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

Comparing libev/ev.3 (file contents):
Revision 1.8 by root, Fri Nov 23 15:26:08 2007 UTC vs.
Revision 1.59 by root, Tue Dec 25 07:16:53 2007 UTC

… …
126. ds Ae AE 126. ds Ae AE
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 "EV 1"
132.TH "<STANDARD INPUT>" 1 "2007-11-23" "perl v5.8.8" "User Contributed Perl Documentation" 132.TH EV 1 "2007-12-25" "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 "\s-1EXAMPLE\s0 \s-1PROGRAM\s0"
141.IX Subsection "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 occurring), 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
147(or thread) by executing the \fIevent loop\fR handler, and will then 210(or thread) by executing the \fIevent loop\fR handler, and will then
148communicate events via a callback mechanism. 211communicate events via a callback mechanism.
149.PP 212.PP
150You register interest in certain events by registering so-called \fIevent 213You register interest in certain events by registering so-called \fIevent
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 "\s-1FEATURES\s0"
155.IX Header "FEATURES" 218.IX Subsection "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 "\s-1CONVENTIONS\s0"
164.IX Header "CONVENTIONS" 234.IX Subsection "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. 241.Sh "\s-1TIME\s0 \s-1REPRESENTATION\s0"
172.SH "TIME REPRESENTATION"
173.IX Header "TIME REPRESENTATION" 242.IX Subsection "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 some floatingpoint value. Unlike the name
249component \f(CW\*(C`stamp\*(C'\fR might indicate, it is also used for time differences
250throughout libev.
179.SH "GLOBAL FUNCTIONS" 251.SH "GLOBAL FUNCTIONS"
180.IX Header "GLOBAL FUNCTIONS" 252.IX Header "GLOBAL FUNCTIONS"
181These functions can be called anytime, even before initialising the 253These functions can be called anytime, even before initialising the
182library in any way. 254library in any way.
183.IP "ev_tstamp ev_time ()" 4 255.IP "ev_tstamp ev_time ()" 4
184.IX Item "ev_tstamp ev_time ()" 256.IX Item "ev_tstamp ev_time ()"
185Returns the current time as libev would use it. Please note that the 257Returns the current time as libev would use it. Please note that the
186\&\f(CW\*(C`ev_now\*(C'\fR function is usually faster and also often returns the timestamp 258\&\f(CW\*(C`ev_now\*(C'\fR function is usually faster and also often returns the timestamp
187you actually want to know. 259you actually want to know.
260.IP "ev_sleep (ev_tstamp interval)" 4
261.IX Item "ev_sleep (ev_tstamp interval)"
262Sleep for the given interval: The current thread will be blocked until
263either it is interrupted or the given time interval has passed. Basically
264this is a subsecond-resolution \f(CW\*(C`sleep ()\*(C'\fR.
188.IP "int ev_version_major ()" 4 265.IP "int ev_version_major ()" 4
189.IX Item "int ev_version_major ()" 266.IX Item "int ev_version_major ()"
190.PD 0 267.PD 0
191.IP "int ev_version_minor ()" 4 268.IP "int ev_version_minor ()" 4
192.IX Item "int ev_version_minor ()" 269.IX Item "int ev_version_minor ()"
193.PD 270.PD
194You can find out the major and minor version numbers of the library 271You 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 272you 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 273\&\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 274symbols \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. 275version of the library your program was compiled against.
199.Sp 276.Sp
277These version numbers refer to the \s-1ABI\s0 version of the library, not the
278release version.
279.Sp
200Usually, it's a good idea to terminate if the major versions mismatch, 280Usually, it's a good idea to terminate if the major versions mismatch,
201as this indicates an incompatible change. Minor versions are usually 281as this indicates an incompatible change. Minor versions are usually
202compatible to older versions, so a larger minor version alone is usually 282compatible to older versions, so a larger minor version alone is usually
203not a problem. 283not a problem.
284.Sp
285Example: Make sure we haven't accidentally been linked against the wrong
286version.
287.Sp
288.Vb 3
289\& assert (("libev version mismatch",
290\& ev_version_major () == EV_VERSION_MAJOR
291\& && ev_version_minor () >= EV_VERSION_MINOR));
292.Ve
204.IP "unsigned int ev_supported_backends ()" 4 293.IP "unsigned int ev_supported_backends ()" 4
205.IX Item "unsigned int ev_supported_backends ()" 294.IX Item "unsigned int ev_supported_backends ()"
206Return the set of all backends (i.e. their corresponding \f(CW\*(C`EV_BACKEND_*\*(C'\fR 295Return the set of all backends (i.e. their corresponding \f(CW\*(C`EV_BACKEND_*\*(C'\fR
207value) compiled into this binary of libev (independent of their 296value) compiled into this binary of libev (independent of their
208availability on the system you are running on). See \f(CW\*(C`ev_default_loop\*(C'\fR for 297availability on the system you are running on). See \f(CW\*(C`ev_default_loop\*(C'\fR for
209a description of the set values. 298a description of the set values.
299.Sp
300Example: make sure we have the epoll method, because yeah this is cool and
301a must have and can we have a torrent of it please!!!11
302.Sp
303.Vb 2
304\& assert (("sorry, no epoll, no sex",
305\& ev_supported_backends () & EVBACKEND_EPOLL));
306.Ve
210.IP "unsigned int ev_recommended_backends ()" 4 307.IP "unsigned int ev_recommended_backends ()" 4
211.IX Item "unsigned int ev_recommended_backends ()" 308.IX Item "unsigned int ev_recommended_backends ()"
212Return the set of all backends compiled into this binary of libev and also 309Return the set of all backends compiled into this binary of libev and also
213recommended for this platform. This set is often smaller than the one 310recommended for this platform. This set is often smaller than the one
214returned by \f(CW\*(C`ev_supported_backends\*(C'\fR, as for example kqueue is broken on 311returned by \f(CW\*(C`ev_supported_backends\*(C'\fR, as for example kqueue is broken on
215most BSDs and will not be autodetected unless you explicitly request it 312most BSDs and will not be autodetected unless you explicitly request it
216(assuming you know what you are doing). This is the set of backends that 313(assuming you know what you are doing). This is the set of backends that
217libev will probe for if you specify no backends explicitly. 314libev will probe for if you specify no backends explicitly.
315.IP "unsigned int ev_embeddable_backends ()" 4
316.IX Item "unsigned int ev_embeddable_backends ()"
317Returns the set of backends that are embeddable in other event loops. This
318is the theoretical, all\-platform, value. To find which backends
319might be supported on the current system, you would need to look at
320\&\f(CW\*(C`ev_embeddable_backends () & ev_supported_backends ()\*(C'\fR, likewise for
321recommended ones.
322.Sp
323See the description of \f(CW\*(C`ev_embed\*(C'\fR watchers for more info.
218.IP "ev_set_allocator (void *(*cb)(void *ptr, long size))" 4 324.IP "ev_set_allocator (void *(*cb)(void *ptr, long size))" 4
219.IX Item "ev_set_allocator (void *(*cb)(void *ptr, long size))" 325.IX Item "ev_set_allocator (void *(*cb)(void *ptr, long size))"
220Sets the allocation function to use (the prototype is similar to the 326Sets the allocation function to use (the prototype is similar \- the
221realloc C function, the semantics are identical). It is used to allocate 327semantics is identical \- to the realloc C function). It is used to
222and free memory (no surprises here). If it returns zero when memory 328allocate and free memory (no surprises here). If it returns zero when
223needs to be allocated, the library might abort or take some potentially 329memory needs to be allocated, the library might abort or take some
224destructive action. The default is your system realloc function. 330potentially destructive action. The default is your system realloc
331function.
225.Sp 332.Sp
226You could override this function in high-availability programs to, say, 333You could override this function in high-availability programs to, say,
227free some memory if it cannot allocate memory, to use a special allocator, 334free some memory if it cannot allocate memory, to use a special allocator,
228or even to sleep a while and retry until some memory is available. 335or even to sleep a while and retry until some memory is available.
336.Sp
337Example: Replace the libev allocator with one that waits a bit and then
338retries).
339.Sp
340.Vb 6
341\& static void *
342\& persistent_realloc (void *ptr, size_t size)
343\& {
344\& for (;;)
345\& {
346\& void *newptr = realloc (ptr, size);
347.Ve
348.Sp
349.Vb 2
350\& if (newptr)
351\& return newptr;
352.Ve
353.Sp
354.Vb 3
355\& sleep (60);
356\& }
357\& }
358.Ve
359.Sp
360.Vb 2
361\& ...
362\& ev_set_allocator (persistent_realloc);
363.Ve
229.IP "ev_set_syserr_cb (void (*cb)(const char *msg));" 4 364.IP "ev_set_syserr_cb (void (*cb)(const char *msg));" 4
230.IX Item "ev_set_syserr_cb (void (*cb)(const char *msg));" 365.IX Item "ev_set_syserr_cb (void (*cb)(const char *msg));"
231Set the callback function to call on a retryable syscall error (such 366Set the callback function to call on a retryable syscall error (such
232as failed select, poll, epoll_wait). The message is a printable string 367as failed select, poll, epoll_wait). The message is a printable string
233indicating the system call or subsystem causing the problem. If this 368indicating the system call or subsystem causing the problem. If this
234callback is set, then libev will expect it to remedy the sitution, no 369callback is set, then libev will expect it to remedy the sitution, no
235matter what, when it returns. That is, libev will generally retry the 370matter what, when it returns. That is, libev will generally retry the
236requested operation, or, if the condition doesn't go away, do bad stuff 371requested operation, or, if the condition doesn't go away, do bad stuff
237(such as abort). 372(such as abort).
373.Sp
374Example: This is basically the same thing that libev does internally, too.
375.Sp
376.Vb 6
377\& static void
378\& fatal_error (const char *msg)
379\& {
380\& perror (msg);
381\& abort ();
382\& }
383.Ve
384.Sp
385.Vb 2
386\& ...
