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

Comparing libev/ev.3 (file contents):
Revision 1.3 by root, Thu Nov 22 12:28:27 2007 UTC vs.
Revision 1.31 by root, Wed Nov 28 11:31:34 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 "2007-11-22" "perl v5.8.8" "User Contributed Perl Documentation" 132.TH "<STANDARD INPUT>" 1 "2007-11-28" "perl v5.8.8" "User Contributed Perl Documentation"
133.SH "NAME" 133.SH "NAME"
134libev \- a high performance full\-featured event loop written in C 134libev \- a high performance full\-featured event loop written in C
135.SH "SYNOPSIS" 135.SH "SYNOPSIS"
136.IX Header "SYNOPSIS" 136.IX Header "SYNOPSIS"
137.Vb 1 137.Vb 1
138\& #include <ev.h> 138\& #include <ev.h>
139.Ve 139.Ve
140.SH "EXAMPLE PROGRAM"
141.IX Header "EXAMPLE PROGRAM"
142.Vb 1
143\& #include <ev.h>
144.Ve
145.PP
146.Vb 2
147\& ev_io stdin_watcher;
148\& ev_timer timeout_watcher;
149.Ve
150.PP
151.Vb 8
152\& /* called when data readable on stdin */
153\& static void
154\& stdin_cb (EV_P_ struct ev_io *w, int revents)
155\& {
156\& /* puts ("stdin ready"); */
157\& ev_io_stop (EV_A_ w); /* just a syntax example */
158\& ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */
159\& }
160.Ve
161.PP
162.Vb 6
163\& static void
164\& timeout_cb (EV_P_ struct ev_timer *w, int revents)
165\& {
166\& /* puts ("timeout"); */
167\& ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */
168\& }
169.Ve
170.PP
171.Vb 4
172\& int
173\& main (void)
174\& {
175\& struct ev_loop *loop = ev_default_loop (0);
176.Ve
177.PP
178.Vb 3
179\& /* initialise an io watcher, then start it */
180\& ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
181\& ev_io_start (loop, &stdin_watcher);
182.Ve
183.PP
184.Vb 3
185\& /* simple non-repeating 5.5 second timeout */
186\& ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
187\& ev_timer_start (loop, &timeout_watcher);
188.Ve
189.PP
190.Vb 2
191\& /* loop till timeout or data ready */
192\& ev_loop (loop, 0);
193.Ve
194.PP
195.Vb 2
196\& return 0;
197\& }
198.Ve
140.SH "DESCRIPTION" 199.SH "DESCRIPTION"
141.IX Header "DESCRIPTION" 200.IX Header "DESCRIPTION"
142Libev is an event loop: you register interest in certain events (such as a 201Libev 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 202file descriptor being readable or a timeout occuring), and it will manage
144these event sources and provide your program with events. 203these event sources and provide your program with events.
… …
151watchers\fR, which are relatively small C structures you initialise with the 210watchers\fR, which are relatively small C structures you initialise with the
152details of the event, and then hand it over to libev by \fIstarting\fR the 211details of the event, and then hand it over to libev by \fIstarting\fR the
153watcher. 212watcher.
154.SH "FEATURES" 213.SH "FEATURES"
155.IX Header "FEATURES" 214.IX Header "FEATURES"
156Libev supports select, poll, the linux-specific epoll and the bsd-specific 215Libev supports \f(CW\*(C`select\*(C'\fR, \f(CW\*(C`poll\*(C'\fR, the Linux-specific \f(CW\*(C`epoll\*(C'\fR, the
157kqueue mechanisms for file descriptor events, relative timers, absolute 216BSD-specific \f(CW\*(C`kqueue\*(C'\fR and the Solaris-specific event port mechanisms
158timers with customised rescheduling, signal events, process status change 217for file descriptor events (\f(CW\*(C`ev_io\*(C'\fR), the Linux \f(CW\*(C`inotify\*(C'\fR interface
159events (related to \s-1SIGCHLD\s0), and event watchers dealing with the event 218(for \f(CW\*(C`ev_stat\*(C'\fR), relative timers (\f(CW\*(C`ev_timer\*(C'\fR), absolute timers
160loop mechanism itself (idle, prepare and check watchers). It also is quite 219with customised rescheduling (\f(CW\*(C`ev_periodic\*(C'\fR), synchronous signals
161fast (see this benchmark comparing 220(\f(CW\*(C`ev_signal\*(C'\fR), process status change events (\f(CW\*(C`ev_child\*(C'\fR), and event
162it to libevent for example). 221watchers dealing with the event loop mechanism itself (\f(CW\*(C`ev_idle\*(C'\fR,
222\&\f(CW\*(C`ev_embed\*(C'\fR, \f(CW\*(C`ev_prepare\*(C'\fR and \f(CW\*(C`ev_check\*(C'\fR watchers) as well as
223file watchers (\f(CW\*(C`ev_stat\*(C'\fR) and even limited support for fork events
224(\f(CW\*(C`ev_fork\*(C'\fR).
225.PP
226It also is quite fast (see this
227benchmark comparing it to libevent
228for example).
163.SH "CONVENTIONS" 229.SH "CONVENTIONS"
164.IX Header "CONVENTIONS" 230.IX Header "CONVENTIONS"
165Libev is very configurable. In this manual the default configuration 231Libev is very configurable. In this manual the default configuration will
166will be described, which supports multiple event loops. For more info 232be described, which supports multiple event loops. For more info about
167about various configuration options please have a look at the file 233various configuration options please have a look at \fB\s-1EMBED\s0\fR section in
168\&\fI\s-1README\s0.embed\fR in the libev distribution. If libev was configured without 234this manual. If libev was configured without support for multiple event
169support for multiple event loops, then all functions taking an initial 235loops, then all functions taking an initial argument of name \f(CW\*(C`loop\*(C'\fR
170argument of name \f(CW\*(C`loop\*(C'\fR (which is always of type \f(CW\*(C`struct ev_loop *\*(C'\fR) 236(which is always of type \f(CW\*(C`struct ev_loop *\*(C'\fR) will not have this argument.
171will not have this argument.
172.SH "TIME REPRESENTATION" 237.SH "TIME REPRESENTATION"
173.IX Header "TIME REPRESENTATION" 238.IX Header "TIME REPRESENTATION"
174Libev represents time as a single floating point number, representing the 239Libev represents time as a single floating point number, representing the
175(fractional) number of seconds since the (\s-1POSIX\s0) epoch (somewhere near 240(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 241the 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 242called \f(CW\*(C`ev_tstamp\*(C'\fR, which is what you should use too. It usually aliases
178to the double type in C. 243to the \f(CW\*(C`double\*(C'\fR type in C, and when you need to do any calculations on
244it, you should treat it as such.
179.SH "GLOBAL FUNCTIONS" 245.SH "GLOBAL FUNCTIONS"
180.IX Header "GLOBAL FUNCTIONS" 246.IX Header "GLOBAL FUNCTIONS"
181These functions can be called anytime, even before initialising the 247These functions can be called anytime, even before initialising the
182library in any way. 248library in any way.
183.IP "ev_tstamp ev_time ()" 4 249.IP "ev_tstamp ev_time ()" 4
… …
199.Sp 265.Sp
200Usually, it's a good idea to terminate if the major versions mismatch, 266Usually, it's a good idea to terminate if the major versions mismatch,
201as this indicates an incompatible change. Minor versions are usually 267as this indicates an incompatible change. Minor versions are usually
202compatible to older versions, so a larger minor version alone is usually 268compatible to older versions, so a larger minor version alone is usually
203not a problem. 269not a problem.
270.Sp
271Example: Make sure we haven't accidentally been linked against the wrong
272version.
273.Sp
274.Vb 3
275\& assert (("libev version mismatch",
276\& ev_version_major () == EV_VERSION_MAJOR
277\& && ev_version_minor () >= EV_VERSION_MINOR));
278.Ve
279.IP "unsigned int ev_supported_backends ()" 4
280.IX Item "unsigned int ev_supported_backends ()"
281Return the set of all backends (i.e. their corresponding \f(CW\*(C`EV_BACKEND_*\*(C'\fR
282value) compiled into this binary of libev (independent of their
283availability on the system you are running on). See \f(CW\*(C`ev_default_loop\*(C'\fR for
284a description of the set values.
285.Sp
286Example: make sure we have the epoll method, because yeah this is cool and
287a must have and can we have a torrent of it please!!!11
288.Sp
289.Vb 2
290\& assert (("sorry, no epoll, no sex",
291\& ev_supported_backends () & EVBACKEND_EPOLL));
292.Ve
293.IP "unsigned int ev_recommended_backends ()" 4
294.IX Item "unsigned int ev_recommended_backends ()"
295Return the set of all backends compiled into this binary of libev and also
296recommended for this platform. This set is often smaller than the one
297returned by \f(CW\*(C`ev_supported_backends\*(C'\fR, as for example kqueue is broken on
298most BSDs and will not be autodetected unless you explicitly request it
299(assuming you know what you are doing). This is the set of backends that
300libev will probe for if you specify no backends explicitly.
301.IP "unsigned int ev_embeddable_backends ()" 4
302.IX Item "unsigned int ev_embeddable_backends ()"
303Returns the set of backends that are embeddable in other event loops. This
304is the theoretical, all\-platform, value. To find which backends
305might be supported on the current system, you would need to look at
306\&\f(CW\*(C`ev_embeddable_backends () & ev_supported_backends ()\*(C'\fR, likewise for
307recommended ones.
308.Sp
309See the description of \f(CW\*(C`ev_embed\*(C'\fR watchers for more info.
204.IP "ev_set_allocator (void *(*cb)(void *ptr, long size))" 4 310.IP "ev_set_allocator (void *(*cb)(void *ptr, size_t size))" 4
205.IX Item "ev_set_allocator (void *(*cb)(void *ptr, long size))" 311.IX Item "ev_set_allocator (void *(*cb)(void *ptr, size_t size))"
206Sets the allocation function to use (the prototype is similar to the 312Sets the allocation function to use (the prototype and semantics are
207realloc C function, the semantics are identical). It is used to allocate 313identical to the realloc C function). It is used to allocate and free
208and free memory (no surprises here). If it returns zero when memory 314memory (no surprises here). If it returns zero when memory needs to be
209needs to be allocated, the library might abort or take some potentially 315allocated, the library might abort or take some potentially destructive
210destructive action. The default is your system realloc function. 316action. The default is your system realloc function.
211.Sp 317.Sp
212You could override this function in high-availability programs to, say, 318You could override this function in high-availability programs to, say,
213free some memory if it cannot allocate memory, to use a special allocator, 319free some memory if it cannot allocate memory, to use a special allocator,
214or even to sleep a while and retry until some memory is available. 320or even to sleep a while and retry until some memory is available.
321.Sp
322Example: Replace the libev allocator with one that waits a bit and then
323retries).
324.Sp
325.Vb 6
326\& static void *
327\& persistent_realloc (void *ptr, size_t size)
328\& {
329\& for (;;)
330\& {
331\& void *newptr = realloc (ptr, size);
332.Ve
333.Sp
334.Vb 2
335\& if (newptr)
336\& return newptr;
337.Ve
338.Sp
339.Vb 3
340\& sleep (60);
341\& }
342\& }
343.Ve
344.Sp
345.Vb 2
346\& ...
347\& ev_set_allocator (persistent_realloc);
348.Ve
215.IP "ev_set_syserr_cb (void (*cb)(const char *msg));" 4 349.IP "ev_set_syserr_cb (void (*cb)(const char *msg));" 4
216.IX Item "ev_set_syserr_cb (void (*cb)(const char *msg));" 350.IX Item "ev_set_syserr_cb (void (*cb)(const char *msg));"
217Set the callback function to call on a retryable syscall error (such 351Set the callback function to call on a retryable syscall error (such
218as failed select, poll, epoll_wait). The message is a printable string 352as failed select, poll, epoll_wait). The message is a printable string
219indicating the system call or subsystem causing the problem. If this 353indicating the system call or subsystem causing the problem. If this
220callback is set, then libev will expect it to remedy the sitution, no 354callback is set, then libev will expect it to remedy the sitution, no
221matter what, when it returns. That is, libev will generally retry the 355matter what, when it returns. That is, libev will generally retry the
222requested operation, or, if the condition doesn't go away, do bad stuff 356requested operation, or, if the condition doesn't go away, do bad stuff
223(such as abort). 357(such as abort).
358.Sp
359Example: This is basically the same thing that libev does internally, too.
360.Sp
361.Vb 6
362\& static void
363\& fatal_error (const char *msg)
364\& {
365\& perror (msg);
366\& abort ();
367\& }
368.Ve
369.Sp
370.Vb 2
371\& ...
372\& ev_set_syserr_cb (fatal_error);
373.Ve
224.SH "FUNCTIONS CONTROLLING THE EVENT LOOP" 374.SH "FUNCTIONS CONTROLLING THE EVENT LOOP"
225.IX Header "FUNCTIONS CONTROLLING THE EVENT LOOP" 375.IX Header "FUNCTIONS CONTROLLING THE EVENT LOOP"
226An event loop is described by a \f(CW\*(C`struct ev_loop *\*(C'\fR. The library knows two 376An event loop is described by a \f(CW\*(C`struct ev_loop *\*(C'\fR. The library knows two
227types of such loops, the \fIdefault\fR loop, which supports signals and child 377types of such loops, the \fIdefault\fR loop, which supports signals and child
228events, and dynamically created loops which do not. 378events, and dynamically created loops which do not.
… …
236.IP "struct ev_loop *ev_default_loop (unsigned int flags)" 4 386.IP "struct ev_loop *ev_default_loop (unsigned int flags)" 4
237.IX Item "struct ev_loop *ev_default_loop (unsigned int flags)" 387.IX Item "struct ev_loop *ev_default_loop (unsigned int flags)"
238This will initialise the default event loop if it hasn't been initialised 388This will initialise the default event loop if it hasn't been initialised
239yet and return it. If the default loop could not be initialised, returns 389yet and return it. If the default loop could not be initialised, returns
240false. If it already was initialised it simply returns it (and ignores the 390false. If it already was initialised it simply returns it (and ignores the
241flags). 391flags. If that is troubling you, check \f(CW\*(C`ev_backend ()\*(C'\fR afterwards).
242.Sp 392.Sp
243If you don't know what event loop to use, use the one returned from this 393If you don't know what event loop to use, use the one returned from this
244function. 394function.
