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

Comparing libev/ev.3 (file contents):
Revision 1.5 by root, Fri Nov 23 04:36:03 2007 UTC vs.
Revision 1.49 by root, Wed Dec 12 04:53:58 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 "2007-11-23" "perl v5.8.8" "User Contributed Perl Documentation" 132.TH "<STANDARD INPUT>" 1 "2007-12-12" "perl v5.8.8" "User Contributed Perl Documentation"
133.SH "NAME" 133.SH "NAME"
134libev \- a high performance full\-featured event loop written in C 134libev \- a high performance full\-featured event loop written in C
135.SH "SYNOPSIS" 135.SH "SYNOPSIS"
136.IX Header "SYNOPSIS" 136.IX Header "SYNOPSIS"
137.Vb 1 137.Vb 1
138\& #include <ev.h> 138\& #include <ev.h>
139.Ve 139.Ve
140.SH "EXAMPLE PROGRAM"
141.IX Header "EXAMPLE PROGRAM"
142.Vb 1
143\& #include <ev.h>
144.Ve
145.PP
146.Vb 2
147\& ev_io stdin_watcher;
148\& ev_timer timeout_watcher;
149.Ve
150.PP
151.Vb 8
152\& /* called when data readable on stdin */
153\& static void
154\& stdin_cb (EV_P_ struct ev_io *w, int revents)
155\& {
156\& /* puts ("stdin ready"); */
157\& ev_io_stop (EV_A_ w); /* just a syntax example */
158\& ev_unloop (EV_A_ EVUNLOOP_ALL); /* leave all loop calls */
159\& }
160.Ve
161.PP
162.Vb 6
163\& static void
164\& timeout_cb (EV_P_ struct ev_timer *w, int revents)
165\& {
166\& /* puts ("timeout"); */
167\& ev_unloop (EV_A_ EVUNLOOP_ONE); /* leave one loop call */
168\& }
169.Ve
170.PP
171.Vb 4
172\& int
173\& main (void)
174\& {
175\& struct ev_loop *loop = ev_default_loop (0);
176.Ve
177.PP
178.Vb 3
179\& /* initialise an io watcher, then start it */
180\& ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
181\& ev_io_start (loop, &stdin_watcher);
182.Ve
183.PP
184.Vb 3
185\& /* simple non-repeating 5.5 second timeout */
186\& ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
187\& ev_timer_start (loop, &timeout_watcher);
188.Ve
189.PP
190.Vb 2
191\& /* loop till timeout or data ready */
192\& ev_loop (loop, 0);
193.Ve
194.PP
195.Vb 2
196\& return 0;
197\& }
198.Ve
140.SH "DESCRIPTION" 199.SH "DESCRIPTION"
141.IX Header "DESCRIPTION" 200.IX Header "DESCRIPTION"
201The newest version of this document is also available as a html-formatted
202web page you might find easier to navigate when reading it for the first
203time: <http://cvs.schmorp.de/libev/ev.html>.
204.PP
142Libev is an event loop: you register interest in certain events (such as a 205Libev is an event loop: you register interest in certain events (such as a
143file descriptor being readable or a timeout occuring), and it will manage 206file descriptor being readable or a timeout occuring), and it will manage
144these event sources and provide your program with events. 207these event sources and provide your program with events.
145.PP 208.PP
146To do this, it must take more or less complete control over your process 209To do this, it must take more or less complete control over your process
… …
151watchers\fR, which are relatively small C structures you initialise with the 214watchers\fR, which are relatively small C structures you initialise with the
152details of the event, and then hand it over to libev by \fIstarting\fR the 215details of the event, and then hand it over to libev by \fIstarting\fR the
153watcher. 216watcher.
154.SH "FEATURES" 217.SH "FEATURES"
155.IX Header "FEATURES" 218.IX Header "FEATURES"
156Libev supports select, poll, the linux-specific epoll and the bsd-specific 219Libev supports \f(CW\*(C`select\*(C'\fR, \f(CW\*(C`poll\*(C'\fR, the Linux-specific \f(CW\*(C`epoll\*(C'\fR, the
157kqueue mechanisms for file descriptor events, relative timers, absolute 220BSD-specific \f(CW\*(C`kqueue\*(C'\fR and the Solaris-specific event port mechanisms
158timers with customised rescheduling, signal events, process status change 221for file descriptor events (\f(CW\*(C`ev_io\*(C'\fR), the Linux \f(CW\*(C`inotify\*(C'\fR interface
159events (related to \s-1SIGCHLD\s0), and event watchers dealing with the event 222(for \f(CW\*(C`ev_stat\*(C'\fR), relative timers (\f(CW\*(C`ev_timer\*(C'\fR), absolute timers
160loop mechanism itself (idle, prepare and check watchers). It also is quite 223with customised rescheduling (\f(CW\*(C`ev_periodic\*(C'\fR), synchronous signals
161fast (see this benchmark comparing 224(\f(CW\*(C`ev_signal\*(C'\fR), process status change events (\f(CW\*(C`ev_child\*(C'\fR), and event
162it to libevent for example). 225watchers dealing with the event loop mechanism itself (\f(CW\*(C`ev_idle\*(C'\fR,
226\&\f(CW\*(C`ev_embed\*(C'\fR, \f(CW\*(C`ev_prepare\*(C'\fR and \f(CW\*(C`ev_check\*(C'\fR watchers) as well as
227file watchers (\f(CW\*(C`ev_stat\*(C'\fR) and even limited support for fork events
228(\f(CW\*(C`ev_fork\*(C'\fR).
229.PP
230It also is quite fast (see this
231benchmark comparing it to libevent
232for example).
163.SH "CONVENTIONS" 233.SH "CONVENTIONS"
164.IX Header "CONVENTIONS" 234.IX Header "CONVENTIONS"
165Libev is very configurable. In this manual the default configuration 235Libev is very configurable. In this manual the default configuration will
166will be described, which supports multiple event loops. For more info 236be described, which supports multiple event loops. For more info about
167about various configuration options please have a look at the file 237various configuration options please have a look at \fB\s-1EMBED\s0\fR section in
168\&\fI\s-1README\s0.embed\fR in the libev distribution. If libev was configured without 238this manual. If libev was configured without support for multiple event
169support for multiple event loops, then all functions taking an initial 239loops, then all functions taking an initial argument of name \f(CW\*(C`loop\*(C'\fR
170argument of name \f(CW\*(C`loop\*(C'\fR (which is always of type \f(CW\*(C`struct ev_loop *\*(C'\fR) 240(which is always of type \f(CW\*(C`struct ev_loop *\*(C'\fR) will not have this argument.
171will not have this argument.
172.SH "TIME REPRESENTATION" 241.SH "TIME REPRESENTATION"
173.IX Header "TIME REPRESENTATION" 242.IX Header "TIME REPRESENTATION"
174Libev represents time as a single floating point number, representing the 243Libev represents time as a single floating point number, representing the
175(fractional) number of seconds since the (\s-1POSIX\s0) epoch (somewhere near 244(fractional) number of seconds since the (\s-1POSIX\s0) epoch (somewhere near
176the beginning of 1970, details are complicated, don't ask). This type is 245the beginning of 1970, details are complicated, don't ask). This type is
177called \f(CW\*(C`ev_tstamp\*(C'\fR, which is what you should use too. It usually aliases 246called \f(CW\*(C`ev_tstamp\*(C'\fR, which is what you should use too. It usually aliases
178to the double type in C. 247to the \f(CW\*(C`double\*(C'\fR type in C, and when you need to do any calculations on
248it, you should treat it as such.
179.SH "GLOBAL FUNCTIONS" 249.SH "GLOBAL FUNCTIONS"
180.IX Header "GLOBAL FUNCTIONS" 250.IX Header "GLOBAL FUNCTIONS"
181These functions can be called anytime, even before initialising the 251These functions can be called anytime, even before initialising the
182library in any way. 252library in any way.
183.IP "ev_tstamp ev_time ()" 4 253.IP "ev_tstamp ev_time ()" 4
… …
189.IX Item "int ev_version_major ()" 259.IX Item "int ev_version_major ()"
190.PD 0 260.PD 0
191.IP "int ev_version_minor ()" 4 261.IP "int ev_version_minor ()" 4
192.IX Item "int ev_version_minor ()" 262.IX Item "int ev_version_minor ()"
193.PD 263.PD
194You can find out the major and minor version numbers of the library 264You can find out the major and minor \s-1ABI\s0 version numbers of the library
195you linked against by calling the functions \f(CW\*(C`ev_version_major\*(C'\fR and 265you linked against by calling the functions \f(CW\*(C`ev_version_major\*(C'\fR and
196\&\f(CW\*(C`ev_version_minor\*(C'\fR. If you want, you can compare against the global 266\&\f(CW\*(C`ev_version_minor\*(C'\fR. If you want, you can compare against the global
197symbols \f(CW\*(C`EV_VERSION_MAJOR\*(C'\fR and \f(CW\*(C`EV_VERSION_MINOR\*(C'\fR, which specify the 267symbols \f(CW\*(C`EV_VERSION_MAJOR\*(C'\fR and \f(CW\*(C`EV_VERSION_MINOR\*(C'\fR, which specify the
198version of the library your program was compiled against. 268version of the library your program was compiled against.
199.Sp 269.Sp
270These version numbers refer to the \s-1ABI\s0 version of the library, not the
271release version.
272.Sp
200Usually, it's a good idea to terminate if the major versions mismatch, 273Usually, it's a good idea to terminate if the major versions mismatch,
201as this indicates an incompatible change. Minor versions are usually 274as this indicates an incompatible change. Minor versions are usually
202compatible to older versions, so a larger minor version alone is usually 275compatible to older versions, so a larger minor version alone is usually
203not a problem. 276not a problem.
277.Sp
278Example: Make sure we haven't accidentally been linked against the wrong
279version.
280.Sp
281.Vb 3
282\& assert (("libev version mismatch",
283\& ev_version_major () == EV_VERSION_MAJOR
284\& && ev_version_minor () >= EV_VERSION_MINOR));
285.Ve
286.IP "unsigned int ev_supported_backends ()" 4
287.IX Item "unsigned int ev_supported_backends ()"
288Return the set of all backends (i.e. their corresponding \f(CW\*(C`EV_BACKEND_*\*(C'\fR
289value) compiled into this binary of libev (independent of their
290availability on the system you are running on). See \f(CW\*(C`ev_default_loop\*(C'\fR for
291a description of the set values.
292.Sp
293Example: make sure we have the epoll method, because yeah this is cool and
294a must have and can we have a torrent of it please!!!11
295.Sp
296.Vb 2
297\& assert (("sorry, no epoll, no sex",
298\& ev_supported_backends () & EVBACKEND_EPOLL));
299.Ve
300.IP "unsigned int ev_recommended_backends ()" 4
301.IX Item "unsigned int ev_recommended_backends ()"
302Return the set of all backends compiled into this binary of libev and also
303recommended for this platform. This set is often smaller than the one
304returned by \f(CW\*(C`ev_supported_backends\*(C'\fR, as for example kqueue is broken on
305most BSDs and will not be autodetected unless you explicitly request it
306(assuming you know what you are doing). This is the set of backends that
307libev will probe for if you specify no backends explicitly.
308.IP "unsigned int ev_embeddable_backends ()" 4
309.IX Item "unsigned int ev_embeddable_backends ()"
310Returns the set of backends that are embeddable in other event loops. This
311is the theoretical, all\-platform, value. To find which backends
312might be supported on the current system, you would need to look at
313\&\f(CW\*(C`ev_embeddable_backends () & ev_supported_backends ()\*(C'\fR, likewise for
314recommended ones.
315.Sp
316See 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))" 4 317.IP "ev_set_allocator (void *(*cb)(void *ptr, long size))" 4
205.IX Item "ev_set_allocator (void *(*cb)(void *ptr, long size))" 318.IX Item "ev_set_allocator (void *(*cb)(void *ptr, long size))"
206Sets the allocation function to use (the prototype is similar to the 319Sets the allocation function to use (the prototype is similar \- the
207realloc C function, the semantics are identical). It is used to allocate 320semantics is identical \- to the realloc C function). It is used to
208and free memory (no surprises here). If it returns zero when memory 321allocate and free memory (no surprises here). If it returns zero when
209needs to be allocated, the library might abort or take some potentially 322memory needs to be allocated, the library might abort or take some
210destructive action. The default is your system realloc function. 323potentially destructive action. The default is your system realloc
324function.
211.Sp 325.Sp
212You could override this function in high-availability programs to, say, 326You could override this function in high-availability programs to, say,
213free some memory if it cannot allocate memory, to use a special allocator, 327free 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. 328or even to sleep a while and retry until some memory is available.
329.Sp
330Example: Replace the libev allocator with one that waits a bit and then
331retries).
332.Sp
333.Vb 6
334\& static void *
335\& persistent_realloc (void *ptr, size_t size)
336\& {
337\& for (;;)
338\& {
339\& void *newptr = realloc (ptr, size);
340.Ve
341.Sp
342.Vb 2
343\& if (newptr)
344\& return newptr;
345.Ve
346.Sp
347.Vb 3
348\& sleep (60);
349\& }
350\& }
351.Ve
352.Sp
353.Vb 2
354\& ...
355\& ev_set_allocator (persistent_realloc);
356.Ve
215.IP "ev_set_syserr_cb (void (*cb)(const char *msg));" 4 357.IP "ev_set_syserr_cb (void (*cb)(const char *msg));" 4
216.IX Item "ev_set_syserr_cb (void (*cb)(const char *msg));" 358.IX Item "ev_set_syserr_cb (void (*cb)(const char *msg));"
217Set the callback function to call on a retryable syscall error (such 359Set the callback function to call on a retryable syscall error (such
218as failed select, poll, epoll_wait). The message is a printable string 360as failed select, poll, epoll_wait). The message is a printable string
219indicating the system call or subsystem causing the problem. If this 361indicating the system call or subsystem causing the problem. If this
220callback is set, then libev will expect it to remedy the sitution, no 362callback is set, then libev will expect it to remedy the sitution, no
221matter what, when it returns. That is, libev will generally retry the 363matter what, when it returns. That is, libev will generally retry the
222requested operation, or, if the condition doesn't go away, do bad stuff 364requested operation, or, if the condition doesn't go away, do bad stuff
223(such as abort). 365(such as abort).
366.Sp
367Example: This is basically the same thing that libev does internally, too.
368.Sp
369.Vb 6
370\& static void
371\& fatal_error (const char *msg)
372\& {
373\& perror (msg);
374\& abort ();
375\& }
376.Ve
377.Sp
378.Vb 2
379\& ...
380\& ev_set_syserr_cb (fatal_error);
381.Ve
224.SH "FUNCTIONS CONTROLLING THE EVENT LOOP" 382.SH "FUNCTIONS CONTROLLING THE EVENT LOOP"
225.IX Header "FUNCTIONS CONTROLLING THE EVENT LOOP" 383.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 384An 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 385types of such loops, the \fIdefault\fR loop, which supports signals and child
228events, and dynamically created loops which do not. 386events, and dynamically created loops which do not.
… …
236.IP "struct ev_loop *ev_default_loop (unsigned int flags)" 4 394.IP "struct ev_loop *ev_default_loop (unsigned int flags)" 4
237.IX Item "struct ev_loop *ev_default_loop (unsigned int flags)" 395.IX Item "struct ev_loop *ev_default_loop (unsigned int flags)"
238This will initialise the default event loop if it hasn't been initialised 396This 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 397yet 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 398false. If it already was initialised it simply returns it (and ignores the
241flags). 399flags. If that is troubling you, check \f(CW\*(C`ev_backend ()\*(C'\fR afterwards).
242.Sp 400.Sp
243If you don't know what event loop to use, use the one returned from this 401If you don't know what event loop to use, use the one returned from this
244function. 402function.
245.Sp 403.Sp
246The flags argument can be used to specify special behaviour or specific 404The 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). 405backends to use, and is usually specified as \f(CW0\fR (or \f(CW\*(C`EVFLAG_AUTO\*(C'\fR).
248.Sp 406.Sp
249It supports the following flags: 407The following flags are supported:
250.RS 4 408.RS 4
251.ie n .IP """EVFLAG_AUTO""" 4 409.ie n .IP """EVFLAG_AUTO""" 4
252.el .IP "\f(CWEVFLAG_AUTO\fR" 4 410.el .IP "\f(CWEVFLAG_AUTO\fR" 4
253.IX Item "EVFLAG_AUTO" 411.IX Item "EVFLAG_AUTO"
254The default flags value. Use this if you have no clue (it's the right 412The 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 418or 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 419\&\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 420override 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 421useful to try out specific backends to test their performance, or to work
264around bugs. 422around bugs.
423.ie n .IP """EVFLAG_FORKCHECK""" 4
424.el .IP "\f(CWEVFLAG_FORKCHECK\fR" 4
425.IX Item "EVFLAG_FORKCHECK"
426Instead of calling \f(CW\*(C`ev_default_fork\*(C'\fR or \f(CW\*(C`ev_loop_fork\*(C'\fR manually after
427a fork, you can also make libev check for a fork in each iteration by
428enabling this flag.