387\& ev_set_syserr_cb (fatal_error);
388.Ve
238.SH "FUNCTIONS CONTROLLING THE EVENT LOOP" 389.SH "FUNCTIONS CONTROLLING THE EVENT LOOP"
239.IX Header "FUNCTIONS CONTROLLING THE EVENT LOOP" 390.IX Header "FUNCTIONS CONTROLLING THE EVENT LOOP"
240An event loop is described by a \f(CW\*(C`struct ev_loop *\*(C'\fR. The library knows two 391An event loop is described by a \f(CW\*(C`struct ev_loop *\*(C'\fR. The library knows two
241types of such loops, the \fIdefault\fR loop, which supports signals and child 392types of such loops, the \fIdefault\fR loop, which supports signals and child
242events, and dynamically created loops which do not. 393events, and dynamically created loops which do not.
… …
274or setgid) then libev will \fInot\fR look at the environment variable 425or setgid) then libev will \fInot\fR look at the environment variable
275\&\f(CW\*(C`LIBEV_FLAGS\*(C'\fR. Otherwise (the default), this environment variable will 426\&\f(CW\*(C`LIBEV_FLAGS\*(C'\fR. Otherwise (the default), this environment variable will
276override the flags completely if it is found in the environment. This is 427override the flags completely if it is found in the environment. This is
277useful to try out specific backends to test their performance, or to work 428useful to try out specific backends to test their performance, or to work
278around bugs. 429around bugs.
430.ie n .IP """EVFLAG_FORKCHECK""" 4
431.el .IP "\f(CWEVFLAG_FORKCHECK\fR" 4
432.IX Item "EVFLAG_FORKCHECK"
433Instead of calling \f(CW\*(C`ev_default_fork\*(C'\fR or \f(CW\*(C`ev_loop_fork\*(C'\fR manually after
434a fork, you can also make libev check for a fork in each iteration by
435enabling this flag.
436.Sp
437This works by calling \f(CW\*(C`getpid ()\*(C'\fR on every iteration of the loop,
438and thus this might slow down your event loop if you do a lot of loop
439iterations and little real work, but is usually not noticeable (on my
440Linux system for example, \f(CW\*(C`getpid\*(C'\fR is actually a simple 5\-insn sequence
441without a syscall and thus \fIvery\fR fast, but my Linux system also has
442\&\f(CW\*(C`pthread_atfork\*(C'\fR which is even faster).
443.Sp
444The big advantage of this flag is that you can forget about fork (and
445forget about forgetting to tell libev about forking) when you use this
446flag.
447.Sp
448This flag setting cannot be overriden or specified in the \f(CW\*(C`LIBEV_FLAGS\*(C'\fR
449environment variable.
279.ie n .IP """EVBACKEND_SELECT"" (value 1, portable select backend)" 4 450.ie n .IP """EVBACKEND_SELECT"" (value 1, portable select backend)" 4
280.el .IP "\f(CWEVBACKEND_SELECT\fR (value 1, portable select backend)" 4 451.el .IP "\f(CWEVBACKEND_SELECT\fR (value 1, portable select backend)" 4
281.IX Item "EVBACKEND_SELECT (value 1, portable select backend)" 452.IX Item "EVBACKEND_SELECT (value 1, portable select backend)"
282This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as 453This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as
283libev tries to roll its own fd_set with no limits on the number of fds, 454libev tries to roll its own fd_set with no limits on the number of fds,
284but if that fails, expect a fairly low limit on the number of fds when 455but if that fails, expect a fairly low limit on the number of fds when
285using this backend. It doesn't scale too well (O(highest_fd)), but its usually 456using this backend. It doesn't scale too well (O(highest_fd)), but its
286the fastest backend for a low number of fds. 457usually the fastest backend for a low number of (low\-numbered :) fds.
458.Sp
459To get good performance out of this backend you need a high amount of
460parallelity (most of the file descriptors should be busy). If you are
461writing a server, you should \f(CW\*(C`accept ()\*(C'\fR in a loop to accept as many
462connections as possible during one iteration. You might also want to have
463a look at \f(CW\*(C`ev_set_io_collect_interval ()\*(C'\fR to increase the amount of
464readyness notifications you get per iteration.
287.ie n .IP """EVBACKEND_POLL"" (value 2, poll backend, available everywhere except on windows)" 4 465.ie n .IP """EVBACKEND_POLL"" (value 2, poll backend, available everywhere except on windows)" 4
288.el .IP "\f(CWEVBACKEND_POLL\fR (value 2, poll backend, available everywhere except on windows)" 4 466.el .IP "\f(CWEVBACKEND_POLL\fR (value 2, poll backend, available everywhere except on windows)" 4
289.IX Item "EVBACKEND_POLL (value 2, poll backend, available everywhere except on windows)" 467.IX Item "EVBACKEND_POLL (value 2, poll backend, available everywhere except on windows)"
290And this is your standard \fIpoll\fR\|(2) backend. It's more complicated than 468And this is your standard \fIpoll\fR\|(2) backend. It's more complicated
291select, but handles sparse fds better and has no artificial limit on the 469than select, but handles sparse fds better and has no artificial
292number of fds you can use (except it will slow down considerably with a 470limit on the number of fds you can use (except it will slow down
293lot of inactive fds). It scales similarly to select, i.e. O(total_fds). 471considerably with a lot of inactive fds). It scales similarly to select,
472i.e. O(total_fds). See the entry for \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR, above, for
473performance tips.
294.ie n .IP """EVBACKEND_EPOLL"" (value 4, Linux)" 4 474.ie n .IP """EVBACKEND_EPOLL"" (value 4, Linux)" 4
295.el .IP "\f(CWEVBACKEND_EPOLL\fR (value 4, Linux)" 4 475.el .IP "\f(CWEVBACKEND_EPOLL\fR (value 4, Linux)" 4
296.IX Item "EVBACKEND_EPOLL (value 4, Linux)" 476.IX Item "EVBACKEND_EPOLL (value 4, Linux)"
297For few fds, this backend is a bit little slower than poll and select, 477For few fds, this backend is a bit little slower than poll and select,
298but it scales phenomenally better. While poll and select usually scale like 478but it scales phenomenally better. While poll and select usually scale
299O(total_fds) where n is the total number of fds (or the highest fd), epoll scales 479like O(total_fds) where n is the total number of fds (or the highest fd),
300either O(1) or O(active_fds). 480epoll scales either O(1) or O(active_fds). The epoll design has a number
481of shortcomings, such as silently dropping events in some hard-to-detect
482cases and rewiring a syscall per fd change, no fork support and bad
483support for dup.
301.Sp 484.Sp
302While stopping and starting an I/O watcher in the same iteration will 485While stopping, setting and starting an I/O watcher in the same iteration
303result in some caching, there is still a syscall per such incident 486will result in some caching, there is still a syscall per such incident
304(because the fd could point to a different file description now), so its 487(because the fd could point to a different file description now), so its
305best to avoid that. Also, \fIdup()\fRed file descriptors might not work very 488best to avoid that. Also, \f(CW\*(C`dup ()\*(C'\fR'ed file descriptors might not work
306well if you register events for both fds. 489very well if you register events for both fds.
307.Sp 490.Sp
308Please note that epoll sometimes generates spurious notifications, so you 491Please note that epoll sometimes generates spurious notifications, so you
309need to use non-blocking I/O or other means to avoid blocking when no data 492need to use non-blocking I/O or other means to avoid blocking when no data
310(or space) is available. 493(or space) is available.
494.Sp
495Best performance from this backend is achieved by not unregistering all
496watchers for a file descriptor until it has been closed, if possible, i.e.
497keep at least one watcher active per fd at all times.
498.Sp
499While nominally embeddeble in other event loops, this feature is broken in
500all kernel versions tested so far.
311.ie n .IP """EVBACKEND_KQUEUE"" (value 8, most \s-1BSD\s0 clones)" 4 501.ie n .IP """EVBACKEND_KQUEUE"" (value 8, most \s-1BSD\s0 clones)" 4
312.el .IP "\f(CWEVBACKEND_KQUEUE\fR (value 8, most \s-1BSD\s0 clones)" 4 502.el .IP "\f(CWEVBACKEND_KQUEUE\fR (value 8, most \s-1BSD\s0 clones)" 4
313.IX Item "EVBACKEND_KQUEUE (value 8, most BSD clones)" 503.IX Item "EVBACKEND_KQUEUE (value 8, most BSD clones)"
314Kqueue deserves special mention, as at the time of this writing, it 504Kqueue deserves special mention, as at the time of this writing, it
315was broken on all BSDs except NetBSD (usually it doesn't work with 505was broken on all BSDs except NetBSD (usually it doesn't work reliably
316anything but sockets and pipes, except on Darwin, where of course its 506with anything but sockets and pipes, except on Darwin, where of course
317completely useless). For this reason its not being \*(L"autodetected\*(R" 507it's completely useless). For this reason it's not being \*(L"autodetected\*(R"
318unless you explicitly specify it explicitly in the flags (i.e. using 508unless you explicitly specify it explicitly in the flags (i.e. using
319\&\f(CW\*(C`EVBACKEND_KQUEUE\*(C'\fR). 509\&\f(CW\*(C`EVBACKEND_KQUEUE\*(C'\fR) or libev was compiled on a known-to-be-good (\-enough)
510system like NetBSD.
511.Sp
512You still can embed kqueue into a normal poll or select backend and use it
513only for sockets (after having made sure that sockets work with kqueue on
514the target platform). See \f(CW\*(C`ev_embed\*(C'\fR watchers for more info.
320.Sp 515.Sp
321It scales in the same way as the epoll backend, but the interface to the 516It scales in the same way as the epoll backend, but the interface to the
322kernel is more efficient (which says nothing about its actual speed, of 517kernel is more efficient (which says nothing about its actual speed, of
323course). While starting and stopping an I/O watcher does not cause an 518course). While stopping, setting and starting an I/O watcher does never
324extra syscall as with epoll, it still adds up to four event changes per 519cause an extra syscall as with \f(CW\*(C`EVBACKEND_EPOLL\*(C'\fR, it still adds up to
325incident, so its best to avoid that. 520two event changes per incident, support for \f(CW\*(C`fork ()\*(C'\fR is very bad and it
521drops fds silently in similarly hard-to-detect cases.
522.Sp
523This backend usually performs well under most conditions.
