ev::timer, ev::periodic only"
1840For \f(CW\*(C`ev::timer\*(C'\fR and \f(CW\*(C`ev::periodic\*(C'\fR, this invokes the corresponding
1841\&\f(CW\*(C`ev_TYPE_again\*(C'\fR function.
1842.ie n .IP "w\->sweep () ""ev::embed"" only" 4
1843.el .IP "w\->sweep () \f(CWev::embed\fR only" 4
1844.IX Item "w->sweep () ev::embed only"
1845Invokes \f(CW\*(C`ev_embed_sweep\*(C'\fR.
1846.ie n .IP "w\->update () ""ev::stat"" only" 4
1847.el .IP "w\->update () \f(CWev::stat\fR only" 4
1848.IX Item "w->update () ev::stat only"
1849Invokes \f(CW\*(C`ev_stat_stat\*(C'\fR.
1850.RE
1851.RS 4
1852.RE
1853.PP
1854Example: Define a class with an \s-1IO\s0 and idle watcher, start one of them in
1855the constructor.
1856.PP
1857.Vb 4
1858\& class myclass
1859\& {
1860\& ev_io io; void io_cb (ev::io &w, int revents);
1861\& ev_idle idle void idle_cb (ev::idle &w, int revents);
1862.Ve
1863.PP
1864.Vb 2
1865\& myclass ();
1866\& }
1867.Ve
1868.PP
1869.Vb 6
1870\& myclass::myclass (int fd)
1871\& : io (this, &myclass::io_cb),
1872\& idle (this, &myclass::idle_cb)
1873\& {
1874\& io.start (fd, ev::READ);
1875\& }
1876.Ve
1877.SH "MACRO MAGIC"
1878.IX Header "MACRO MAGIC"
1879Libev can be compiled with a variety of options, the most fundemantal is
1880\&\f(CW\*(C`EV_MULTIPLICITY\*(C'\fR. This option determines wether (most) functions and
1881callbacks have an initial \f(CW\*(C`struct ev_loop *\*(C'\fR argument.
1882.PP
1883To make it easier to write programs that cope with either variant, the
1884following macros are defined:
1885.ie n .IP """EV_A""\fR, \f(CW""EV_A_""" 4
1886.el .IP "\f(CWEV_A\fR, \f(CWEV_A_\fR" 4
1887.IX Item "EV_A, EV_A_"
1888This provides the loop \fIargument\fR for functions, if one is required (\*(L"ev
1889loop argument\*(R"). The \f(CW\*(C`EV_A\*(C'\fR form is used when this is the sole argument,
1890\&\f(CW\*(C`EV_A_\*(C'\fR is used when other arguments are following. Example:
1891.Sp
1892.Vb 3
1893\& ev_unref (EV_A);
1894\& ev_timer_add (EV_A_ watcher);
1895\& ev_loop (EV_A_ 0);
1896.Ve
1897.Sp
1898It assumes the variable \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR is in scope,
1899which is often provided by the following macro.
1900.ie n .IP """EV_P""\fR, \f(CW""EV_P_""" 4
1901.el .IP "\f(CWEV_P\fR, \f(CWEV_P_\fR" 4
1902.IX Item "EV_P, EV_P_"
1903This provides the loop \fIparameter\fR for functions, if one is required (\*(L"ev
1904loop parameter\*(R"). The \f(CW\*(C`EV_P\*(C'\fR form is used when this is the sole parameter,
1905\&\f(CW\*(C`EV_P_\*(C'\fR is used when other parameters are following. Example:
1906.Sp
1907.Vb 2
1908\& // this is how ev_unref is being declared
1909\& static void ev_unref (EV_P);
1910.Ve
1911.Sp
1912.Vb 2
1913\& // this is how you can declare your typical callback
1914\& static void cb (EV_P_ ev_timer *w, int revents)
1915.Ve
1916.Sp
1917It declares a parameter \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR, quite
1918suitable for use with \f(CW\*(C`EV_A\*(C'\fR.