429.Sp
430This works by calling \f(CW\*(C`getpid ()\*(C'\fR on every iteration of the loop,
431and thus this might slow down your event loop if you do a lot of loop
432iterations and little real work, but is usually not noticeable (on my
433Linux system for example, \f(CW\*(C`getpid\*(C'\fR is actually a simple 5\-insn sequence
434without a syscall and thus \fIvery\fR fast, but my Linux system also has
435\&\f(CW\*(C`pthread_atfork\*(C'\fR which is even faster).
436.Sp
437The big advantage of this flag is that you can forget about fork (and
438forget about forgetting to tell libev about forking) when you use this
439flag.
440.Sp
441This flag setting cannot be overriden or specified in the \f(CW\*(C`LIBEV_FLAGS\*(C'\fR
442environment variable.
265.ie n .IP """EVMETHOD_SELECT"" (value 1, portable select backend)" 4 443.ie n .IP """EVBACKEND_SELECT"" (value 1, portable select backend)" 4
266.el .IP "\f(CWEVMETHOD_SELECT\fR (value 1, portable select backend)" 4 444.el .IP "\f(CWEVBACKEND_SELECT\fR (value 1, portable select backend)" 4
267.IX Item "EVMETHOD_SELECT (value 1, portable select backend)" 445.IX Item "EVBACKEND_SELECT (value 1, portable select backend)"
268This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as 446This 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, 447libev 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 448but 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 449using this backend. It doesn't scale too well (O(highest_fd)), but its usually
272the fastest backend for a low number of fds. 450the fastest backend for a low number of fds.
273.ie n .IP """EVMETHOD_POLL"" (value 2, poll backend, available everywhere except on windows)" 4 451.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)" 4 452.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)" 453.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 454And 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 455select, 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 456number 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). 457lot of inactive fds). It scales similarly to select, i.e. O(total_fds).
280.ie n .IP """EVMETHOD_EPOLL"" (value 4, Linux)" 4 458.ie n .IP """EVBACKEND_EPOLL"" (value 4, Linux)" 4
281.el .IP "\f(CWEVMETHOD_EPOLL\fR (value 4, Linux)" 4 459.el .IP "\f(CWEVBACKEND_EPOLL\fR (value 4, Linux)" 4
282.IX Item "EVMETHOD_EPOLL (value 4, Linux)" 460.IX Item "EVBACKEND_EPOLL (value 4, Linux)"
283For few fds, this backend is a bit little slower than poll and select, 461For 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 462but 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 463O(total_fds) where n is the total number of fds (or the highest fd), epoll scales
286either O(1) or O(active_fds). 464either O(1) or O(active_fds).
287.Sp 465.Sp
288While stopping and starting an I/O watcher in the same iteration will 466While stopping and starting an I/O watcher in the same iteration will
289result in some caching, there is still a syscall per such incident 467result 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 468(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 469best to avoid that. Also, \fIdup()\fRed file descriptors might not work very
292well if you register events for both fds. 470well if you register events for both fds.
471.Sp
472Please note that epoll sometimes generates spurious notifications, so you
473need to use non-blocking I/O or other means to avoid blocking when no data
474(or space) is available.
293.ie n .IP """EVMETHOD_KQUEUE"" (value 8, most \s-1BSD\s0 clones)" 4 475.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)" 4 476.el .IP "\f(CWEVBACKEND_KQUEUE\fR (value 8, most \s-1BSD\s0 clones)" 4
295.IX Item "EVMETHOD_KQUEUE (value 8, most BSD clones)" 477.IX Item "EVBACKEND_KQUEUE (value 8, most BSD clones)"
296Kqueue deserves special mention, as at the time of this writing, it 478Kqueue deserves special mention, as at the time of this writing, it
297was broken on all BSDs except NetBSD (usually it doesn't work with 479was broken on all BSDs except NetBSD (usually it doesn't work with
298anything but sockets and pipes, except on Darwin, where of course its 480anything but sockets and pipes, except on Darwin, where of course its
299completely useless). For this reason its not being \*(L"autodetected\*(R" unless 481completely 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). 482unless you explicitly specify it explicitly in the flags (i.e. using
483\&\f(CW\*(C`EVBACKEND_KQUEUE\*(C'\fR).
301.Sp 484.Sp
302It scales in the same way as the epoll backend, but the interface to the 485It 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 486kernel is more efficient (which says nothing about its actual speed, of
304course). While starting and stopping an I/O watcher does not cause an 487course). 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 488extra syscall as with epoll, it still adds up to four event changes per
306incident, so its best to avoid that. 489incident, so its best to avoid that.
307.ie n .IP """EVMETHOD_DEVPOLL"" (value 16, Solaris 8)" 4 490.ie n .IP """EVBACKEND_DEVPOLL"" (value 16, Solaris 8)" 4
308.el .IP "\f(CWEVMETHOD_DEVPOLL\fR (value 16, Solaris 8)" 4 491.el .IP "\f(CWEVBACKEND_DEVPOLL\fR (value 16, Solaris 8)" 4
309.IX Item "EVMETHOD_DEVPOLL (value 16, Solaris 8)" 492.IX Item "EVBACKEND_DEVPOLL (value 16, Solaris 8)"
310This is not implemented yet (and might never be). 493This is not implemented yet (and might never be).
311.ie n .IP """EVMETHOD_PORT"" (value 32, Solaris 10)" 4 494.ie n .IP """EVBACKEND_PORT"" (value 32, Solaris 10)" 4
312.el .IP "\f(CWEVMETHOD_PORT\fR (value 32, Solaris 10)" 4 495.el .IP "\f(CWEVBACKEND_PORT\fR (value 32, Solaris 10)" 4
313.IX Item "EVMETHOD_PORT (value 32, Solaris 10)" 496.IX Item "EVBACKEND_PORT (value 32, Solaris 10)"
314This uses the Solaris 10 port mechanism. As with everything on Solaris, 497This uses the Solaris 10 port mechanism. As with everything on Solaris,
315it's really slow, but it still scales very well (O(active_fds)). 498it's really slow, but it still scales very well (O(active_fds)).
499.Sp
500Please note that solaris ports can result in a lot of spurious
501notifications, so you need to use non-blocking I/O or other means to avoid
502blocking when no data (or space) is available.
316.ie n .IP """EVMETHOD_ALL""" 4 503.ie n .IP """EVBACKEND_ALL""" 4
317.el .IP "\f(CWEVMETHOD_ALL\fR" 4 504.el .IP "\f(CWEVBACKEND_ALL\fR" 4
318.IX Item "EVMETHOD_ALL" 505.IX Item "EVBACKEND_ALL"
319Try all backends (even potentially broken ones that wouldn't be tried 506Try all backends (even potentially broken ones that wouldn't be tried
320with \f(CW\*(C`EVFLAG_AUTO\*(C'\fR). Since this is a mask, you can do stuff such as 507with \f(CW\*(C`EVFLAG_AUTO\*(C'\fR). Since this is a mask, you can do stuff such as
321\&\f(CW\*(C`EVMETHOD_ALL & ~EVMETHOD_KQUEUE\*(C'\fR. 508\&\f(CW\*(C`EVBACKEND_ALL & ~EVBACKEND_KQUEUE\*(C'\fR.
322.RE 509.RE
323.RS 4 510.RS 4
324.Sp 511.Sp
325If one or more of these are ored into the flags value, then only these 512If one or more of these are ored into the flags value, then only these
326backends will be tried (in the reverse order as given here). If none are 513backends will be tried (in the reverse order as given here). If none are
327specified, most compiled-in backend will be tried, usually in reverse 514specified, most compiled-in backend will be tried, usually in reverse
328order of their flag values :) 515order of their flag values :)
516.Sp
517The most typical usage is like this:
518.Sp
519.Vb 2
520\& if (!ev_default_loop (0))
521\& fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
522.Ve
523.Sp
524Restrict libev to the select and poll backends, and do not allow
525environment settings to be taken into account:
526.Sp
527.Vb 1
528\& ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
529.Ve
530.Sp
531Use whatever libev has to offer, but make sure that kqueue is used if
532available (warning, breaks stuff, best use only with your own private
533event loop and only if you know the \s-1OS\s0 supports your types of fds):
534.Sp
535.Vb 1
536\& ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
537.Ve
329.RE 538.RE
330.IP "struct ev_loop *ev_loop_new (unsigned int flags)" 4 539.IP "struct ev_loop *ev_loop_new (unsigned int flags)" 4
331.IX Item "struct ev_loop *ev_loop_new (unsigned int flags)" 540.IX Item "struct ev_loop *ev_loop_new (unsigned int flags)"
332Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is 541Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is
333always distinct from the default loop. Unlike the default loop, it cannot 542always distinct from the default loop. Unlike the default loop, it cannot
334handle signal and child watchers, and attempts to do so will be greeted by 543handle signal and child watchers, and attempts to do so will be greeted by
335undefined behaviour (or a failed assertion if assertions are enabled). 544undefined behaviour (or a failed assertion if assertions are enabled).
545.Sp
546Example: Try to create a event loop that uses epoll and nothing else.
547.Sp
548.Vb 3
549\& struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
550\& if (!epoller)
551\& fatal ("no epoll found here, maybe it hides under your chair");
552.Ve
336.IP "ev_default_destroy ()" 4 553.IP "ev_default_destroy ()" 4
337.IX Item "ev_default_destroy ()" 554.IX Item "ev_default_destroy ()"
338Destroys the default loop again (frees all memory and kernel state 555Destroys the default loop again (frees all memory and kernel state
339etc.). This stops all registered event watchers (by not touching them in 556etc.). None of the active event watchers will be stopped in the normal
340any way whatsoever, although you cannot rely on this :). 557sense, so e.g. \f(CW\*(C`ev_is_active\*(C'\fR might still return true. It is your
558responsibility to either stop all watchers cleanly yoursef \fIbefore\fR
559calling this function, or cope with the fact afterwards (which is usually
560the easiest thing, youc na just ignore the watchers and/or \f(CW\*(C`free ()\*(C'\fR them
561for example).
341.IP "ev_loop_destroy (loop)" 4 562.IP "ev_loop_destroy (loop)" 4
342.IX Item "ev_loop_destroy (loop)" 563.IX Item "ev_loop_destroy (loop)"
343Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an 564Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an
344earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR. 565earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR.
345.IP "ev_default_fork ()" 4 566.IP "ev_default_fork ()" 4
… …
358quite nicely into a call to \f(CW\*(C`pthread_atfork\*(C'\fR: 579quite nicely into a call to \f(CW\*(C`pthread_atfork\*(C'\fR:
359.Sp 580.Sp
360.Vb 1 581.Vb 1
361\& pthread_atfork (0, 0, ev_default_fork); 582\& pthread_atfork (0, 0, ev_default_fork);
362.Ve 583.Ve
584.Sp
585At the moment, \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and \f(CW\*(C`EVBACKEND_POLL\*(C'\fR are safe to use
586without calling this function, so if you force one of those backends you
587do not need to care.
363.IP "ev_loop_fork (loop)" 4 588.IP "ev_loop_fork (loop)" 4
364.IX Item "ev_loop_fork (loop)" 589.IX Item "ev_loop_fork (loop)"
365Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by 590Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by
366\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop 591\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop
367after fork, and how you do this is entirely your own problem. 592after fork, and how you do this is entirely your own problem.
593.IP "unsigned int ev_loop_count (loop)" 4
594.IX Item "unsigned int ev_loop_count (loop)"
595Returns the count of loop iterations for the loop, which is identical to
596the number of times libev did poll for new events. It starts at \f(CW0\fR and
597happily wraps around with enough iterations.
598.Sp
599This value can sometimes be useful as a generation counter of sorts (it
600\&\*(L"ticks\*(R" the number of loop iterations), as it roughly corresponds with
601\&\f(CW\*(C`ev_prepare\*(C'\fR and \f(CW\*(C`ev_check\*(C'\fR calls.
368.IP "unsigned int ev_method (loop)" 4 602.IP "unsigned int ev_backend (loop)" 4
369.IX Item "unsigned int ev_method (loop)" 603.IX Item "unsigned int ev_backend (loop)"
370Returns one of the \f(CW\*(C`EVMETHOD_*\*(C'\fR flags indicating the event backend in 604Returns one of the \f(CW\*(C`EVBACKEND_*\*(C'\fR flags indicating the event backend in
371use. 605use.
372.IP "ev_tstamp ev_now (loop)" 4 606.IP "ev_tstamp ev_now (loop)" 4
373.IX Item "ev_tstamp ev_now (loop)" 607.IX Item "ev_tstamp ev_now (loop)"
374Returns the current \*(L"event loop time\*(R", which is the time the event loop 608Returns the current \*(L"event loop time\*(R", which is the time the event loop
375got events and started processing them. This timestamp does not change 609received events and started processing them. This timestamp does not
376as long as callbacks are being processed, and this is also the base time 610change as long as callbacks are being processed, and this is also the base
377used for relative timers. You can treat it as the timestamp of the event 611time used for relative timers. You can treat it as the timestamp of the
378occuring (or more correctly, the mainloop finding out about it). 612event occuring (or more correctly, libev finding out about it).
379.IP "ev_loop (loop, int flags)" 4 613.IP "ev_loop (loop, int flags)" 4
380.IX Item "ev_loop (loop, int flags)" 614.IX Item "ev_loop (loop, int flags)"
381Finally, this is it, the event handler. This function usually is called 615Finally, this is it, the event handler. This function usually is called
382after you initialised all your watchers and you want to start handling 616after you initialised all your watchers and you want to start handling
383events. 617events.
384.Sp 618.Sp
385If the flags argument is specified as 0, it will not return until either 619If the flags argument is specified as \f(CW0\fR, it will not return until
386no event watchers are active anymore or \f(CW\*(C`ev_unloop\*(C'\fR was called. 620either no event watchers are active anymore or \f(CW\*(C`ev_unloop\*(C'\fR was called.
621.Sp
622Please note that an explicit \f(CW\*(C`ev_unloop\*(C'\fR is usually better than
623relying on all watchers to be stopped when deciding when a program has
624finished (especially in interactive programs), but having a program that
625automatically loops as long as it has to and no longer by virtue of
626relying on its watchers stopping correctly is a thing of beauty.
387.Sp 627.Sp
388A flags value of \f(CW\*(C`EVLOOP_NONBLOCK\*(C'\fR will look for new events, will handle 628A flags value of \f(CW\*(C`EVLOOP_NONBLOCK\*(C'\fR will look for new events, will handle
389those events and any outstanding ones, but will not block your process in 629those events and any outstanding ones, but will not block your process in
390case there are no events and will return after one iteration of the loop. 630case there are no events and will return after one iteration of the loop.
391.Sp 631.Sp
392A flags value of \f(CW\*(C`EVLOOP_ONESHOT\*(C'\fR will look for new events (waiting if 632A flags value of \f(CW\*(C`EVLOOP_ONESHOT\*(C'\fR will look for new events (waiting if
393neccessary) and will handle those and any outstanding ones. It will block 633neccessary) and will handle those and any outstanding ones. It will block
394your process until at least one new event arrives, and will return after 634your process until at least one new event arrives, and will return after
395one iteration of the loop. 635one iteration of the loop. This is useful if you are waiting for some
636external event in conjunction with something not expressible using other
637libev watchers. However, a pair of \f(CW\*(C`ev_prepare\*(C'\fR/\f(CW\*(C`ev_check\*(C'\fR watchers is
638usually a better approach for this kind of thing.
396.Sp 639.Sp
397This flags value could be used to implement alternative looping
398constructs, but the \f(CW\*(C`prepare\*(C'\fR and \f(CW\*(C`check\*(C'\fR watchers provide a better and
399more generic mechanism.
400.Sp
401Here are the gory details of what ev_loop does: 640Here are the gory details of what \f(CW\*(C`ev_loop\*(C'\fR does:
402.Sp 641.Sp
403.Vb 15 642.Vb 19
643\& - Before the first iteration, call any pending watchers.
404\& 1. If there are no active watchers (reference count is zero), return. 644\& * If there are no active watchers (reference count is zero), return.
405\& 2. Queue and immediately call all prepare watchers. 645\& - Queue all prepare watchers and then call all outstanding watchers.
406\& 3. If we have been forked, recreate the kernel state. 646\& - If we have been forked, recreate the kernel state.
407\& 4. Update the kernel state with all outstanding changes. 647\& - Update the kernel state with all outstanding changes.
408\& 5. Update the "event loop time". 648\& - Update the "event loop time".
409\& 6. Calculate for how long to block. 649\& - Calculate for how long to block.
410\& 7. Block the process, waiting for events. 650\& - Block the process, waiting for any events.
651\& - Queue all outstanding I/O (fd) events.
411\& 8. Update the "event loop time" and do time jump handling. 652\& - Update the "event loop time" and do time jump handling.
412\& 9. Queue all outstanding timers. 653\& - Queue all outstanding timers.
413\& 10. Queue all outstanding periodics. 654\& - Queue all outstanding periodics.
414\& 11. If no events are pending now, queue all idle watchers. 655\& - If no events are pending now, queue all idle watchers.
415\& 12. Queue all check watchers. 656\& - Queue all check watchers.
416\& 13. Call all queued watchers in reverse order (i.e. check watchers first). 657\& - Call all queued watchers in reverse order (i.e. check watchers first).