245.Sp 395.Sp
246The flags argument can be used to specify special behaviour or specific 396The flags argument can be used to specify special behaviour or specific
247backends to use, and is usually specified as 0 (or \s-1EVFLAG_AUTO\s0). 397backends to use, and is usually specified as \f(CW0\fR (or \f(CW\*(C`EVFLAG_AUTO\*(C'\fR).
248.Sp 398.Sp
249It supports the following flags: 399The following flags are supported:
250.RS 4 400.RS 4
251.ie n .IP """EVFLAG_AUTO""" 4 401.ie n .IP """EVFLAG_AUTO""" 4
252.el .IP "\f(CWEVFLAG_AUTO\fR" 4 402.el .IP "\f(CWEVFLAG_AUTO\fR" 4
253.IX Item "EVFLAG_AUTO" 403.IX Item "EVFLAG_AUTO"
254The default flags value. Use this if you have no clue (it's the right 404The default flags value. Use this if you have no clue (it's the right
… …
260or setgid) then libev will \fInot\fR look at the environment variable 410or setgid) then libev will \fInot\fR look at the environment variable
261\&\f(CW\*(C`LIBEV_FLAGS\*(C'\fR. Otherwise (the default), this environment variable will 411\&\f(CW\*(C`LIBEV_FLAGS\*(C'\fR. Otherwise (the default), this environment variable will
262override the flags completely if it is found in the environment. This is 412override the flags completely if it is found in the environment. This is
263useful to try out specific backends to test their performance, or to work 413useful to try out specific backends to test their performance, or to work
264around bugs. 414around bugs.
265.ie n .IP """EVMETHOD_SELECT"" (value 1, portable select backend)" 4 415.ie n .IP """EVBACKEND_SELECT"" (value 1, portable select backend)" 4
266.el .IP "\f(CWEVMETHOD_SELECT\fR (value 1, portable select backend)" 4 416.el .IP "\f(CWEVBACKEND_SELECT\fR (value 1, portable select backend)" 4
267.IX Item "EVMETHOD_SELECT (value 1, portable select backend)" 417.IX Item "EVBACKEND_SELECT (value 1, portable select backend)"
268This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as 418This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as
269libev tries to roll its own fd_set with no limits on the number of fds, 419libev tries to roll its own fd_set with no limits on the number of fds,
270but if that fails, expect a fairly low limit on the number of fds when 420but if that fails, expect a fairly low limit on the number of fds when
271using this backend. It doesn't scale too well (O(highest_fd)), but its usually 421using this backend. It doesn't scale too well (O(highest_fd)), but its usually
272the fastest backend for a low number of fds. 422the fastest backend for a low number of fds.
273.ie n .IP """EVMETHOD_POLL"" (value 2, poll backend, available everywhere except on windows)" 4 423.ie n .IP """EVBACKEND_POLL"" (value 2, poll backend, available everywhere except on windows)" 4
274.el .IP "\f(CWEVMETHOD_POLL\fR (value 2, poll backend, available everywhere except on windows)" 4 424.el .IP "\f(CWEVBACKEND_POLL\fR (value 2, poll backend, available everywhere except on windows)" 4
275.IX Item "EVMETHOD_POLL (value 2, poll backend, available everywhere except on windows)" 425.IX Item "EVBACKEND_POLL (value 2, poll backend, available everywhere except on windows)"
276And this is your standard \fIpoll\fR\|(2) backend. It's more complicated than 426And this is your standard \fIpoll\fR\|(2) backend. It's more complicated than
277select, but handles sparse fds better and has no artificial limit on the 427select, but handles sparse fds better and has no artificial limit on the
278number of fds you can use (except it will slow down considerably with a 428number of fds you can use (except it will slow down considerably with a
279lot of inactive fds). It scales similarly to select, i.e. O(total_fds). 429lot of inactive fds). It scales similarly to select, i.e. O(total_fds).
280.ie n .IP """EVMETHOD_EPOLL"" (value 4, Linux)" 4 430.ie n .IP """EVBACKEND_EPOLL"" (value 4, Linux)" 4
281.el .IP "\f(CWEVMETHOD_EPOLL\fR (value 4, Linux)" 4 431.el .IP "\f(CWEVBACKEND_EPOLL\fR (value 4, Linux)" 4
282.IX Item "EVMETHOD_EPOLL (value 4, Linux)" 432.IX Item "EVBACKEND_EPOLL (value 4, Linux)"
283For few fds, this backend is a bit little slower than poll and select, 433For few fds, this backend is a bit little slower than poll and select,
284but it scales phenomenally better. While poll and select usually scale like 434but it scales phenomenally better. While poll and select usually scale like
285O(total_fds) where n is the total number of fds (or the highest fd), epoll scales 435O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
286either O(1) or O(active_fds). 436either O(1) or O(active_fds).
287.Sp 437.Sp
288While stopping and starting an I/O watcher in the same iteration will 438While stopping and starting an I/O watcher in the same iteration will
289result in some caching, there is still a syscall per such incident 439result in some caching, there is still a syscall per such incident
290(because the fd could point to a different file description now), so its 440(because the fd could point to a different file description now), so its
291best to avoid that. Also, \fIdup()\fRed file descriptors might not work very 441best to avoid that. Also, \fIdup()\fRed file descriptors might not work very
292well if you register events for both fds. 442well if you register events for both fds.
443.Sp
444Please note that epoll sometimes generates spurious notifications, so you
445need to use non-blocking I/O or other means to avoid blocking when no data
446(or space) is available.
293.ie n .IP """EVMETHOD_KQUEUE"" (value 8, most \s-1BSD\s0 clones)" 4 447.ie n .IP """EVBACKEND_KQUEUE"" (value 8, most \s-1BSD\s0 clones)" 4
294.el .IP "\f(CWEVMETHOD_KQUEUE\fR (value 8, most \s-1BSD\s0 clones)" 4 448.el .IP "\f(CWEVBACKEND_KQUEUE\fR (value 8, most \s-1BSD\s0 clones)" 4
295.IX Item "EVMETHOD_KQUEUE (value 8, most BSD clones)" 449.IX Item "EVBACKEND_KQUEUE (value 8, most BSD clones)"
296Kqueue deserves special mention, as at the time of this writing, it 450Kqueue deserves special mention, as at the time of this writing, it
297was broken on all BSDs except NetBSD (usually it doesn't work with 451was broken on all BSDs except NetBSD (usually it doesn't work with
298anything but sockets and pipes, except on Darwin, where of course its 452anything but sockets and pipes, except on Darwin, where of course its
299completely useless). For this reason its not being \*(L"autodetected\*(R" unless 453completely useless). For this reason its not being \*(L"autodetected\*(R"
300you explicitly specify the flags (i.e. you don't use \s-1EVFLAG_AUTO\s0). 454unless you explicitly specify it explicitly in the flags (i.e. using
455\&\f(CW\*(C`EVBACKEND_KQUEUE\*(C'\fR).
301.Sp 456.Sp
302It scales in the same way as the epoll backend, but the interface to the 457It scales in the same way as the epoll backend, but the interface to the
303kernel is more efficient (which says nothing about its actual speed, of 458kernel is more efficient (which says nothing about its actual speed, of
304course). While starting and stopping an I/O watcher does not cause an 459course). While starting and stopping an I/O watcher does not cause an
305extra syscall as with epoll, it still adds up to four event changes per 460extra syscall as with epoll, it still adds up to four event changes per
306incident, so its best to avoid that. 461incident, so its best to avoid that.
307.ie n .IP """EVMETHOD_DEVPOLL"" (value 16, Solaris 8)" 4 462.ie n .IP """EVBACKEND_DEVPOLL"" (value 16, Solaris 8)" 4
308.el .IP "\f(CWEVMETHOD_DEVPOLL\fR (value 16, Solaris 8)" 4 463.el .IP "\f(CWEVBACKEND_DEVPOLL\fR (value 16, Solaris 8)" 4
309.IX Item "EVMETHOD_DEVPOLL (value 16, Solaris 8)" 464.IX Item "EVBACKEND_DEVPOLL (value 16, Solaris 8)"
310This is not implemented yet (and might never be). 465This is not implemented yet (and might never be).
311.ie n .IP """EVMETHOD_PORT"" (value 32, Solaris 10)" 4 466.ie n .IP """EVBACKEND_PORT"" (value 32, Solaris 10)" 4
312.el .IP "\f(CWEVMETHOD_PORT\fR (value 32, Solaris 10)" 4 467.el .IP "\f(CWEVBACKEND_PORT\fR (value 32, Solaris 10)" 4
313.IX Item "EVMETHOD_PORT (value 32, Solaris 10)" 468.IX Item "EVBACKEND_PORT (value 32, Solaris 10)"
314This uses the Solaris 10 port mechanism. As with everything on Solaris, 469This uses the Solaris 10 port mechanism. As with everything on Solaris,
315it's really slow, but it still scales very well (O(active_fds)). 470it's really slow, but it still scales very well (O(active_fds)).
471.Sp
472Please note that solaris ports can result in a lot of spurious
473notifications, so you need to use non-blocking I/O or other means to avoid
474blocking when no data (or space) is available.
316.ie n .IP """EVMETHOD_ALL""" 4 475.ie n .IP """EVBACKEND_ALL""" 4
317.el .IP "\f(CWEVMETHOD_ALL\fR" 4 476.el .IP "\f(CWEVBACKEND_ALL\fR" 4
318.IX Item "EVMETHOD_ALL" 477.IX Item "EVBACKEND_ALL"
319Try all backends (even potentially broken ones). Since this is a mask, you 478Try all backends (even potentially broken ones that wouldn't be tried
320can do stuff like \f(CW\*(C`EVMETHOD_ALL & ~EVMETHOD_KQUEUE\*(C'\fR. 479with \f(CW\*(C`EVFLAG_AUTO\*(C'\fR). Since this is a mask, you can do stuff such as
480\&\f(CW\*(C`EVBACKEND_ALL & ~EVBACKEND_KQUEUE\*(C'\fR.
321.RE 481.RE
322.RS 4 482.RS 4
323.Sp 483.Sp
324If one or more of these are ored into the flags value, then only these 484If one or more of these are ored into the flags value, then only these
325backends will be tried (in the reverse order as given here). If none are 485backends will be tried (in the reverse order as given here). If none are
326specified, most compiled-in backend will be tried, usually in reverse 486specified, most compiled-in backend will be tried, usually in reverse
327order of their flag values :) 487order of their flag values :)
488.Sp
489The most typical usage is like this:
490.Sp
491.Vb 2
492\& if (!ev_default_loop (0))
493\& fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
494.Ve
495.Sp
496Restrict libev to the select and poll backends, and do not allow
497environment settings to be taken into account:
498.Sp
499.Vb 1
500\& ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
501.Ve
502.Sp
503Use whatever libev has to offer, but make sure that kqueue is used if
504available (warning, breaks stuff, best use only with your own private
505event loop and only if you know the \s-1OS\s0 supports your types of fds):
506.Sp
507.Vb 1
508\& ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
509.Ve
328.RE 510.RE
329.IP "struct ev_loop *ev_loop_new (unsigned int flags)" 4 511.IP "struct ev_loop *ev_loop_new (unsigned int flags)" 4
330.IX Item "struct ev_loop *ev_loop_new (unsigned int flags)" 512.IX Item "struct ev_loop *ev_loop_new (unsigned int flags)"
331Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is 513Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is
332always distinct from the default loop. Unlike the default loop, it cannot 514always distinct from the default loop. Unlike the default loop, it cannot
333handle signal and child watchers, and attempts to do so will be greeted by 515handle signal and child watchers, and attempts to do so will be greeted by
334undefined behaviour (or a failed assertion if assertions are enabled). 516undefined behaviour (or a failed assertion if assertions are enabled).
517.Sp
518Example: Try to create a event loop that uses epoll and nothing else.
519.Sp
520.Vb 3
521\& struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
522\& if (!epoller)
523\& fatal ("no epoll found here, maybe it hides under your chair");
524.Ve
335.IP "ev_default_destroy ()" 4 525.IP "ev_default_destroy ()" 4
336.IX Item "ev_default_destroy ()" 526.IX Item "ev_default_destroy ()"
337Destroys the default loop again (frees all memory and kernel state 527Destroys the default loop again (frees all memory and kernel state
338etc.). This stops all registered event watchers (by not touching them in 528etc.). None of the active event watchers will be stopped in the normal
339any way whatsoever, although you cannot rely on this :). 529sense, so e.g. \f(CW\*(C`ev_is_active\*(C'\fR might still return true. It is your
530responsibility to either stop all watchers cleanly yoursef \fIbefore\fR
531calling this function, or cope with the fact afterwards (which is usually
532the easiest thing, youc na just ignore the watchers and/or \f(CW\*(C`free ()\*(C'\fR them
533for example).
340.IP "ev_loop_destroy (loop)" 4 534.IP "ev_loop_destroy (loop)" 4
341.IX Item "ev_loop_destroy (loop)" 535.IX Item "ev_loop_destroy (loop)"
342Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an 536Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an
343earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR. 537earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR.
344.IP "ev_default_fork ()" 4 538.IP "ev_default_fork ()" 4
… …
346This function reinitialises the kernel state for backends that have 540This function reinitialises the kernel state for backends that have
347one. Despite the name, you can call it anytime, but it makes most sense 541one. Despite the name, you can call it anytime, but it makes most sense
348after forking, in either the parent or child process (or both, but that 542after forking, in either the parent or child process (or both, but that
349again makes little sense). 543again makes little sense).
350.Sp 544.Sp
351You \fImust\fR call this function after forking if and only if you want to 545You \fImust\fR call this function in the child process after forking if and
352use the event library in both processes. If you just fork+exec, you don't 546only if you want to use the event library in both processes. If you just
353have to call it. 547fork+exec, you don't have to call it.
354.Sp 548.Sp
355The function itself is quite fast and it's usually not a problem to call 549The function itself is quite fast and it's usually not a problem to call
356it just in case after a fork. To make this easy, the function will fit in 550it just in case after a fork. To make this easy, the function will fit in
357quite nicely into a call to \f(CW\*(C`pthread_atfork\*(C'\fR: 551quite nicely into a call to \f(CW\*(C`pthread_atfork\*(C'\fR:
358.Sp 552.Sp
359.Vb 1 553.Vb 1
360\& pthread_atfork (0, 0, ev_default_fork); 554\& pthread_atfork (0, 0, ev_default_fork);
361.Ve 555.Ve
556.Sp
557At the moment, \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and \f(CW\*(C`EVBACKEND_POLL\*(C'\fR are safe to use
558without calling this function, so if you force one of those backends you
559do not need to care.