1919.ie n .IP """EV_DEFAULT""\fR, \f(CW""EV_DEFAULT_""" 4
1920.el .IP "\f(CWEV_DEFAULT\fR, \f(CWEV_DEFAULT_\fR" 4
1921.IX Item "EV_DEFAULT, EV_DEFAULT_"
1922Similar to the other two macros, this gives you the value of the default
1923loop, if multiple loops are supported (\*(L"ev loop default\*(R").
1924.PP
1925Example: Declare and initialise a check watcher, working regardless of
1926wether multiple loops are supported or not.
1927.PP
1928.Vb 5
1929\& static void
1930\& check_cb (EV_P_ ev_timer *w, int revents)
1931\& {
1932\& ev_check_stop (EV_A_ w);
1933\& }
1934.Ve
1935.PP
1936.Vb 4
1937\& ev_check check;
1938\& ev_check_init (&check, check_cb);
1939\& ev_check_start (EV_DEFAULT_ &check);
1940\& ev_loop (EV_DEFAULT_ 0);
1941.Ve
1942.SH "EMBEDDING"
1943.IX Header "EMBEDDING"
1944Libev can (and often is) directly embedded into host
1945applications. Examples of applications that embed it include the Deliantra
1946Game Server, the \s-1EV\s0 perl module, the \s-1GNU\s0 Virtual Private Ethernet (gvpe)
1947and rxvt\-unicode.
1948.PP
1949The goal is to enable you to just copy the neecssary files into your
1950source directory without having to change even a single line in them, so
1951you can easily upgrade by simply copying (or having a checked-out copy of
1952libev somewhere in your source tree).
1953.Sh "\s-1FILESETS\s0"
1954.IX Subsection "FILESETS"
1955Depending on what features you need you need to include one or more sets of files
1956in your app.
1957.PP
1958\fI\s-1CORE\s0 \s-1EVENT\s0 \s-1LOOP\s0\fR
1959.IX Subsection "CORE EVENT LOOP"
1960.PP
1961To include only the libev core (all the \f(CW\*(C`ev_*\*(C'\fR functions), with manual
1962configuration (no autoconf):
1963.PP
1964.Vb 2
1965\& #define EV_STANDALONE 1
1966\& #include "ev.c"
1967.Ve
1968.PP
1969This will automatically include \fIev.h\fR, too, and should be done in a
1970single C source file only to provide the function implementations. To use
1971it, do the same for \fIev.h\fR in all files wishing to use this \s-1API\s0 (best
1972done by writing a wrapper around \fIev.h\fR that you can include instead and
1973where you can put other configuration options):
1974.PP
1975.Vb 2
1976\& #define EV_STANDALONE 1
1977\& #include "ev.h"
1978.Ve
1979.PP
1980Both header files and implementation files can be compiled with a \*(C+
1981compiler (at least, thats a stated goal, and breakage will be treated
1982as a bug).
1983.PP
1984You need the following files in your source tree, or in a directory
1985in your include path (e.g. in libev/ when using \-Ilibev):
1986.PP
1987.Vb 4
1988\& ev.h
1989\& ev.c
1990\& ev_vars.h
1991\& ev_wrap.h
1992.Ve
1993.PP
1994.Vb 1
1995\& ev_win32.c required on win32 platforms only
1996.Ve
1997.PP
1998.Vb 5
1999\& ev_select.c only when select backend is enabled (which is by default)
2000\& ev_poll.c only when poll backend is enabled (disabled by default)
2001\& ev_epoll.c only when the epoll backend is enabled (disabled by default)
2002\& ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2003\& ev_port.c only when the solaris port backend is enabled (disabled by default)
2004.Ve
2005.PP
2006\&\fIev.c\fR includes the backend files directly when enabled, so you only need
2007to compile this single file.