658\& Signals and child watchers are implemented as I/O watchers, and will
659\& be handled here by queueing them when their watcher gets executed.
417\& 14. If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK 660\& - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
418\& was used, return, otherwise continue with step #1. 661\& were used, return, otherwise continue with step *.
662.Ve
663.Sp
664Example: Queue some jobs and then loop until no events are outsanding
665anymore.
666.Sp
667.Vb 4
668\& ... queue jobs here, make sure they register event watchers as long
669\& ... as they still have work to do (even an idle watcher will do..)
670\& ev_loop (my_loop, 0);
671\& ... jobs done. yeah!
419.Ve 672.Ve
420.IP "ev_unloop (loop, how)" 4 673.IP "ev_unloop (loop, how)" 4
421.IX Item "ev_unloop (loop, how)" 674.IX Item "ev_unloop (loop, how)"
422Can be used to make a call to \f(CW\*(C`ev_loop\*(C'\fR return early (but only after it 675Can be used to make a call to \f(CW\*(C`ev_loop\*(C'\fR return early (but only after it
423has processed all outstanding events). The \f(CW\*(C`how\*(C'\fR argument must be either 676has processed all outstanding events). The \f(CW\*(C`how\*(C'\fR argument must be either
… …
437example, libev itself uses this for its internal signal pipe: It is not 690example, libev itself uses this for its internal signal pipe: It is not
438visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if 691visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if
439no event watchers registered by it are active. It is also an excellent 692no event watchers registered by it are active. It is also an excellent
440way to do this for generic recurring timers or from within third-party 693way to do this for generic recurring timers or from within third-party
441libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR. 694libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR.
695.Sp
696Example: Create a signal watcher, but keep it from keeping \f(CW\*(C`ev_loop\*(C'\fR
697running when nothing else is active.
698.Sp
699.Vb 4
700\& struct ev_signal exitsig;
701\& ev_signal_init (&exitsig, sig_cb, SIGINT);
702\& ev_signal_start (loop, &exitsig);
703\& evf_unref (loop);
704.Ve
705.Sp
706Example: For some weird reason, unregister the above signal handler again.
707.Sp
708.Vb 2
709\& ev_ref (loop);
710\& ev_signal_stop (loop, &exitsig);
711.Ve
442.SH "ANATOMY OF A WATCHER" 712.SH "ANATOMY OF A WATCHER"
443.IX Header "ANATOMY OF A WATCHER" 713.IX Header "ANATOMY OF A WATCHER"
444A watcher is a structure that you create and register to record your 714A watcher is a structure that you create and register to record your
445interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to 715interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to
446become readable, you would create an \f(CW\*(C`ev_io\*(C'\fR watcher for that: 716become readable, you would create an \f(CW\*(C`ev_io\*(C'\fR watcher for that:
… …
482*)\*(C'\fR), and you can stop watching for events at any time by calling the 752*)\*(C'\fR), and you can stop watching for events at any time by calling the
483corresponding stop function (\f(CW\*(C`ev_<type>_stop (loop, watcher *)\*(C'\fR. 753corresponding stop function (\f(CW\*(C`ev_<type>_stop (loop, watcher *)\*(C'\fR.
484.PP 754.PP
485As long as your watcher is active (has been started but not stopped) you 755As long as your watcher is active (has been started but not stopped) you
486must not touch the values stored in it. Most specifically you must never 756must not touch the values stored in it. Most specifically you must never
487reinitialise it or call its set method. 757reinitialise it or call its \f(CW\*(C`set\*(C'\fR macro.
488.PP
489You can check whether an event is active by calling the \f(CW\*(C`ev_is_active
490(watcher *)\*(C'\fR macro. To see whether an event is outstanding (but the
491callback for it has not been called yet) you can use the \f(CW\*(C`ev_is_pending
492(watcher *)\*(C'\fR macro.
493.PP 758.PP
494Each and every callback receives the event loop pointer as first, the 759Each and every callback receives the event loop pointer as first, the
495registered watcher structure as second, and a bitset of received events as 760registered watcher structure as second, and a bitset of received events as
496third argument. 761third argument.
497.PP 762.PP
… …
522The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread. 787The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread.
523.ie n .IP """EV_CHILD""" 4 788.ie n .IP """EV_CHILD""" 4
524.el .IP "\f(CWEV_CHILD\fR" 4 789.el .IP "\f(CWEV_CHILD\fR" 4
525.IX Item "EV_CHILD" 790.IX Item "EV_CHILD"
526The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change. 791The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change.
792.ie n .IP """EV_STAT""" 4
793.el .IP "\f(CWEV_STAT\fR" 4
794.IX Item "EV_STAT"
795The path specified in the \f(CW\*(C`ev_stat\*(C'\fR watcher changed its attributes somehow.
527.ie n .IP """EV_IDLE""" 4 796.ie n .IP """EV_IDLE""" 4
528.el .IP "\f(CWEV_IDLE\fR" 4 797.el .IP "\f(CWEV_IDLE\fR" 4
529.IX Item "EV_IDLE" 798.IX Item "EV_IDLE"
530The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do. 799The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do.
531.ie n .IP """EV_PREPARE""" 4 800.ie n .IP """EV_PREPARE""" 4
… …
541\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any 810\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any
542received events. Callbacks of both watcher types can start and stop as 811received events. Callbacks of both watcher types can start and stop as
543many watchers as they want, and all of them will be taken into account 812many watchers as they want, and all of them will be taken into account
544(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep 813(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep
545\&\f(CW\*(C`ev_loop\*(C'\fR from blocking). 814\&\f(CW\*(C`ev_loop\*(C'\fR from blocking).
815.ie n .IP """EV_EMBED""" 4
816.el .IP "\f(CWEV_EMBED\fR" 4
817.IX Item "EV_EMBED"
818The embedded event loop specified in the \f(CW\*(C`ev_embed\*(C'\fR watcher needs attention.
819.ie n .IP """EV_FORK""" 4
820.el .IP "\f(CWEV_FORK\fR" 4
821.IX Item "EV_FORK"
822The event loop has been resumed in the child process after fork (see
823\&\f(CW\*(C`ev_fork\*(C'\fR).
546.ie n .IP """EV_ERROR""" 4 824.ie n .IP """EV_ERROR""" 4
547.el .IP "\f(CWEV_ERROR\fR" 4 825.el .IP "\f(CWEV_ERROR\fR" 4
548.IX Item "EV_ERROR" 826.IX Item "EV_ERROR"
549An unspecified error has occured, the watcher has been stopped. This might 827An unspecified error has occured, the watcher has been stopped. This might
550happen because the watcher could not be properly started because libev 828happen because the watcher could not be properly started because libev
… …
555Libev will usually signal a few \*(L"dummy\*(R" events together with an error, 833Libev will usually signal a few \*(L"dummy\*(R" events together with an error,
556for example it might indicate that a fd is readable or writable, and if 834for example it might indicate that a fd is readable or writable, and if
557your callbacks is well-written it can just attempt the operation and cope 835your callbacks is well-written it can just attempt the operation and cope
558with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded 836with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded
559programs, though, so beware. 837programs, though, so beware.
838.Sh "\s-1GENERIC\s0 \s-1WATCHER\s0 \s-1FUNCTIONS\s0"
839.IX Subsection "GENERIC WATCHER FUNCTIONS"
840In the following description, \f(CW\*(C`TYPE\*(C'\fR stands for the watcher type,
841e.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.
842.ie n .IP """ev_init"" (ev_TYPE *watcher, callback)" 4
843.el .IP "\f(CWev_init\fR (ev_TYPE *watcher, callback)" 4
844.IX Item "ev_init (ev_TYPE *watcher, callback)"
845This macro initialises the generic portion of a watcher. The contents
846of the watcher object can be arbitrary (so \f(CW\*(C`malloc\*(C'\fR will do). Only
847the generic parts of the watcher are initialised, you \fIneed\fR to call
848the type-specific \f(CW\*(C`ev_TYPE_set\*(C'\fR macro afterwards to initialise the
849type-specific parts. For each type there is also a \f(CW\*(C`ev_TYPE_init\*(C'\fR macro
850which rolls both calls into one.
851.Sp
852You can reinitialise a watcher at any time as long as it has been stopped
853(or never started) and there are no pending events outstanding.
854.Sp
855The callback is always of type \f(CW\*(C`void (*)(ev_loop *loop, ev_TYPE *watcher,
856int revents)\*(C'\fR.
857.ie n .IP """ev_TYPE_set"" (ev_TYPE *, [args])" 4
858.el .IP "\f(CWev_TYPE_set\fR (ev_TYPE *, [args])" 4
859.IX Item "ev_TYPE_set (ev_TYPE *, [args])"
860This macro initialises the type-specific parts of a watcher. You need to
861call \f(CW\*(C`ev_init\*(C'\fR at least once before you call this macro, but you can
862call \f(CW\*(C`ev_TYPE_set\*(C'\fR any number of times. You must not, however, call this
863macro on a watcher that is active (it can be pending, however, which is a
864difference to the \f(CW\*(C`ev_init\*(C'\fR macro).
865.Sp
866Although some watcher types do not have type-specific arguments
867(e.g. \f(CW\*(C`ev_prepare\*(C'\fR) you still need to call its \f(CW\*(C`set\*(C'\fR macro.
868.ie n .IP """ev_TYPE_init"" (ev_TYPE *watcher, callback, [args])" 4
869.el .IP "\f(CWev_TYPE_init\fR (ev_TYPE *watcher, callback, [args])" 4
870.IX Item "ev_TYPE_init (ev_TYPE *watcher, callback, [args])"
871This convinience macro rolls both \f(CW\*(C`ev_init\*(C'\fR and \f(CW\*(C`ev_TYPE_set\*(C'\fR macro
872calls into a single call. This is the most convinient method to initialise
873a watcher. The same limitations apply, of course.
874.ie n .IP """ev_TYPE_start"" (loop *, ev_TYPE *watcher)" 4
875.el .IP "\f(CWev_TYPE_start\fR (loop *, ev_TYPE *watcher)" 4
876.IX Item "ev_TYPE_start (loop *, ev_TYPE *watcher)"
877Starts (activates) the given watcher. Only active watchers will receive
878events. If the watcher is already active nothing will happen.
879.ie n .IP """ev_TYPE_stop"" (loop *, ev_TYPE *watcher)" 4
880.el .IP "\f(CWev_TYPE_stop\fR (loop *, ev_TYPE *watcher)" 4
881.IX Item "ev_TYPE_stop (loop *, ev_TYPE *watcher)"
882Stops the given watcher again (if active) and clears the pending
883status. It is possible that stopped watchers are pending (for example,
884non-repeating timers are being stopped when they become pending), but
885\&\f(CW\*(C`ev_TYPE_stop\*(C'\fR ensures that the watcher is neither active nor pending. If
886you want to free or reuse the memory used by the watcher it is therefore a
887good idea to always call its \f(CW\*(C`ev_TYPE_stop\*(C'\fR function.
888.IP "bool ev_is_active (ev_TYPE *watcher)" 4
889.IX Item "bool ev_is_active (ev_TYPE *watcher)"
890Returns a true value iff the watcher is active (i.e. it has been started
891and not yet been stopped). As long as a watcher is active you must not modify
892it.
893.IP "bool ev_is_pending (ev_TYPE *watcher)" 4
894.IX Item "bool ev_is_pending (ev_TYPE *watcher)"
895Returns a true value iff the watcher is pending, (i.e. it has outstanding
896events but its callback has not yet been invoked). As long as a watcher
897is pending (but not active) you must not call an init function on it (but
898\&\f(CW\*(C`ev_TYPE_set\*(C'\fR is safe), you must not change its priority, and you must
899make sure the watcher is available to libev (e.g. you cannot \f(CW\*(C`free ()\*(C'\fR
900it).
901.IP "callback ev_cb (ev_TYPE *watcher)" 4
902.IX Item "callback ev_cb (ev_TYPE *watcher)"
903Returns the callback currently set on the watcher.
904.IP "ev_cb_set (ev_TYPE *watcher, callback)" 4
905.IX Item "ev_cb_set (ev_TYPE *watcher, callback)"
906Change the callback. You can change the callback at virtually any time
907(modulo threads).
908.IP "ev_set_priority (ev_TYPE *watcher, priority)" 4
909.IX Item "ev_set_priority (ev_TYPE *watcher, priority)"
910.PD 0
911.IP "int ev_priority (ev_TYPE *watcher)" 4
912.IX Item "int ev_priority (ev_TYPE *watcher)"
913.PD
914Set and query the priority of the watcher. The priority is a small
915integer between \f(CW\*(C`EV_MAXPRI\*(C'\fR (default: \f(CW2\fR) and \f(CW\*(C`EV_MINPRI\*(C'\fR
916(default: \f(CW\*(C`\-2\*(C'\fR). Pending watchers with higher priority will be invoked
917before watchers with lower priority, but priority will not keep watchers
918from being executed (except for \f(CW\*(C`ev_idle\*(C'\fR watchers).
919.Sp
920This means that priorities are \fIonly\fR used for ordering callback
921invocation after new events have been received. This is useful, for
922example, to reduce latency after idling, or more often, to bind two
923watchers on the same event and make sure one is called first.
924.Sp
925If you need to suppress invocation when higher priority events are pending
926you need to look at \f(CW\*(C`ev_idle\*(C'\fR watchers, which provide this functionality.
927.Sp
928You \fImust not\fR change the priority of a watcher as long as it is active or
929pending.
930.Sp
931The default priority used by watchers when no priority has been set is
932always \f(CW0\fR, which is supposed to not be too high and not be too low :).
933.Sp
934Setting a priority outside the range of \f(CW\*(C`EV_MINPRI\*(C'\fR to \f(CW\*(C`EV_MAXPRI\*(C'\fR is
935fine, as long as you do not mind that the priority value you query might
936or might not have been adjusted to be within valid range.
937.IP "ev_invoke (loop, ev_TYPE *watcher, int revents)" 4
938.IX Item "ev_invoke (loop, ev_TYPE *watcher, int revents)"
939Invoke the \f(CW\*(C`watcher\*(C'\fR with the given \f(CW\*(C`loop\*(C'\fR and \f(CW\*(C`revents\*(C'\fR. Neither
940\&\f(CW\*(C`loop\*(C'\fR nor \f(CW\*(C`revents\*(C'\fR need to be valid as long as the watcher callback
941can deal with that fact.
942.IP "int ev_clear_pending (loop, ev_TYPE *watcher)" 4
943.IX Item "int ev_clear_pending (loop, ev_TYPE *watcher)"
944If the watcher is pending, this function returns clears its pending status
945and returns its \f(CW\*(C`revents\*(C'\fR bitset (as if its callback was invoked). If the
946watcher isn't pending it does nothing and returns \f(CW0\fR.
560.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0" 947.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0"
561.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER" 948.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER"
562Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change 949Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change
563and read at any time, libev will completely ignore it. This can be used 950and read at any time, libev will completely ignore it. This can be used
564to associate arbitrary data with your watcher. If you need more data and 951to associate arbitrary data with your watcher. If you need more data and
… …
585\& struct my_io *w = (struct my_io *)w_; 972\& struct my_io *w = (struct my_io *)w_;
586\& ... 973\& ...
587\& } 974\& }
588.Ve 975.Ve
589.PP 976.PP
590More interesting and less C\-conformant ways of catsing your callback type 977More interesting and less C\-conformant ways of casting your callback type
591have been omitted.... 978instead have been omitted.
979.PP
980Another common scenario is having some data structure with multiple
981watchers:
982.PP
983.Vb 6
984\& struct my_biggy
985\& {
986\& int some_data;
987\& ev_timer t1;
988\& ev_timer t2;
989\& }
990.Ve
991.PP
992In this case getting the pointer to \f(CW\*(C`my_biggy\*(C'\fR is a bit more complicated,
993you need to use \f(CW\*(C`offsetof\*(C'\fR:
994.PP
995.Vb 1
996\& #include <stddef.h>
997.Ve
998.PP
999.Vb 6
1000\& static void
1001\& t1_cb (EV_P_ struct ev_timer *w, int revents)
1002\& {
1003\& struct my_biggy big = (struct my_biggy *
1004\& (((char *)w) - offsetof (struct my_biggy, t1));
1005\& }
1006.Ve
1007.PP
1008.Vb 6
1009\& static void
1010\& t2_cb (EV_P_ struct ev_timer *w, int revents)
1011\& {
1012\& struct my_biggy big = (struct my_biggy *
1013\& (((char *)w) - offsetof (struct my_biggy, t2));
1014\& }
1015.Ve
592.SH "WATCHER TYPES" 1016.SH "WATCHER TYPES"
593.IX Header "WATCHER TYPES" 1017.IX Header "WATCHER TYPES"
594This section describes each watcher in detail, but will not repeat 1018This section describes each watcher in detail, but will not repeat
595information given in the last section. 1019information given in the last section. Any initialisation/set macros,
1020functions and members specific to the watcher type are explained.
1021.PP
1022Members are additionally marked with either \fI[read\-only]\fR, meaning that,
1023while the watcher is active, you can look at the member and expect some
1024sensible content, but you must not modify it (you can modify it while the
1025watcher is stopped to your hearts content), or \fI[read\-write]\fR, which
1026means you can expect it to have some sensible content while the watcher
1027is active, but you can also modify it. Modifying it may not do something
1028sensible or take immediate effect (or do anything at all), but libev will
1029not crash or malfunction in any way.