524.Sp
525While nominally embeddable in other event loops, this doesn't work
526everywhere, so you might need to test for this. And since it is broken
527almost everywhere, you should only use it when you have a lot of sockets
528(for which it usually works), by embedding it into another event loop
529(e.g. \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR or \f(CW\*(C`EVBACKEND_POLL\*(C'\fR) and using it only for
530sockets.
326.ie n .IP """EVBACKEND_DEVPOLL"" (value 16, Solaris 8)" 4 531.ie n .IP """EVBACKEND_DEVPOLL"" (value 16, Solaris 8)" 4
327.el .IP "\f(CWEVBACKEND_DEVPOLL\fR (value 16, Solaris 8)" 4 532.el .IP "\f(CWEVBACKEND_DEVPOLL\fR (value 16, Solaris 8)" 4
328.IX Item "EVBACKEND_DEVPOLL (value 16, Solaris 8)" 533.IX Item "EVBACKEND_DEVPOLL (value 16, Solaris 8)"
329This is not implemented yet (and might never be). 534This is not implemented yet (and might never be, unless you send me an
535implementation). According to reports, \f(CW\*(C`/dev/poll\*(C'\fR only supports sockets
536and is not embeddable, which would limit the usefulness of this backend
537immensely.
330.ie n .IP """EVBACKEND_PORT"" (value 32, Solaris 10)" 4 538.ie n .IP """EVBACKEND_PORT"" (value 32, Solaris 10)" 4
331.el .IP "\f(CWEVBACKEND_PORT\fR (value 32, Solaris 10)" 4 539.el .IP "\f(CWEVBACKEND_PORT\fR (value 32, Solaris 10)" 4
332.IX Item "EVBACKEND_PORT (value 32, Solaris 10)" 540.IX Item "EVBACKEND_PORT (value 32, Solaris 10)"
333This uses the Solaris 10 port mechanism. As with everything on Solaris, 541This uses the Solaris 10 event port mechanism. As with everything on Solaris,
334it's really slow, but it still scales very well (O(active_fds)). 542it's really slow, but it still scales very well (O(active_fds)).
335.Sp 543.Sp
336Please note that solaris ports can result in a lot of spurious 544Please note that solaris event ports can deliver a lot of spurious
337notifications, so you need to use non-blocking I/O or other means to avoid 545notifications, so you need to use non-blocking I/O or other means to avoid
338blocking when no data (or space) is available. 546blocking when no data (or space) is available.
547.Sp
548While this backend scales well, it requires one system call per active
549file descriptor per loop iteration. For small and medium numbers of file
550descriptors a \*(L"slow\*(R" \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR or \f(CW\*(C`EVBACKEND_POLL\*(C'\fR backend
551might perform better.
339.ie n .IP """EVBACKEND_ALL""" 4 552.ie n .IP """EVBACKEND_ALL""" 4
340.el .IP "\f(CWEVBACKEND_ALL\fR" 4 553.el .IP "\f(CWEVBACKEND_ALL\fR" 4
341.IX Item "EVBACKEND_ALL" 554.IX Item "EVBACKEND_ALL"
342Try all backends (even potentially broken ones that wouldn't be tried 555Try all backends (even potentially broken ones that wouldn't be tried
343with \f(CW\*(C`EVFLAG_AUTO\*(C'\fR). Since this is a mask, you can do stuff such as 556with \f(CW\*(C`EVFLAG_AUTO\*(C'\fR). Since this is a mask, you can do stuff such as
344\&\f(CW\*(C`EVBACKEND_ALL & ~EVBACKEND_KQUEUE\*(C'\fR. 557\&\f(CW\*(C`EVBACKEND_ALL & ~EVBACKEND_KQUEUE\*(C'\fR.
558.Sp
559It is definitely not recommended to use this flag.
345.RE 560.RE
346.RS 4 561.RS 4
347.Sp 562.Sp
348If one or more of these are ored into the flags value, then only these 563If one or more of these are ored into the flags value, then only these
349backends will be tried (in the reverse order as given here). If none are 564backends will be tried (in the reverse order as given here). If none are
… …
376.IX Item "struct ev_loop *ev_loop_new (unsigned int flags)" 591.IX Item "struct ev_loop *ev_loop_new (unsigned int flags)"
377Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is 592Similar to \f(CW\*(C`ev_default_loop\*(C'\fR, but always creates a new event loop that is
378always distinct from the default loop. Unlike the default loop, it cannot 593always distinct from the default loop. Unlike the default loop, it cannot
379handle signal and child watchers, and attempts to do so will be greeted by 594handle signal and child watchers, and attempts to do so will be greeted by
380undefined behaviour (or a failed assertion if assertions are enabled). 595undefined behaviour (or a failed assertion if assertions are enabled).
596.Sp
597Example: Try to create a event loop that uses epoll and nothing else.
598.Sp
599.Vb 3
600\& struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
601\& if (!epoller)
602\& fatal ("no epoll found here, maybe it hides under your chair");
603.Ve
381.IP "ev_default_destroy ()" 4 604.IP "ev_default_destroy ()" 4
382.IX Item "ev_default_destroy ()" 605.IX Item "ev_default_destroy ()"
383Destroys the default loop again (frees all memory and kernel state 606Destroys the default loop again (frees all memory and kernel state
384etc.). This stops all registered event watchers (by not touching them in 607etc.). None of the active event watchers will be stopped in the normal
385any way whatsoever, although you cannot rely on this :). 608sense, so e.g. \f(CW\*(C`ev_is_active\*(C'\fR might still return true. It is your
609responsibility to either stop all watchers cleanly yoursef \fIbefore\fR
610calling this function, or cope with the fact afterwards (which is usually
611the easiest thing, you can just ignore the watchers and/or \f(CW\*(C`free ()\*(C'\fR them
612for example).
613.Sp
614Note that certain global state, such as signal state, will not be freed by
615this function, and related watchers (such as signal and child watchers)
616would need to be stopped manually.
617.Sp
618In general it is not advisable to call this function except in the
619rare occasion where you really need to free e.g. the signal handling
620pipe fds. If you need dynamically allocated loops it is better to use
621\&\f(CW\*(C`ev_loop_new\*(C'\fR and \f(CW\*(C`ev_loop_destroy\*(C'\fR).
386.IP "ev_loop_destroy (loop)" 4 622.IP "ev_loop_destroy (loop)" 4
387.IX Item "ev_loop_destroy (loop)" 623.IX Item "ev_loop_destroy (loop)"
388Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an 624Like \f(CW\*(C`ev_default_destroy\*(C'\fR, but destroys an event loop created by an
389earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR. 625earlier call to \f(CW\*(C`ev_loop_new\*(C'\fR.
390.IP "ev_default_fork ()" 4 626.IP "ev_default_fork ()" 4
… …
412.IP "ev_loop_fork (loop)" 4 648.IP "ev_loop_fork (loop)" 4
413.IX Item "ev_loop_fork (loop)" 649.IX Item "ev_loop_fork (loop)"
414Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by 650Like \f(CW\*(C`ev_default_fork\*(C'\fR, but acts on an event loop created by
415\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop 651\&\f(CW\*(C`ev_loop_new\*(C'\fR. Yes, you have to call this on every allocated event loop
416after fork, and how you do this is entirely your own problem. 652after fork, and how you do this is entirely your own problem.
653.IP "unsigned int ev_loop_count (loop)" 4
654.IX Item "unsigned int ev_loop_count (loop)"
655Returns the count of loop iterations for the loop, which is identical to
656the number of times libev did poll for new events. It starts at \f(CW0\fR and
657happily wraps around with enough iterations.
658.Sp
659This value can sometimes be useful as a generation counter of sorts (it
660\&\*(L"ticks\*(R" the number of loop iterations), as it roughly corresponds with
661\&\f(CW\*(C`ev_prepare\*(C'\fR and \f(CW\*(C`ev_check\*(C'\fR calls.
417.IP "unsigned int ev_backend (loop)" 4 662.IP "unsigned int ev_backend (loop)" 4
418.IX Item "unsigned int ev_backend (loop)" 663.IX Item "unsigned int ev_backend (loop)"
419Returns one of the \f(CW\*(C`EVBACKEND_*\*(C'\fR flags indicating the event backend in 664Returns one of the \f(CW\*(C`EVBACKEND_*\*(C'\fR flags indicating the event backend in
420use. 665use.
421.IP "ev_tstamp ev_now (loop)" 4 666.IP "ev_tstamp ev_now (loop)" 4
422.IX Item "ev_tstamp ev_now (loop)" 667.IX Item "ev_tstamp ev_now (loop)"
423Returns the current \*(L"event loop time\*(R", which is the time the event loop 668Returns the current \*(L"event loop time\*(R", which is the time the event loop
424got events and started processing them. This timestamp does not change 669received events and started processing them. This timestamp does not
425as long as callbacks are being processed, and this is also the base time 670change as long as callbacks are being processed, and this is also the base
426used for relative timers. You can treat it as the timestamp of the event 671time used for relative timers. You can treat it as the timestamp of the
427occuring (or more correctly, the mainloop finding out about it). 672event occurring (or more correctly, libev finding out about it).
428.IP "ev_loop (loop, int flags)" 4 673.IP "ev_loop (loop, int flags)" 4
429.IX Item "ev_loop (loop, int flags)" 674.IX Item "ev_loop (loop, int flags)"
430Finally, this is it, the event handler. This function usually is called 675Finally, this is it, the event handler. This function usually is called
431after you initialised all your watchers and you want to start handling 676after you initialised all your watchers and you want to start handling
432events. 677events.
433.Sp 678.Sp
434If the flags argument is specified as \f(CW0\fR, it will not return until 679If the flags argument is specified as \f(CW0\fR, it will not return until
435either no event watchers are active anymore or \f(CW\*(C`ev_unloop\*(C'\fR was called. 680either no event watchers are active anymore or \f(CW\*(C`ev_unloop\*(C'\fR was called.
681.Sp
682Please note that an explicit \f(CW\*(C`ev_unloop\*(C'\fR is usually better than
683relying on all watchers to be stopped when deciding when a program has
684finished (especially in interactive programs), but having a program that
685automatically loops as long as it has to and no longer by virtue of
686relying on its watchers stopping correctly is a thing of beauty.