362.IP "ev_loop_fork (loop)" 4 560.IP "ev_loop_fork (loop)" 4
363.IX Item "ev_loop_fork (loop)" 561.IX Item "ev_loop_fork (loop)"
364Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by 562Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by
365\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop 563\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop
366after fork, and how you do this is entirely your own problem. 564after fork, and how you do this is entirely your own problem.
367.IP "unsigned int ev_method (loop)" 4 565.IP "unsigned int ev_backend (loop)" 4
368.IX Item "unsigned int ev_method (loop)" 566.IX Item "unsigned int ev_backend (loop)"
369Returns one of the \f(CW\*(C`EVMETHOD_*\*(C'\fR flags indicating the event backend in 567Returns one of the \f(CW\*(C`EVBACKEND_*\*(C'\fR flags indicating the event backend in
370use. 568use.
371.IP "ev_tstamp ev_now (loop)" 4 569.IP "ev_tstamp ev_now (loop)" 4
372.IX Item "ev_tstamp ev_now (loop)" 570.IX Item "ev_tstamp ev_now (loop)"
373Returns the current \*(L"event loop time\*(R", which is the time the event loop 571Returns the current \*(L"event loop time\*(R", which is the time the event loop
374got events and started processing them. This timestamp does not change 572received events and started processing them. This timestamp does not
375as long as callbacks are being processed, and this is also the base time 573change as long as callbacks are being processed, and this is also the base
376used for relative timers. You can treat it as the timestamp of the event 574time used for relative timers. You can treat it as the timestamp of the
377occuring (or more correctly, the mainloop finding out about it). 575event occuring (or more correctly, libev finding out about it).
378.IP "ev_loop (loop, int flags)" 4 576.IP "ev_loop (loop, int flags)" 4
379.IX Item "ev_loop (loop, int flags)" 577.IX Item "ev_loop (loop, int flags)"
380Finally, this is it, the event handler. This function usually is called 578Finally, this is it, the event handler. This function usually is called
381after you initialised all your watchers and you want to start handling 579after you initialised all your watchers and you want to start handling
382events. 580events.
383.Sp 581.Sp
384If the flags argument is specified as 0, it will not return until either 582If the flags argument is specified as \f(CW0\fR, it will not return until
385no event watchers are active anymore or \f(CW\*(C`ev_unloop\*(C'\fR was called. 583either no event watchers are active anymore or \f(CW\*(C`ev_unloop\*(C'\fR was called.
584.Sp
585Please note that an explicit \f(CW\*(C`ev_unloop\*(C'\fR is usually better than
586relying on all watchers to be stopped when deciding when a program has
587finished (especially in interactive programs), but having a program that
588automatically loops as long as it has to and no longer by virtue of
589relying on its watchers stopping correctly is a thing of beauty.
386.Sp 590.Sp
387A flags value of \f(CW\*(C`EVLOOP_NONBLOCK\*(C'\fR will look for new events, will handle 591A flags value of \f(CW\*(C`EVLOOP_NONBLOCK\*(C'\fR will look for new events, will handle
388those events and any outstanding ones, but will not block your process in 592those events and any outstanding ones, but will not block your process in
389case there are no events and will return after one iteration of the loop. 593case there are no events and will return after one iteration of the loop.
390.Sp 594.Sp
391A flags value of \f(CW\*(C`EVLOOP_ONESHOT\*(C'\fR will look for new events (waiting if 595A flags value of \f(CW\*(C`EVLOOP_ONESHOT\*(C'\fR will look for new events (waiting if
392neccessary) and will handle those and any outstanding ones. It will block 596neccessary) and will handle those and any outstanding ones. It will block
393your process until at least one new event arrives, and will return after 597your process until at least one new event arrives, and will return after
394one iteration of the loop. 598one iteration of the loop. This is useful if you are waiting for some
599external event in conjunction with something not expressible using other
600libev watchers. However, a pair of \f(CW\*(C`ev_prepare\*(C'\fR/\f(CW\*(C`ev_check\*(C'\fR watchers is
601usually a better approach for this kind of thing.
395.Sp 602.Sp
396This flags value could be used to implement alternative looping
397constructs, but the \f(CW\*(C`prepare\*(C'\fR and \f(CW\*(C`check\*(C'\fR watchers provide a better and
398more generic mechanism.
399.Sp
400Here are the gory details of what ev_loop does: 603Here are the gory details of what \f(CW\*(C`ev_loop\*(C'\fR does:
401.Sp 604.Sp
402.Vb 15 605.Vb 18
403\& 1. If there are no active watchers (reference count is zero), return. 606\& * If there are no active watchers (reference count is zero), return.
404\& 2. Queue and immediately call all prepare watchers. 607\& - Queue prepare watchers and then call all outstanding watchers.
405\& 3. If we have been forked, recreate the kernel state. 608\& - If we have been forked, recreate the kernel state.
406\& 4. Update the kernel state with all outstanding changes. 609\& - Update the kernel state with all outstanding changes.
407\& 5. Update the "event loop time". 610\& - Update the "event loop time".
408\& 6. Calculate for how long to block. 611\& - Calculate for how long to block.
409\& 7. Block the process, waiting for events. 612\& - Block the process, waiting for any events.
613\& - Queue all outstanding I/O (fd) events.
410\& 8. Update the "event loop time" and do time jump handling. 614\& - Update the "event loop time" and do time jump handling.
411\& 9. Queue all outstanding timers. 615\& - Queue all outstanding timers.
412\& 10. Queue all outstanding periodics. 616\& - Queue all outstanding periodics.
413\& 11. If no events are pending now, queue all idle watchers. 617\& - If no events are pending now, queue all idle watchers.
414\& 12. Queue all check watchers. 618\& - Queue all check watchers.
415\& 13. Call all queued watchers in reverse order (i.e. check watchers first). 619\& - Call all queued watchers in reverse order (i.e. check watchers first).
620\& Signals and child watchers are implemented as I/O watchers, and will
621\& be handled here by queueing them when their watcher gets executed.
416\& 14. If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK 622\& - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
417\& was used, return, otherwise continue with step #1. 623\& were used, return, otherwise continue with step *.
624.Ve
625.Sp
626Example: Queue some jobs and then loop until no events are outsanding
627anymore.
628.Sp
629.Vb 4
630\& ... queue jobs here, make sure they register event watchers as long
631\& ... as they still have work to do (even an idle watcher will do..)
632\& ev_loop (my_loop, 0);
633\& ... jobs done. yeah!
418.Ve 634.Ve
419.IP "ev_unloop (loop, how)" 4 635.IP "ev_unloop (loop, how)" 4
420.IX Item "ev_unloop (loop, how)" 636.IX Item "ev_unloop (loop, how)"
421Can be used to make a call to \f(CW\*(C`ev_loop\*(C'\fR return early (but only after it 637Can be used to make a call to \f(CW\*(C`ev_loop\*(C'\fR return early (but only after it
422has processed all outstanding events). The \f(CW\*(C`how\*(C'\fR argument must be either 638has processed all outstanding events). The \f(CW\*(C`how\*(C'\fR argument must be either
… …
436example, libev itself uses this for its internal signal pipe: It is not 652example, libev itself uses this for its internal signal pipe: It is not
437visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if 653visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if
438no event watchers registered by it are active. It is also an excellent 654no event watchers registered by it are active. It is also an excellent
439way to do this for generic recurring timers or from within third-party 655way to do this for generic recurring timers or from within third-party
440libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR. 656libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR.
657.Sp
658Example: Create a signal watcher, but keep it from keeping \f(CW\*(C`ev_loop\*(C'\fR
659running when nothing else is active.
660.Sp
661.Vb 4
662\& struct ev_signal exitsig;
663\& ev_signal_init (&exitsig, sig_cb, SIGINT);
664\& ev_signal_start (loop, &exitsig);
665\& evf_unref (loop);
666.Ve
667.Sp
668Example: For some weird reason, unregister the above signal handler again.
669.Sp
670.Vb 2
671\& ev_ref (loop);
672\& ev_signal_stop (loop, &exitsig);
673.Ve
441.SH "ANATOMY OF A WATCHER" 674.SH "ANATOMY OF A WATCHER"
442.IX Header "ANATOMY OF A WATCHER" 675.IX Header "ANATOMY OF A WATCHER"
443A watcher is a structure that you create and register to record your 676A watcher is a structure that you create and register to record your
444interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to 677interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to
445become readable, you would create an \f(CW\*(C`ev_io\*(C'\fR watcher for that: 678become readable, you would create an \f(CW\*(C`ev_io\*(C'\fR watcher for that:
… …
481*)\*(C'\fR), and you can stop watching for events at any time by calling the 714*)\*(C'\fR), and you can stop watching for events at any time by calling the
482corresponding stop function (\f(CW\*(C`ev_<type>_stop (loop, watcher *)\*(C'\fR. 715corresponding stop function (\f(CW\*(C`ev_<type>_stop (loop, watcher *)\*(C'\fR.
483.PP 716.PP
484As long as your watcher is active (has been started but not stopped) you 717As long as your watcher is active (has been started but not stopped) you
485must not touch the values stored in it. Most specifically you must never 718must not touch the values stored in it. Most specifically you must never
486reinitialise it or call its set method. 719reinitialise it or call its \f(CW\*(C`set\*(C'\fR macro.
487.PP
488You can check whether an event is active by calling the \f(CW\*(C`ev_is_active
489(watcher *)\*(C'\fR macro. To see whether an event is outstanding (but the
490callback for it has not been called yet) you can use the \f(CW\*(C`ev_is_pending
491(watcher *)\*(C'\fR macro.
492.PP 720.PP
493Each and every callback receives the event loop pointer as first, the 721Each and every callback receives the event loop pointer as first, the
494registered watcher structure as second, and a bitset of received events as 722registered watcher structure as second, and a bitset of received events as
495third argument. 723third argument.
496.PP 724.PP
… …
521The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread. 749The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread.
522.ie n .IP """EV_CHILD""" 4 750.ie n .IP """EV_CHILD""" 4
523.el .IP "\f(CWEV_CHILD\fR" 4 751.el .IP "\f(CWEV_CHILD\fR" 4
524.IX Item "EV_CHILD" 752.IX Item "EV_CHILD"
525The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change. 753The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change.
754.ie n .IP """EV_STAT""" 4
755.el .IP "\f(CWEV_STAT\fR" 4
756.IX Item "EV_STAT"
757The path specified in the \f(CW\*(C`ev_stat\*(C'\fR watcher changed its attributes somehow.
526.ie n .IP """EV_IDLE""" 4 758.ie n .IP """EV_IDLE""" 4
527.el .IP "\f(CWEV_IDLE\fR" 4 759.el .IP "\f(CWEV_IDLE\fR" 4
528.IX Item "EV_IDLE" 760.IX Item "EV_IDLE"
529The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do. 761The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do.
530.ie n .IP """EV_PREPARE""" 4 762.ie n .IP """EV_PREPARE""" 4
… …
540\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any 772\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any
541received events. Callbacks of both watcher types can start and stop as 773received events. Callbacks of both watcher types can start and stop as
542many watchers as they want, and all of them will be taken into account 774many watchers as they want, and all of them will be taken into account
543(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep 775(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep
544\&\f(CW\*(C`ev_loop\*(C'\fR from blocking). 776\&\f(CW\*(C`ev_loop\*(C'\fR from blocking).
777.ie n .IP """EV_EMBED""" 4
778.el .IP "\f(CWEV_EMBED\fR" 4
779.IX Item "EV_EMBED"
780The embedded event loop specified in the \f(CW\*(C`ev_embed\*(C'\fR watcher needs attention.
781.ie n .IP """EV_FORK""" 4
782.el .IP "\f(CWEV_FORK\fR" 4
783.IX Item "EV_FORK"
784The event loop has been resumed in the child process after fork (see
785\&\f(CW\*(C`ev_fork\*(C'\fR).
545.ie n .IP """EV_ERROR""" 4 786.ie n .IP """EV_ERROR""" 4
546.el .IP "\f(CWEV_ERROR\fR" 4 787.el .IP "\f(CWEV_ERROR\fR" 4
547.IX Item "EV_ERROR" 788.IX Item "EV_ERROR"
548An unspecified error has occured, the watcher has been stopped. This might 789An unspecified error has occured, the watcher has been stopped. This might
549happen because the watcher could not be properly started because libev 790happen because the watcher could not be properly started because libev
… …
554Libev will usually signal a few \*(L"dummy\*(R" events together with an error, 795Libev will usually signal a few \*(L"dummy\*(R" events together with an error,
555for example it might indicate that a fd is readable or writable, and if 796for example it might indicate that a fd is readable or writable, and if
556your callbacks is well-written it can just attempt the operation and cope 797your callbacks is well-written it can just attempt the operation and cope
557with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded 798with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded
558programs, though, so beware. 799programs, though, so beware.
800.Sh "\s-1GENERIC\s0 \s-1WATCHER\s0 \s-1FUNCTIONS\s0"
801.IX Subsection "GENERIC WATCHER FUNCTIONS"
802In the following description, \f(CW\*(C`TYPE\*(C'\fR stands for the watcher type,
803e.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.
804.ie n .IP """ev_init"" (ev_TYPE *watcher, callback)" 4
805.el .IP "\f(CWev_init\fR (ev_TYPE *watcher, callback)" 4
806.IX Item "ev_init (ev_TYPE *watcher, callback)"
807This macro initialises the generic portion of a watcher. The contents
808of the watcher object can be arbitrary (so \f(CW\*(C`malloc\*(C'\fR will do). Only
809the generic parts of the watcher are initialised, you \fIneed\fR to call
810the type-specific \f(CW\*(C`ev_TYPE_set\*(C'\fR macro afterwards to initialise the
811type-specific parts. For each type there is also a \f(CW\*(C`ev_TYPE_init\*(C'\fR macro
812which rolls both calls into one.
813.Sp
814You can reinitialise a watcher at any time as long as it has been stopped
815(or never started) and there are no pending events outstanding.
816.Sp
817The callback is always of type \f(CW\*(C`void (*)(ev_loop *loop, ev_TYPE *watcher,
818int revents)\*(C'\fR.