596.ie n .Sh """ev_io"" \- is this file descriptor readable or writable" 1030.ie n .Sh """ev_io"" \- is this file descriptor readable or writable?"
597.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable" 1031.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable?"
598.IX Subsection "ev_io - is this file descriptor readable or writable" 1032.IX Subsection "ev_io - is this file descriptor readable or writable?"
599I/O watchers check whether a file descriptor is readable or writable 1033I/O watchers check whether a file descriptor is readable or writable
600in each iteration of the event loop (This behaviour is called 1034in each iteration of the event loop, or, more precisely, when reading
601level-triggering because you keep receiving events as long as the 1035would not block the process and writing would at least be able to write
602condition persists. Remember you can stop the watcher if you don't want to 1036some data. This behaviour is called level-triggering because you keep
603act on the event and neither want to receive future events). 1037receiving events as long as the condition persists. Remember you can stop
1038the watcher if you don't want to act on the event and neither want to
1039receive future events.
604.PP 1040.PP
605In general you can register as many read and/or write event watchers per 1041In general you can register as many read and/or write event watchers per
606fd as you want (as long as you don't confuse yourself). Setting all file 1042fd as you want (as long as you don't confuse yourself). Setting all file
607descriptors to non-blocking mode is also usually a good idea (but not 1043descriptors to non-blocking mode is also usually a good idea (but not
608required if you know what you are doing). 1044required if you know what you are doing).
609.PP 1045.PP
610You have to be careful with dup'ed file descriptors, though. Some backends 1046You have to be careful with dup'ed file descriptors, though. Some backends
611(the linux epoll backend is a notable example) cannot handle dup'ed file 1047(the linux epoll backend is a notable example) cannot handle dup'ed file
612descriptors correctly if you register interest in two or more fds pointing 1048descriptors correctly if you register interest in two or more fds pointing
613to the same underlying file/socket etc. description (that is, they share 1049to the same underlying file/socket/etc. description (that is, they share
614the same underlying \*(L"file open\*(R"). 1050the same underlying \*(L"file open\*(R").
615.PP 1051.PP
616If you must do this, then force the use of a known-to-be-good backend 1052If you must do this, then force the use of a known-to-be-good backend
617(at the time of this writing, this includes only \s-1EVMETHOD_SELECT\s0 and 1053(at the time of this writing, this includes only \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and
618\&\s-1EVMETHOD_POLL\s0). 1054\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR).
1055.PP
1056Another thing you have to watch out for is that it is quite easy to
1057receive \*(L"spurious\*(R" readyness notifications, that is your callback might
1058be called with \f(CW\*(C`EV_READ\*(C'\fR but a subsequent \f(CW\*(C`read\*(C'\fR(2) will actually block
1059because there is no data. Not only are some backends known to create a
1060lot of those (for example solaris ports), it is very easy to get into
1061this situation even with a relatively standard program structure. Thus
1062it is best to always use non-blocking I/O: An extra \f(CW\*(C`read\*(C'\fR(2) returning
1063\&\f(CW\*(C`EAGAIN\*(C'\fR is far preferable to a program hanging until some data arrives.
1064.PP
1065If you cannot run the fd in non-blocking mode (for example you should not
1066play around with an Xlib connection), then you have to seperately re-test
1067whether a file descriptor is really ready with a known-to-be good interface
1068such as poll (fortunately in our Xlib example, Xlib already does this on
1069its own, so its quite safe to use).
1070.PP
1071\fIThe special problem of disappearing file descriptors\fR
1072.IX Subsection "The special problem of disappearing file descriptors"
1073.PP
1074Some backends (e.g kqueue, epoll) need to be told about closing a file
1075descriptor (either by calling \f(CW\*(C`close\*(C'\fR explicitly or by any other means,
1076such as \f(CW\*(C`dup\*(C'\fR). The reason is that you register interest in some file
1077descriptor, but when it goes away, the operating system will silently drop
1078this interest. If another file descriptor with the same number then is
1079registered with libev, there is no efficient way to see that this is, in
1080fact, a different file descriptor.
1081.PP
1082To avoid having to explicitly tell libev about such cases, libev follows
1083the following policy: Each time \f(CW\*(C`ev_io_set\*(C'\fR is being called, libev
1084will assume that this is potentially a new file descriptor, otherwise
1085it is assumed that the file descriptor stays the same. That means that
1086you \fIhave\fR to call \f(CW\*(C`ev_io_set\*(C'\fR (or \f(CW\*(C`ev_io_init\*(C'\fR) when you change the
1087descriptor even if the file descriptor number itself did not change.
1088.PP
1089This is how one would do it normally anyway, the important point is that
1090the libev application should not optimise around libev but should leave
1091optimisations to libev.
619.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4 1092.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4
620.IX Item "ev_io_init (ev_io *, callback, int fd, int events)" 1093.IX Item "ev_io_init (ev_io *, callback, int fd, int events)"
621.PD 0 1094.PD 0
622.IP "ev_io_set (ev_io *, int fd, int events)" 4 1095.IP "ev_io_set (ev_io *, int fd, int events)" 4
623.IX Item "ev_io_set (ev_io *, int fd, int events)" 1096.IX Item "ev_io_set (ev_io *, int fd, int events)"
624.PD 1097.PD
625Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The fd is the file descriptor to rceeive 1098Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The \f(CW\*(C`fd\*(C'\fR is the file descriptor to
626events 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 | 1099rceeive events for and events is either \f(CW\*(C`EV_READ\*(C'\fR, \f(CW\*(C`EV_WRITE\*(C'\fR or
627EV_WRITE\*(C'\fR to receive the given events. 1100\&\f(CW\*(C`EV_READ | EV_WRITE\*(C'\fR to receive the given events.
1101.IP "int fd [read\-only]" 4
1102.IX Item "int fd [read-only]"
1103The file descriptor being watched.
1104.IP "int events [read\-only]" 4
1105.IX Item "int events [read-only]"
1106The events being watched.
1107.PP
1108Example: Call \f(CW\*(C`stdin_readable_cb\*(C'\fR when \s-1STDIN_FILENO\s0 has become, well
1109readable, but only once. Since it is likely line\-buffered, you could
1110attempt to read a whole line in the callback.
1111.PP
1112.Vb 6
1113\& static void
1114\& stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1115\& {
1116\& ev_io_stop (loop, w);
1117\& .. read from stdin here (or from w->fd) and haqndle any I/O errors
1118\& }
1119.Ve
1120.PP
1121.Vb 6
1122\& ...
1123\& struct ev_loop *loop = ev_default_init (0);
1124\& struct ev_io stdin_readable;
1125\& ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1126\& ev_io_start (loop, &stdin_readable);
1127\& ev_loop (loop, 0);
1128.Ve
628.ie n .Sh """ev_timer"" \- relative and optionally recurring timeouts" 1129.ie n .Sh """ev_timer"" \- relative and optionally repeating timeouts"
629.el .Sh "\f(CWev_timer\fP \- relative and optionally recurring timeouts" 1130.el .Sh "\f(CWev_timer\fP \- relative and optionally repeating timeouts"
630.IX Subsection "ev_timer - relative and optionally recurring timeouts" 1131.IX Subsection "ev_timer - relative and optionally repeating timeouts"
631Timer watchers are simple relative timers that generate an event after a 1132Timer watchers are simple relative timers that generate an event after a
632given time, and optionally repeating in regular intervals after that. 1133given time, and optionally repeating in regular intervals after that.
633.PP 1134.PP
634The timers are based on real time, that is, if you register an event that 1135The timers are based on real time, that is, if you register an event that
635times out after an hour and you reset your system clock to last years 1136times out after an hour and you reset your system clock to last years
… …
669.IP "ev_timer_again (loop)" 4 1170.IP "ev_timer_again (loop)" 4
670.IX Item "ev_timer_again (loop)" 1171.IX Item "ev_timer_again (loop)"
671This will act as if the timer timed out and restart it again if it is 1172This will act as if the timer timed out and restart it again if it is
672repeating. The exact semantics are: 1173repeating. The exact semantics are:
673.Sp 1174.Sp
1175If the timer is pending, its pending status is cleared.
1176.Sp
674If the timer is started but nonrepeating, stop it. 1177If the timer is started but nonrepeating, stop it (as if it timed out).
675.Sp 1178.Sp
676If the timer is repeating, either start it if necessary (with the repeat 1179If the timer is repeating, either start it if necessary (with the
677value), or reset the running timer to the repeat value. 1180\&\f(CW\*(C`repeat\*(C'\fR value), or reset the running timer to the \f(CW\*(C`repeat\*(C'\fR value.
678.Sp 1181.Sp
679This sounds a bit complicated, but here is a useful and typical 1182This sounds a bit complicated, but here is a useful and typical
680example: Imagine you have a tcp connection and you want a so-called idle 1183example: Imagine you have a tcp connection and you want a so-called idle
681timeout, that is, you want to be called when there have been, say, 60 1184timeout, that is, you want to be called when there have been, say, 60
682seconds of inactivity on the socket. The easiest way to do this is to 1185seconds of inactivity on the socket. The easiest way to do this is to
683configure an \f(CW\*(C`ev_timer\*(C'\fR with after=repeat=60 and calling ev_timer_again each 1186configure an \f(CW\*(C`ev_timer\*(C'\fR with a \f(CW\*(C`repeat\*(C'\fR value of \f(CW60\fR and then call
684time you successfully read or write some data. If you go into an idle 1187\&\f(CW\*(C`ev_timer_again\*(C'\fR each time you successfully read or write some data. If
685state where you do not expect data to travel on the socket, you can stop 1188you go into an idle state where you do not expect data to travel on the
686the timer, and again will automatically restart it if need be. 1189socket, you can \f(CW\*(C`ev_timer_stop\*(C'\fR the timer, and \f(CW\*(C`ev_timer_again\*(C'\fR will
1190automatically restart it if need be.
1191.Sp
1192That means you can ignore the \f(CW\*(C`after\*(C'\fR value and \f(CW\*(C`ev_timer_start\*(C'\fR
1193altogether and only ever use the \f(CW\*(C`repeat\*(C'\fR value and \f(CW\*(C`ev_timer_again\*(C'\fR:
1194.Sp
1195.Vb 8
1196\& ev_timer_init (timer, callback, 0., 5.);
1197\& ev_timer_again (loop, timer);
1198\& ...
1199\& timer->again = 17.;
1200\& ev_timer_again (loop, timer);
1201\& ...
1202\& timer->again = 10.;
1203\& ev_timer_again (loop, timer);
1204.Ve
1205.Sp
1206This is more slightly efficient then stopping/starting the timer each time
1207you want to modify its timeout value.
1208.IP "ev_tstamp repeat [read\-write]" 4
1209.IX Item "ev_tstamp repeat [read-write]"
1210The current \f(CW\*(C`repeat\*(C'\fR value. Will be used each time the watcher times out
1211or \f(CW\*(C`ev_timer_again\*(C'\fR is called and determines the next timeout (if any),
1212which is also when any modifications are taken into account.
1213.PP
1214Example: Create a timer that fires after 60 seconds.
1215.PP
1216.Vb 5
1217\& static void
1218\& one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1219\& {
1220\& .. one minute over, w is actually stopped right here
1221\& }
1222.Ve
1223.PP
1224.Vb 3
1225\& struct ev_timer mytimer;
1226\& ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1227\& ev_timer_start (loop, &mytimer);
1228.Ve
1229.PP
1230Example: Create a timeout timer that times out after 10 seconds of
1231inactivity.
1232.PP
1233.Vb 5
1234\& static void
1235\& timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1236\& {
1237\& .. ten seconds without any activity
1238\& }
1239.Ve
1240.PP
1241.Vb 4
1242\& struct ev_timer mytimer;
1243\& ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1244\& ev_timer_again (&mytimer); /* start timer */
1245\& ev_loop (loop, 0);
1246.Ve
1247.PP
1248.Vb 3
1249\& // and in some piece of code that gets executed on any "activity":
1250\& // reset the timeout to start ticking again at 10 seconds
1251\& ev_timer_again (&mytimer);
1252.Ve
687.ie n .Sh """ev_periodic"" \- to cron or not to cron" 1253.ie n .Sh """ev_periodic"" \- to cron or not to cron?"
688.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron" 1254.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron?"
689.IX Subsection "ev_periodic - to cron or not to cron" 1255.IX Subsection "ev_periodic - to cron or not to cron?"
690Periodic watchers are also timers of a kind, but they are very versatile 1256Periodic watchers are also timers of a kind, but they are very versatile
691(and unfortunately a bit complex). 1257(and unfortunately a bit complex).
692.PP 1258.PP
693Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time) 1259Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time)
694but on wallclock time (absolute time). You can tell a periodic watcher 1260but on wallclock time (absolute time). You can tell a periodic watcher
695to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a 1261to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a
696periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now () 1262periodic watcher to trigger in 10 seconds (by specifiying e.g. \f(CW\*(C`ev_now ()
697+ 10.>) and then reset your system clock to the last year, then it will 1263+ 10.\*(C'\fR) and then reset your system clock to the last year, then it will
698take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger 1264take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger
699roughly 10 seconds later and of course not if you reset your system time 1265roughly 10 seconds later).
700again).
701.PP 1266.PP
702They can also be used to implement vastly more complex timers, such as 1267They can also be used to implement vastly more complex timers, such as
703triggering an event on eahc midnight, local time. 1268triggering an event on each midnight, local time or other, complicated,
1269rules.
704.PP 1270.PP
705As with timers, the callback is guarenteed to be invoked only when the 1271As with timers, the callback is guarenteed to be invoked only when the
706time (\f(CW\*(C`at\*(C'\fR) has been passed, but if multiple periodic timers become ready 1272time (\f(CW\*(C`at\*(C'\fR) has been passed, but if multiple periodic timers become ready
707during the same loop iteration then order of execution is undefined. 1273during the same loop iteration then order of execution is undefined.
708.IP "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)" 4 1274.IP "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)" 4
… …
712.IX Item "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)" 1278.IX Item "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)"
713.PD 1279.PD
714Lots of arguments, lets sort it out... There are basically three modes of 1280Lots of arguments, lets sort it out... There are basically three modes of
715operation, and we will explain them from simplest to complex: 1281operation, and we will explain them from simplest to complex:
716.RS 4 1282.RS 4
717.IP "* absolute timer (interval = reschedule_cb = 0)" 4 1283.IP "* absolute timer (at = time, interval = reschedule_cb = 0)" 4
718.IX Item "absolute timer (interval = reschedule_cb = 0)" 1284.IX Item "absolute timer (at = time, interval = reschedule_cb = 0)"
719In this configuration the watcher triggers an event at the wallclock time 1285In this configuration the watcher triggers an event at the wallclock time
720\&\f(CW\*(C`at\*(C'\fR and doesn't repeat. It will not adjust when a time jump occurs, 1286\&\f(CW\*(C`at\*(C'\fR and doesn't repeat. It will not adjust when a time jump occurs,
721that is, if it is to be run at January 1st 2011 then it will run when the 1287that is, if it is to be run at January 1st 2011 then it will run when the
722system time reaches or surpasses this time. 1288system time reaches or surpasses this time.
723.IP "* non-repeating interval timer (interval > 0, reschedule_cb = 0)" 4 1289.IP "* non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)" 4
724.IX Item "non-repeating interval timer (interval > 0, reschedule_cb = 0)" 1290.IX Item "non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)"
725In this mode the watcher will always be scheduled to time out at the next 1291In this mode the watcher will always be scheduled to time out at the next
726\&\f(CW\*(C`at + N * interval\*(C'\fR time (for some integer N) and then repeat, regardless 1292\&\f(CW\*(C`at + N * interval\*(C'\fR time (for some integer N, which can also be negative)
727of any time jumps. 1293and then repeat, regardless of any time jumps.
728.Sp 1294.Sp
729This can be used to create timers that do not drift with respect to system 1295This can be used to create timers that do not drift with respect to system
730time: 1296time:
731.Sp 1297.Sp
732.Vb 1 1298.Vb 1
… …
739by 3600. 1305by 3600.
740.Sp 1306.Sp
741Another way to think about it (for the mathematically inclined) is that 1307Another way to think about it (for the mathematically inclined) is that
742\&\f(CW\*(C`ev_periodic\*(C'\fR will try to run the callback in this mode at the next possible 1308\&\f(CW\*(C`ev_periodic\*(C'\fR will try to run the callback in this mode at the next possible
743time where \f(CW\*(C`time = at (mod interval)\*(C'\fR, regardless of any time jumps. 1309time where \f(CW\*(C`time = at (mod interval)\*(C'\fR, regardless of any time jumps.
1310.Sp
1311For numerical stability it is preferable that the \f(CW\*(C`at\*(C'\fR value is near
1312\&\f(CW\*(C`ev_now ()\*(C'\fR (the current time), but there is no range requirement for
1313this value.