436.Sp 687.Sp
437A flags value of \f(CW\*(C`EVLOOP_NONBLOCK\*(C'\fR will look for new events, will handle 688A flags value of \f(CW\*(C`EVLOOP_NONBLOCK\*(C'\fR will look for new events, will handle
438those events and any outstanding ones, but will not block your process in 689those events and any outstanding ones, but will not block your process in
439case there are no events and will return after one iteration of the loop. 690case there are no events and will return after one iteration of the loop.
440.Sp 691.Sp
… …
446libev watchers. However, a pair of \f(CW\*(C`ev_prepare\*(C'\fR/\f(CW\*(C`ev_check\*(C'\fR watchers is 697libev watchers. However, a pair of \f(CW\*(C`ev_prepare\*(C'\fR/\f(CW\*(C`ev_check\*(C'\fR watchers is
447usually a better approach for this kind of thing. 698usually a better approach for this kind of thing.
448.Sp 699.Sp
449Here are the gory details of what \f(CW\*(C`ev_loop\*(C'\fR does: 700Here are the gory details of what \f(CW\*(C`ev_loop\*(C'\fR does:
450.Sp 701.Sp
451.Vb 18 702.Vb 19
703\& - Before the first iteration, call any pending watchers.
452\& * If there are no active watchers (reference count is zero), return. 704\& * If there are no active watchers (reference count is zero), return.
453\& - Queue prepare watchers and then call all outstanding watchers. 705\& - Queue all prepare watchers and then call all outstanding watchers.
454\& - If we have been forked, recreate the kernel state. 706\& - If we have been forked, recreate the kernel state.
455\& - Update the kernel state with all outstanding changes. 707\& - Update the kernel state with all outstanding changes.
456\& - Update the "event loop time". 708\& - Update the "event loop time".
457\& - Calculate for how long to block. 709\& - Calculate for how long to block.
458\& - Block the process, waiting for any events. 710\& - Block the process, waiting for any events.
… …
465\& - Call all queued watchers in reverse order (i.e. check watchers first). 717\& - Call all queued watchers in reverse order (i.e. check watchers first).
466\& Signals and child watchers are implemented as I/O watchers, and will 718\& Signals and child watchers are implemented as I/O watchers, and will
467\& be handled here by queueing them when their watcher gets executed. 719\& be handled here by queueing them when their watcher gets executed.
468\& - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK 720\& - If ev_unloop has been called or EVLOOP_ONESHOT or EVLOOP_NONBLOCK
469\& were used, return, otherwise continue with step *. 721\& were used, return, otherwise continue with step *.
722.Ve
723.Sp
724Example: Queue some jobs and then loop until no events are outsanding
725anymore.
726.Sp
727.Vb 4
728\& ... queue jobs here, make sure they register event watchers as long
729\& ... as they still have work to do (even an idle watcher will do..)
730\& ev_loop (my_loop, 0);
731\& ... jobs done. yeah!
470.Ve 732.Ve
471.IP "ev_unloop (loop, how)" 4 733.IP "ev_unloop (loop, how)" 4
472.IX Item "ev_unloop (loop, how)" 734.IX Item "ev_unloop (loop, how)"
473Can be used to make a call to \f(CW\*(C`ev_loop\*(C'\fR return early (but only after it 735Can be used to make a call to \f(CW\*(C`ev_loop\*(C'\fR return early (but only after it
474has processed all outstanding events). The \f(CW\*(C`how\*(C'\fR argument must be either 736has processed all outstanding events). The \f(CW\*(C`how\*(C'\fR argument must be either
… …
488example, libev itself uses this for its internal signal pipe: It is not 750example, libev itself uses this for its internal signal pipe: It is not
489visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if 751visible to the libev user and should not keep \f(CW\*(C`ev_loop\*(C'\fR from exiting if
490no event watchers registered by it are active. It is also an excellent 752no event watchers registered by it are active. It is also an excellent
491way to do this for generic recurring timers or from within third-party 753way to do this for generic recurring timers or from within third-party
492libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR. 754libraries. Just remember to \fIunref after start\fR and \fIref before stop\fR.
755.Sp
756Example: Create a signal watcher, but keep it from keeping \f(CW\*(C`ev_loop\*(C'\fR
757running when nothing else is active.
758.Sp
759.Vb 4
760\& struct ev_signal exitsig;
761\& ev_signal_init (&exitsig, sig_cb, SIGINT);
762\& ev_signal_start (loop, &exitsig);
763\& evf_unref (loop);
764.Ve
765.Sp
766Example: For some weird reason, unregister the above signal handler again.
767.Sp
768.Vb 2
769\& ev_ref (loop);
770\& ev_signal_stop (loop, &exitsig);
771.Ve
772.IP "ev_set_io_collect_interval (loop, ev_tstamp interval)" 4
773.IX Item "ev_set_io_collect_interval (loop, ev_tstamp interval)"
774.PD 0
775.IP "ev_set_timeout_collect_interval (loop, ev_tstamp interval)" 4
776.IX Item "ev_set_timeout_collect_interval (loop, ev_tstamp interval)"
777.PD
778These advanced functions influence the time that libev will spend waiting
779for events. Both are by default \f(CW0\fR, meaning that libev will try to
780invoke timer/periodic callbacks and I/O callbacks with minimum latency.
781.Sp
782Setting these to a higher value (the \f(CW\*(C`interval\*(C'\fR \fImust\fR be >= \f(CW0\fR)
783allows libev to delay invocation of I/O and timer/periodic callbacks to
784increase efficiency of loop iterations.
785.Sp
786The background is that sometimes your program runs just fast enough to
787handle one (or very few) event(s) per loop iteration. While this makes
788the program responsive, it also wastes a lot of \s-1CPU\s0 time to poll for new
789events, especially with backends like \f(CW\*(C`select ()\*(C'\fR which have a high
790overhead for the actual polling but can deliver many events at once.
791.Sp
792By setting a higher \fIio collect interval\fR you allow libev to spend more
793time collecting I/O events, so you can handle more events per iteration,
794at the cost of increasing latency. Timeouts (both \f(CW\*(C`ev_periodic\*(C'\fR and
795\&\f(CW\*(C`ev_timer\*(C'\fR) will be not affected. Setting this to a non-null value will
796introduce an additional \f(CW\*(C`ev_sleep ()\*(C'\fR call into most loop iterations.
797.Sp
798Likewise, by setting a higher \fItimeout collect interval\fR you allow libev
799to spend more time collecting timeouts, at the expense of increased
800latency (the watcher callback will be called later). \f(CW\*(C`ev_io\*(C'\fR watchers
801will not be affected. Setting this to a non-null value will not introduce
802any overhead in libev.
803.Sp
804Many (busy) programs can usually benefit by setting the io collect
805interval to a value near \f(CW0.1\fR or so, which is often enough for
806interactive servers (of course not for games), likewise for timeouts. It
807usually doesn't make much sense to set it to a lower value than \f(CW0.01\fR,
808as this approsaches the timing granularity of most systems.
493.SH "ANATOMY OF A WATCHER" 809.SH "ANATOMY OF A WATCHER"
494.IX Header "ANATOMY OF A WATCHER" 810.IX Header "ANATOMY OF A WATCHER"
495A watcher is a structure that you create and register to record your 811A watcher is a structure that you create and register to record your
496interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to 812interest in some event. For instance, if you want to wait for \s-1STDIN\s0 to
497become readable, you would create an \f(CW\*(C`ev_io\*(C'\fR watcher for that: 813become readable, you would create an \f(CW\*(C`ev_io\*(C'\fR watcher for that:
… …
533*)\*(C'\fR), and you can stop watching for events at any time by calling the 849*)\*(C'\fR), and you can stop watching for events at any time by calling the
534corresponding stop function (\f(CW\*(C`ev_<type>_stop (loop, watcher *)\*(C'\fR. 850corresponding stop function (\f(CW\*(C`ev_<type>_stop (loop, watcher *)\*(C'\fR.
535.PP 851.PP
536As long as your watcher is active (has been started but not stopped) you 852As long as your watcher is active (has been started but not stopped) you
537must not touch the values stored in it. Most specifically you must never 853must not touch the values stored in it. Most specifically you must never
538reinitialise it or call its set macro. 854reinitialise it or call its \f(CW\*(C`set\*(C'\fR macro.
539.PP
540You can check whether an event is active by calling the \f(CW\*(C`ev_is_active
541(watcher *)\*(C'\fR macro. To see whether an event is outstanding (but the
542callback for it has not been called yet) you can use the \f(CW\*(C`ev_is_pending
543(watcher *)\*(C'\fR macro.
544.PP 855.PP
545Each and every callback receives the event loop pointer as first, the 856Each and every callback receives the event loop pointer as first, the
546registered watcher structure as second, and a bitset of received events as 857registered watcher structure as second, and a bitset of received events as
547third argument. 858third argument.
548.PP 859.PP
… …
573The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread. 884The signal specified in the \f(CW\*(C`ev_signal\*(C'\fR watcher has been received by a thread.
574.ie n .IP """EV_CHILD""" 4 885.ie n .IP """EV_CHILD""" 4
575.el .IP "\f(CWEV_CHILD\fR" 4 886.el .IP "\f(CWEV_CHILD\fR" 4
576.IX Item "EV_CHILD" 887.IX Item "EV_CHILD"
577The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change. 888The pid specified in the \f(CW\*(C`ev_child\*(C'\fR watcher has received a status change.
889.ie n .IP """EV_STAT""" 4
890.el .IP "\f(CWEV_STAT\fR" 4
891.IX Item "EV_STAT"
892The path specified in the \f(CW\*(C`ev_stat\*(C'\fR watcher changed its attributes somehow.
578.ie n .IP """EV_IDLE""" 4 893.ie n .IP """EV_IDLE""" 4
579.el .IP "\f(CWEV_IDLE\fR" 4 894.el .IP "\f(CWEV_IDLE\fR" 4
580.IX Item "EV_IDLE" 895.IX Item "EV_IDLE"
581The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do. 896The \f(CW\*(C`ev_idle\*(C'\fR watcher has determined that you have nothing better to do.