819.ie n .IP """ev_TYPE_set"" (ev_TYPE *, [args])" 4
820.el .IP "\f(CWev_TYPE_set\fR (ev_TYPE *, [args])" 4
821.IX Item "ev_TYPE_set (ev_TYPE *, [args])"
822This macro initialises the type-specific parts of a watcher. You need to
823call \f(CW\*(C`ev_init\*(C'\fR at least once before you call this macro, but you can
824call \f(CW\*(C`ev_TYPE_set\*(C'\fR any number of times. You must not, however, call this
825macro on a watcher that is active (it can be pending, however, which is a
826difference to the \f(CW\*(C`ev_init\*(C'\fR macro).
827.Sp
828Although some watcher types do not have type-specific arguments
829(e.g. \f(CW\*(C`ev_prepare\*(C'\fR) you still need to call its \f(CW\*(C`set\*(C'\fR macro.
830.ie n .IP """ev_TYPE_init"" (ev_TYPE *watcher, callback, [args])" 4
831.el .IP "\f(CWev_TYPE_init\fR (ev_TYPE *watcher, callback, [args])" 4
832.IX Item "ev_TYPE_init (ev_TYPE *watcher, callback, [args])"
833This convinience macro rolls both \f(CW\*(C`ev_init\*(C'\fR and \f(CW\*(C`ev_TYPE_set\*(C'\fR macro
834calls into a single call. This is the most convinient method to initialise
835a watcher. The same limitations apply, of course.
836.ie n .IP """ev_TYPE_start"" (loop *, ev_TYPE *watcher)" 4
837.el .IP "\f(CWev_TYPE_start\fR (loop *, ev_TYPE *watcher)" 4
838.IX Item "ev_TYPE_start (loop *, ev_TYPE *watcher)"
839Starts (activates) the given watcher. Only active watchers will receive
840events. If the watcher is already active nothing will happen.
841.ie n .IP """ev_TYPE_stop"" (loop *, ev_TYPE *watcher)" 4
842.el .IP "\f(CWev_TYPE_stop\fR (loop *, ev_TYPE *watcher)" 4
843.IX Item "ev_TYPE_stop (loop *, ev_TYPE *watcher)"
844Stops the given watcher again (if active) and clears the pending
845status. It is possible that stopped watchers are pending (for example,
846non-repeating timers are being stopped when they become pending), but
847\&\f(CW\*(C`ev_TYPE_stop\*(C'\fR ensures that the watcher is neither active nor pending. If
848you want to free or reuse the memory used by the watcher it is therefore a
849good idea to always call its \f(CW\*(C`ev_TYPE_stop\*(C'\fR function.
850.IP "bool ev_is_active (ev_TYPE *watcher)" 4
851.IX Item "bool ev_is_active (ev_TYPE *watcher)"
852Returns a true value iff the watcher is active (i.e. it has been started
853and not yet been stopped). As long as a watcher is active you must not modify
854it.
855.IP "bool ev_is_pending (ev_TYPE *watcher)" 4
856.IX Item "bool ev_is_pending (ev_TYPE *watcher)"
857Returns a true value iff the watcher is pending, (i.e. it has outstanding
858events but its callback has not yet been invoked). As long as a watcher
859is pending (but not active) you must not call an init function on it (but
860\&\f(CW\*(C`ev_TYPE_set\*(C'\fR is safe) and you must make sure the watcher is available to
861libev (e.g. you cnanot \f(CW\*(C`free ()\*(C'\fR it).
862.IP "callback ev_cb (ev_TYPE *watcher)" 4
863.IX Item "callback ev_cb (ev_TYPE *watcher)"
864Returns the callback currently set on the watcher.
865.IP "ev_cb_set (ev_TYPE *watcher, callback)" 4
866.IX Item "ev_cb_set (ev_TYPE *watcher, callback)"
867Change the callback. You can change the callback at virtually any time
868(modulo threads).
559.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0" 869.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0"
560.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER" 870.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER"
561Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change 871Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change
562and read at any time, libev will completely ignore it. This can be used 872and read at any time, libev will completely ignore it. This can be used
563to associate arbitrary data with your watcher. If you need more data and 873to associate arbitrary data with your watcher. If you need more data and
… …
584\& struct my_io *w = (struct my_io *)w_; 894\& struct my_io *w = (struct my_io *)w_;
585\& ... 895\& ...
586\& } 896\& }
587.Ve 897.Ve
588.PP 898.PP
589More interesting and less C\-conformant ways of catsing your callback type 899More interesting and less C\-conformant ways of casting your callback type
590have been omitted.... 900instead have been omitted.
901.PP
902Another common scenario is having some data structure with multiple
903watchers:
904.PP
905.Vb 6
906\& struct my_biggy
907\& {
908\& int some_data;
909\& ev_timer t1;
910\& ev_timer t2;
911\& }
912.Ve
913.PP
914In this case getting the pointer to \f(CW\*(C`my_biggy\*(C'\fR is a bit more complicated,
915you need to use \f(CW\*(C`offsetof\*(C'\fR:
916.PP
917.Vb 1
918\& #include <stddef.h>
919.Ve
920.PP
921.Vb 6
922\& static void
923\& t1_cb (EV_P_ struct ev_timer *w, int revents)
924\& {
925\& struct my_biggy big = (struct my_biggy *
926\& (((char *)w) - offsetof (struct my_biggy, t1));
927\& }
928.Ve
929.PP
930.Vb 6
931\& static void
932\& t2_cb (EV_P_ struct ev_timer *w, int revents)
933\& {
934\& struct my_biggy big = (struct my_biggy *
935\& (((char *)w) - offsetof (struct my_biggy, t2));
936\& }
937.Ve
591.SH "WATCHER TYPES" 938.SH "WATCHER TYPES"
592.IX Header "WATCHER TYPES" 939.IX Header "WATCHER TYPES"
593This section describes each watcher in detail, but will not repeat 940This section describes each watcher in detail, but will not repeat
594information given in the last section. 941information given in the last section. Any initialisation/set macros,
942functions and members specific to the watcher type are explained.
943.PP
944Members are additionally marked with either \fI[read\-only]\fR, meaning that,
945while the watcher is active, you can look at the member and expect some
946sensible content, but you must not modify it (you can modify it while the
947watcher is stopped to your hearts content), or \fI[read\-write]\fR, which
948means you can expect it to have some sensible content while the watcher
949is active, but you can also modify it. Modifying it may not do something
950sensible or take immediate effect (or do anything at all), but libev will
951not crash or malfunction in any way.
595.ie n .Sh """ev_io"" \- is this file descriptor readable or writable" 952.ie n .Sh """ev_io"" \- is this file descriptor readable or writable?"
596.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable" 953.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable?"
597.IX Subsection "ev_io - is this file descriptor readable or writable" 954.IX Subsection "ev_io - is this file descriptor readable or writable?"
598I/O watchers check whether a file descriptor is readable or writable 955I/O watchers check whether a file descriptor is readable or writable
599in each iteration of the event loop (This behaviour is called 956in each iteration of the event loop, or, more precisely, when reading
600level-triggering because you keep receiving events as long as the 957would not block the process and writing would at least be able to write
601condition persists. Remember you can stop the watcher if you don't want to 958some data. This behaviour is called level-triggering because you keep
602act on the event and neither want to receive future events). 959receiving events as long as the condition persists. Remember you can stop
960the watcher if you don't want to act on the event and neither want to
961receive future events.
603.PP 962.PP
604In general you can register as many read and/or write event watchers per 963In general you can register as many read and/or write event watchers per
605fd as you want (as long as you don't confuse yourself). Setting all file 964fd as you want (as long as you don't confuse yourself). Setting all file
606descriptors to non-blocking mode is also usually a good idea (but not 965descriptors to non-blocking mode is also usually a good idea (but not
607required if you know what you are doing). 966required if you know what you are doing).
608.PP 967.PP
609You have to be careful with dup'ed file descriptors, though. Some backends 968You have to be careful with dup'ed file descriptors, though. Some backends
610(the linux epoll backend is a notable example) cannot handle dup'ed file 969(the linux epoll backend is a notable example) cannot handle dup'ed file
611descriptors correctly if you register interest in two or more fds pointing 970descriptors correctly if you register interest in two or more fds pointing
612to the same underlying file/socket etc. description (that is, they share 971to the same underlying file/socket/etc. description (that is, they share
613the same underlying \*(L"file open\*(R"). 972the same underlying \*(L"file open\*(R").
614.PP 973.PP
615If you must do this, then force the use of a known-to-be-good backend 974If you must do this, then force the use of a known-to-be-good backend
616(at the time of this writing, this includes only \s-1EVMETHOD_SELECT\s0 and 975(at the time of this writing, this includes only \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and
617\&\s-1EVMETHOD_POLL\s0). 976\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR).
977.PP
978Another thing you have to watch out for is that it is quite easy to
979receive \*(L"spurious\*(R" readyness notifications, that is your callback might
980be called with \f(CW\*(C`EV_READ\*(C'\fR but a subsequent \f(CW\*(C`read\*(C'\fR(2) will actually block
981because there is no data. Not only are some backends known to create a
982lot of those (for example solaris ports), it is very easy to get into
983this situation even with a relatively standard program structure. Thus
984it is best to always use non-blocking I/O: An extra \f(CW\*(C`read\*(C'\fR(2) returning
985\&\f(CW\*(C`EAGAIN\*(C'\fR is far preferable to a program hanging until some data arrives.
986.PP
987If you cannot run the fd in non-blocking mode (for example you should not
988play around with an Xlib connection), then you have to seperately re-test
989wether a file descriptor is really ready with a known-to-be good interface
990such as poll (fortunately in our Xlib example, Xlib already does this on
991its own, so its quite safe to use).
618.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4 992.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4
619.IX Item "ev_io_init (ev_io *, callback, int fd, int events)" 993.IX Item "ev_io_init (ev_io *, callback, int fd, int events)"
620.PD 0 994.PD 0
621.IP "ev_io_set (ev_io *, int fd, int events)" 4 995.IP "ev_io_set (ev_io *, int fd, int events)" 4
622.IX Item "ev_io_set (ev_io *, int fd, int events)" 996.IX Item "ev_io_set (ev_io *, int fd, int events)"
623.PD 997.PD
624Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The fd is the file descriptor to rceeive 998Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The \f(CW\*(C`fd\*(C'\fR is the file descriptor to
625events 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 | 999rceeive events for and events is either \f(CW\*(C`EV_READ\*(C'\fR, \f(CW\*(C`EV_WRITE\*(C'\fR or
626EV_WRITE\*(C'\fR to receive the given events. 1000\&\f(CW\*(C`EV_READ | EV_WRITE\*(C'\fR to receive the given events.
1001.IP "int fd [read\-only]" 4
1002.IX Item "int fd [read-only]"
1003The file descriptor being watched.
1004.IP "int events [read\-only]" 4
1005.IX Item "int events [read-only]"
1006The events being watched.
1007.PP
1008Example: Call \f(CW\*(C`stdin_readable_cb\*(C'\fR when \s-1STDIN_FILENO\s0 has become, well
1009readable, but only once. Since it is likely line\-buffered, you could
1010attempt to read a whole line in the callback.
1011.PP
1012.Vb 6
1013\& static void
1014\& stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1015\& {
1016\& ev_io_stop (loop, w);
1017\& .. read from stdin here (or from w->fd) and haqndle any I/O errors
1018\& }
1019.Ve
1020.PP
1021.Vb 6
1022\& ...
1023\& struct ev_loop *loop = ev_default_init (0);
1024\& struct ev_io stdin_readable;
1025\& ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1026\& ev_io_start (loop, &stdin_readable);
1027\& ev_loop (loop, 0);
1028.Ve
627.ie n .Sh """ev_timer"" \- relative and optionally recurring timeouts" 1029.ie n .Sh """ev_timer"" \- relative and optionally repeating timeouts"
628.el .Sh "\f(CWev_timer\fP \- relative and optionally recurring timeouts" 1030.el .Sh "\f(CWev_timer\fP \- relative and optionally repeating timeouts"
629.IX Subsection "ev_timer - relative and optionally recurring timeouts" 1031.IX Subsection "ev_timer - relative and optionally repeating timeouts"
630Timer watchers are simple relative timers that generate an event after a 1032Timer watchers are simple relative timers that generate an event after a
631given time, and optionally repeating in regular intervals after that. 1033given time, and optionally repeating in regular intervals after that.
632.PP 1034.PP
633The timers are based on real time, that is, if you register an event that 1035The timers are based on real time, that is, if you register an event that
634times out after an hour and you reset your system clock to last years 1036times out after an hour and you reset your system clock to last years
… …
674.Sp 1076.Sp
675If the timer is repeating, either start it if necessary (with the repeat 1077If the timer is repeating, either start it if necessary (with the repeat
676value), or reset the running timer to the repeat value. 1078value), or reset the running timer to the repeat value.
677.Sp 1079.Sp
678This sounds a bit complicated, but here is a useful and typical 1080This sounds a bit complicated, but here is a useful and typical
679example: Imagine you have a tcp connection and you want a so-called idle 1081example: Imagine you have a tcp connection and you want a so-called
680timeout, that is, you want to be called when there have been, say, 60 1082idle timeout, that is, you want to be called when there have been,
681seconds of inactivity on the socket. The easiest way to do this is to 1083say, 60 seconds of inactivity on the socket. The easiest way to do
682configure an \f(CW\*(C`ev_timer\*(C'\fR with after=repeat=60 and calling ev_timer_again each 1084this 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
683time you successfully read or write some data. If you go into an idle 1085\&\f(CW\*(C`ev_timer_again\*(C'\fR each time you successfully read or write some data. If
684state where you do not expect data to travel on the socket, you can stop 1086you go into an idle state where you do not expect data to travel on the
685the timer, and again will automatically restart it if need be. 1087socket, you can stop the timer, and again will automatically restart it if
1088need be.
1089.Sp
1090You can also ignore the \f(CW\*(C`after\*(C'\fR value and \f(CW\*(C`ev_timer_start\*(C'\fR altogether
1091and only ever use the \f(CW\*(C`repeat\*(C'\fR value:
1092.Sp
1093.Vb 8
1094\& ev_timer_init (timer, callback, 0., 5.);
1095\& ev_timer_again (loop, timer);
1096\& ...
1097\& timer->again = 17.;
1098\& ev_timer_again (loop, timer);
1099\& ...
1100\& timer->again = 10.;
1101\& ev_timer_again (loop, timer);
1102.Ve
1103.Sp
1104This is more efficient then stopping/starting the timer eahc time you want
1105to modify its timeout value.