744.IP "* manual reschedule mode (reschedule_cb = callback)" 4 1314.IP "* manual reschedule mode (at and interval ignored, reschedule_cb = callback)" 4
745.IX Item "manual reschedule mode (reschedule_cb = callback)" 1315.IX Item "manual reschedule mode (at and interval ignored, reschedule_cb = callback)"
746In this mode the values for \f(CW\*(C`interval\*(C'\fR and \f(CW\*(C`at\*(C'\fR are both being 1316In this mode the values for \f(CW\*(C`interval\*(C'\fR and \f(CW\*(C`at\*(C'\fR are both being
747ignored. Instead, each time the periodic watcher gets scheduled, the 1317ignored. Instead, each time the periodic watcher gets scheduled, the
748reschedule callback will be called with the watcher as first, and the 1318reschedule callback will be called with the watcher as first, and the
749current time as second argument. 1319current time as second argument.
750.Sp 1320.Sp
751\&\s-1NOTE:\s0 \fIThis callback \s-1MUST\s0 \s-1NOT\s0 stop or destroy any periodic watcher, 1321\&\s-1NOTE:\s0 \fIThis callback \s-1MUST\s0 \s-1NOT\s0 stop or destroy any periodic watcher,
752ever, or make any event loop modifications\fR. If you need to stop it, 1322ever, or make any event loop modifications\fR. If you need to stop it,
753return \f(CW\*(C`now + 1e30\*(C'\fR (or so, fudge fudge) and stop it afterwards (e.g. by 1323return \f(CW\*(C`now + 1e30\*(C'\fR (or so, fudge fudge) and stop it afterwards (e.g. by
754starting a prepare watcher). 1324starting an \f(CW\*(C`ev_prepare\*(C'\fR watcher, which is legal).
755.Sp 1325.Sp
756Its prototype is \f(CW\*(C`ev_tstamp (*reschedule_cb)(struct ev_periodic *w, 1326Its prototype is \f(CW\*(C`ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
757ev_tstamp now)\*(C'\fR, e.g.: 1327ev_tstamp now)\*(C'\fR, e.g.:
758.Sp 1328.Sp
759.Vb 4 1329.Vb 4
… …
783.IX Item "ev_periodic_again (loop, ev_periodic *)" 1353.IX Item "ev_periodic_again (loop, ev_periodic *)"
784Simply stops and restarts the periodic watcher again. This is only useful 1354Simply stops and restarts the periodic watcher again. This is only useful
785when you changed some parameters or the reschedule callback would return 1355when you changed some parameters or the reschedule callback would return
786a different time than the last time it was called (e.g. in a crond like 1356a different time than the last time it was called (e.g. in a crond like
787program when the crontabs have changed). 1357program when the crontabs have changed).
1358.IP "ev_tstamp offset [read\-write]" 4
1359.IX Item "ev_tstamp offset [read-write]"
1360When repeating, this contains the offset value, otherwise this is the
1361absolute point in time (the \f(CW\*(C`at\*(C'\fR value passed to \f(CW\*(C`ev_periodic_set\*(C'\fR).
1362.Sp
1363Can be modified any time, but changes only take effect when the periodic
1364timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being called.
1365.IP "ev_tstamp interval [read\-write]" 4
1366.IX Item "ev_tstamp interval [read-write]"
1367The current interval value. Can be modified any time, but changes only
1368take effect when the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being
1369called.
1370.IP "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read\-write]" 4
1371.IX Item "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]"
1372The current reschedule callback, or \f(CW0\fR, if this functionality is
1373switched off. Can be changed any time, but changes only take effect when
1374the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being called.
1375.PP
1376Example: Call a callback every hour, or, more precisely, whenever the
1377system clock is divisible by 3600. The callback invocation times have
1378potentially a lot of jittering, but good long-term stability.
1379.PP
1380.Vb 5
1381\& static void
1382\& clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1383\& {
1384\& ... its now a full hour (UTC, or TAI or whatever your clock follows)
1385\& }
1386.Ve
1387.PP
1388.Vb 3
1389\& struct ev_periodic hourly_tick;
1390\& ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1391\& ev_periodic_start (loop, &hourly_tick);
1392.Ve
1393.PP
1394Example: The same as above, but use a reschedule callback to do it:
1395.PP
1396.Vb 1
1397\& #include <math.h>
1398.Ve
1399.PP
1400.Vb 5
1401\& static ev_tstamp
1402\& my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1403\& {
1404\& return fmod (now, 3600.) + 3600.;
1405\& }
1406.Ve
1407.PP
1408.Vb 1
1409\& ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1410.Ve
1411.PP
1412Example: Call a callback every hour, starting now:
1413.PP
1414.Vb 4
1415\& struct ev_periodic hourly_tick;
1416\& ev_periodic_init (&hourly_tick, clock_cb,
1417\& fmod (ev_now (loop), 3600.), 3600., 0);
1418\& ev_periodic_start (loop, &hourly_tick);
1419.Ve
788.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled" 1420.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled!"
789.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled" 1421.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled!"
790.IX Subsection "ev_signal - signal me when a signal gets signalled" 1422.IX Subsection "ev_signal - signal me when a signal gets signalled!"
791Signal watchers will trigger an event when the process receives a specific 1423Signal watchers will trigger an event when the process receives a specific
792signal one or more times. Even though signals are very asynchronous, libev 1424signal one or more times. Even though signals are very asynchronous, libev
793will try it's best to deliver signals synchronously, i.e. as part of the 1425will try it's best to deliver signals synchronously, i.e. as part of the
794normal event processing, like any other event. 1426normal event processing, like any other event.
795.PP 1427.PP
… …
805.IP "ev_signal_set (ev_signal *, int signum)" 4 1437.IP "ev_signal_set (ev_signal *, int signum)" 4
806.IX Item "ev_signal_set (ev_signal *, int signum)" 1438.IX Item "ev_signal_set (ev_signal *, int signum)"
807.PD 1439.PD
808Configures the watcher to trigger on the given signal number (usually one 1440Configures the watcher to trigger on the given signal number (usually one
809of the \f(CW\*(C`SIGxxx\*(C'\fR constants). 1441of the \f(CW\*(C`SIGxxx\*(C'\fR constants).
1442.IP "int signum [read\-only]" 4
1443.IX Item "int signum [read-only]"
1444The signal the watcher watches out for.
810.ie n .Sh """ev_child"" \- wait for pid status changes" 1445.ie n .Sh """ev_child"" \- watch out for process status changes"
811.el .Sh "\f(CWev_child\fP \- wait for pid status changes" 1446.el .Sh "\f(CWev_child\fP \- watch out for process status changes"
812.IX Subsection "ev_child - wait for pid status changes" 1447.IX Subsection "ev_child - watch out for process status changes"
813Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to 1448Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to
814some child status changes (most typically when a child of yours dies). 1449some child status changes (most typically when a child of yours dies).
815.IP "ev_child_init (ev_child *, callback, int pid)" 4 1450.IP "ev_child_init (ev_child *, callback, int pid)" 4
816.IX Item "ev_child_init (ev_child *, callback, int pid)" 1451.IX Item "ev_child_init (ev_child *, callback, int pid)"
817.PD 0 1452.PD 0
… …
822\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look 1457\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look
823at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see 1458at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see
824the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems 1459the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems
825\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the 1460\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the
826process causing the status change. 1461process causing the status change.
1462.IP "int pid [read\-only]" 4
1463.IX Item "int pid [read-only]"
1464The process id this watcher watches out for, or \f(CW0\fR, meaning any process id.
1465.IP "int rpid [read\-write]" 4
1466.IX Item "int rpid [read-write]"
1467The process id that detected a status change.
1468.IP "int rstatus [read\-write]" 4
1469.IX Item "int rstatus [read-write]"
1470The process exit/trace status caused by \f(CW\*(C`rpid\*(C'\fR (see your systems
1471\&\f(CW\*(C`waitpid\*(C'\fR and \f(CW\*(C`sys/wait.h\*(C'\fR documentation for details).
1472.PP
1473Example: Try to exit cleanly on \s-1SIGINT\s0 and \s-1SIGTERM\s0.
1474.PP
1475.Vb 5
1476\& static void
1477\& sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1478\& {
1479\& ev_unloop (loop, EVUNLOOP_ALL);
1480\& }
1481.Ve
1482.PP
1483.Vb 3
1484\& struct ev_signal signal_watcher;
1485\& ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1486\& ev_signal_start (loop, &sigint_cb);
1487.Ve
1488.ie n .Sh """ev_stat"" \- did the file attributes just change?"
1489.el .Sh "\f(CWev_stat\fP \- did the file attributes just change?"
1490.IX Subsection "ev_stat - did the file attributes just change?"
1491This watches a filesystem path for attribute changes. That is, it calls
1492\&\f(CW\*(C`stat\*(C'\fR regularly (or when the \s-1OS\s0 says it changed) and sees if it changed
1493compared to the last time, invoking the callback if it did.
1494.PP
1495The path does not need to exist: changing from \*(L"path exists\*(R" to \*(L"path does
1496not exist\*(R" is a status change like any other. The condition \*(L"path does
1497not exist\*(R" is signified by the \f(CW\*(C`st_nlink\*(C'\fR field being zero (which is
1498otherwise always forced to be at least one) and all the other fields of
1499the stat buffer having unspecified contents.
1500.PP
1501The path \fIshould\fR be absolute and \fImust not\fR end in a slash. If it is
1502relative and your working directory changes, the behaviour is undefined.
1503.PP
1504Since there is no standard to do this, the portable implementation simply
1505calls \f(CW\*(C`stat (2)\*(C'\fR regularly on the path to see if it changed somehow. You
1506can specify a recommended polling interval for this case. If you specify
1507a polling interval of \f(CW0\fR (highly recommended!) then a \fIsuitable,
1508unspecified default\fR value will be used (which you can expect to be around
1509five seconds, although this might change dynamically). Libev will also
1510impose a minimum interval which is currently around \f(CW0.1\fR, but thats
1511usually overkill.
1512.PP
1513This watcher type is not meant for massive numbers of stat watchers,
1514as even with OS-supported change notifications, this can be
1515resource\-intensive.
1516.PP
1517At the time of this writing, only the Linux inotify interface is
1518implemented (implementing kqueue support is left as an exercise for the
1519reader). Inotify will be used to give hints only and should not change the
1520semantics of \f(CW\*(C`ev_stat\*(C'\fR watchers, which means that libev sometimes needs
1521to fall back to regular polling again even with inotify, but changes are
1522usually detected immediately, and if the file exists there will be no
1523polling.
1524.IP "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)" 4
1525.IX Item "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)"
1526.PD 0
1527.IP "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)" 4
1528.IX Item "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)"
1529.PD
1530Configures the watcher to wait for status changes of the given
1531\&\f(CW\*(C`path\*(C'\fR. The \f(CW\*(C`interval\*(C'\fR is a hint on how quickly a change is expected to
1532be detected and should normally be specified as \f(CW0\fR to let libev choose
1533a suitable value. The memory pointed to by \f(CW\*(C`path\*(C'\fR must point to the same
1534path for as long as the watcher is active.
1535.Sp
1536The callback will be receive \f(CW\*(C`EV_STAT\*(C'\fR when a change was detected,
1537relative to the attributes at the time the watcher was started (or the
1538last change was detected).
1539.IP "ev_stat_stat (ev_stat *)" 4
1540.IX Item "ev_stat_stat (ev_stat *)"
1541Updates the stat buffer immediately with new values. If you change the
1542watched path in your callback, you could call this fucntion to avoid
1543detecting this change (while introducing a race condition). Can also be
1544useful simply to find out the new values.
1545.IP "ev_statdata attr [read\-only]" 4
1546.IX Item "ev_statdata attr [read-only]"
1547The most-recently detected attributes of the file. Although the type is of
1548\&\f(CW\*(C`ev_statdata\*(C'\fR, this is usually the (or one of the) \f(CW\*(C`struct stat\*(C'\fR types
1549suitable for your system. If the \f(CW\*(C`st_nlink\*(C'\fR member is \f(CW0\fR, then there
1550was some error while \f(CW\*(C`stat\*(C'\fRing the file.
1551.IP "ev_statdata prev [read\-only]" 4
1552.IX Item "ev_statdata prev [read-only]"
1553The previous attributes of the file. The callback gets invoked whenever
1554\&\f(CW\*(C`prev\*(C'\fR != \f(CW\*(C`attr\*(C'\fR.
1555.IP "ev_tstamp interval [read\-only]" 4
1556.IX Item "ev_tstamp interval [read-only]"
1557The specified interval.
1558.IP "const char *path [read\-only]" 4
1559.IX Item "const char *path [read-only]"
1560The filesystem path that is being watched.
1561.PP
1562Example: Watch \f(CW\*(C`/etc/passwd\*(C'\fR for attribute changes.
1563.PP
1564.Vb 15
1565\& static void
1566\& passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1567\& {
1568\& /* /etc/passwd changed in some way */
1569\& if (w->attr.st_nlink)
1570\& {
1571\& printf ("passwd current size %ld\en", (long)w->attr.st_size);
1572\& printf ("passwd current atime %ld\en", (long)w->attr.st_mtime);
1573\& printf ("passwd current mtime %ld\en", (long)w->attr.st_mtime);
1574\& }
1575\& else
1576\& /* you shalt not abuse printf for puts */
1577\& puts ("wow, /etc/passwd is not there, expect problems. "
1578\& "if this is windows, they already arrived\en");
1579\& }
1580.Ve
1581.PP
1582.Vb 2
1583\& ...
1584\& ev_stat passwd;
1585.Ve
1586.PP
1587.Vb 2
1588\& ev_stat_init (&passwd, passwd_cb, "/etc/passwd");
1589\& ev_stat_start (loop, &passwd);
1590.Ve
827.ie n .Sh """ev_idle"" \- when you've got nothing better to do" 1591.ie n .Sh """ev_idle"" \- when you've got nothing better to do..."
828.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do" 1592.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do..."
829.IX Subsection "ev_idle - when you've got nothing better to do" 1593.IX Subsection "ev_idle - when you've got nothing better to do..."
830Idle watchers trigger events when there are no other events are pending 1594Idle watchers trigger events when no other events of the same or higher
831(prepare, check and other idle watchers do not count). That is, as long 1595priority are pending (prepare, check and other idle watchers do not
832as your process is busy handling sockets or timeouts (or even signals, 1596count).
833imagine) it will not be triggered. But when your process is idle all idle 1597.PP
834watchers are being called again and again, once per event loop iteration \- 1598That is, as long as your process is busy handling sockets or timeouts
1599(or even signals, imagine) of the same or higher priority it will not be
1600triggered. But when your process is idle (or only lower-priority watchers
1601are pending), the idle watchers are being called once per event loop
835until stopped, that is, or your process receives more events and becomes 1602iteration \- until stopped, that is, or your process receives more events
836busy. 1603and becomes busy again with higher priority stuff.
837.PP 1604.PP
838The most noteworthy effect is that as long as any idle watchers are 1605The most noteworthy effect is that as long as any idle watchers are
839active, the process will not block when waiting for new events. 1606active, the process will not block when waiting for new events.
840.PP 1607.PP
841Apart from keeping your process non-blocking (which is a useful 1608Apart from keeping your process non-blocking (which is a useful
… …
845.IP "ev_idle_init (ev_signal *, callback)" 4 1612.IP "ev_idle_init (ev_signal *, callback)" 4
846.IX Item "ev_idle_init (ev_signal *, callback)" 1613.IX Item "ev_idle_init (ev_signal *, callback)"
847Initialises and configures the idle watcher \- it has no parameters of any 1614Initialises and configures the idle watcher \- it has no parameters of any
848kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless, 1615kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless,
849believe me. 1616believe me.
1617.PP
1618Example: Dynamically allocate an \f(CW\*(C`ev_idle\*(C'\fR watcher, start it, and in the
1619callback, free it. Also, use no error checking, as usual.
1620.PP
1621.Vb 7
1622\& static void
1623\& idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1624\& {
1625\& free (w);
1626\& // now do something you wanted to do when the program has
1627\& // no longer asnything immediate to do.
1628\& }
1629.Ve
1630.PP
1631.Vb 3
1632\& struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1633\& ev_idle_init (idle_watcher, idle_cb);
1634\& ev_idle_start (loop, idle_cb);
1635.Ve
850.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop" 1636.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop!"
851.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop" 1637.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop!"
852.IX Subsection "ev_prepare and ev_check - customise your event loop" 1638.IX Subsection "ev_prepare and ev_check - customise your event loop!"
853Prepare and check watchers are usually (but not always) used in tandem: 1639Prepare and check watchers are usually (but not always) used in tandem:
854prepare watchers get invoked before the process blocks and check watchers 1640prepare watchers get invoked before the process blocks and check watchers
855afterwards. 1641afterwards.
856.PP 1642.PP
1643You \fImust not\fR call \f(CW\*(C`ev_loop\*(C'\fR or similar functions that enter
1644the current event loop from either \f(CW\*(C`ev_prepare\*(C'\fR or \f(CW\*(C`ev_check\*(C'\fR
1645watchers. Other loops than the current one are fine, however. The
1646rationale behind this is that you do not need to check for recursion in
1647those watchers, i.e. the sequence will always be \f(CW\*(C`ev_prepare\*(C'\fR, blocking,
1648\&\f(CW\*(C`ev_check\*(C'\fR so if you have one watcher of each kind they will always be
1649called in pairs bracketing the blocking call.