582.ie n .IP """EV_PREPARE""" 4 897.ie n .IP """EV_PREPARE""" 4
… …
592\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any 907\&\f(CW\*(C`ev_loop\*(C'\fR has gathered them, but before it invokes any callbacks for any
593received events. Callbacks of both watcher types can start and stop as 908received events. Callbacks of both watcher types can start and stop as
594many watchers as they want, and all of them will be taken into account 909many watchers as they want, and all of them will be taken into account
595(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep 910(for example, a \f(CW\*(C`ev_prepare\*(C'\fR watcher might start an idle watcher to keep
596\&\f(CW\*(C`ev_loop\*(C'\fR from blocking). 911\&\f(CW\*(C`ev_loop\*(C'\fR from blocking).
912.ie n .IP """EV_EMBED""" 4
913.el .IP "\f(CWEV_EMBED\fR" 4
914.IX Item "EV_EMBED"
915The embedded event loop specified in the \f(CW\*(C`ev_embed\*(C'\fR watcher needs attention.
916.ie n .IP """EV_FORK""" 4
917.el .IP "\f(CWEV_FORK\fR" 4
918.IX Item "EV_FORK"
919The event loop has been resumed in the child process after fork (see
920\&\f(CW\*(C`ev_fork\*(C'\fR).
597.ie n .IP """EV_ERROR""" 4 921.ie n .IP """EV_ERROR""" 4
598.el .IP "\f(CWEV_ERROR\fR" 4 922.el .IP "\f(CWEV_ERROR\fR" 4
599.IX Item "EV_ERROR" 923.IX Item "EV_ERROR"
600An unspecified error has occured, the watcher has been stopped. This might 924An unspecified error has occured, the watcher has been stopped. This might
601happen because the watcher could not be properly started because libev 925happen because the watcher could not be properly started because libev
… …
606Libev will usually signal a few \*(L"dummy\*(R" events together with an error, 930Libev will usually signal a few \*(L"dummy\*(R" events together with an error,
607for example it might indicate that a fd is readable or writable, and if 931for example it might indicate that a fd is readable or writable, and if
608your callbacks is well-written it can just attempt the operation and cope 932your callbacks is well-written it can just attempt the operation and cope
609with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded 933with the error from \fIread()\fR or \fIwrite()\fR. This will not work in multithreaded
610programs, though, so beware. 934programs, though, so beware.
935.Sh "\s-1GENERIC\s0 \s-1WATCHER\s0 \s-1FUNCTIONS\s0"
936.IX Subsection "GENERIC WATCHER FUNCTIONS"
937In the following description, \f(CW\*(C`TYPE\*(C'\fR stands for the watcher type,
938e.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.
939.ie n .IP """ev_init"" (ev_TYPE *watcher, callback)" 4
940.el .IP "\f(CWev_init\fR (ev_TYPE *watcher, callback)" 4
941.IX Item "ev_init (ev_TYPE *watcher, callback)"
942This macro initialises the generic portion of a watcher. The contents
943of the watcher object can be arbitrary (so \f(CW\*(C`malloc\*(C'\fR will do). Only
944the generic parts of the watcher are initialised, you \fIneed\fR to call
945the type-specific \f(CW\*(C`ev_TYPE_set\*(C'\fR macro afterwards to initialise the
946type-specific parts. For each type there is also a \f(CW\*(C`ev_TYPE_init\*(C'\fR macro
947which rolls both calls into one.
948.Sp
949You can reinitialise a watcher at any time as long as it has been stopped
950(or never started) and there are no pending events outstanding.
951.Sp
952The callback is always of type \f(CW\*(C`void (*)(ev_loop *loop, ev_TYPE *watcher,
953int revents)\*(C'\fR.
954.ie n .IP """ev_TYPE_set"" (ev_TYPE *, [args])" 4
955.el .IP "\f(CWev_TYPE_set\fR (ev_TYPE *, [args])" 4
956.IX Item "ev_TYPE_set (ev_TYPE *, [args])"
957This macro initialises the type-specific parts of a watcher. You need to
958call \f(CW\*(C`ev_init\*(C'\fR at least once before you call this macro, but you can
959call \f(CW\*(C`ev_TYPE_set\*(C'\fR any number of times. You must not, however, call this
960macro on a watcher that is active (it can be pending, however, which is a
961difference to the \f(CW\*(C`ev_init\*(C'\fR macro).
962.Sp
963Although some watcher types do not have type-specific arguments
964(e.g. \f(CW\*(C`ev_prepare\*(C'\fR) you still need to call its \f(CW\*(C`set\*(C'\fR macro.
965.ie n .IP """ev_TYPE_init"" (ev_TYPE *watcher, callback, [args])" 4
966.el .IP "\f(CWev_TYPE_init\fR (ev_TYPE *watcher, callback, [args])" 4
967.IX Item "ev_TYPE_init (ev_TYPE *watcher, callback, [args])"
968This convinience macro rolls both \f(CW\*(C`ev_init\*(C'\fR and \f(CW\*(C`ev_TYPE_set\*(C'\fR macro
969calls into a single call. This is the most convinient method to initialise
970a watcher. The same limitations apply, of course.
971.ie n .IP """ev_TYPE_start"" (loop *, ev_TYPE *watcher)" 4
972.el .IP "\f(CWev_TYPE_start\fR (loop *, ev_TYPE *watcher)" 4
973.IX Item "ev_TYPE_start (loop *, ev_TYPE *watcher)"
974Starts (activates) the given watcher. Only active watchers will receive
975events. If the watcher is already active nothing will happen.
976.ie n .IP """ev_TYPE_stop"" (loop *, ev_TYPE *watcher)" 4
977.el .IP "\f(CWev_TYPE_stop\fR (loop *, ev_TYPE *watcher)" 4
978.IX Item "ev_TYPE_stop (loop *, ev_TYPE *watcher)"
979Stops the given watcher again (if active) and clears the pending
980status. It is possible that stopped watchers are pending (for example,
981non-repeating timers are being stopped when they become pending), but
982\&\f(CW\*(C`ev_TYPE_stop\*(C'\fR ensures that the watcher is neither active nor pending. If
983you want to free or reuse the memory used by the watcher it is therefore a
984good idea to always call its \f(CW\*(C`ev_TYPE_stop\*(C'\fR function.
985.IP "bool ev_is_active (ev_TYPE *watcher)" 4
986.IX Item "bool ev_is_active (ev_TYPE *watcher)"
987Returns a true value iff the watcher is active (i.e. it has been started
988and not yet been stopped). As long as a watcher is active you must not modify
989it.
990.IP "bool ev_is_pending (ev_TYPE *watcher)" 4
991.IX Item "bool ev_is_pending (ev_TYPE *watcher)"
992Returns a true value iff the watcher is pending, (i.e. it has outstanding
993events but its callback has not yet been invoked). As long as a watcher
994is pending (but not active) you must not call an init function on it (but
995\&\f(CW\*(C`ev_TYPE_set\*(C'\fR is safe), you must not change its priority, and you must
996make sure the watcher is available to libev (e.g. you cannot \f(CW\*(C`free ()\*(C'\fR
997it).
998.IP "callback ev_cb (ev_TYPE *watcher)" 4
999.IX Item "callback ev_cb (ev_TYPE *watcher)"
1000Returns the callback currently set on the watcher.
1001.IP "ev_cb_set (ev_TYPE *watcher, callback)" 4
1002.IX Item "ev_cb_set (ev_TYPE *watcher, callback)"
1003Change the callback. You can change the callback at virtually any time
1004(modulo threads).
1005.IP "ev_set_priority (ev_TYPE *watcher, priority)" 4
1006.IX Item "ev_set_priority (ev_TYPE *watcher, priority)"
1007.PD 0
1008.IP "int ev_priority (ev_TYPE *watcher)" 4
1009.IX Item "int ev_priority (ev_TYPE *watcher)"
1010.PD
1011Set and query the priority of the watcher. The priority is a small
1012integer between \f(CW\*(C`EV_MAXPRI\*(C'\fR (default: \f(CW2\fR) and \f(CW\*(C`EV_MINPRI\*(C'\fR
1013(default: \f(CW\*(C`\-2\*(C'\fR). Pending watchers with higher priority will be invoked
1014before watchers with lower priority, but priority will not keep watchers
1015from being executed (except for \f(CW\*(C`ev_idle\*(C'\fR watchers).
1016.Sp
1017This means that priorities are \fIonly\fR used for ordering callback
1018invocation after new events have been received. This is useful, for
1019example, to reduce latency after idling, or more often, to bind two
1020watchers on the same event and make sure one is called first.
1021.Sp
1022If you need to suppress invocation when higher priority events are pending
1023you need to look at \f(CW\*(C`ev_idle\*(C'\fR watchers, which provide this functionality.
1024.Sp
1025You \fImust not\fR change the priority of a watcher as long as it is active or
1026pending.
1027.Sp
1028The default priority used by watchers when no priority has been set is
1029always \f(CW0\fR, which is supposed to not be too high and not be too low :).
1030.Sp
1031Setting a priority outside the range of \f(CW\*(C`EV_MINPRI\*(C'\fR to \f(CW\*(C`EV_MAXPRI\*(C'\fR is
1032fine, as long as you do not mind that the priority value you query might
1033or might not have been adjusted to be within valid range.
1034.IP "ev_invoke (loop, ev_TYPE *watcher, int revents)" 4
1035.IX Item "ev_invoke (loop, ev_TYPE *watcher, int revents)"
1036Invoke the \f(CW\*(C`watcher\*(C'\fR with the given \f(CW\*(C`loop\*(C'\fR and \f(CW\*(C`revents\*(C'\fR. Neither
1037\&\f(CW\*(C`loop\*(C'\fR nor \f(CW\*(C`revents\*(C'\fR need to be valid as long as the watcher callback
1038can deal with that fact.
1039.IP "int ev_clear_pending (loop, ev_TYPE *watcher)" 4
1040.IX Item "int ev_clear_pending (loop, ev_TYPE *watcher)"
1041If the watcher is pending, this function returns clears its pending status
1042and returns its \f(CW\*(C`revents\*(C'\fR bitset (as if its callback was invoked). If the
1043watcher isn't pending it does nothing and returns \f(CW0\fR.