1106.IP "ev_tstamp repeat [read\-write]" 4
1107.IX Item "ev_tstamp repeat [read-write]"
1108The current \f(CW\*(C`repeat\*(C'\fR value. Will be used each time the watcher times out
1109or \f(CW\*(C`ev_timer_again\*(C'\fR is called and determines the next timeout (if any),
1110which is also when any modifications are taken into account.
1111.PP
1112Example: Create a timer that fires after 60 seconds.
1113.PP
1114.Vb 5
1115\& static void
1116\& one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1117\& {
1118\& .. one minute over, w is actually stopped right here
1119\& }
1120.Ve
1121.PP
1122.Vb 3
1123\& struct ev_timer mytimer;
1124\& ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1125\& ev_timer_start (loop, &mytimer);
1126.Ve
1127.PP
1128Example: Create a timeout timer that times out after 10 seconds of
1129inactivity.
1130.PP
1131.Vb 5
1132\& static void
1133\& timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1134\& {
1135\& .. ten seconds without any activity
1136\& }
1137.Ve
1138.PP
1139.Vb 4
1140\& struct ev_timer mytimer;
1141\& ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1142\& ev_timer_again (&mytimer); /* start timer */
1143\& ev_loop (loop, 0);
1144.Ve
1145.PP
1146.Vb 3
1147\& // and in some piece of code that gets executed on any "activity":
1148\& // reset the timeout to start ticking again at 10 seconds
1149\& ev_timer_again (&mytimer);
1150.Ve
686.ie n .Sh """ev_periodic"" \- to cron or not to cron" 1151.ie n .Sh """ev_periodic"" \- to cron or not to cron?"
687.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron" 1152.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron?"
688.IX Subsection "ev_periodic - to cron or not to cron" 1153.IX Subsection "ev_periodic - to cron or not to cron?"
689Periodic watchers are also timers of a kind, but they are very versatile 1154Periodic watchers are also timers of a kind, but they are very versatile
690(and unfortunately a bit complex). 1155(and unfortunately a bit complex).
691.PP 1156.PP
692Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time) 1157Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time)
693but on wallclock time (absolute time). You can tell a periodic watcher 1158but on wallclock time (absolute time). You can tell a periodic watcher
694to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a 1159to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a
695periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now () 1160periodic watcher to trigger in 10 seconds (by specifiying e.g. \f(CW\*(C`ev_now ()
696+ 10.>) and then reset your system clock to the last year, then it will 1161+ 10.\*(C'\fR) and then reset your system clock to the last year, then it will
697take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger 1162take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger
698roughly 10 seconds later and of course not if you reset your system time 1163roughly 10 seconds later and of course not if you reset your system time
699again). 1164again).
700.PP 1165.PP
701They can also be used to implement vastly more complex timers, such as 1166They can also be used to implement vastly more complex timers, such as
… …
782.IX Item "ev_periodic_again (loop, ev_periodic *)" 1247.IX Item "ev_periodic_again (loop, ev_periodic *)"
783Simply stops and restarts the periodic watcher again. This is only useful 1248Simply stops and restarts the periodic watcher again. This is only useful
784when you changed some parameters or the reschedule callback would return 1249when you changed some parameters or the reschedule callback would return
785a different time than the last time it was called (e.g. in a crond like 1250a different time than the last time it was called (e.g. in a crond like
786program when the crontabs have changed). 1251program when the crontabs have changed).
1252.IP "ev_tstamp interval [read\-write]" 4
1253.IX Item "ev_tstamp interval [read-write]"
1254The current interval value. Can be modified any time, but changes only
1255take effect when the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being
1256called.
1257.IP "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read\-write]" 4
1258.IX Item "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]"
1259The current reschedule callback, or \f(CW0\fR, if this functionality is
1260switched off. Can be changed any time, but changes only take effect when
1261the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being called.
1262.PP
1263Example: Call a callback every hour, or, more precisely, whenever the
1264system clock is divisible by 3600. The callback invocation times have
1265potentially a lot of jittering, but good long-term stability.
1266.PP
1267.Vb 5
1268\& static void
1269\& clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1270\& {
1271\& ... its now a full hour (UTC, or TAI or whatever your clock follows)
1272\& }
1273.Ve
1274.PP
1275.Vb 3
1276\& struct ev_periodic hourly_tick;
1277\& ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1278\& ev_periodic_start (loop, &hourly_tick);
1279.Ve
1280.PP
1281Example: The same as above, but use a reschedule callback to do it:
1282.PP
1283.Vb 1
1284\& #include <math.h>
1285.Ve
1286.PP
1287.Vb 5
1288\& static ev_tstamp
1289\& my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1290\& {
1291\& return fmod (now, 3600.) + 3600.;
1292\& }
1293.Ve
1294.PP
1295.Vb 1
1296\& ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1297.Ve
1298.PP
1299Example: Call a callback every hour, starting now:
1300.PP
1301.Vb 4
1302\& struct ev_periodic hourly_tick;
1303\& ev_periodic_init (&hourly_tick, clock_cb,
1304\& fmod (ev_now (loop), 3600.), 3600., 0);
1305\& ev_periodic_start (loop, &hourly_tick);
1306.Ve
787.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled" 1307.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled!"
788.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled" 1308.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled!"
789.IX Subsection "ev_signal - signal me when a signal gets signalled" 1309.IX Subsection "ev_signal - signal me when a signal gets signalled!"
790Signal watchers will trigger an event when the process receives a specific 1310Signal watchers will trigger an event when the process receives a specific
791signal one or more times. Even though signals are very asynchronous, libev 1311signal one or more times. Even though signals are very asynchronous, libev
792will try it's best to deliver signals synchronously, i.e. as part of the 1312will try it's best to deliver signals synchronously, i.e. as part of the
793normal event processing, like any other event. 1313normal event processing, like any other event.
794.PP 1314.PP
… …
804.IP "ev_signal_set (ev_signal *, int signum)" 4 1324.IP "ev_signal_set (ev_signal *, int signum)" 4
805.IX Item "ev_signal_set (ev_signal *, int signum)" 1325.IX Item "ev_signal_set (ev_signal *, int signum)"
806.PD 1326.PD
807Configures the watcher to trigger on the given signal number (usually one 1327Configures the watcher to trigger on the given signal number (usually one
808of the \f(CW\*(C`SIGxxx\*(C'\fR constants). 1328of the \f(CW\*(C`SIGxxx\*(C'\fR constants).
1329.IP "int signum [read\-only]" 4
1330.IX Item "int signum [read-only]"
1331The signal the watcher watches out for.
809.ie n .Sh """ev_child"" \- wait for pid status changes" 1332.ie n .Sh """ev_child"" \- watch out for process status changes"
810.el .Sh "\f(CWev_child\fP \- wait for pid status changes" 1333.el .Sh "\f(CWev_child\fP \- watch out for process status changes"
811.IX Subsection "ev_child - wait for pid status changes" 1334.IX Subsection "ev_child - watch out for process status changes"
812Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to 1335Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to
813some child status changes (most typically when a child of yours dies). 1336some child status changes (most typically when a child of yours dies).
814.IP "ev_child_init (ev_child *, callback, int pid)" 4 1337.IP "ev_child_init (ev_child *, callback, int pid)" 4
815.IX Item "ev_child_init (ev_child *, callback, int pid)" 1338.IX Item "ev_child_init (ev_child *, callback, int pid)"
816.PD 0 1339.PD 0
… …
821\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look 1344\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look
822at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see 1345at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see
823the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems 1346the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems
824\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the 1347\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the
825process causing the status change. 1348process causing the status change.
1349.IP "int pid [read\-only]" 4
1350.IX Item "int pid [read-only]"
1351The process id this watcher watches out for, or \f(CW0\fR, meaning any process id.
1352.IP "int rpid [read\-write]" 4
1353.IX Item "int rpid [read-write]"
1354The process id that detected a status change.
1355.IP "int rstatus [read\-write]" 4
1356.IX Item "int rstatus [read-write]"
1357The process exit/trace status caused by \f(CW\*(C`rpid\*(C'\fR (see your systems
1358\&\f(CW\*(C`waitpid\*(C'\fR and \f(CW\*(C`sys/wait.h\*(C'\fR documentation for details).
1359.PP
1360Example: Try to exit cleanly on \s-1SIGINT\s0 and \s-1SIGTERM\s0.
1361.PP
1362.Vb 5
1363\& static void
1364\& sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1365\& {
1366\& ev_unloop (loop, EVUNLOOP_ALL);
1367\& }
1368.Ve
1369.PP
1370.Vb 3
1371\& struct ev_signal signal_watcher;
1372\& ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1373\& ev_signal_start (loop, &sigint_cb);
1374.Ve
1375.ie n .Sh """ev_stat"" \- did the file attributes just change?"
1376.el .Sh "\f(CWev_stat\fP \- did the file attributes just change?"
1377.IX Subsection "ev_stat - did the file attributes just change?"
1378This watches a filesystem path for attribute changes. That is, it calls
1379\&\f(CW\*(C`stat\*(C'\fR regularly (or when the \s-1OS\s0 says it changed) and sees if it changed
1380compared to the last time, invoking the callback if it did.
1381.PP
1382The path does not need to exist: changing from \*(L"path exists\*(R" to \*(L"path does
1383not exist\*(R" is a status change like any other. The condition \*(L"path does
1384not exist\*(R" is signified by the \f(CW\*(C`st_nlink\*(C'\fR field being zero (which is
1385otherwise always forced to be at least one) and all the other fields of
1386the stat buffer having unspecified contents.
1387.PP
1388Since there is no standard to do this, the portable implementation simply
1389calls \f(CW\*(C`stat (2)\*(C'\fR regularly on the path to see if it changed somehow. You
1390can specify a recommended polling interval for this case. If you specify
1391a polling interval of \f(CW0\fR (highly recommended!) then a \fIsuitable,
1392unspecified default\fR value will be used (which you can expect to be around
1393five seconds, although this might change dynamically). Libev will also
1394impose a minimum interval which is currently around \f(CW0.1\fR, but thats
1395usually overkill.
1396.PP
1397This watcher type is not meant for massive numbers of stat watchers,
1398as even with OS-supported change notifications, this can be
1399resource\-intensive.
1400.PP
1401At the time of this writing, only the Linux inotify interface is
1402implemented (implementing kqueue support is left as an exercise for the
1403reader). Inotify will be used to give hints only and should not change the
1404semantics of \f(CW\*(C`ev_stat\*(C'\fR watchers, which means that libev sometimes needs
1405to fall back to regular polling again even with inotify, but changes are
1406usually detected immediately, and if the file exists there will be no
1407polling.
1408.IP "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)" 4
1409.IX Item "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)"
1410.PD 0
1411.IP "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)" 4
1412.IX Item "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)"
1413.PD
1414Configures the watcher to wait for status changes of the given
1415\&\f(CW\*(C`path\*(C'\fR. The \f(CW\*(C`interval\*(C'\fR is a hint on how quickly a change is expected to
1416be detected and should normally be specified as \f(CW0\fR to let libev choose
1417a suitable value. The memory pointed to by \f(CW\*(C`path\*(C'\fR must point to the same
1418path for as long as the watcher is active.
1419.Sp
1420The callback will be receive \f(CW\*(C`EV_STAT\*(C'\fR when a change was detected,
1421relative to the attributes at the time the watcher was started (or the
1422last change was detected).
1423.IP "ev_stat_stat (ev_stat *)" 4
1424.IX Item "ev_stat_stat (ev_stat *)"
1425Updates the stat buffer immediately with new values. If you change the
1426watched path in your callback, you could call this fucntion to avoid
1427detecting this change (while introducing a race condition). Can also be
1428useful simply to find out the new values.
1429.IP "ev_statdata attr [read\-only]" 4
1430.IX Item "ev_statdata attr [read-only]"
1431The most-recently detected attributes of the file. Although the type is of
1432\&\f(CW\*(C`ev_statdata\*(C'\fR, this is usually the (or one of the) \f(CW\*(C`struct stat\*(C'\fR types
1433suitable for your system. If the \f(CW\*(C`st_nlink\*(C'\fR member is \f(CW0\fR, then there
1434was some error while \f(CW\*(C`stat\*(C'\fRing the file.
1435.IP "ev_statdata prev [read\-only]" 4
1436.IX Item "ev_statdata prev [read-only]"
1437The previous attributes of the file. The callback gets invoked whenever
1438\&\f(CW\*(C`prev\*(C'\fR != \f(CW\*(C`attr\*(C'\fR.
1439.IP "ev_tstamp interval [read\-only]" 4
1440.IX Item "ev_tstamp interval [read-only]"
1441The specified interval.
1442.IP "const char *path [read\-only]" 4
1443.IX Item "const char *path [read-only]"
1444The filesystem path that is being watched.
1445.PP
1446Example: Watch \f(CW\*(C`/etc/passwd\*(C'\fR for attribute changes.
1447.PP
1448.Vb 15
1449\& static void
1450\& passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1451\& {
1452\& /* /etc/passwd changed in some way */
1453\& if (w->attr.st_nlink)
1454\& {
1455\& printf ("passwd current size %ld\en", (long)w->attr.st_size);
1456\& printf ("passwd current atime %ld\en", (long)w->attr.st_mtime);
1457\& printf ("passwd current mtime %ld\en", (long)w->attr.st_mtime);
1458\& }
1459\& else
1460\& /* you shalt not abuse printf for puts */
1461\& puts ("wow, /etc/passwd is not there, expect problems. "
1462\& "if this is windows, they already arrived\en");
1463\& }
1464.Ve
1465.PP
1466.Vb 2
1467\& ...
1468\& ev_stat passwd;
1469.Ve
1470.PP
1471.Vb 2
1472\& ev_stat_init (&passwd, passwd_cb, "/etc/passwd");
1473\& ev_stat_start (loop, &passwd);
1474.Ve
826.ie n .Sh """ev_idle"" \- when you've got nothing better to do" 1475.ie n .Sh """ev_idle"" \- when you've got nothing better to do..."
827.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do" 1476.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do..."
828.IX Subsection "ev_idle - when you've got nothing better to do" 1477.IX Subsection "ev_idle - when you've got nothing better to do..."