1650.PP
857Their main purpose is to integrate other event mechanisms into libev. This 1651Their main purpose is to integrate other event mechanisms into libev and
858could be used, for example, to track variable changes, implement your own 1652their use is somewhat advanced. This could be used, for example, to track
859watchers, integrate net-snmp or a coroutine library and lots more. 1653variable changes, implement your own watchers, integrate net-snmp or a
1654coroutine library and lots more. They are also occasionally useful if
1655you cache some data and want to flush it before blocking (for example,
1656in X programs you might want to do an \f(CW\*(C`XFlush ()\*(C'\fR in an \f(CW\*(C`ev_prepare\*(C'\fR
1657watcher).
860.PP 1658.PP
861This is done by examining in each prepare call which file descriptors need 1659This is done by examining in each prepare call which file descriptors need
862to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for 1660to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for
863them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries 1661them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries
864provide just this functionality). Then, in the check watcher you check for 1662provide just this functionality). Then, in the check watcher you check for
… …
873are ready to run (it's actually more complicated: it only runs coroutines 1671are ready to run (it's actually more complicated: it only runs coroutines
874with priority higher than or equal to the event loop and one coroutine 1672with priority higher than or equal to the event loop and one coroutine
875of lower priority, but only once, using idle watchers to keep the event 1673of lower priority, but only once, using idle watchers to keep the event
876loop from blocking if lower-priority coroutines are active, thus mapping 1674loop from blocking if lower-priority coroutines are active, thus mapping
877low-priority coroutines to idle/background tasks). 1675low-priority coroutines to idle/background tasks).
1676.PP
1677It is recommended to give \f(CW\*(C`ev_check\*(C'\fR watchers highest (\f(CW\*(C`EV_MAXPRI\*(C'\fR)
1678priority, to ensure that they are being run before any other watchers
1679after the poll. Also, \f(CW\*(C`ev_check\*(C'\fR watchers (and \f(CW\*(C`ev_prepare\*(C'\fR watchers,
1680too) should not activate (\*(L"feed\*(R") events into libev. While libev fully
1681supports this, they will be called before other \f(CW\*(C`ev_check\*(C'\fR watchers did
1682their job. As \f(CW\*(C`ev_check\*(C'\fR watchers are often used to embed other event
1683loops those other event loops might be in an unusable state until their
1684\&\f(CW\*(C`ev_check\*(C'\fR watcher ran (always remind yourself to coexist peacefully with
1685others).
878.IP "ev_prepare_init (ev_prepare *, callback)" 4 1686.IP "ev_prepare_init (ev_prepare *, callback)" 4
879.IX Item "ev_prepare_init (ev_prepare *, callback)" 1687.IX Item "ev_prepare_init (ev_prepare *, callback)"
880.PD 0 1688.PD 0
881.IP "ev_check_init (ev_check *, callback)" 4 1689.IP "ev_check_init (ev_check *, callback)" 4
882.IX Item "ev_check_init (ev_check *, callback)" 1690.IX Item "ev_check_init (ev_check *, callback)"
883.PD 1691.PD
884Initialises and configures the prepare or check watcher \- they have no 1692Initialises and configures the prepare or check watcher \- they have no
885parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR 1693parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR
886macros, but using them is utterly, utterly and completely pointless. 1694macros, but using them is utterly, utterly and completely pointless.
1695.PP
1696There are a number of principal ways to embed other event loops or modules
1697into libev. Here are some ideas on how to include libadns into libev
1698(there is a Perl module named \f(CW\*(C`EV::ADNS\*(C'\fR that does this, which you could
1699use for an actually working example. Another Perl module named \f(CW\*(C`EV::Glib\*(C'\fR
1700embeds a Glib main context into libev, and finally, \f(CW\*(C`Glib::EV\*(C'\fR embeds \s-1EV\s0
1701into the Glib event loop).
1702.PP
1703Method 1: Add \s-1IO\s0 watchers and a timeout watcher in a prepare handler,
1704and in a check watcher, destroy them and call into libadns. What follows
1705is pseudo-code only of course. This requires you to either use a low
1706priority for the check watcher or use \f(CW\*(C`ev_clear_pending\*(C'\fR explicitly, as
1707the callbacks for the IO/timeout watchers might not have been called yet.
1708.PP
1709.Vb 2
1710\& static ev_io iow [nfd];
1711\& static ev_timer tw;
1712.Ve
1713.PP
1714.Vb 4
1715\& static void
1716\& io_cb (ev_loop *loop, ev_io *w, int revents)
1717\& {
1718\& }
1719.Ve
1720.PP
1721.Vb 8
1722\& // create io watchers for each fd and a timer before blocking
1723\& static void
1724\& adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1725\& {
1726\& int timeout = 3600000;
1727\& struct pollfd fds [nfd];
1728\& // actual code will need to loop here and realloc etc.
1729\& adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1730.Ve
1731.PP
1732.Vb 3
1733\& /* the callback is illegal, but won't be called as we stop during check */
1734\& ev_timer_init (&tw, 0, timeout * 1e-3);
1735\& ev_timer_start (loop, &tw);
1736.Ve
1737.PP
1738.Vb 6
1739\& // create one ev_io per pollfd
1740\& for (int i = 0; i < nfd; ++i)
1741\& {
1742\& ev_io_init (iow + i, io_cb, fds [i].fd,
1743\& ((fds [i].events & POLLIN ? EV_READ : 0)
1744\& | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1745.Ve
1746.PP
1747.Vb 4
1748\& fds [i].revents = 0;
1749\& ev_io_start (loop, iow + i);
1750\& }
1751\& }
1752.Ve
1753.PP
1754.Vb 5
1755\& // stop all watchers after blocking
1756\& static void
1757\& adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1758\& {
1759\& ev_timer_stop (loop, &tw);
1760.Ve
1761.PP
1762.Vb 8
1763\& for (int i = 0; i < nfd; ++i)
1764\& {
1765\& // set the relevant poll flags
1766\& // could also call adns_processreadable etc. here
1767\& struct pollfd *fd = fds + i;
1768\& int revents = ev_clear_pending (iow + i);
1769\& if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1770\& if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1771.Ve
1772.PP
1773.Vb 3
1774\& // now stop the watcher
1775\& ev_io_stop (loop, iow + i);
1776\& }
1777.Ve
1778.PP
1779.Vb 2
1780\& adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
1781\& }
1782.Ve
1783.PP
1784Method 2: This would be just like method 1, but you run \f(CW\*(C`adns_afterpoll\*(C'\fR
1785in the prepare watcher and would dispose of the check watcher.
1786.PP
1787Method 3: If the module to be embedded supports explicit event
1788notification (adns does), you can also make use of the actual watcher
1789callbacks, and only destroy/create the watchers in the prepare watcher.
1790.PP
1791.Vb 5
1792\& static void
1793\& timer_cb (EV_P_ ev_timer *w, int revents)
1794\& {
1795\& adns_state ads = (adns_state)w->data;
1796\& update_now (EV_A);
1797.Ve
1798.PP
1799.Vb 2
1800\& adns_processtimeouts (ads, &tv_now);
1801\& }
1802.Ve
1803.PP
1804.Vb 5
1805\& static void
1806\& io_cb (EV_P_ ev_io *w, int revents)
1807\& {
1808\& adns_state ads = (adns_state)w->data;
1809\& update_now (EV_A);
1810.Ve
1811.PP
1812.Vb 3
1813\& if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
1814\& if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
1815\& }
1816.Ve
1817.PP
1818.Vb 1
1819\& // do not ever call adns_afterpoll
1820.Ve
1821.PP
1822Method 4: Do not use a prepare or check watcher because the module you
1823want to embed is too inflexible to support it. Instead, youc na override
1824their poll function. The drawback with this solution is that the main
1825loop is now no longer controllable by \s-1EV\s0. The \f(CW\*(C`Glib::EV\*(C'\fR module does
1826this.
1827.PP
1828.Vb 4
1829\& static gint
1830\& event_poll_func (GPollFD *fds, guint nfds, gint timeout)
1831\& {
1832\& int got_events = 0;
1833.Ve
1834.PP
1835.Vb 2
1836\& for (n = 0; n < nfds; ++n)
1837\& // create/start io watcher that sets the relevant bits in fds[n] and increment got_events
1838.Ve
1839.PP
1840.Vb 2
1841\& if (timeout >= 0)
1842\& // create/start timer
1843.Ve
1844.PP
1845.Vb 2
1846\& // poll
1847\& ev_loop (EV_A_ 0);
1848.Ve
1849.PP
1850.Vb 3
1851\& // stop timer again
1852\& if (timeout >= 0)
1853\& ev_timer_stop (EV_A_ &to);
1854.Ve
1855.PP
1856.Vb 3
1857\& // stop io watchers again - their callbacks should have set
1858\& for (n = 0; n < nfds; ++n)
1859\& ev_io_stop (EV_A_ iow [n]);
1860.Ve
1861.PP
1862.Vb 2
1863\& return got_events;
1864\& }
1865.Ve
1866.ie n .Sh """ev_embed"" \- when one backend isn't enough..."
1867.el .Sh "\f(CWev_embed\fP \- when one backend isn't enough..."
1868.IX Subsection "ev_embed - when one backend isn't enough..."
1869This is a rather advanced watcher type that lets you embed one event loop
1870into another (currently only \f(CW\*(C`ev_io\*(C'\fR events are supported in the embedded
1871loop, other types of watchers might be handled in a delayed or incorrect
1872fashion and must not be used).
1873.PP
1874There are primarily two reasons you would want that: work around bugs and
1875prioritise I/O.
1876.PP
1877As an example for a bug workaround, the kqueue backend might only support
1878sockets on some platform, so it is unusable as generic backend, but you
1879still want to make use of it because you have many sockets and it scales
1880so nicely. In this case, you would create a kqueue-based loop and embed it
1881into your default loop (which might use e.g. poll). Overall operation will
1882be a bit slower because first libev has to poll and then call kevent, but
1883at least you can use both at what they are best.
1884.PP
1885As for prioritising I/O: rarely you have the case where some fds have
1886to be watched and handled very quickly (with low latency), and even
1887priorities and idle watchers might have too much overhead. In this case
1888you would put all the high priority stuff in one loop and all the rest in
1889a second one, and embed the second one in the first.
1890.PP
1891As long as the watcher is active, the callback will be invoked every time
1892there might be events pending in the embedded loop. The callback must then
1893call \f(CW\*(C`ev_embed_sweep (mainloop, watcher)\*(C'\fR to make a single sweep and invoke
1894their callbacks (you could also start an idle watcher to give the embedded
1895loop strictly lower priority for example). You can also set the callback
1896to \f(CW0\fR, in which case the embed watcher will automatically execute the
1897embedded loop sweep.
1898.PP
1899As long as the watcher is started it will automatically handle events. The
1900callback will be invoked whenever some events have been handled. You can
1901set the callback to \f(CW0\fR to avoid having to specify one if you are not
1902interested in that.
1903.PP
1904Also, there have not currently been made special provisions for forking:
1905when you fork, you not only have to call \f(CW\*(C`ev_loop_fork\*(C'\fR on both loops,
1906but you will also have to stop and restart any \f(CW\*(C`ev_embed\*(C'\fR watchers
1907yourself.
1908.PP
1909Unfortunately, not all backends are embeddable, only the ones returned by
1910\&\f(CW\*(C`ev_embeddable_backends\*(C'\fR are, which, unfortunately, does not include any
1911portable one.
1912.PP
1913So when you want to use this feature you will always have to be prepared
1914that you cannot get an embeddable loop. The recommended way to get around
1915this is to have a separate variables for your embeddable loop, try to
1916create it, and if that fails, use the normal loop for everything:
1917.PP
1918.Vb 3
1919\& struct ev_loop *loop_hi = ev_default_init (0);
1920\& struct ev_loop *loop_lo = 0;
1921\& struct ev_embed embed;
1922.Ve
1923.PP
1924.Vb 5
1925\& // see if there is a chance of getting one that works
1926\& // (remember that a flags value of 0 means autodetection)
1927\& loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
1928\& ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
1929\& : 0;
1930.Ve
1931.PP
1932.Vb 8
1933\& // if we got one, then embed it, otherwise default to loop_hi
1934\& if (loop_lo)
1935\& {
1936\& ev_embed_init (&embed, 0, loop_lo);
1937\& ev_embed_start (loop_hi, &embed);
1938\& }
1939\& else
1940\& loop_lo = loop_hi;
1941.Ve
1942.IP "ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)" 4
1943.IX Item "ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)"
1944.PD 0
1945.IP "ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)" 4
1946.IX Item "ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)"
1947.PD
1948Configures the watcher to embed the given loop, which must be
1949embeddable. If the callback is \f(CW0\fR, then \f(CW\*(C`ev_embed_sweep\*(C'\fR will be
1950invoked automatically, otherwise it is the responsibility of the callback
1951to invoke it (it will continue to be called until the sweep has been done,
1952if you do not want thta, you need to temporarily stop the embed watcher).
1953.IP "ev_embed_sweep (loop, ev_embed *)" 4
1954.IX Item "ev_embed_sweep (loop, ev_embed *)"
1955Make a single, non-blocking sweep over the embedded loop. This works
1956similarly to \f(CW\*(C`ev_loop (embedded_loop, EVLOOP_NONBLOCK)\*(C'\fR, but in the most
1957apropriate way for embedded loops.
1958.IP "struct ev_loop *loop [read\-only]" 4
1959.IX Item "struct ev_loop *loop [read-only]"
1960The embedded event loop.
1961.ie n .Sh """ev_fork"" \- the audacity to resume the event loop after a fork"
1962.el .Sh "\f(CWev_fork\fP \- the audacity to resume the event loop after a fork"
1963.IX Subsection "ev_fork - the audacity to resume the event loop after a fork"
1964Fork watchers are called when a \f(CW\*(C`fork ()\*(C'\fR was detected (usually because
1965whoever is a good citizen cared to tell libev about it by calling
1966\&\f(CW\*(C`ev_default_fork\*(C'\fR or \f(CW\*(C`ev_loop_fork\*(C'\fR). The invocation is done before the
1967event loop blocks next and before \f(CW\*(C`ev_check\*(C'\fR watchers are being called,
1968and only in the child after the fork. If whoever good citizen calling
1969\&\f(CW\*(C`ev_default_fork\*(C'\fR cheats and calls it in the wrong process, the fork
1970handlers will be invoked, too, of course.
1971.IP "ev_fork_init (ev_signal *, callback)" 4
1972.IX Item "ev_fork_init (ev_signal *, callback)"
1973Initialises and configures the fork watcher \- it has no parameters of any
1974kind. There is a \f(CW\*(C`ev_fork_set\*(C'\fR macro, but using it is utterly pointless,
1975believe me.
887.SH "OTHER FUNCTIONS" 1976.SH "OTHER FUNCTIONS"
888.IX Header "OTHER FUNCTIONS" 1977.IX Header "OTHER FUNCTIONS"
889There are some other functions of possible interest. Described. Here. Now. 1978There are some other functions of possible interest. Described. Here. Now.
890.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4 1979.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4
891.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 1980.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)"
… …
920.Ve 2009.Ve
921.Sp 2010.Sp
922.Vb 1 2011.Vb 1
923\& ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0); 2012\& ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
924.Ve 2013.Ve
925.IP "ev_feed_event (loop, watcher, int events)" 4 2014.IP "ev_feed_event (ev_loop *, watcher *, int revents)" 4
926.IX Item "ev_feed_event (loop, watcher, int events)" 2015.IX Item "ev_feed_event (ev_loop *, watcher *, int revents)"
927Feeds the given event set into the event loop, as if the specified event 2016Feeds the given event set into the event loop, as if the specified event
928had happened for the specified watcher (which must be a pointer to an 2017had happened for the specified watcher (which must be a pointer to an
929initialised but not necessarily started event watcher). 2018initialised but not necessarily started event watcher).
930.IP "ev_feed_fd_event (loop, int fd, int revents)" 4 2019.IP "ev_feed_fd_event (ev_loop *, int fd, int revents)" 4
931.IX Item "ev_feed_fd_event (loop, int fd, int revents)" 2020.IX Item "ev_feed_fd_event (ev_loop *, int fd, int revents)"
932Feed an event on the given fd, as if a file descriptor backend detected 2021Feed an event on the given fd, as if a file descriptor backend detected
933the given events it. 2022the given events it.
934.IP "ev_feed_signal_event (loop, int signum)" 4 2023.IP "ev_feed_signal_event (ev_loop *loop, int signum)" 4
935.IX Item "ev_feed_signal_event (loop, int signum)" 2024.IX Item "ev_feed_signal_event (ev_loop *loop, int signum)"
936Feed an event as if the given signal occured (loop must be the default loop!). 2025Feed an event as if the given signal occured (\f(CW\*(C`loop\*(C'\fR must be the default
2026loop!).