611.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0" 1044.Sh "\s-1ASSOCIATING\s0 \s-1CUSTOM\s0 \s-1DATA\s0 \s-1WITH\s0 A \s-1WATCHER\s0"
612.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER" 1045.IX Subsection "ASSOCIATING CUSTOM DATA WITH A WATCHER"
613Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change 1046Each watcher has, by default, a member \f(CW\*(C`void *data\*(C'\fR that you can change
614and read at any time, libev will completely ignore it. This can be used 1047and read at any time, libev will completely ignore it. This can be used
615to associate arbitrary data with your watcher. If you need more data and 1048to associate arbitrary data with your watcher. If you need more data and
… …
636\& struct my_io *w = (struct my_io *)w_; 1069\& struct my_io *w = (struct my_io *)w_;
637\& ... 1070\& ...
638\& } 1071\& }
639.Ve 1072.Ve
640.PP 1073.PP
641More interesting and less C\-conformant ways of catsing your callback type 1074More interesting and less C\-conformant ways of casting your callback type
642have been omitted.... 1075instead have been omitted.
1076.PP
1077Another common scenario is having some data structure with multiple
1078watchers:
1079.PP
1080.Vb 6
1081\& struct my_biggy
1082\& {
1083\& int some_data;
1084\& ev_timer t1;
1085\& ev_timer t2;
1086\& }
1087.Ve
1088.PP
1089In this case getting the pointer to \f(CW\*(C`my_biggy\*(C'\fR is a bit more complicated,
1090you need to use \f(CW\*(C`offsetof\*(C'\fR:
1091.PP
1092.Vb 1
1093\& #include <stddef.h>
1094.Ve
1095.PP
1096.Vb 6
1097\& static void
1098\& t1_cb (EV_P_ struct ev_timer *w, int revents)
1099\& {
1100\& struct my_biggy big = (struct my_biggy *
1101\& (((char *)w) - offsetof (struct my_biggy, t1));
1102\& }
1103.Ve
1104.PP
1105.Vb 6
1106\& static void
1107\& t2_cb (EV_P_ struct ev_timer *w, int revents)
1108\& {
1109\& struct my_biggy big = (struct my_biggy *
1110\& (((char *)w) - offsetof (struct my_biggy, t2));
1111\& }
1112.Ve
643.SH "WATCHER TYPES" 1113.SH "WATCHER TYPES"
644.IX Header "WATCHER TYPES" 1114.IX Header "WATCHER TYPES"
645This section describes each watcher in detail, but will not repeat 1115This section describes each watcher in detail, but will not repeat
646information given in the last section. 1116information given in the last section. Any initialisation/set macros,
1117functions and members specific to the watcher type are explained.
1118.PP
1119Members are additionally marked with either \fI[read\-only]\fR, meaning that,
1120while the watcher is active, you can look at the member and expect some
1121sensible content, but you must not modify it (you can modify it while the
1122watcher is stopped to your hearts content), or \fI[read\-write]\fR, which
1123means you can expect it to have some sensible content while the watcher
1124is active, but you can also modify it. Modifying it may not do something
1125sensible or take immediate effect (or do anything at all), but libev will
1126not crash or malfunction in any way.
647.ie n .Sh """ev_io"" \- is this file descriptor readable or writable" 1127.ie n .Sh """ev_io"" \- is this file descriptor readable or writable?"
648.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable" 1128.el .Sh "\f(CWev_io\fP \- is this file descriptor readable or writable?"
649.IX Subsection "ev_io - is this file descriptor readable or writable" 1129.IX Subsection "ev_io - is this file descriptor readable or writable?"
650I/O watchers check whether a file descriptor is readable or writable 1130I/O watchers check whether a file descriptor is readable or writable
651in each iteration of the event loop (This behaviour is called 1131in each iteration of the event loop, or, more precisely, when reading
652level-triggering because you keep receiving events as long as the 1132would not block the process and writing would at least be able to write
653condition persists. Remember you can stop the watcher if you don't want to 1133some data. This behaviour is called level-triggering because you keep
654act on the event and neither want to receive future events). 1134receiving events as long as the condition persists. Remember you can stop
1135the watcher if you don't want to act on the event and neither want to
1136receive future events.
655.PP 1137.PP
656In general you can register as many read and/or write event watchers per 1138In general you can register as many read and/or write event watchers per
657fd as you want (as long as you don't confuse yourself). Setting all file 1139fd as you want (as long as you don't confuse yourself). Setting all file
658descriptors to non-blocking mode is also usually a good idea (but not 1140descriptors to non-blocking mode is also usually a good idea (but not
659required if you know what you are doing). 1141required if you know what you are doing).
660.PP 1142.PP
661You have to be careful with dup'ed file descriptors, though. Some backends
662(the linux epoll backend is a notable example) cannot handle dup'ed file
663descriptors correctly if you register interest in two or more fds pointing
664to the same underlying file/socket etc. description (that is, they share
665the same underlying \*(L"file open\*(R").
666.PP
667If you must do this, then force the use of a known-to-be-good backend 1143If you must do this, then force the use of a known-to-be-good backend
668(at the time of this writing, this includes only \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and 1144(at the time of this writing, this includes only \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR and
669\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR). 1145\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR).
1146.PP
1147Another thing you have to watch out for is that it is quite easy to
1148receive \*(L"spurious\*(R" readyness notifications, that is your callback might
1149be called with \f(CW\*(C`EV_READ\*(C'\fR but a subsequent \f(CW\*(C`read\*(C'\fR(2) will actually block
1150because there is no data. Not only are some backends known to create a
1151lot of those (for example solaris ports), it is very easy to get into
1152this situation even with a relatively standard program structure. Thus
1153it is best to always use non-blocking I/O: An extra \f(CW\*(C`read\*(C'\fR(2) returning
1154\&\f(CW\*(C`EAGAIN\*(C'\fR is far preferable to a program hanging until some data arrives.
1155.PP
1156If you cannot run the fd in non-blocking mode (for example you should not
1157play around with an Xlib connection), then you have to seperately re-test
1158whether a file descriptor is really ready with a known-to-be good interface
1159such as poll (fortunately in our Xlib example, Xlib already does this on
1160its own, so its quite safe to use).
1161.PP
1162\fIThe special problem of disappearing file descriptors\fR
1163.IX Subsection "The special problem of disappearing file descriptors"
1164.PP
1165Some backends (e.g. kqueue, epoll) need to be told about closing a file
1166descriptor (either by calling \f(CW\*(C`close\*(C'\fR explicitly or by any other means,
1167such as \f(CW\*(C`dup\*(C'\fR). The reason is that you register interest in some file
1168descriptor, but when it goes away, the operating system will silently drop
1169this interest. If another file descriptor with the same number then is
1170registered with libev, there is no efficient way to see that this is, in
1171fact, a different file descriptor.
1172.PP
1173To avoid having to explicitly tell libev about such cases, libev follows
1174the following policy: Each time \f(CW\*(C`ev_io_set\*(C'\fR is being called, libev
1175will assume that this is potentially a new file descriptor, otherwise
1176it is assumed that the file descriptor stays the same. That means that
1177you \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
1178descriptor even if the file descriptor number itself did not change.
1179.PP
1180This is how one would do it normally anyway, the important point is that
1181the libev application should not optimise around libev but should leave
1182optimisations to libev.
1183.PP
1184\fIThe special problem of dup'ed file descriptors\fR
1185.IX Subsection "The special problem of dup'ed file descriptors"
1186.PP
1187Some backends (e.g. epoll), cannot register events for file descriptors,
1188but only events for the underlying file descriptions. That means when you
1189have \f(CW\*(C`dup ()\*(C'\fR'ed file descriptors or weirder constellations, and register
1190events for them, only one file descriptor might actually receive events.
1191.PP
1192There is no workaround possible except not registering events
1193for potentially \f(CW\*(C`dup ()\*(C'\fR'ed file descriptors, or to resort to
1194\&\f(CW\*(C`EVBACKEND_SELECT\*(C'\fR or \f(CW\*(C`EVBACKEND_POLL\*(C'\fR.
1195.PP
1196\fIThe special problem of fork\fR
1197.IX Subsection "The special problem of fork"
1198.PP
1199Some backends (epoll, kqueue) do not support \f(CW\*(C`fork ()\*(C'\fR at all or exhibit
1200useless behaviour. Libev fully supports fork, but needs to be told about
1201it in the child.
1202.PP
1203To support fork in your programs, you either have to call
1204\&\f(CW\*(C`ev_default_fork ()\*(C'\fR or \f(CW\*(C`ev_loop_fork ()\*(C'\fR after a fork in the child,
1205enable \f(CW\*(C`EVFLAG_FORKCHECK\*(C'\fR, or resort to \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR or
1206\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR.
1207.PP
1208\fIWatcher-Specific Functions\fR
1209.IX Subsection "Watcher-Specific Functions"
670.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4 1210.IP "ev_io_init (ev_io *, callback, int fd, int events)" 4
671.IX Item "ev_io_init (ev_io *, callback, int fd, int events)" 1211.IX Item "ev_io_init (ev_io *, callback, int fd, int events)"
672.PD 0 1212.PD 0
673.IP "ev_io_set (ev_io *, int fd, int events)" 4 1213.IP "ev_io_set (ev_io *, int fd, int events)" 4
674.IX Item "ev_io_set (ev_io *, int fd, int events)" 1214.IX Item "ev_io_set (ev_io *, int fd, int events)"
675.PD 1215.PD
676Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The fd is the file descriptor to rceeive 1216Configures an \f(CW\*(C`ev_io\*(C'\fR watcher. The \f(CW\*(C`fd\*(C'\fR is the file descriptor to
677events 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 | 1217rceeive events for and events is either \f(CW\*(C`EV_READ\*(C'\fR, \f(CW\*(C`EV_WRITE\*(C'\fR or
678EV_WRITE\*(C'\fR to receive the given events. 1218\&\f(CW\*(C`EV_READ | EV_WRITE\*(C'\fR to receive the given events.
679.Sp 1219.IP "int fd [read\-only]" 4
680Please note that most of the more scalable backend mechanisms (for example 1220.IX Item "int fd [read-only]"
681epoll and solaris ports) can result in spurious readyness notifications 1221The file descriptor being watched.
682for file descriptors, so you practically need to use non-blocking I/O (and 1222.IP "int events [read\-only]" 4
683treat callback invocation as hint only), or retest separately with a safe 1223.IX Item "int events [read-only]"
684interface before doing I/O (XLib can do this), or force the use of either 1224The events being watched.