829Idle watchers trigger events when there are no other events are pending 1478Idle watchers trigger events when there are no other events are pending
830(prepare, check and other idle watchers do not count). That is, as long 1479(prepare, check and other idle watchers do not count). That is, as long
831as your process is busy handling sockets or timeouts (or even signals, 1480as your process is busy handling sockets or timeouts (or even signals,
832imagine) it will not be triggered. But when your process is idle all idle 1481imagine) it will not be triggered. But when your process is idle all idle
833watchers are being called again and again, once per event loop iteration \- 1482watchers are being called again and again, once per event loop iteration \-
… …
844.IP "ev_idle_init (ev_signal *, callback)" 4 1493.IP "ev_idle_init (ev_signal *, callback)" 4
845.IX Item "ev_idle_init (ev_signal *, callback)" 1494.IX Item "ev_idle_init (ev_signal *, callback)"
846Initialises and configures the idle watcher \- it has no parameters of any 1495Initialises and configures the idle watcher \- it has no parameters of any
847kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless, 1496kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless,
848believe me. 1497believe me.
1498.PP
1499Example: Dynamically allocate an \f(CW\*(C`ev_idle\*(C'\fR watcher, start it, and in the
1500callback, free it. Also, use no error checking, as usual.
1501.PP
1502.Vb 7
1503\& static void
1504\& idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1505\& {
1506\& free (w);
1507\& // now do something you wanted to do when the program has
1508\& // no longer asnything immediate to do.
1509\& }
1510.Ve
1511.PP
1512.Vb 3
1513\& struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1514\& ev_idle_init (idle_watcher, idle_cb);
1515\& ev_idle_start (loop, idle_cb);
1516.Ve
849.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop" 1517.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop!"
850.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop" 1518.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop!"
851.IX Subsection "ev_prepare and ev_check - customise your event loop" 1519.IX Subsection "ev_prepare and ev_check - customise your event loop!"
852Prepare and check watchers are usually (but not always) used in tandem: 1520Prepare and check watchers are usually (but not always) used in tandem:
853prepare watchers get invoked before the process blocks and check watchers 1521prepare watchers get invoked before the process blocks and check watchers
854afterwards. 1522afterwards.
855.PP 1523.PP
1524You \fImust not\fR call \f(CW\*(C`ev_loop\*(C'\fR or similar functions that enter
1525the current event loop from either \f(CW\*(C`ev_prepare\*(C'\fR or \f(CW\*(C`ev_check\*(C'\fR
1526watchers. Other loops than the current one are fine, however. The
1527rationale behind this is that you do not need to check for recursion in
1528those watchers, i.e. the sequence will always be \f(CW\*(C`ev_prepare\*(C'\fR, blocking,
1529\&\f(CW\*(C`ev_check\*(C'\fR so if you have one watcher of each kind they will always be
1530called in pairs bracketing the blocking call.
1531.PP
856Their main purpose is to integrate other event mechanisms into libev. This 1532Their main purpose is to integrate other event mechanisms into libev and
857could be used, for example, to track variable changes, implement your own 1533their use is somewhat advanced. This could be used, for example, to track
858watchers, integrate net-snmp or a coroutine library and lots more. 1534variable changes, implement your own watchers, integrate net-snmp or a
1535coroutine library and lots more. They are also occasionally useful if
1536you cache some data and want to flush it before blocking (for example,
1537in X programs you might want to do an \f(CW\*(C`XFlush ()\*(C'\fR in an \f(CW\*(C`ev_prepare\*(C'\fR
1538watcher).
859.PP 1539.PP
860This is done by examining in each prepare call which file descriptors need 1540This is done by examining in each prepare call which file descriptors need
861to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for 1541to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for
862them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries 1542them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries
863provide just this functionality). Then, in the check watcher you check for 1543provide just this functionality). Then, in the check watcher you check for
… …
881.IX Item "ev_check_init (ev_check *, callback)" 1561.IX Item "ev_check_init (ev_check *, callback)"
882.PD 1562.PD
883Initialises and configures the prepare or check watcher \- they have no 1563Initialises and configures the prepare or check watcher \- they have no
884parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR 1564parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR
885macros, but using them is utterly, utterly and completely pointless. 1565macros, but using them is utterly, utterly and completely pointless.
1566.PP
1567Example: To include a library such as adns, you would add \s-1IO\s0 watchers
1568and a timeout watcher in a prepare handler, as required by libadns, and
1569in a check watcher, destroy them and call into libadns. What follows is
1570pseudo-code only of course:
1571.PP
1572.Vb 2
1573\& static ev_io iow [nfd];
1574\& static ev_timer tw;
1575.Ve
1576.PP
1577.Vb 9
1578\& static void
1579\& io_cb (ev_loop *loop, ev_io *w, int revents)
1580\& {
1581\& // set the relevant poll flags
1582\& // could also call adns_processreadable etc. here
1583\& struct pollfd *fd = (struct pollfd *)w->data;
1584\& if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1585\& if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1586\& }
1587.Ve
1588.PP
1589.Vb 7
1590\& // create io watchers for each fd and a timer before blocking
1591\& static void
1592\& adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1593\& {
1594\& int timeout = 3600000;truct pollfd fds [nfd];
1595\& // actual code will need to loop here and realloc etc.
1596\& adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1597.Ve
1598.PP
1599.Vb 3
1600\& /* the callback is illegal, but won't be called as we stop during check */
1601\& ev_timer_init (&tw, 0, timeout * 1e-3);
1602\& ev_timer_start (loop, &tw);
1603.Ve
1604.PP
1605.Vb 6
1606\& // create on ev_io per pollfd
1607\& for (int i = 0; i < nfd; ++i)
1608\& {
1609\& ev_io_init (iow + i, io_cb, fds [i].fd,
1610\& ((fds [i].events & POLLIN ? EV_READ : 0)
1611\& | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1612.Ve
1613.PP
1614.Vb 5
1615\& fds [i].revents = 0;
1616\& iow [i].data = fds + i;
1617\& ev_io_start (loop, iow + i);
1618\& }
1619\& }
1620.Ve
1621.PP
1622.Vb 5
1623\& // stop all watchers after blocking
1624\& static void
1625\& adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1626\& {
1627\& ev_timer_stop (loop, &tw);
1628.Ve
1629.PP
1630.Vb 2
1631\& for (int i = 0; i < nfd; ++i)
1632\& ev_io_stop (loop, iow + i);
1633.Ve
1634.PP
1635.Vb 2
1636\& adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1637\& }
1638.Ve
1639.ie n .Sh """ev_embed"" \- when one backend isn't enough..."
1640.el .Sh "\f(CWev_embed\fP \- when one backend isn't enough..."
1641.IX Subsection "ev_embed - when one backend isn't enough..."
1642This is a rather advanced watcher type that lets you embed one event loop
1643into another (currently only \f(CW\*(C`ev_io\*(C'\fR events are supported in the embedded
1644loop, other types of watchers might be handled in a delayed or incorrect
1645fashion and must not be used).
1646.PP
1647There are primarily two reasons you would want that: work around bugs and
1648prioritise I/O.
1649.PP
1650As an example for a bug workaround, the kqueue backend might only support
1651sockets on some platform, so it is unusable as generic backend, but you
1652still want to make use of it because you have many sockets and it scales
1653so nicely. In this case, you would create a kqueue-based loop and embed it
1654into your default loop (which might use e.g. poll). Overall operation will
1655be a bit slower because first libev has to poll and then call kevent, but
1656at least you can use both at what they are best.
1657.PP
1658As for prioritising I/O: rarely you have the case where some fds have
1659to be watched and handled very quickly (with low latency), and even
1660priorities and idle watchers might have too much overhead. In this case
1661you would put all the high priority stuff in one loop and all the rest in
1662a second one, and embed the second one in the first.
1663.PP
1664As long as the watcher is active, the callback will be invoked every time
1665there might be events pending in the embedded loop. The callback must then
1666call \f(CW\*(C`ev_embed_sweep (mainloop, watcher)\*(C'\fR to make a single sweep and invoke
1667their callbacks (you could also start an idle watcher to give the embedded
1668loop strictly lower priority for example). You can also set the callback
1669to \f(CW0\fR, in which case the embed watcher will automatically execute the
1670embedded loop sweep.
1671.PP
1672As long as the watcher is started it will automatically handle events. The
1673callback will be invoked whenever some events have been handled. You can
1674set the callback to \f(CW0\fR to avoid having to specify one if you are not
1675interested in that.
1676.PP
1677Also, there have not currently been made special provisions for forking:
1678when you fork, you not only have to call \f(CW\*(C`ev_loop_fork\*(C'\fR on both loops,
1679but you will also have to stop and restart any \f(CW\*(C`ev_embed\*(C'\fR watchers
1680yourself.
1681.PP
1682Unfortunately, not all backends are embeddable, only the ones returned by
1683\&\f(CW\*(C`ev_embeddable_backends\*(C'\fR are, which, unfortunately, does not include any
1684portable one.
1685.PP
1686So when you want to use this feature you will always have to be prepared
1687that you cannot get an embeddable loop. The recommended way to get around
1688this is to have a separate variables for your embeddable loop, try to
1689create it, and if that fails, use the normal loop for everything:
1690.PP
1691.Vb 3
1692\& struct ev_loop *loop_hi = ev_default_init (0);
1693\& struct ev_loop *loop_lo = 0;
1694\& struct ev_embed embed;
1695.Ve
1696.PP
1697.Vb 5
1698\& // see if there is a chance of getting one that works
1699\& // (remember that a flags value of 0 means autodetection)
1700\& loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1701\& ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1702\& : 0;
1703.Ve
1704.PP
1705.Vb 8
1706\& // if we got one, then embed it, otherwise default to loop_hi
1707\& if (loop_lo)
1708\& {
1709\& ev_embed_init (&embed, 0, loop_lo);
1710\& ev_embed_start (loop_hi, &embed);
1711\& }
1712\& else
1713\& loop_lo = loop_hi;
1714.Ve
1715.IP "ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)" 4
1716.IX Item "ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)"
1717.PD 0
1718.IP "ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)" 4
1719.IX Item "ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)"
1720.PD
1721Configures the watcher to embed the given loop, which must be
1722embeddable. If the callback is \f(CW0\fR, then \f(CW\*(C`ev_embed_sweep\*(C'\fR will be
1723invoked automatically, otherwise it is the responsibility of the callback
1724to invoke it (it will continue to be called until the sweep has been done,
1725if you do not want thta, you need to temporarily stop the embed watcher).
1726.IP "ev_embed_sweep (loop, ev_embed *)" 4
1727.IX Item "ev_embed_sweep (loop, ev_embed *)"
1728Make a single, non-blocking sweep over the embedded loop. This works
1729similarly to \f(CW\*(C`ev_loop (embedded_loop, EVLOOP_NONBLOCK)\*(C'\fR, but in the most
1730apropriate way for embedded loops.
1731.IP "struct ev_loop *loop [read\-only]" 4
1732.IX Item "struct ev_loop *loop [read-only]"
1733The embedded event loop.
1734.ie n .Sh """ev_fork"" \- the audacity to resume the event loop after a fork"
1735.el .Sh "\f(CWev_fork\fP \- the audacity to resume the event loop after a fork"
1736.IX Subsection "ev_fork - the audacity to resume the event loop after a fork"
1737Fork watchers are called when a \f(CW\*(C`fork ()\*(C'\fR was detected (usually because
1738whoever is a good citizen cared to tell libev about it by calling
1739\&\f(CW\*(C`ev_default_fork\*(C'\fR or \f(CW\*(C`ev_loop_fork\*(C'\fR). The invocation is done before the
1740event loop blocks next and before \f(CW\*(C`ev_check\*(C'\fR watchers are being called,
1741and only in the child after the fork. If whoever good citizen calling
1742\&\f(CW\*(C`ev_default_fork\*(C'\fR cheats and calls it in the wrong process, the fork
1743handlers will be invoked, too, of course.
1744.IP "ev_fork_init (ev_signal *, callback)" 4
1745.IX Item "ev_fork_init (ev_signal *, callback)"
1746Initialises and configures the fork watcher \- it has no parameters of any
1747kind. There is a \f(CW\*(C`ev_fork_set\*(C'\fR macro, but using it is utterly pointless,
1748believe me.
886.SH "OTHER FUNCTIONS" 1749.SH "OTHER FUNCTIONS"
887.IX Header "OTHER FUNCTIONS" 1750.IX Header "OTHER FUNCTIONS"
888There are some other functions of possible interest. Described. Here. Now. 1751There are some other functions of possible interest. Described. Here. Now.
889.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4 1752.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4
890.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 1753.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)"
… …
919.Ve 1782.Ve
920.Sp 1783.Sp
921.Vb 1 1784.Vb 1
922\& ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0); 1785\& ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
923.Ve 1786.Ve
924.IP "ev_feed_event (loop, watcher, int events)" 4 1787.IP "ev_feed_event (ev_loop *, watcher *, int revents)" 4
925.IX Item "ev_feed_event (loop, watcher, int events)" 1788.IX Item "ev_feed_event (ev_loop *, watcher *, int revents)"
926Feeds the given event set into the event loop, as if the specified event 1789Feeds the given event set into the event loop, as if the specified event
927had happened for the specified watcher (which must be a pointer to an 1790had happened for the specified watcher (which must be a pointer to an
928initialised but not necessarily started event watcher). 1791initialised but not necessarily started event watcher).
929.IP "ev_feed_fd_event (loop, int fd, int revents)" 4 1792.IP "ev_feed_fd_event (ev_loop *, int fd, int revents)" 4
930.IX Item "ev_feed_fd_event (loop, int fd, int revents)" 1793.IX Item "ev_feed_fd_event (ev_loop *, int fd, int revents)"
931Feed an event on the given fd, as if a file descriptor backend detected 1794Feed an event on the given fd, as if a file descriptor backend detected
932the given events it. 1795the given events it.
933.IP "ev_feed_signal_event (loop, int signum)" 4 1796.IP "ev_feed_signal_event (ev_loop *loop, int signum)" 4
934.IX Item "ev_feed_signal_event (loop, int signum)" 1797.IX Item "ev_feed_signal_event (ev_loop *loop, int signum)"
935Feed an event as if the given signal occured (loop must be the default loop!). 1798Feed an event as if the given signal occured (\f(CW\*(C`loop\*(C'\fR must be the default
1799loop!).