937.SH "LIBEVENT EMULATION" 2027.SH "LIBEVENT EMULATION"
938.IX Header "LIBEVENT EMULATION" 2028.IX Header "LIBEVENT EMULATION"
939Libev offers a compatibility emulation layer for libevent. It cannot 2029Libev offers a compatibility emulation layer for libevent. It cannot
940emulate the internals of libevent, so here are some usage hints: 2030emulate the internals of libevent, so here are some usage hints:
941.IP "* Use it by including <event.h>, as usual." 4 2031.IP "* Use it by including <event.h>, as usual." 4
… …
952.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4 2042.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4
953.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library." 2043.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library."
954.PD 2044.PD
955.SH "\*(C+ SUPPORT" 2045.SH "\*(C+ SUPPORT"
956.IX Header " SUPPORT" 2046.IX Header " SUPPORT"
957\&\s-1TBD\s0. 2047Libev comes with some simplistic wrapper classes for \*(C+ that mainly allow
2048you to use some convinience methods to start/stop watchers and also change
2049the callback model to a model using method callbacks on objects.
2050.PP
2051To use it,
2052.PP
2053.Vb 1
2054\& #include <ev++.h>
2055.Ve
2056.PP
2057This automatically includes \fIev.h\fR and puts all of its definitions (many
2058of them macros) into the global namespace. All \*(C+ specific things are
2059put into the \f(CW\*(C`ev\*(C'\fR namespace. It should support all the same embedding
2060options as \fIev.h\fR, most notably \f(CW\*(C`EV_MULTIPLICITY\*(C'\fR.
2061.PP
2062Care has been taken to keep the overhead low. The only data member the \*(C+
2063classes add (compared to plain C\-style watchers) is the event loop pointer
2064that the watcher is associated with (or no additional members at all if
2065you disable \f(CW\*(C`EV_MULTIPLICITY\*(C'\fR when embedding libev).
2066.PP
2067Currently, functions, and static and non-static member functions can be
2068used as callbacks. Other types should be easy to add as long as they only
2069need one additional pointer for context. If you need support for other
2070types of functors please contact the author (preferably after implementing
2071it).
2072.PP
2073Here is a list of things available in the \f(CW\*(C`ev\*(C'\fR namespace:
2074.ie n .IP """ev::READ""\fR, \f(CW""ev::WRITE"" etc." 4
2075.el .IP "\f(CWev::READ\fR, \f(CWev::WRITE\fR etc." 4
2076.IX Item "ev::READ, ev::WRITE etc."
2077These are just enum values with the same values as the \f(CW\*(C`EV_READ\*(C'\fR etc.
2078macros from \fIev.h\fR.
2079.ie n .IP """ev::tstamp""\fR, \f(CW""ev::now""" 4
2080.el .IP "\f(CWev::tstamp\fR, \f(CWev::now\fR" 4
2081.IX Item "ev::tstamp, ev::now"
2082Aliases to the same types/functions as with the \f(CW\*(C`ev_\*(C'\fR prefix.
2083.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
2084.el .IP "\f(CWev::io\fR, \f(CWev::timer\fR, \f(CWev::periodic\fR, \f(CWev::idle\fR, \f(CWev::sig\fR etc." 4
2085.IX Item "ev::io, ev::timer, ev::periodic, ev::idle, ev::sig etc."
2086For each \f(CW\*(C`ev_TYPE\*(C'\fR watcher in \fIev.h\fR there is a corresponding class of
2087the same name in the \f(CW\*(C`ev\*(C'\fR namespace, with the exception of \f(CW\*(C`ev_signal\*(C'\fR
2088which is called \f(CW\*(C`ev::sig\*(C'\fR to avoid clashes with the \f(CW\*(C`signal\*(C'\fR macro
2089defines by many implementations.
2090.Sp
2091All of those classes have these methods:
2092.RS 4
2093.IP "ev::TYPE::TYPE ()" 4
2094.IX Item "ev::TYPE::TYPE ()"
2095.PD 0
2096.IP "ev::TYPE::TYPE (struct ev_loop *)" 4
2097.IX Item "ev::TYPE::TYPE (struct ev_loop *)"
2098.IP "ev::TYPE::~TYPE" 4
2099.IX Item "ev::TYPE::~TYPE"
2100.PD
2101The constructor (optionally) takes an event loop to associate the watcher
2102with. If it is omitted, it will use \f(CW\*(C`EV_DEFAULT\*(C'\fR.
2103.Sp
2104The constructor calls \f(CW\*(C`ev_init\*(C'\fR for you, which means you have to call the
2105\&\f(CW\*(C`set\*(C'\fR method before starting it.
2106.Sp
2107It will not set a callback, however: You have to call the templated \f(CW\*(C`set\*(C'\fR
2108method to set a callback before you can start the watcher.
2109.Sp
2110(The reason why you have to use a method is a limitation in \*(C+ which does
2111not allow explicit template arguments for constructors).
2112.Sp
2113The destructor automatically stops the watcher if it is active.
2114.IP "w\->set<class, &class::method> (object *)" 4
2115.IX Item "w->set<class, &class::method> (object *)"
2116This method sets the callback method to call. The method has to have a
2117signature of \f(CW\*(C`void (*)(ev_TYPE &, int)\*(C'\fR, it receives the watcher as
2118first argument and the \f(CW\*(C`revents\*(C'\fR as second. The object must be given as
2119parameter and is stored in the \f(CW\*(C`data\*(C'\fR member of the watcher.
2120.Sp
2121This method synthesizes efficient thunking code to call your method from
2122the C callback that libev requires. If your compiler can inline your
2123callback (i.e. it is visible to it at the place of the \f(CW\*(C`set\*(C'\fR call and
2124your compiler is good :), then the method will be fully inlined into the
2125thunking function, making it as fast as a direct C callback.
2126.Sp
2127Example: simple class declaration and watcher initialisation
2128.Sp
2129.Vb 4
2130\& struct myclass
2131\& {
2132\& void io_cb (ev::io &w, int revents) { }
2133\& }
2134.Ve
2135.Sp
2136.Vb 3
2137\& myclass obj;
2138\& ev::io iow;
2139\& iow.set <myclass, &myclass::io_cb> (&obj);
2140.Ve
2141.IP "w\->set<function> (void *data = 0)" 4
2142.IX Item "w->set<function> (void *data = 0)"
2143Also sets a callback, but uses a static method or plain function as
2144callback. The optional \f(CW\*(C`data\*(C'\fR argument will be stored in the watcher's
2145\&\f(CW\*(C`data\*(C'\fR member and is free for you to use.
2146.Sp
2147The prototype of the \f(CW\*(C`function\*(C'\fR must be \f(CW\*(C`void (*)(ev::TYPE &w, int)\*(C'\fR.
2148.Sp
2149See the method\-\f(CW\*(C`set\*(C'\fR above for more details.
2150.Sp
2151Example:
2152.Sp
2153.Vb 2
2154\& static void io_cb (ev::io &w, int revents) { }
2155\& iow.set <io_cb> ();
2156.Ve
2157.IP "w\->set (struct ev_loop *)" 4
2158.IX Item "w->set (struct ev_loop *)"
2159Associates a different \f(CW\*(C`struct ev_loop\*(C'\fR with this watcher. You can only
2160do this when the watcher is inactive (and not pending either).
2161.IP "w\->set ([args])" 4
2162.IX Item "w->set ([args])"
2163Basically the same as \f(CW\*(C`ev_TYPE_set\*(C'\fR, with the same args. Must be
2164called at least once. Unlike the C counterpart, an active watcher gets
2165automatically stopped and restarted when reconfiguring it with this
2166method.
2167.IP "w\->start ()" 4
2168.IX Item "w->start ()"
2169Starts the watcher. Note that there is no \f(CW\*(C`loop\*(C'\fR argument, as the
2170constructor already stores the event loop.
2171.IP "w\->stop ()" 4
2172.IX Item "w->stop ()"
2173Stops the watcher if it is active. Again, no \f(CW\*(C`loop\*(C'\fR argument.
2174.ie n .IP "w\->again () ""ev::timer""\fR, \f(CW""ev::periodic"" only" 4
2175.el .IP "w\->again () \f(CWev::timer\fR, \f(CWev::periodic\fR only" 4
2176.IX Item "w->again () ev::timer, ev::periodic only"
2177For \f(CW\*(C`ev::timer\*(C'\fR and \f(CW\*(C`ev::periodic\*(C'\fR, this invokes the corresponding
2178\&\f(CW\*(C`ev_TYPE_again\*(C'\fR function.
2179.ie n .IP "w\->sweep () ""ev::embed"" only" 4
2180.el .IP "w\->sweep () \f(CWev::embed\fR only" 4
2181.IX Item "w->sweep () ev::embed only"
2182Invokes \f(CW\*(C`ev_embed_sweep\*(C'\fR.
2183.ie n .IP "w\->update () ""ev::stat"" only" 4
2184.el .IP "w\->update () \f(CWev::stat\fR only" 4
2185.IX Item "w->update () ev::stat only"
2186Invokes \f(CW\*(C`ev_stat_stat\*(C'\fR.
2187.RE
2188.RS 4
2189.RE
2190.PP
2191Example: Define a class with an \s-1IO\s0 and idle watcher, start one of them in
2192the constructor.
2193.PP
2194.Vb 4
2195\& class myclass
2196\& {
2197\& ev_io io; void io_cb (ev::io &w, int revents);
2198\& ev_idle idle void idle_cb (ev::idle &w, int revents);
2199.Ve
2200.PP
2201.Vb 2
2202\& myclass ();
2203\& }
2204.Ve
2205.PP
2206.Vb 4
2207\& myclass::myclass (int fd)
2208\& {
2209\& io .set <myclass, &myclass::io_cb > (this);
2210\& idle.set <myclass, &myclass::idle_cb> (this);
2211.Ve
2212.PP
2213.Vb 2
2214\& io.start (fd, ev::READ);
2215\& }
2216.Ve
2217.SH "MACRO MAGIC"
2218.IX Header "MACRO MAGIC"
2219Libev can be compiled with a variety of options, the most fundemantal is
2220\&\f(CW\*(C`EV_MULTIPLICITY\*(C'\fR. This option determines whether (most) functions and
2221callbacks have an initial \f(CW\*(C`struct ev_loop *\*(C'\fR argument.
2222.PP
2223To make it easier to write programs that cope with either variant, the
2224following macros are defined:
2225.ie n .IP """EV_A""\fR, \f(CW""EV_A_""" 4
2226.el .IP "\f(CWEV_A\fR, \f(CWEV_A_\fR" 4
2227.IX Item "EV_A, EV_A_"
2228This provides the loop \fIargument\fR for functions, if one is required (\*(L"ev
2229loop argument\*(R"). The \f(CW\*(C`EV_A\*(C'\fR form is used when this is the sole argument,
2230\&\f(CW\*(C`EV_A_\*(C'\fR is used when other arguments are following. Example:
2231.Sp
2232.Vb 3
2233\& ev_unref (EV_A);
2234\& ev_timer_add (EV_A_ watcher);
2235\& ev_loop (EV_A_ 0);
2236.Ve
2237.Sp
2238It assumes the variable \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR is in scope,
2239which is often provided by the following macro.
2240.ie n .IP """EV_P""\fR, \f(CW""EV_P_""" 4
2241.el .IP "\f(CWEV_P\fR, \f(CWEV_P_\fR" 4
2242.IX Item "EV_P, EV_P_"
2243This provides the loop \fIparameter\fR for functions, if one is required (\*(L"ev
2244loop parameter\*(R"). The \f(CW\*(C`EV_P\*(C'\fR form is used when this is the sole parameter,
2245\&\f(CW\*(C`EV_P_\*(C'\fR is used when other parameters are following. Example:
2246.Sp
2247.Vb 2
2248\& // this is how ev_unref is being declared
2249\& static void ev_unref (EV_P);
2250.Ve
2251.Sp
2252.Vb 2
2253\& // this is how you can declare your typical callback
2254\& static void cb (EV_P_ ev_timer *w, int revents)
2255.Ve
2256.Sp
2257It declares a parameter \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR, quite
2258suitable for use with \f(CW\*(C`EV_A\*(C'\fR.
2259.ie n .IP """EV_DEFAULT""\fR, \f(CW""EV_DEFAULT_""" 4
2260.el .IP "\f(CWEV_DEFAULT\fR, \f(CWEV_DEFAULT_\fR" 4
2261.IX Item "EV_DEFAULT, EV_DEFAULT_"
2262Similar to the other two macros, this gives you the value of the default
2263loop, if multiple loops are supported (\*(L"ev loop default\*(R").
2264.PP
2265Example: Declare and initialise a check watcher, utilising the above
2266macros so it will work regardless of whether multiple loops are supported
2267or not.
2268.PP
2269.Vb 5
2270\& static void
2271\& check_cb (EV_P_ ev_timer *w, int revents)
2272\& {
2273\& ev_check_stop (EV_A_ w);
2274\& }
2275.Ve
2276.PP
2277.Vb 4
2278\& ev_check check;
2279\& ev_check_init (&check, check_cb);
2280\& ev_check_start (EV_DEFAULT_ &check);
2281\& ev_loop (EV_DEFAULT_ 0);
2282.Ve
2283.SH "EMBEDDING"
2284.IX Header "EMBEDDING"
2285Libev can (and often is) directly embedded into host
2286applications. Examples of applications that embed it include the Deliantra
2287Game Server, the \s-1EV\s0 perl module, the \s-1GNU\s0 Virtual Private Ethernet (gvpe)
2288and rxvt\-unicode.
2289.PP
2290The goal is to enable you to just copy the neecssary files into your
2291source directory without having to change even a single line in them, so
2292you can easily upgrade by simply copying (or having a checked-out copy of
2293libev somewhere in your source tree).
2294.Sh "\s-1FILESETS\s0"
2295.IX Subsection "FILESETS"
2296Depending on what features you need you need to include one or more sets of files
2297in your app.
2298.PP
2299\fI\s-1CORE\s0 \s-1EVENT\s0 \s-1LOOP\s0\fR
2300.IX Subsection "CORE EVENT LOOP"
2301.PP
2302To include only the libev core (all the \f(CW\*(C`ev_*\*(C'\fR functions), with manual
2303configuration (no autoconf):
2304.PP
2305.Vb 2
2306\& #define EV_STANDALONE 1
2307\& #include "ev.c"
2308.Ve
2309.PP
2310This will automatically include \fIev.h\fR, too, and should be done in a
2311single C source file only to provide the function implementations. To use
2312it, do the same for \fIev.h\fR in all files wishing to use this \s-1API\s0 (best
2313done by writing a wrapper around \fIev.h\fR that you can include instead and
2314where you can put other configuration options):
2315.PP
2316.Vb 2
2317\& #define EV_STANDALONE 1
2318\& #include "ev.h"
2319.Ve
2320.PP
2321Both header files and implementation files can be compiled with a \*(C+
2322compiler (at least, thats a stated goal, and breakage will be treated
2323as a bug).
2324.PP
2325You need the following files in your source tree, or in a directory
2326in your include path (e.g. in libev/ when using \-Ilibev):
2327.PP
2328.Vb 4
2329\& ev.h
2330\& ev.c
2331\& ev_vars.h
2332\& ev_wrap.h
2333.Ve
2334.PP
2335.Vb 1
2336\& ev_win32.c required on win32 platforms only
2337.Ve
2338.PP
2339.Vb 5
2340\& ev_select.c only when select backend is enabled (which is enabled by default)
2341\& ev_poll.c only when poll backend is enabled (disabled by default)
2342\& ev_epoll.c only when the epoll backend is enabled (disabled by default)
2343\& ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2344\& ev_port.c only when the solaris port backend is enabled (disabled by default)
2345.Ve
2346.PP
2347\&\fIev.c\fR includes the backend files directly when enabled, so you only need
2348to compile this single file.
2349.PP
2350\fI\s-1LIBEVENT\s0 \s-1COMPATIBILITY\s0 \s-1API\s0\fR
2351.IX Subsection "LIBEVENT COMPATIBILITY API"
2352.PP
2353To include the libevent compatibility \s-1API\s0, also include:
2354.PP
2355.Vb 1
2356\& #include "event.c"
2357.Ve
2358.PP
2359in the file including \fIev.c\fR, and:
2360.PP
2361.Vb 1
2362\& #include "event.h"
2363.Ve
2364.PP
2365in the files that want to use the libevent \s-1API\s0. This also includes \fIev.h\fR.
2366.PP
2367You need the following additional files for this:
2368.PP
2369.Vb 2
2370\& event.h
2371\& event.c
2372.Ve
2373.PP
2374\fI\s-1AUTOCONF\s0 \s-1SUPPORT\s0\fR
2375.IX Subsection "AUTOCONF SUPPORT"
2376.PP
2377Instead of using \f(CW\*(C`EV_STANDALONE=1\*(C'\fR and providing your config in
2378whatever way you want, you can also \f(CW\*(C`m4_include([libev.m4])\*(C'\fR in your
2379\&\fIconfigure.ac\fR and leave \f(CW\*(C`EV_STANDALONE\*(C'\fR undefined. \fIev.c\fR will then
2380include \fIconfig.h\fR and configure itself accordingly.
2381.PP
2382For this of course you need the m4 file:
2383.PP
2384.Vb 1
2385\& libev.m4
2386.Ve
2387.Sh "\s-1PREPROCESSOR\s0 \s-1SYMBOLS/MACROS\s0"
2388.IX Subsection "PREPROCESSOR SYMBOLS/MACROS"
2389Libev can be configured via a variety of preprocessor symbols you have to define
2390before including any of its files. The default is not to build for multiplicity
2391and only include the select backend.