685\&\f(CW\*(C`EVBACKEND_SELECT\*(C'\fR or \f(CW\*(C`EVBACKEND_POLL\*(C'\fR, which don't suffer from this 1225.PP
686problem. Also note that it is quite easy to have your callback invoked 1226Example: Call \f(CW\*(C`stdin_readable_cb\*(C'\fR when \s-1STDIN_FILENO\s0 has become, well
687when the readyness condition is no longer valid even when employing 1227readable, but only once. Since it is likely line\-buffered, you could
688typical ways of handling events, so its a good idea to use non-blocking 1228attempt to read a whole line in the callback.
689I/O unconditionally. 1229.PP
1230.Vb 6
1231\& static void
1232\& stdin_readable_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1233\& {
1234\& ev_io_stop (loop, w);
1235\& .. read from stdin here (or from w->fd) and haqndle any I/O errors
1236\& }
1237.Ve
1238.PP
1239.Vb 6
1240\& ...
1241\& struct ev_loop *loop = ev_default_init (0);
1242\& struct ev_io stdin_readable;
1243\& ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1244\& ev_io_start (loop, &stdin_readable);
1245\& ev_loop (loop, 0);
1246.Ve
690.ie n .Sh """ev_timer"" \- relative and optionally recurring timeouts" 1247.ie n .Sh """ev_timer"" \- relative and optionally repeating timeouts"
691.el .Sh "\f(CWev_timer\fP \- relative and optionally recurring timeouts" 1248.el .Sh "\f(CWev_timer\fP \- relative and optionally repeating timeouts"
692.IX Subsection "ev_timer - relative and optionally recurring timeouts" 1249.IX Subsection "ev_timer - relative and optionally repeating timeouts"
693Timer watchers are simple relative timers that generate an event after a 1250Timer watchers are simple relative timers that generate an event after a
694given time, and optionally repeating in regular intervals after that. 1251given time, and optionally repeating in regular intervals after that.
695.PP 1252.PP
696The timers are based on real time, that is, if you register an event that 1253The timers are based on real time, that is, if you register an event that
697times out after an hour and you reset your system clock to last years 1254times out after an hour and you reset your system clock to last years
… …
710.Ve 1267.Ve
711.PP 1268.PP
712The callback is guarenteed to be invoked only when its timeout has passed, 1269The callback is guarenteed to be invoked only when its timeout has passed,
713but if multiple timers become ready during the same loop iteration then 1270but if multiple timers become ready during the same loop iteration then
714order of execution is undefined. 1271order of execution is undefined.
1272.PP
1273\fIWatcher-Specific Functions and Data Members\fR
1274.IX Subsection "Watcher-Specific Functions and Data Members"
715.IP "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)" 4 1275.IP "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)" 4
716.IX Item "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)" 1276.IX Item "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)"
717.PD 0 1277.PD 0
718.IP "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 4 1278.IP "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 4
719.IX Item "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 1279.IX Item "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)"
… …
731.IP "ev_timer_again (loop)" 4 1291.IP "ev_timer_again (loop)" 4
732.IX Item "ev_timer_again (loop)" 1292.IX Item "ev_timer_again (loop)"
733This will act as if the timer timed out and restart it again if it is 1293This will act as if the timer timed out and restart it again if it is
734repeating. The exact semantics are: 1294repeating. The exact semantics are:
735.Sp 1295.Sp
1296If the timer is pending, its pending status is cleared.
1297.Sp
736If the timer is started but nonrepeating, stop it. 1298If the timer is started but nonrepeating, stop it (as if it timed out).
737.Sp 1299.Sp
738If the timer is repeating, either start it if necessary (with the repeat 1300If the timer is repeating, either start it if necessary (with the
739value), or reset the running timer to the repeat value. 1301\&\f(CW\*(C`repeat\*(C'\fR value), or reset the running timer to the \f(CW\*(C`repeat\*(C'\fR value.
740.Sp 1302.Sp
741This sounds a bit complicated, but here is a useful and typical 1303This sounds a bit complicated, but here is a useful and typical
742example: Imagine you have a tcp connection and you want a so-called idle 1304example: Imagine you have a tcp connection and you want a so-called idle
743timeout, that is, you want to be called when there have been, say, 60 1305timeout, that is, you want to be called when there have been, say, 60
744seconds of inactivity on the socket. The easiest way to do this is to 1306seconds of inactivity on the socket. The easiest way to do this is to
745configure an \f(CW\*(C`ev_timer\*(C'\fR with after=repeat=60 and calling ev_timer_again each 1307configure an \f(CW\*(C`ev_timer\*(C'\fR with a \f(CW\*(C`repeat\*(C'\fR value of \f(CW60\fR and then call
746time you successfully read or write some data. If you go into an idle 1308\&\f(CW\*(C`ev_timer_again\*(C'\fR each time you successfully read or write some data. If
747state where you do not expect data to travel on the socket, you can stop 1309you go into an idle state where you do not expect data to travel on the
748the timer, and again will automatically restart it if need be. 1310socket, you can \f(CW\*(C`ev_timer_stop\*(C'\fR the timer, and \f(CW\*(C`ev_timer_again\*(C'\fR will
1311automatically restart it if need be.
1312.Sp
1313That means you can ignore the \f(CW\*(C`after\*(C'\fR value and \f(CW\*(C`ev_timer_start\*(C'\fR
1314altogether and only ever use the \f(CW\*(C`repeat\*(C'\fR value and \f(CW\*(C`ev_timer_again\*(C'\fR:
1315.Sp
1316.Vb 8
1317\& ev_timer_init (timer, callback, 0., 5.);
1318\& ev_timer_again (loop, timer);
1319\& ...
1320\& timer->again = 17.;
1321\& ev_timer_again (loop, timer);
1322\& ...
1323\& timer->again = 10.;
1324\& ev_timer_again (loop, timer);
1325.Ve
1326.Sp
1327This is more slightly efficient then stopping/starting the timer each time
1328you want to modify its timeout value.
1329.IP "ev_tstamp repeat [read\-write]" 4
1330.IX Item "ev_tstamp repeat [read-write]"
1331The current \f(CW\*(C`repeat\*(C'\fR value. Will be used each time the watcher times out
1332or \f(CW\*(C`ev_timer_again\*(C'\fR is called and determines the next timeout (if any),
1333which is also when any modifications are taken into account.
1334.PP
1335Example: Create a timer that fires after 60 seconds.
1336.PP
1337.Vb 5
1338\& static void
1339\& one_minute_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1340\& {
1341\& .. one minute over, w is actually stopped right here
1342\& }
1343.Ve
1344.PP
1345.Vb 3
1346\& struct ev_timer mytimer;
1347\& ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
1348\& ev_timer_start (loop, &mytimer);
1349.Ve
1350.PP
1351Example: Create a timeout timer that times out after 10 seconds of
1352inactivity.
1353.PP
1354.Vb 5
1355\& static void
1356\& timeout_cb (struct ev_loop *loop, struct ev_timer *w, int revents)
1357\& {
1358\& .. ten seconds without any activity
1359\& }
1360.Ve
1361.PP
1362.Vb 4
1363\& struct ev_timer mytimer;
1364\& ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
1365\& ev_timer_again (&mytimer); /* start timer */
1366\& ev_loop (loop, 0);
1367.Ve
1368.PP
1369.Vb 3
1370\& // and in some piece of code that gets executed on any "activity":
1371\& // reset the timeout to start ticking again at 10 seconds
1372\& ev_timer_again (&mytimer);
1373.Ve
749.ie n .Sh """ev_periodic"" \- to cron or not to cron" 1374.ie n .Sh """ev_periodic"" \- to cron or not to cron?"
750.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron" 1375.el .Sh "\f(CWev_periodic\fP \- to cron or not to cron?"
751.IX Subsection "ev_periodic - to cron or not to cron" 1376.IX Subsection "ev_periodic - to cron or not to cron?"
752Periodic watchers are also timers of a kind, but they are very versatile 1377Periodic watchers are also timers of a kind, but they are very versatile
753(and unfortunately a bit complex). 1378(and unfortunately a bit complex).
754.PP 1379.PP
755Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time) 1380Unlike \f(CW\*(C`ev_timer\*(C'\fR's, they are not based on real time (or relative time)
756but on wallclock time (absolute time). You can tell a periodic watcher 1381but on wallclock time (absolute time). You can tell a periodic watcher
757to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a 1382to trigger \*(L"at\*(R" some specific point in time. For example, if you tell a
758periodic watcher to trigger in 10 seconds (by specifiying e.g. c<ev_now () 1383periodic watcher to trigger in 10 seconds (by specifiying e.g. \f(CW\*(C`ev_now ()
759+ 10.>) and then reset your system clock to the last year, then it will 1384+ 10.\*(C'\fR) and then reset your system clock to the last year, then it will
760take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger 1385take a year to trigger the event (unlike an \f(CW\*(C`ev_timer\*(C'\fR, which would trigger
761roughly 10 seconds later and of course not if you reset your system time 1386roughly 10 seconds later).
762again).
763.PP 1387.PP
764They can also be used to implement vastly more complex timers, such as 1388They can also be used to implement vastly more complex timers, such as
765triggering an event on eahc midnight, local time. 1389triggering an event on each midnight, local time or other, complicated,
1390rules.
766.PP 1391.PP
767As with timers, the callback is guarenteed to be invoked only when the 1392As with timers, the callback is guarenteed to be invoked only when the
768time (\f(CW\*(C`at\*(C'\fR) has been passed, but if multiple periodic timers become ready 1393time (\f(CW\*(C`at\*(C'\fR) has been passed, but if multiple periodic timers become ready
769during the same loop iteration then order of execution is undefined. 1394during the same loop iteration then order of execution is undefined.