936.SH "LIBEVENT EMULATION" 1800.SH "LIBEVENT EMULATION"
937.IX Header "LIBEVENT EMULATION" 1801.IX Header "LIBEVENT EMULATION"
938Libev offers a compatibility emulation layer for libevent. It cannot 1802Libev offers a compatibility emulation layer for libevent. It cannot
939emulate the internals of libevent, so here are some usage hints: 1803emulate the internals of libevent, so here are some usage hints:
940.IP "* Use it by including <event.h>, as usual." 4 1804.IP "* Use it by including <event.h>, as usual." 4
… …
951.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4 1815.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4
952.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library." 1816.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library."
953.PD 1817.PD
954.SH "\*(C+ SUPPORT" 1818.SH "\*(C+ SUPPORT"
955.IX Header " SUPPORT" 1819.IX Header " SUPPORT"
956\&\s-1TBD\s0. 1820Libev comes with some simplistic wrapper classes for \*(C+ that mainly allow
1821you to use some convinience methods to start/stop watchers and also change
1822the callback model to a model using method callbacks on objects.
1823.PP
1824To use it,
1825.PP
1826.Vb 1
1827\& #include <ev++.h>
1828.Ve
1829.PP
1830(it is not installed by default). This automatically includes \fIev.h\fR
1831and puts all of its definitions (many of them macros) into the global
1832namespace. All \*(C+ specific things are put into the \f(CW\*(C`ev\*(C'\fR namespace.
1833.PP
1834It should support all the same embedding options as \fIev.h\fR, most notably
1835\&\f(CW\*(C`EV_MULTIPLICITY\*(C'\fR.
1836.PP
1837Here is a list of things available in the \f(CW\*(C`ev\*(C'\fR namespace:
1838.ie n .IP """ev::READ""\fR, \f(CW""ev::WRITE"" etc." 4
1839.el .IP "\f(CWev::READ\fR, \f(CWev::WRITE\fR etc." 4
1840.IX Item "ev::READ, ev::WRITE etc."
1841These are just enum values with the same values as the \f(CW\*(C`EV_READ\*(C'\fR etc.
1842macros from \fIev.h\fR.
1843.ie n .IP """ev::tstamp""\fR, \f(CW""ev::now""" 4
1844.el .IP "\f(CWev::tstamp\fR, \f(CWev::now\fR" 4
1845.IX Item "ev::tstamp, ev::now"
1846Aliases to the same types/functions as with the \f(CW\*(C`ev_\*(C'\fR prefix.
1847.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
1848.el .IP "\f(CWev::io\fR, \f(CWev::timer\fR, \f(CWev::periodic\fR, \f(CWev::idle\fR, \f(CWev::sig\fR etc." 4
1849.IX Item "ev::io, ev::timer, ev::periodic, ev::idle, ev::sig etc."
1850For each \f(CW\*(C`ev_TYPE\*(C'\fR watcher in \fIev.h\fR there is a corresponding class of
1851the same name in the \f(CW\*(C`ev\*(C'\fR namespace, with the exception of \f(CW\*(C`ev_signal\*(C'\fR
1852which is called \f(CW\*(C`ev::sig\*(C'\fR to avoid clashes with the \f(CW\*(C`signal\*(C'\fR macro
1853defines by many implementations.
1854.Sp
1855All of those classes have these methods:
1856.RS 4
1857.IP "ev::TYPE::TYPE (object *, object::method *)" 4
1858.IX Item "ev::TYPE::TYPE (object *, object::method *)"
1859.PD 0
1860.IP "ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)" 4
1861.IX Item "ev::TYPE::TYPE (object *, object::method *, struct ev_loop *)"
1862.IP "ev::TYPE::~TYPE" 4
1863.IX Item "ev::TYPE::~TYPE"
1864.PD
1865The constructor takes a pointer to an object and a method pointer to
1866the event handler callback to call in this class. The constructor calls
1867\&\f(CW\*(C`ev_init\*(C'\fR for you, which means you have to call the \f(CW\*(C`set\*(C'\fR method
1868before starting it. If you do not specify a loop then the constructor
1869automatically associates the default loop with this watcher.
1870.Sp
1871The destructor automatically stops the watcher if it is active.
1872.IP "w\->set (struct ev_loop *)" 4
1873.IX Item "w->set (struct ev_loop *)"
1874Associates a different \f(CW\*(C`struct ev_loop\*(C'\fR with this watcher. You can only
1875do this when the watcher is inactive (and not pending either).
1876.IP "w\->set ([args])" 4
1877.IX Item "w->set ([args])"
1878Basically the same as \f(CW\*(C`ev_TYPE_set\*(C'\fR, with the same args. Must be
1879called at least once. Unlike the C counterpart, an active watcher gets
1880automatically stopped and restarted.
1881.IP "w\->start ()" 4
1882.IX Item "w->start ()"
1883Starts the watcher. Note that there is no \f(CW\*(C`loop\*(C'\fR argument as the
1884constructor already takes the loop.
1885.IP "w\->stop ()" 4
1886.IX Item "w->stop ()"
1887Stops the watcher if it is active. Again, no \f(CW\*(C`loop\*(C'\fR argument.
1888.ie n .IP "w\->again () ""ev::timer""\fR, \f(CW""ev::periodic"" only" 4
1889.el .IP "w\->again () \f(CWev::timer\fR, \f(CWev::periodic\fR only" 4
1890.IX Item "w->again () ev::timer, ev::periodic only"
1891For \f(CW\*(C`ev::timer\*(C'\fR and \f(CW\*(C`ev::periodic\*(C'\fR, this invokes the corresponding
1892\&\f(CW\*(C`ev_TYPE_again\*(C'\fR function.
1893.ie n .IP "w\->sweep () ""ev::embed"" only" 4
1894.el .IP "w\->sweep () \f(CWev::embed\fR only" 4
1895.IX Item "w->sweep () ev::embed only"
1896Invokes \f(CW\*(C`ev_embed_sweep\*(C'\fR.
1897.ie n .IP "w\->update () ""ev::stat"" only" 4
1898.el .IP "w\->update () \f(CWev::stat\fR only" 4
1899.IX Item "w->update () ev::stat only"
1900Invokes \f(CW\*(C`ev_stat_stat\*(C'\fR.
1901.RE
1902.RS 4
1903.RE
1904.PP
1905Example: Define a class with an \s-1IO\s0 and idle watcher, start one of them in
1906the constructor.
1907.PP
1908.Vb 4
1909\& class myclass
1910\& {
1911\& ev_io io; void io_cb (ev::io &w, int revents);
1912\& ev_idle idle void idle_cb (ev::idle &w, int revents);
1913.Ve
1914.PP
1915.Vb 2
1916\& myclass ();
1917\& }
1918.Ve
1919.PP
1920.Vb 6
1921\& myclass::myclass (int fd)
1922\& : io (this, &myclass::io_cb),
1923\& idle (this, &myclass::idle_cb)
1924\& {
1925\& io.start (fd, ev::READ);
1926\& }
1927.Ve
1928.SH "MACRO MAGIC"
1929.IX Header "MACRO MAGIC"
1930Libev can be compiled with a variety of options, the most fundemantal is
1931\&\f(CW\*(C`EV_MULTIPLICITY\*(C'\fR. This option determines wether (most) functions and
1932callbacks have an initial \f(CW\*(C`struct ev_loop *\*(C'\fR argument.
1933.PP
1934To make it easier to write programs that cope with either variant, the
1935following macros are defined:
1936.ie n .IP """EV_A""\fR, \f(CW""EV_A_""" 4
1937.el .IP "\f(CWEV_A\fR, \f(CWEV_A_\fR" 4
1938.IX Item "EV_A, EV_A_"
1939This provides the loop \fIargument\fR for functions, if one is required (\*(L"ev
1940loop argument\*(R"). The \f(CW\*(C`EV_A\*(C'\fR form is used when this is the sole argument,
1941\&\f(CW\*(C`EV_A_\*(C'\fR is used when other arguments are following. Example:
1942.Sp
1943.Vb 3
1944\& ev_unref (EV_A);
1945\& ev_timer_add (EV_A_ watcher);
1946\& ev_loop (EV_A_ 0);
1947.Ve
1948.Sp
1949It assumes the variable \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR is in scope,
1950which is often provided by the following macro.
1951.ie n .IP """EV_P""\fR, \f(CW""EV_P_""" 4
1952.el .IP "\f(CWEV_P\fR, \f(CWEV_P_\fR" 4
1953.IX Item "EV_P, EV_P_"
1954This provides the loop \fIparameter\fR for functions, if one is required (\*(L"ev
1955loop parameter\*(R"). The \f(CW\*(C`EV_P\*(C'\fR form is used when this is the sole parameter,
1956\&\f(CW\*(C`EV_P_\*(C'\fR is used when other parameters are following. Example:
1957.Sp
1958.Vb 2
1959\& // this is how ev_unref is being declared
1960\& static void ev_unref (EV_P);
1961.Ve
1962.Sp
1963.Vb 2
1964\& // this is how you can declare your typical callback
1965\& static void cb (EV_P_ ev_timer *w, int revents)
1966.Ve
1967.Sp
1968It declares a parameter \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR, quite
1969suitable for use with \f(CW\*(C`EV_A\*(C'\fR.
1970.ie n .IP """EV_DEFAULT""\fR, \f(CW""EV_DEFAULT_""" 4
1971.el .IP "\f(CWEV_DEFAULT\fR, \f(CWEV_DEFAULT_\fR" 4
1972.IX Item "EV_DEFAULT, EV_DEFAULT_"
1973Similar to the other two macros, this gives you the value of the default
1974loop, if multiple loops are supported (\*(L"ev loop default\*(R").
1975.PP
1976Example: Declare and initialise a check watcher, working regardless of
1977wether multiple loops are supported or not.
1978.PP
1979.Vb 5
1980\& static void
1981\& check_cb (EV_P_ ev_timer *w, int revents)
1982\& {
1983\& ev_check_stop (EV_A_ w);
1984\& }
1985.Ve
1986.PP
1987.Vb 4
1988\& ev_check check;
1989\& ev_check_init (&check, check_cb);
1990\& ev_check_start (EV_DEFAULT_ &check);
1991\& ev_loop (EV_DEFAULT_ 0);
1992.Ve
1993.SH "EMBEDDING"
1994.IX Header "EMBEDDING"
1995Libev can (and often is) directly embedded into host
1996applications. Examples of applications that embed it include the Deliantra
1997Game Server, the \s-1EV\s0 perl module, the \s-1GNU\s0 Virtual Private Ethernet (gvpe)
1998and rxvt\-unicode.
1999.PP
2000The goal is to enable you to just copy the neecssary files into your
2001source directory without having to change even a single line in them, so
2002you can easily upgrade by simply copying (or having a checked-out copy of
2003libev somewhere in your source tree).
2004.Sh "\s-1FILESETS\s0"
2005.IX Subsection "FILESETS"
2006Depending on what features you need you need to include one or more sets of files
2007in your app.
2008.PP
2009\fI\s-1CORE\s0 \s-1EVENT\s0 \s-1LOOP\s0\fR
2010.IX Subsection "CORE EVENT LOOP"
2011.PP
2012To include only the libev core (all the \f(CW\*(C`ev_*\*(C'\fR functions), with manual
2013configuration (no autoconf):
2014.PP
2015.Vb 2
2016\& #define EV_STANDALONE 1
2017\& #include "ev.c"
2018.Ve
2019.PP
2020This will automatically include \fIev.h\fR, too, and should be done in a
2021single C source file only to provide the function implementations. To use
2022it, do the same for \fIev.h\fR in all files wishing to use this \s-1API\s0 (best
2023done by writing a wrapper around \fIev.h\fR that you can include instead and
2024where you can put other configuration options):
2025.PP
2026.Vb 2
2027\& #define EV_STANDALONE 1
2028\& #include "ev.h"
2029.Ve
2030.PP
2031Both header files and implementation files can be compiled with a \*(C+
2032compiler (at least, thats a stated goal, and breakage will be treated
2033as a bug).
2034.PP
2035You need the following files in your source tree, or in a directory
2036in your include path (e.g. in libev/ when using \-Ilibev):
2037.PP
2038.Vb 4
2039\& ev.h
2040\& ev.c
2041\& ev_vars.h
2042\& ev_wrap.h
2043.Ve
2044.PP
2045.Vb 1
2046\& ev_win32.c required on win32 platforms only
2047.Ve
2048.PP
2049.Vb 5
2050\& ev_select.c only when select backend is enabled (which is by default)
2051\& ev_poll.c only when poll backend is enabled (disabled by default)
2052\& ev_epoll.c only when the epoll backend is enabled (disabled by default)
2053\& ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2054\& ev_port.c only when the solaris port backend is enabled (disabled by default)
2055.Ve
2056.PP
2057\&\fIev.c\fR includes the backend files directly when enabled, so you only need
2058to compile this single file.
2059.PP
2060\fI\s-1LIBEVENT\s0 \s-1COMPATIBILITY\s0 \s-1API\s0\fR
2061.IX Subsection "LIBEVENT COMPATIBILITY API"
2062.PP
2063To include the libevent compatibility \s-1API\s0, also include:
2064.PP
2065.Vb 1
2066\& #include "event.c"
2067.Ve
2068.PP
2069in the file including \fIev.c\fR, and:
2070.PP
2071.Vb 1
2072\& #include "event.h"
2073.Ve
2074.PP
2075in the files that want to use the libevent \s-1API\s0. This also includes \fIev.h\fR.
2076.PP
2077You need the following additional files for this:
2078.PP
2079.Vb 2
2080\& event.h
2081\& event.c
2082.Ve
2083.PP
2084\fI\s-1AUTOCONF\s0 \s-1SUPPORT\s0\fR
2085.IX Subsection "AUTOCONF SUPPORT"
2086.PP
2087Instead of using \f(CW\*(C`EV_STANDALONE=1\*(C'\fR and providing your config in
2088whatever way you want, you can also \f(CW\*(C`m4_include([libev.m4])\*(C'\fR in your
2089\&\fIconfigure.ac\fR and leave \f(CW\*(C`EV_STANDALONE\*(C'\fR undefined. \fIev.c\fR will then
2090include \fIconfig.h\fR and configure itself accordingly.
2091.PP
2092For this of course you need the m4 file:
2093.PP
2094.Vb 1
2095\& libev.m4
2096.Ve
2097.Sh "\s-1PREPROCESSOR\s0 \s-1SYMBOLS/MACROS\s0"
2098.IX Subsection "PREPROCESSOR SYMBOLS/MACROS"
2099Libev can be configured via a variety of preprocessor symbols you have to define
2100before including any of its files. The default is not to build for multiplicity
2101and only include the select backend.