2008.PP
2009\fI\s-1LIBEVENT\s0 \s-1COMPATIBILITY\s0 \s-1API\s0\fR
2010.IX Subsection "LIBEVENT COMPATIBILITY API"
2011.PP
2012To include the libevent compatibility \s-1API\s0, also include:
2013.PP
2014.Vb 1
2015\& #include "event.c"
2016.Ve
2017.PP
2018in the file including \fIev.c\fR, and:
2019.PP
2020.Vb 1
2021\& #include "event.h"
2022.Ve
2023.PP
2024in the files that want to use the libevent \s-1API\s0. This also includes \fIev.h\fR.
2025.PP
2026You need the following additional files for this:
2027.PP
2028.Vb 2
2029\& event.h
2030\& event.c
2031.Ve
2032.PP
2033\fI\s-1AUTOCONF\s0 \s-1SUPPORT\s0\fR
2034.IX Subsection "AUTOCONF SUPPORT"
2035.PP
2036Instead of using \f(CW\*(C`EV_STANDALONE=1\*(C'\fR and providing your config in
2037whatever way you want, you can also \f(CW\*(C`m4_include([libev.m4])\*(C'\fR in your
2038\&\fIconfigure.ac\fR and leave \f(CW\*(C`EV_STANDALONE\*(C'\fR undefined. \fIev.c\fR will then
2039include \fIconfig.h\fR and configure itself accordingly.
2040.PP
2041For this of course you need the m4 file:
2042.PP
2043.Vb 1
2044\& libev.m4
2045.Ve
2046.Sh "\s-1PREPROCESSOR\s0 \s-1SYMBOLS/MACROS\s0"
2047.IX Subsection "PREPROCESSOR SYMBOLS/MACROS"
2048Libev can be configured via a variety of preprocessor symbols you have to define
2049before including any of its files. The default is not to build for multiplicity
2050and only include the select backend.
2051.IP "\s-1EV_STANDALONE\s0" 4
2052.IX Item "EV_STANDALONE"
2053Must always be \f(CW1\fR if you do not use autoconf configuration, which
2054keeps libev from including \fIconfig.h\fR, and it also defines dummy
2055implementations for some libevent functions (such as logging, which is not
2056supported). It will also not define any of the structs usually found in
2057\&\fIevent.h\fR that are not directly supported by the libev core alone.
2058.IP "\s-1EV_USE_MONOTONIC\s0" 4
2059.IX Item "EV_USE_MONOTONIC"
2060If defined to be \f(CW1\fR, libev will try to detect the availability of the
2061monotonic clock option at both compiletime and runtime. Otherwise no use
2062of the monotonic clock option will be attempted. If you enable this, you
2063usually have to link against librt or something similar. Enabling it when
2064the functionality isn't available is safe, though, althoguh you have
2065to make sure you link against any libraries where the \f(CW\*(C`clock_gettime\*(C'\fR
2066function is hiding in (often \fI\-lrt\fR).
2067.IP "\s-1EV_USE_REALTIME\s0" 4
2068.IX Item "EV_USE_REALTIME"
2069If defined to be \f(CW1\fR, libev will try to detect the availability of the
2070realtime clock option at compiletime (and assume its availability at
2071runtime if successful). Otherwise no use of the realtime clock option will
2072be attempted. This effectively replaces \f(CW\*(C`gettimeofday\*(C'\fR by \f(CW\*(C`clock_get
2073(CLOCK_REALTIME, ...)\*(C'\fR and will not normally affect correctness. See tzhe note about libraries
2074in the description of \f(CW\*(C`EV_USE_MONOTONIC\*(C'\fR, though.
2075.IP "\s-1EV_USE_SELECT\s0" 4
2076.IX Item "EV_USE_SELECT"
2077If undefined or defined to be \f(CW1\fR, libev will compile in support for the
2078\&\f(CW\*(C`select\*(C'\fR(2) backend. No attempt at autodetection will be done: if no
2079other method takes over, select will be it. Otherwise the select backend
2080will not be compiled in.