2392.IP "\s-1EV_STANDALONE\s0" 4
2393.IX Item "EV_STANDALONE"
2394Must always be \f(CW1\fR if you do not use autoconf configuration, which
2395keeps libev from including \fIconfig.h\fR, and it also defines dummy
2396implementations for some libevent functions (such as logging, which is not
2397supported). It will also not define any of the structs usually found in
2398\&\fIevent.h\fR that are not directly supported by the libev core alone.
2399.IP "\s-1EV_USE_MONOTONIC\s0" 4
2400.IX Item "EV_USE_MONOTONIC"
2401If defined to be \f(CW1\fR, libev will try to detect the availability of the
2402monotonic clock option at both compiletime and runtime. Otherwise no use
2403of the monotonic clock option will be attempted. If you enable this, you
2404usually have to link against librt or something similar. Enabling it when
2405the functionality isn't available is safe, though, althoguh you have
2406to make sure you link against any libraries where the \f(CW\*(C`clock_gettime\*(C'\fR
2407function is hiding in (often \fI\-lrt\fR).
2408.IP "\s-1EV_USE_REALTIME\s0" 4
2409.IX Item "EV_USE_REALTIME"
2410If defined to be \f(CW1\fR, libev will try to detect the availability of the
2411realtime clock option at compiletime (and assume its availability at
2412runtime if successful). Otherwise no use of the realtime clock option will
2413be attempted. This effectively replaces \f(CW\*(C`gettimeofday\*(C'\fR by \f(CW\*(C`clock_get
2414(CLOCK_REALTIME, ...)\*(C'\fR and will not normally affect correctness. See tzhe note about libraries
2415in the description of \f(CW\*(C`EV_USE_MONOTONIC\*(C'\fR, though.
2416.IP "\s-1EV_USE_SELECT\s0" 4
2417.IX Item "EV_USE_SELECT"
2418If undefined or defined to be \f(CW1\fR, libev will compile in support for the
2419\&\f(CW\*(C`select\*(C'\fR(2) backend. No attempt at autodetection will be done: if no
2420other method takes over, select will be it. Otherwise the select backend
2421will not be compiled in.
2422.IP "\s-1EV_SELECT_USE_FD_SET\s0" 4
2423.IX Item "EV_SELECT_USE_FD_SET"
2424If defined to \f(CW1\fR, then the select backend will use the system \f(CW\*(C`fd_set\*(C'\fR
2425structure. This is useful if libev doesn't compile due to a missing
2426\&\f(CW\*(C`NFDBITS\*(C'\fR or \f(CW\*(C`fd_mask\*(C'\fR definition or it misguesses the bitset layout on
2427exotic systems. This usually limits the range of file descriptors to some
2428low limit such as 1024 or might have other limitations (winsocket only
2429allows 64 sockets). The \f(CW\*(C`FD_SETSIZE\*(C'\fR macro, set before compilation, might
2430influence the size of the \f(CW\*(C`fd_set\*(C'\fR used.
2431.IP "\s-1EV_SELECT_IS_WINSOCKET\s0" 4
2432.IX Item "EV_SELECT_IS_WINSOCKET"
2433When defined to \f(CW1\fR, the select backend will assume that
2434select/socket/connect etc. don't understand file descriptors but
2435wants osf handles on win32 (this is the case when the select to
2436be used is the winsock select). This means that it will call
2437\&\f(CW\*(C`_get_osfhandle\*(C'\fR on the fd to convert it to an \s-1OS\s0 handle. Otherwise,
2438it is assumed that all these functions actually work on fds, even
2439on win32. Should not be defined on non\-win32 platforms.
2440.IP "\s-1EV_USE_POLL\s0" 4
2441.IX Item "EV_USE_POLL"
2442If defined to be \f(CW1\fR, libev will compile in support for the \f(CW\*(C`poll\*(C'\fR(2)
2443backend. Otherwise it will be enabled on non\-win32 platforms. It
2444takes precedence over select.
2445.IP "\s-1EV_USE_EPOLL\s0" 4
2446.IX Item "EV_USE_EPOLL"
2447If defined to be \f(CW1\fR, libev will compile in support for the Linux
2448\&\f(CW\*(C`epoll\*(C'\fR(7) backend. Its availability will be detected at runtime,
2449otherwise another method will be used as fallback. This is the
2450preferred backend for GNU/Linux systems.
2451.IP "\s-1EV_USE_KQUEUE\s0" 4
2452.IX Item "EV_USE_KQUEUE"
2453If defined to be \f(CW1\fR, libev will compile in support for the \s-1BSD\s0 style
2454\&\f(CW\*(C`kqueue\*(C'\fR(2) backend. Its actual availability will be detected at runtime,
2455otherwise another method will be used as fallback. This is the preferred
2456backend for \s-1BSD\s0 and BSD-like systems, although on most BSDs kqueue only
2457supports some types of fds correctly (the only platform we found that
2458supports ptys for example was NetBSD), so kqueue might be compiled in, but
2459not be used unless explicitly requested. The best way to use it is to find
2460out whether kqueue supports your type of fd properly and use an embedded
2461kqueue loop.
2462.IP "\s-1EV_USE_PORT\s0" 4
2463.IX Item "EV_USE_PORT"
2464If defined to be \f(CW1\fR, libev will compile in support for the Solaris
246510 port style backend. Its availability will be detected at runtime,
2466otherwise another method will be used as fallback. This is the preferred
2467backend for Solaris 10 systems.
2468.IP "\s-1EV_USE_DEVPOLL\s0" 4
2469.IX Item "EV_USE_DEVPOLL"
2470reserved for future expansion, works like the \s-1USE\s0 symbols above.
2471.IP "\s-1EV_USE_INOTIFY\s0" 4
2472.IX Item "EV_USE_INOTIFY"
2473If defined to be \f(CW1\fR, libev will compile in support for the Linux inotify
2474interface to speed up \f(CW\*(C`ev_stat\*(C'\fR watchers. Its actual availability will
2475be detected at runtime.
2476.IP "\s-1EV_H\s0" 4
2477.IX Item "EV_H"
2478The name of the \fIev.h\fR header file used to include it. The default if
2479undefined is \f(CW\*(C`<ev.h>\*(C'\fR in \fIevent.h\fR and \f(CW"ev.h"\fR in \fIev.c\fR. This
2480can be used to virtually rename the \fIev.h\fR header file in case of conflicts.
2481.IP "\s-1EV_CONFIG_H\s0" 4
2482.IX Item "EV_CONFIG_H"
2483If \f(CW\*(C`EV_STANDALONE\*(C'\fR isn't \f(CW1\fR, this variable can be used to override
2484\&\fIev.c\fR's idea of where to find the \fIconfig.h\fR file, similarly to
2485\&\f(CW\*(C`EV_H\*(C'\fR, above.
2486.IP "\s-1EV_EVENT_H\s0" 4
2487.IX Item "EV_EVENT_H"
2488Similarly to \f(CW\*(C`EV_H\*(C'\fR, this macro can be used to override \fIevent.c\fR's idea
2489of how the \fIevent.h\fR header can be found.
2490.IP "\s-1EV_PROTOTYPES\s0" 4
2491.IX Item "EV_PROTOTYPES"
2492If defined to be \f(CW0\fR, then \fIev.h\fR will not define any function
2493prototypes, but still define all the structs and other symbols. This is
2494occasionally useful if you want to provide your own wrapper functions
2495around libev functions.
2496.IP "\s-1EV_MULTIPLICITY\s0" 4
2497.IX Item "EV_MULTIPLICITY"
2498If undefined or defined to \f(CW1\fR, then all event-loop-specific functions
2499will have the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument, and you can create
2500additional independent event loops. Otherwise there will be no support
2501for multiple event loops and there is no first event loop pointer
2502argument. Instead, all functions act on the single default loop.
2503.IP "\s-1EV_MINPRI\s0" 4
2504.IX Item "EV_MINPRI"
2505.PD 0
2506.IP "\s-1EV_MAXPRI\s0" 4
2507.IX Item "EV_MAXPRI"
2508.PD
2509The range of allowed priorities. \f(CW\*(C`EV_MINPRI\*(C'\fR must be smaller or equal to
2510\&\f(CW\*(C`EV_MAXPRI\*(C'\fR, but otherwise there are no non-obvious limitations. You can
2511provide for more priorities by overriding those symbols (usually defined
2512to be \f(CW\*(C`\-2\*(C'\fR and \f(CW2\fR, respectively).
2513.Sp
2514When doing priority-based operations, libev usually has to linearly search
2515all the priorities, so having many of them (hundreds) uses a lot of space
2516and time, so using the defaults of five priorities (\-2 .. +2) is usually
2517fine.
2518.Sp
2519If your embedding app does not need any priorities, defining these both to
2520\&\f(CW0\fR will save some memory and cpu.
2521.IP "\s-1EV_PERIODIC_ENABLE\s0" 4
2522.IX Item "EV_PERIODIC_ENABLE"
2523If undefined or defined to be \f(CW1\fR, then periodic timers are supported. If
2524defined to be \f(CW0\fR, then they are not. Disabling them saves a few kB of
2525code.
2526.IP "\s-1EV_IDLE_ENABLE\s0" 4
2527.IX Item "EV_IDLE_ENABLE"
2528If undefined or defined to be \f(CW1\fR, then idle watchers are supported. If
2529defined to be \f(CW0\fR, then they are not. Disabling them saves a few kB of
2530code.
2531.IP "\s-1EV_EMBED_ENABLE\s0" 4
2532.IX Item "EV_EMBED_ENABLE"
2533If undefined or defined to be \f(CW1\fR, then embed watchers are supported. If
2534defined to be \f(CW0\fR, then they are not.
2535.IP "\s-1EV_STAT_ENABLE\s0" 4
2536.IX Item "EV_STAT_ENABLE"
2537If undefined or defined to be \f(CW1\fR, then stat watchers are supported. If
2538defined to be \f(CW0\fR, then they are not.
2539.IP "\s-1EV_FORK_ENABLE\s0" 4
2540.IX Item "EV_FORK_ENABLE"
2541If undefined or defined to be \f(CW1\fR, then fork watchers are supported. If
2542defined to be \f(CW0\fR, then they are not.
2543.IP "\s-1EV_MINIMAL\s0" 4
2544.IX Item "EV_MINIMAL"
2545If you need to shave off some kilobytes of code at the expense of some
2546speed, define this symbol to \f(CW1\fR. Currently only used for gcc to override
2547some inlining decisions, saves roughly 30% codesize of amd64.
2548.IP "\s-1EV_PID_HASHSIZE\s0" 4
2549.IX Item "EV_PID_HASHSIZE"
2550\&\f(CW\*(C`ev_child\*(C'\fR watchers use a small hash table to distribute workload by
2551pid. The default size is \f(CW16\fR (or \f(CW1\fR with \f(CW\*(C`EV_MINIMAL\*(C'\fR), usually more
2552than enough. If you need to manage thousands of children you might want to
2553increase this value (\fImust\fR be a power of two).
2554.IP "\s-1EV_INOTIFY_HASHSIZE\s0" 4
2555.IX Item "EV_INOTIFY_HASHSIZE"
2556\&\f(CW\*(C`ev_staz\*(C'\fR watchers use a small hash table to distribute workload by
2557inotify watch id. The default size is \f(CW16\fR (or \f(CW1\fR with \f(CW\*(C`EV_MINIMAL\*(C'\fR),
2558usually more than enough. If you need to manage thousands of \f(CW\*(C`ev_stat\*(C'\fR
2559watchers you might want to increase this value (\fImust\fR be a power of
2560two).
2561.IP "\s-1EV_COMMON\s0" 4
2562.IX Item "EV_COMMON"
2563By default, all watchers have a \f(CW\*(C`void *data\*(C'\fR member. By redefining
2564this macro to a something else you can include more and other types of
2565members. You have to define it each time you include one of the files,
2566though, and it must be identical each time.
2567.Sp
2568For example, the perl \s-1EV\s0 module uses something like this:
2569.Sp
2570.Vb 3
2571\& #define EV_COMMON \e
2572\& SV *self; /* contains this struct */ \e
2573\& SV *cb_sv, *fh /* note no trailing ";" */
2574.Ve
2575.IP "\s-1EV_CB_DECLARE\s0 (type)" 4
2576.IX Item "EV_CB_DECLARE (type)"
2577.PD 0
2578.IP "\s-1EV_CB_INVOKE\s0 (watcher, revents)" 4
2579.IX Item "EV_CB_INVOKE (watcher, revents)"
2580.IP "ev_set_cb (ev, cb)" 4
2581.IX Item "ev_set_cb (ev, cb)"
2582.PD
2583Can be used to change the callback member declaration in each watcher,
2584and the way callbacks are invoked and set. Must expand to a struct member
2585definition and a statement, respectively. See the \fIev.v\fR header file for
2586their default definitions. One possible use for overriding these is to
2587avoid the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument in all cases, or to use
2588method calls instead of plain function calls in \*(C+.
2589.Sh "\s-1EXAMPLES\s0"
2590.IX Subsection "EXAMPLES"
2591For a real-world example of a program the includes libev
2592verbatim, you can have a look at the \s-1EV\s0 perl module
2593(<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
2594the \fIlibev/\fR subdirectory and includes them in the \fI\s-1EV/EVAPI\s0.h\fR (public
2595interface) and \fI\s-1EV\s0.xs\fR (implementation) files. Only the \fI\s-1EV\s0.xs\fR file
2596will be compiled. It is pretty complex because it provides its own header
2597file.
2598.Sp
2599The usage in rxvt-unicode is simpler. It has a \fIev_cpp.h\fR header file
2600that everybody includes and which overrides some configure choices:
2601.Sp
2602.Vb 9
2603\& #define EV_MINIMAL 1
2604\& #define EV_USE_POLL 0
2605\& #define EV_MULTIPLICITY 0
2606\& #define EV_PERIODIC_ENABLE 0
2607\& #define EV_STAT_ENABLE 0
2608\& #define EV_FORK_ENABLE 0
2609\& #define EV_CONFIG_H <config.h>
2610\& #define EV_MINPRI 0
2611\& #define EV_MAXPRI 0
2612.Ve
2613.Sp
2614.Vb 1
2615\& #include "ev++.h"
2616.Ve
2617.Sp
2618And a \fIev_cpp.C\fR implementation file that contains libev proper and is compiled:
2619.Sp
2620.Vb 2
2621\& #include "ev_cpp.h"
2622\& #include "ev.c"
2623.Ve
2624.SH "COMPLEXITIES"
2625.IX Header "COMPLEXITIES"
2626In this section the complexities of (many of) the algorithms used inside
2627libev will be explained. For complexity discussions about backends see the
2628documentation for \f(CW\*(C`ev_default_init\*(C'\fR.
2629.Sp
2630All of the following are about amortised time: If an array needs to be
2631extended, libev needs to realloc and move the whole array, but this
2632happens asymptotically never with higher number of elements, so O(1) might
2633mean it might do a lengthy realloc operation in rare cases, but on average
2634it is much faster and asymptotically approaches constant time.
2635.RS 4
2636.IP "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)" 4
2637.IX Item "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)"
2638This means that, when you have a watcher that triggers in one hour and
2639there are 100 watchers that would trigger before that then inserting will
2640have to skip those 100 watchers.
2641.IP "Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)" 4
2642.IX Item "Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers)"
2643That means that for changing a timer costs less than removing/adding them
2644as only the relative motion in the event queue has to be paid for.
2645.IP "Starting io/check/prepare/idle/signal/child watchers: O(1)" 4
2646.IX Item "Starting io/check/prepare/idle/signal/child watchers: O(1)"
2647These just add the watcher into an array or at the head of a list.
2648=item Stopping check/prepare/idle watchers: O(1)
2649.IP "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % \s-1EV_PID_HASHSIZE\s0))" 4
2650.IX Item "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))"
2651These watchers are stored in lists then need to be walked to find the
2652correct watcher to remove. The lists are usually short (you don't usually
2653have many watchers waiting for the same fd or signal).
2654.IP "Finding the next timer per loop iteration: O(1)" 4
2655.IX Item "Finding the next timer per loop iteration: O(1)"
2656.PD 0
2657.IP "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)" 4
2658.IX Item "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)"
2659.PD
2660A change means an I/O watcher gets started or stopped, which requires
2661libev to recalculate its status (and possibly tell the kernel).
2662.IP "Activating one watcher: O(1)" 4
2663.IX Item "Activating one watcher: O(1)"
2664.PD 0
2665.IP "Priority handling: O(number_of_priorities)" 4
2666.IX Item "Priority handling: O(number_of_priorities)"
2667.PD
2668Priorities are implemented by allocating some space for each
2669priority. When doing priority-based operations, libev usually has to
2670linearly search all the priorities.
2671.RE
2672.RS 4
958.SH "AUTHOR" 2673.SH "AUTHOR"
959.IX Header "AUTHOR" 2674.IX Header "AUTHOR"
960Marc Lehmann <libev@schmorp.de>. 2675Marc Lehmann <libev@schmorp.de>.

Diff Legend

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