2102.IP "\s-1EV_STANDALONE\s0" 4
2103.IX Item "EV_STANDALONE"
2104Must always be \f(CW1\fR if you do not use autoconf configuration, which
2105keeps libev from including \fIconfig.h\fR, and it also defines dummy
2106implementations for some libevent functions (such as logging, which is not
2107supported). It will also not define any of the structs usually found in
2108\&\fIevent.h\fR that are not directly supported by the libev core alone.
2109.IP "\s-1EV_USE_MONOTONIC\s0" 4
2110.IX Item "EV_USE_MONOTONIC"
2111If defined to be \f(CW1\fR, libev will try to detect the availability of the
2112monotonic clock option at both compiletime and runtime. Otherwise no use
2113of the monotonic clock option will be attempted. If you enable this, you
2114usually have to link against librt or something similar. Enabling it when
2115the functionality isn't available is safe, though, althoguh you have
2116to make sure you link against any libraries where the \f(CW\*(C`clock_gettime\*(C'\fR
2117function is hiding in (often \fI\-lrt\fR).
2118.IP "\s-1EV_USE_REALTIME\s0" 4
2119.IX Item "EV_USE_REALTIME"
2120If defined to be \f(CW1\fR, libev will try to detect the availability of the
2121realtime clock option at compiletime (and assume its availability at
2122runtime if successful). Otherwise no use of the realtime clock option will
2123be attempted. This effectively replaces \f(CW\*(C`gettimeofday\*(C'\fR by \f(CW\*(C`clock_get
2124(CLOCK_REALTIME, ...)\*(C'\fR and will not normally affect correctness. See tzhe note about libraries
2125in the description of \f(CW\*(C`EV_USE_MONOTONIC\*(C'\fR, though.
2126.IP "\s-1EV_USE_SELECT\s0" 4
2127.IX Item "EV_USE_SELECT"
2128If undefined or defined to be \f(CW1\fR, libev will compile in support for the
2129\&\f(CW\*(C`select\*(C'\fR(2) backend. No attempt at autodetection will be done: if no
2130other method takes over, select will be it. Otherwise the select backend
2131will not be compiled in.
2132.IP "\s-1EV_SELECT_USE_FD_SET\s0" 4
2133.IX Item "EV_SELECT_USE_FD_SET"
2134If defined to \f(CW1\fR, then the select backend will use the system \f(CW\*(C`fd_set\*(C'\fR
2135structure. This is useful if libev doesn't compile due to a missing
2136\&\f(CW\*(C`NFDBITS\*(C'\fR or \f(CW\*(C`fd_mask\*(C'\fR definition or it misguesses the bitset layout on
2137exotic systems. This usually limits the range of file descriptors to some
2138low limit such as 1024 or might have other limitations (winsocket only
2139allows 64 sockets). The \f(CW\*(C`FD_SETSIZE\*(C'\fR macro, set before compilation, might
2140influence the size of the \f(CW\*(C`fd_set\*(C'\fR used.
2141.IP "\s-1EV_SELECT_IS_WINSOCKET\s0" 4
2142.IX Item "EV_SELECT_IS_WINSOCKET"
2143When defined to \f(CW1\fR, the select backend will assume that
2144select/socket/connect etc. don't understand file descriptors but
2145wants osf handles on win32 (this is the case when the select to
2146be used is the winsock select). This means that it will call
2147\&\f(CW\*(C`_get_osfhandle\*(C'\fR on the fd to convert it to an \s-1OS\s0 handle. Otherwise,
2148it is assumed that all these functions actually work on fds, even
2149on win32. Should not be defined on non\-win32 platforms.
2150.IP "\s-1EV_USE_POLL\s0" 4
2151.IX Item "EV_USE_POLL"
2152If defined to be \f(CW1\fR, libev will compile in support for the \f(CW\*(C`poll\*(C'\fR(2)
2153backend. Otherwise it will be enabled on non\-win32 platforms. It
2154takes precedence over select.
2155.IP "\s-1EV_USE_EPOLL\s0" 4
2156.IX Item "EV_USE_EPOLL"
2157If defined to be \f(CW1\fR, libev will compile in support for the Linux
2158\&\f(CW\*(C`epoll\*(C'\fR(7) backend. Its availability will be detected at runtime,
2159otherwise another method will be used as fallback. This is the
2160preferred backend for GNU/Linux systems.
2161.IP "\s-1EV_USE_KQUEUE\s0" 4
2162.IX Item "EV_USE_KQUEUE"
2163If defined to be \f(CW1\fR, libev will compile in support for the \s-1BSD\s0 style
2164\&\f(CW\*(C`kqueue\*(C'\fR(2) backend. Its actual availability will be detected at runtime,
2165otherwise another method will be used as fallback. This is the preferred
2166backend for \s-1BSD\s0 and BSD-like systems, although on most BSDs kqueue only
2167supports some types of fds correctly (the only platform we found that
2168supports ptys for example was NetBSD), so kqueue might be compiled in, but
2169not be used unless explicitly requested. The best way to use it is to find
2170out whether kqueue supports your type of fd properly and use an embedded
2171kqueue loop.
2172.IP "\s-1EV_USE_PORT\s0" 4
2173.IX Item "EV_USE_PORT"
2174If defined to be \f(CW1\fR, libev will compile in support for the Solaris
217510 port style backend. Its availability will be detected at runtime,
2176otherwise another method will be used as fallback. This is the preferred
2177backend for Solaris 10 systems.
2178.IP "\s-1EV_USE_DEVPOLL\s0" 4
2179.IX Item "EV_USE_DEVPOLL"
2180reserved for future expansion, works like the \s-1USE\s0 symbols above.
2181.IP "\s-1EV_USE_INOTIFY\s0" 4
2182.IX Item "EV_USE_INOTIFY"
2183If defined to be \f(CW1\fR, libev will compile in support for the Linux inotify
2184interface to speed up \f(CW\*(C`ev_stat\*(C'\fR watchers. Its actual availability will
2185be detected at runtime.
2186.IP "\s-1EV_H\s0" 4
2187.IX Item "EV_H"
2188The name of the \fIev.h\fR header file used to include it. The default if
2189undefined is \f(CW\*(C`<ev.h>\*(C'\fR in \fIevent.h\fR and \f(CW"ev.h"\fR in \fIev.c\fR. This
2190can be used to virtually rename the \fIev.h\fR header file in case of conflicts.
2191.IP "\s-1EV_CONFIG_H\s0" 4
2192.IX Item "EV_CONFIG_H"
2193If \f(CW\*(C`EV_STANDALONE\*(C'\fR isn't \f(CW1\fR, this variable can be used to override
2194\&\fIev.c\fR's idea of where to find the \fIconfig.h\fR file, similarly to
2195\&\f(CW\*(C`EV_H\*(C'\fR, above.
2196.IP "\s-1EV_EVENT_H\s0" 4
2197.IX Item "EV_EVENT_H"
2198Similarly to \f(CW\*(C`EV_H\*(C'\fR, this macro can be used to override \fIevent.c\fR's idea
2199of how the \fIevent.h\fR header can be found.
2200.IP "\s-1EV_PROTOTYPES\s0" 4
2201.IX Item "EV_PROTOTYPES"
2202If defined to be \f(CW0\fR, then \fIev.h\fR will not define any function
2203prototypes, but still define all the structs and other symbols. This is
2204occasionally useful if you want to provide your own wrapper functions
2205around libev functions.
2206.IP "\s-1EV_MULTIPLICITY\s0" 4
2207.IX Item "EV_MULTIPLICITY"
2208If undefined or defined to \f(CW1\fR, then all event-loop-specific functions
2209will have the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument, and you can create
2210additional independent event loops. Otherwise there will be no support
2211for multiple event loops and there is no first event loop pointer
2212argument. Instead, all functions act on the single default loop.
2213.IP "\s-1EV_PERIODIC_ENABLE\s0" 4
2214.IX Item "EV_PERIODIC_ENABLE"
2215If undefined or defined to be \f(CW1\fR, then periodic timers are supported. If
2216defined to be \f(CW0\fR, then they are not. Disabling them saves a few kB of
2217code.
2218.IP "\s-1EV_EMBED_ENABLE\s0" 4
2219.IX Item "EV_EMBED_ENABLE"
2220If undefined or defined to be \f(CW1\fR, then embed watchers are supported. If
2221defined to be \f(CW0\fR, then they are not.
2222.IP "\s-1EV_STAT_ENABLE\s0" 4
2223.IX Item "EV_STAT_ENABLE"
2224If undefined or defined to be \f(CW1\fR, then stat watchers are supported. If
2225defined to be \f(CW0\fR, then they are not.
2226.IP "\s-1EV_FORK_ENABLE\s0" 4
2227.IX Item "EV_FORK_ENABLE"
2228If undefined or defined to be \f(CW1\fR, then fork watchers are supported. If
2229defined to be \f(CW0\fR, then they are not.
2230.IP "\s-1EV_MINIMAL\s0" 4
2231.IX Item "EV_MINIMAL"
2232If you need to shave off some kilobytes of code at the expense of some
2233speed, define this symbol to \f(CW1\fR. Currently only used for gcc to override
2234some inlining decisions, saves roughly 30% codesize of amd64.
2235.IP "\s-1EV_PID_HASHSIZE\s0" 4
2236.IX Item "EV_PID_HASHSIZE"
2237\&\f(CW\*(C`ev_child\*(C'\fR watchers use a small hash table to distribute workload by
2238pid. The default size is \f(CW16\fR (or \f(CW1\fR with \f(CW\*(C`EV_MINIMAL\*(C'\fR), usually more
2239than enough. If you need to manage thousands of children you might want to
2240increase this value (\fImust\fR be a power of two).
2241.IP "\s-1EV_INOTIFY_HASHSIZE\s0" 4
2242.IX Item "EV_INOTIFY_HASHSIZE"
2243\&\f(CW\*(C`ev_staz\*(C'\fR watchers use a small hash table to distribute workload by
2244inotify watch id. The default size is \f(CW16\fR (or \f(CW1\fR with \f(CW\*(C`EV_MINIMAL\*(C'\fR),
2245usually more than enough. If you need to manage thousands of \f(CW\*(C`ev_stat\*(C'\fR
2246watchers you might want to increase this value (\fImust\fR be a power of
2247two).
2248.IP "\s-1EV_COMMON\s0" 4
2249.IX Item "EV_COMMON"
2250By default, all watchers have a \f(CW\*(C`void *data\*(C'\fR member. By redefining
2251this macro to a something else you can include more and other types of
2252members. You have to define it each time you include one of the files,
2253though, and it must be identical each time.
2254.Sp
2255For example, the perl \s-1EV\s0 module uses something like this:
2256.Sp
2257.Vb 3
2258\& #define EV_COMMON \e
2259\& SV *self; /* contains this struct */ \e
2260\& SV *cb_sv, *fh /* note no trailing ";" */
2261.Ve
2262.IP "\s-1EV_CB_DECLARE\s0 (type)" 4
2263.IX Item "EV_CB_DECLARE (type)"
2264.PD 0
2265.IP "\s-1EV_CB_INVOKE\s0 (watcher, revents)" 4
2266.IX Item "EV_CB_INVOKE (watcher, revents)"
2267.IP "ev_set_cb (ev, cb)" 4
2268.IX Item "ev_set_cb (ev, cb)"
2269.PD
2270Can be used to change the callback member declaration in each watcher,
2271and the way callbacks are invoked and set. Must expand to a struct member
2272definition and a statement, respectively. See the \fIev.v\fR header file for
2273their default definitions. One possible use for overriding these is to
2274avoid the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument in all cases, or to use
2275method calls instead of plain function calls in \*(C+.
2276.Sh "\s-1EXAMPLES\s0"
2277.IX Subsection "EXAMPLES"
2278For a real-world example of a program the includes libev
2279verbatim, you can have a look at the \s-1EV\s0 perl module
2280(<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
2281the \fIlibev/\fR subdirectory and includes them in the \fI\s-1EV/EVAPI\s0.h\fR (public
2282interface) and \fI\s-1EV\s0.xs\fR (implementation) files. Only the \fI\s-1EV\s0.xs\fR file
2283will be compiled. It is pretty complex because it provides its own header
2284file.
2285.Sp
2286The usage in rxvt-unicode is simpler. It has a \fIev_cpp.h\fR header file
2287that everybody includes and which overrides some autoconf choices:
2288.Sp
2289.Vb 4
2290\& #define EV_USE_POLL 0
2291\& #define EV_MULTIPLICITY 0
2292\& #define EV_PERIODICS 0
2293\& #define EV_CONFIG_H <config.h>
2294.Ve
2295.Sp
2296.Vb 1
2297\& #include "ev++.h"
2298.Ve
2299.Sp
2300And a \fIev_cpp.C\fR implementation file that contains libev proper and is compiled:
2301.Sp
2302.Vb 2
2303\& #include "ev_cpp.h"
2304\& #include "ev.c"
2305.Ve
2306.SH "COMPLEXITIES"
2307.IX Header "COMPLEXITIES"
2308In this section the complexities of (many of) the algorithms used inside
2309libev will be explained. For complexity discussions about backends see the
2310documentation for \f(CW\*(C`ev_default_init\*(C'\fR.
2311.RS 4
2312.IP "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)" 4
2313.IX Item "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)"
2314.PD 0
2315.IP "Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)" 4
2316.IX Item "Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)"
2317.IP "Starting io/check/prepare/idle/signal/child watchers: O(1)" 4
2318.IX Item "Starting io/check/prepare/idle/signal/child watchers: O(1)"
2319.IP "Stopping check/prepare/idle watchers: O(1)" 4
2320.IX Item "Stopping check/prepare/idle watchers: O(1)"
2321.IP "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % \s-1EV_PID_HASHSIZE\s0))" 4
2322.IX Item "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))"
2323.IP "Finding the next timer per loop iteration: O(1)" 4
2324.IX Item "Finding the next timer per loop iteration: O(1)"
2325.IP "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)" 4
2326.IX Item "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)"
2327.IP "Activating one watcher: O(1)" 4
2328.IX Item "Activating one watcher: O(1)"
2329.RE
2330.RS 4
2331.PD
957.SH "AUTHOR" 2332.SH "AUTHOR"
958.IX Header "AUTHOR" 2333.IX Header "AUTHOR"
959Marc Lehmann <libev@schmorp.de>. 2334Marc Lehmann <libev@schmorp.de>.

Diff Legend

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