2081.IP "\s-1EV_SELECT_USE_FD_SET\s0" 4
2082.IX Item "EV_SELECT_USE_FD_SET"
2083If defined to \f(CW1\fR, then the select backend will use the system \f(CW\*(C`fd_set\*(C'\fR
2084structure. This is useful if libev doesn't compile due to a missing
2085\&\f(CW\*(C`NFDBITS\*(C'\fR or \f(CW\*(C`fd_mask\*(C'\fR definition or it misguesses the bitset layout on
2086exotic systems. This usually limits the range of file descriptors to some
2087low limit such as 1024 or might have other limitations (winsocket only
2088allows 64 sockets). The \f(CW\*(C`FD_SETSIZE\*(C'\fR macro, set before compilation, might
2089influence the size of the \f(CW\*(C`fd_set\*(C'\fR used.
2090.IP "\s-1EV_SELECT_IS_WINSOCKET\s0" 4
2091.IX Item "EV_SELECT_IS_WINSOCKET"
2092When defined to \f(CW1\fR, the select backend will assume that
2093select/socket/connect etc. don't understand file descriptors but
2094wants osf handles on win32 (this is the case when the select to
2095be used is the winsock select). This means that it will call
2096\&\f(CW\*(C`_get_osfhandle\*(C'\fR on the fd to convert it to an \s-1OS\s0 handle. Otherwise,
2097it is assumed that all these functions actually work on fds, even
2098on win32. Should not be defined on non\-win32 platforms.
2099.IP "\s-1EV_USE_POLL\s0" 4
2100.IX Item "EV_USE_POLL"
2101If defined to be \f(CW1\fR, libev will compile in support for the \f(CW\*(C`poll\*(C'\fR(2)
2102backend. Otherwise it will be enabled on non\-win32 platforms. It
2103takes precedence over select.
2104.IP "\s-1EV_USE_EPOLL\s0" 4
2105.IX Item "EV_USE_EPOLL"
2106If defined to be \f(CW1\fR, libev will compile in support for the Linux
2107\&\f(CW\*(C`epoll\*(C'\fR(7) backend. Its availability will be detected at runtime,
2108otherwise another method will be used as fallback. This is the
2109preferred backend for GNU/Linux systems.
2110.IP "\s-1EV_USE_KQUEUE\s0" 4
2111.IX Item "EV_USE_KQUEUE"
2112If defined to be \f(CW1\fR, libev will compile in support for the \s-1BSD\s0 style
2113\&\f(CW\*(C`kqueue\*(C'\fR(2) backend. Its actual availability will be detected at runtime,
2114otherwise another method will be used as fallback. This is the preferred
2115backend for \s-1BSD\s0 and BSD-like systems, although on most BSDs kqueue only
2116supports some types of fds correctly (the only platform we found that
2117supports ptys for example was NetBSD), so kqueue might be compiled in, but
2118not be used unless explicitly requested. The best way to use it is to find
2119out whether kqueue supports your type of fd properly and use an embedded
2120kqueue loop.
2121.IP "\s-1EV_USE_PORT\s0" 4
2122.IX Item "EV_USE_PORT"
2123If defined to be \f(CW1\fR, libev will compile in support for the Solaris
212410 port style backend. Its availability will be detected at runtime,
2125otherwise another method will be used as fallback. This is the preferred
2126backend for Solaris 10 systems.
2127.IP "\s-1EV_USE_DEVPOLL\s0" 4
2128.IX Item "EV_USE_DEVPOLL"
2129reserved for future expansion, works like the \s-1USE\s0 symbols above.
2130.IP "\s-1EV_H\s0" 4
2131.IX Item "EV_H"
2132The name of the \fIev.h\fR header file used to include it. The default if
2133undefined is \f(CW\*(C`<ev.h>\*(C'\fR in \fIevent.h\fR and \f(CW"ev.h"\fR in \fIev.c\fR. This
2134can be used to virtually rename the \fIev.h\fR header file in case of conflicts.
2135.IP "\s-1EV_CONFIG_H\s0" 4
2136.IX Item "EV_CONFIG_H"
2137If \f(CW\*(C`EV_STANDALONE\*(C'\fR isn't \f(CW1\fR, this variable can be used to override
2138\&\fIev.c\fR's idea of where to find the \fIconfig.h\fR file, similarly to
2139\&\f(CW\*(C`EV_H\*(C'\fR, above.
2140.IP "\s-1EV_EVENT_H\s0" 4
2141.IX Item "EV_EVENT_H"
2142Similarly to \f(CW\*(C`EV_H\*(C'\fR, this macro can be used to override \fIevent.c\fR's idea
2143of how the \fIevent.h\fR header can be found.
2144.IP "\s-1EV_PROTOTYPES\s0" 4
2145.IX Item "EV_PROTOTYPES"
2146If defined to be \f(CW0\fR, then \fIev.h\fR will not define any function
2147prototypes, but still define all the structs and other symbols. This is
2148occasionally useful if you want to provide your own wrapper functions
2149around libev functions.
2150.IP "\s-1EV_MULTIPLICITY\s0" 4
2151.IX Item "EV_MULTIPLICITY"
2152If undefined or defined to \f(CW1\fR, then all event-loop-specific functions
2153will have the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument, and you can create
2154additional independent event loops. Otherwise there will be no support
2155for multiple event loops and there is no first event loop pointer
2156argument. Instead, all functions act on the single default loop.
2157.IP "\s-1EV_PERIODIC_ENABLE\s0" 4
2158.IX Item "EV_PERIODIC_ENABLE"
2159If undefined or defined to be \f(CW1\fR, then periodic timers are supported. If
2160defined to be \f(CW0\fR, then they are not. Disabling them saves a few kB of
2161code.
2162.IP "\s-1EV_EMBED_ENABLE\s0" 4
2163.IX Item "EV_EMBED_ENABLE"
2164If undefined or defined to be \f(CW1\fR, then embed watchers are supported. If
2165defined to be \f(CW0\fR, then they are not.
2166.IP "\s-1EV_STAT_ENABLE\s0" 4
2167.IX Item "EV_STAT_ENABLE"
2168If undefined or defined to be \f(CW1\fR, then stat watchers are supported. If
2169defined to be \f(CW0\fR, then they are not.
2170.IP "\s-1EV_FORK_ENABLE\s0" 4
2171.IX Item "EV_FORK_ENABLE"
2172If undefined or defined to be \f(CW1\fR, then fork watchers are supported. If
2173defined to be \f(CW0\fR, then they are not.
2174.IP "\s-1EV_MINIMAL\s0" 4
2175.IX Item "EV_MINIMAL"
2176If you need to shave off some kilobytes of code at the expense of some
2177speed, define this symbol to \f(CW1\fR. Currently only used for gcc to override
2178some inlining decisions, saves roughly 30% codesize of amd64.
2179.IP "\s-1EV_PID_HASHSIZE\s0" 4
2180.IX Item "EV_PID_HASHSIZE"
2181\&\f(CW\*(C`ev_child\*(C'\fR watchers use a small hash table to distribute workload by
2182pid. The default size is \f(CW16\fR (or \f(CW1\fR with \f(CW\*(C`EV_MINIMAL\*(C'\fR), usually more
2183than enough. If you need to manage thousands of children you might want to
2184increase this value.
2185.IP "\s-1EV_COMMON\s0" 4
2186.IX Item "EV_COMMON"
2187By default, all watchers have a \f(CW\*(C`void *data\*(C'\fR member. By redefining
2188this macro to a something else you can include more and other types of
2189members. You have to define it each time you include one of the files,
2190though, and it must be identical each time.
2191.Sp
2192For example, the perl \s-1EV\s0 module uses something like this:
2193.Sp
2194.Vb 3
2195\& #define EV_COMMON \e
2196\& SV *self; /* contains this struct */ \e
2197\& SV *cb_sv, *fh /* note no trailing ";" */
2198.Ve
2199.IP "\s-1EV_CB_DECLARE\s0 (type)" 4
2200.IX Item "EV_CB_DECLARE (type)"
2201.PD 0
2202.IP "\s-1EV_CB_INVOKE\s0 (watcher, revents)" 4
2203.IX Item "EV_CB_INVOKE (watcher, revents)"
2204.IP "ev_set_cb (ev, cb)" 4
2205.IX Item "ev_set_cb (ev, cb)"
2206.PD
2207Can be used to change the callback member declaration in each watcher,
2208and the way callbacks are invoked and set. Must expand to a struct member
2209definition and a statement, respectively. See the \fIev.v\fR header file for
2210their default definitions. One possible use for overriding these is to
2211avoid the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument in all cases, or to use
2212method calls instead of plain function calls in \*(C+.
2213.Sh "\s-1EXAMPLES\s0"
2214.IX Subsection "EXAMPLES"
2215For a real-world example of a program the includes libev
2216verbatim, you can have a look at the \s-1EV\s0 perl module
2217(<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
2218the \fIlibev/\fR subdirectory and includes them in the \fI\s-1EV/EVAPI\s0.h\fR (public
2219interface) and \fI\s-1EV\s0.xs\fR (implementation) files. Only the \fI\s-1EV\s0.xs\fR file
2220will be compiled. It is pretty complex because it provides its own header
2221file.
2222.Sp
2223The usage in rxvt-unicode is simpler. It has a \fIev_cpp.h\fR header file
2224that everybody includes and which overrides some autoconf choices:
2225.Sp
2226.Vb 4
2227\& #define EV_USE_POLL 0
2228\& #define EV_MULTIPLICITY 0
2229\& #define EV_PERIODICS 0
2230\& #define EV_CONFIG_H <config.h>
2231.Ve
2232.Sp
2233.Vb 1
2234\& #include "ev++.h"
2235.Ve
2236.Sp
2237And a \fIev_cpp.C\fR implementation file that contains libev proper and is compiled:
2238.Sp
2239.Vb 2
2240\& #include "ev_cpp.h"
2241\& #include "ev.c"
2242.Ve
2243.SH "COMPLEXITIES"
2244.IX Header "COMPLEXITIES"
2245In this section the complexities of (many of) the algorithms used inside
2246libev will be explained. For complexity discussions about backends see the
2247documentation for \f(CW\*(C`ev_default_init\*(C'\fR.
2248.RS 4
2249.IP "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)" 4
2250.IX Item "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)"
2251.PD 0
2252.IP "Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)" 4
2253.IX Item "Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)"
2254.IP "Starting io/check/prepare/idle/signal/child watchers: O(1)" 4
2255.IX Item "Starting io/check/prepare/idle/signal/child watchers: O(1)"
2256.IP "Stopping check/prepare/idle watchers: O(1)" 4
2257.IX Item "Stopping check/prepare/idle watchers: O(1)"
2258.IP "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % 16))" 4
2259.IX Item "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % 16))"
2260.IP "Finding the next timer per loop iteration: O(1)" 4
2261.IX Item "Finding the next timer per loop iteration: O(1)"
2262.IP "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)" 4
2263.IX Item "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)"
2264.IP "Activating one watcher: O(1)" 4
2265.IX Item "Activating one watcher: O(1)"
2266.RE
2267.RS 4
2268.PD
889.SH "AUTHOR" 2269.SH "AUTHOR"
890.IX Header "AUTHOR" 2270.IX Header "AUTHOR"
891Marc Lehmann <libev@schmorp.de>. 2271Marc Lehmann <libev@schmorp.de>.

Diff Legend

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