1395.PP
1396\fIWatcher-Specific Functions and Data Members\fR
1397.IX Subsection "Watcher-Specific Functions and Data Members"
770.IP "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)" 4 1398.IP "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)" 4
771.IX Item "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)" 1399.IX Item "ev_periodic_init (ev_periodic *, callback, ev_tstamp at, ev_tstamp interval, reschedule_cb)"
772.PD 0 1400.PD 0
773.IP "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)" 4 1401.IP "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)" 4
774.IX Item "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)" 1402.IX Item "ev_periodic_set (ev_periodic *, ev_tstamp after, ev_tstamp repeat, reschedule_cb)"
775.PD 1403.PD
776Lots of arguments, lets sort it out... There are basically three modes of 1404Lots of arguments, lets sort it out... There are basically three modes of
777operation, and we will explain them from simplest to complex: 1405operation, and we will explain them from simplest to complex:
778.RS 4 1406.RS 4
779.IP "* absolute timer (interval = reschedule_cb = 0)" 4 1407.IP "* absolute timer (at = time, interval = reschedule_cb = 0)" 4
780.IX Item "absolute timer (interval = reschedule_cb = 0)" 1408.IX Item "absolute timer (at = time, interval = reschedule_cb = 0)"
781In this configuration the watcher triggers an event at the wallclock time 1409In this configuration the watcher triggers an event at the wallclock time
782\&\f(CW\*(C`at\*(C'\fR and doesn't repeat. It will not adjust when a time jump occurs, 1410\&\f(CW\*(C`at\*(C'\fR and doesn't repeat. It will not adjust when a time jump occurs,
783that is, if it is to be run at January 1st 2011 then it will run when the 1411that is, if it is to be run at January 1st 2011 then it will run when the
784system time reaches or surpasses this time. 1412system time reaches or surpasses this time.
785.IP "* non-repeating interval timer (interval > 0, reschedule_cb = 0)" 4 1413.IP "* non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)" 4
786.IX Item "non-repeating interval timer (interval > 0, reschedule_cb = 0)" 1414.IX Item "non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0)"
787In this mode the watcher will always be scheduled to time out at the next 1415In this mode the watcher will always be scheduled to time out at the next
788\&\f(CW\*(C`at + N * interval\*(C'\fR time (for some integer N) and then repeat, regardless 1416\&\f(CW\*(C`at + N * interval\*(C'\fR time (for some integer N, which can also be negative)
789of any time jumps. 1417and then repeat, regardless of any time jumps.
790.Sp 1418.Sp
791This can be used to create timers that do not drift with respect to system 1419This can be used to create timers that do not drift with respect to system
792time: 1420time:
793.Sp 1421.Sp
794.Vb 1 1422.Vb 1
… …
801by 3600. 1429by 3600.
802.Sp 1430.Sp
803Another way to think about it (for the mathematically inclined) is that 1431Another way to think about it (for the mathematically inclined) is that
804\&\f(CW\*(C`ev_periodic\*(C'\fR will try to run the callback in this mode at the next possible 1432\&\f(CW\*(C`ev_periodic\*(C'\fR will try to run the callback in this mode at the next possible
805time where \f(CW\*(C`time = at (mod interval)\*(C'\fR, regardless of any time jumps. 1433time where \f(CW\*(C`time = at (mod interval)\*(C'\fR, regardless of any time jumps.
1434.Sp
1435For numerical stability it is preferable that the \f(CW\*(C`at\*(C'\fR value is near
1436\&\f(CW\*(C`ev_now ()\*(C'\fR (the current time), but there is no range requirement for
1437this value.
806.IP "* manual reschedule mode (reschedule_cb = callback)" 4 1438.IP "* manual reschedule mode (at and interval ignored, reschedule_cb = callback)" 4
807.IX Item "manual reschedule mode (reschedule_cb = callback)" 1439.IX Item "manual reschedule mode (at and interval ignored, reschedule_cb = callback)"
808In this mode the values for \f(CW\*(C`interval\*(C'\fR and \f(CW\*(C`at\*(C'\fR are both being 1440In this mode the values for \f(CW\*(C`interval\*(C'\fR and \f(CW\*(C`at\*(C'\fR are both being
809ignored. Instead, each time the periodic watcher gets scheduled, the 1441ignored. Instead, each time the periodic watcher gets scheduled, the
810reschedule callback will be called with the watcher as first, and the 1442reschedule callback will be called with the watcher as first, and the
811current time as second argument. 1443current time as second argument.
812.Sp 1444.Sp
813\&\s-1NOTE:\s0 \fIThis callback \s-1MUST\s0 \s-1NOT\s0 stop or destroy any periodic watcher, 1445\&\s-1NOTE:\s0 \fIThis callback \s-1MUST\s0 \s-1NOT\s0 stop or destroy any periodic watcher,
814ever, or make any event loop modifications\fR. If you need to stop it, 1446ever, or make any event loop modifications\fR. If you need to stop it,
815return \f(CW\*(C`now + 1e30\*(C'\fR (or so, fudge fudge) and stop it afterwards (e.g. by 1447return \f(CW\*(C`now + 1e30\*(C'\fR (or so, fudge fudge) and stop it afterwards (e.g. by
816starting a prepare watcher). 1448starting an \f(CW\*(C`ev_prepare\*(C'\fR watcher, which is legal).
817.Sp 1449.Sp
818Its prototype is \f(CW\*(C`ev_tstamp (*reschedule_cb)(struct ev_periodic *w, 1450Its prototype is \f(CW\*(C`ev_tstamp (*reschedule_cb)(struct ev_periodic *w,
819ev_tstamp now)\*(C'\fR, e.g.: 1451ev_tstamp now)\*(C'\fR, e.g.:
820.Sp 1452.Sp
821.Vb 4 1453.Vb 4
… …
845.IX Item "ev_periodic_again (loop, ev_periodic *)" 1477.IX Item "ev_periodic_again (loop, ev_periodic *)"
846Simply stops and restarts the periodic watcher again. This is only useful 1478Simply stops and restarts the periodic watcher again. This is only useful
847when you changed some parameters or the reschedule callback would return 1479when you changed some parameters or the reschedule callback would return
848a different time than the last time it was called (e.g. in a crond like 1480a different time than the last time it was called (e.g. in a crond like
849program when the crontabs have changed). 1481program when the crontabs have changed).
1482.IP "ev_tstamp offset [read\-write]" 4
1483.IX Item "ev_tstamp offset [read-write]"
1484When repeating, this contains the offset value, otherwise this is the
1485absolute point in time (the \f(CW\*(C`at\*(C'\fR value passed to \f(CW\*(C`ev_periodic_set\*(C'\fR).
1486.Sp
1487Can be modified any time, but changes only take effect when the periodic
1488timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being called.
1489.IP "ev_tstamp interval [read\-write]" 4
1490.IX Item "ev_tstamp interval [read-write]"
1491The current interval value. Can be modified any time, but changes only
1492take effect when the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being
1493called.
1494.IP "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read\-write]" 4
1495.IX Item "ev_tstamp (*reschedule_cb)(struct ev_periodic *w, ev_tstamp now) [read-write]"
1496The current reschedule callback, or \f(CW0\fR, if this functionality is
1497switched off. Can be changed any time, but changes only take effect when
1498the periodic timer fires or \f(CW\*(C`ev_periodic_again\*(C'\fR is being called.
1499.IP "ev_tstamp at [read\-only]" 4
1500.IX Item "ev_tstamp at [read-only]"
1501When active, contains the absolute time that the watcher is supposed to
1502trigger next.
1503.PP
1504Example: Call a callback every hour, or, more precisely, whenever the
1505system clock is divisible by 3600. The callback invocation times have
1506potentially a lot of jittering, but good long-term stability.
1507.PP
1508.Vb 5
1509\& static void
1510\& clock_cb (struct ev_loop *loop, struct ev_io *w, int revents)
1511\& {
1512\& ... its now a full hour (UTC, or TAI or whatever your clock follows)
1513\& }
1514.Ve
1515.PP
1516.Vb 3
1517\& struct ev_periodic hourly_tick;
1518\& ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
1519\& ev_periodic_start (loop, &hourly_tick);
1520.Ve
1521.PP
1522Example: The same as above, but use a reschedule callback to do it:
1523.PP
1524.Vb 1
1525\& #include <math.h>
1526.Ve
1527.PP
1528.Vb 5
1529\& static ev_tstamp
1530\& my_scheduler_cb (struct ev_periodic *w, ev_tstamp now)
1531\& {
1532\& return fmod (now, 3600.) + 3600.;
1533\& }
1534.Ve
1535.PP
1536.Vb 1
1537\& ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
1538.Ve
1539.PP
1540Example: Call a callback every hour, starting now:
1541.PP
1542.Vb 4
1543\& struct ev_periodic hourly_tick;
1544\& ev_periodic_init (&hourly_tick, clock_cb,
1545\& fmod (ev_now (loop), 3600.), 3600., 0);
1546\& ev_periodic_start (loop, &hourly_tick);
1547.Ve
850.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled" 1548.ie n .Sh """ev_signal"" \- signal me when a signal gets signalled!"
851.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled" 1549.el .Sh "\f(CWev_signal\fP \- signal me when a signal gets signalled!"
852.IX Subsection "ev_signal - signal me when a signal gets signalled" 1550.IX Subsection "ev_signal - signal me when a signal gets signalled!"
853Signal watchers will trigger an event when the process receives a specific 1551Signal watchers will trigger an event when the process receives a specific
854signal one or more times. Even though signals are very asynchronous, libev 1552signal one or more times. Even though signals are very asynchronous, libev
855will try it's best to deliver signals synchronously, i.e. as part of the 1553will try it's best to deliver signals synchronously, i.e. as part of the
856normal event processing, like any other event. 1554normal event processing, like any other event.
857.PP 1555.PP
… …
859first watcher gets started will libev actually register a signal watcher 1557first watcher gets started will libev actually register a signal watcher
860with the kernel (thus it coexists with your own signal handlers as long 1558with the kernel (thus it coexists with your own signal handlers as long
861as you don't register any with libev). Similarly, when the last signal 1559as you don't register any with libev). Similarly, when the last signal
862watcher for a signal is stopped libev will reset the signal handler to 1560watcher for a signal is stopped libev will reset the signal handler to
863\&\s-1SIG_DFL\s0 (regardless of what it was set to before). 1561\&\s-1SIG_DFL\s0 (regardless of what it was set to before).
1562.PP
1563\fIWatcher-Specific Functions and Data Members\fR
1564.IX Subsection "Watcher-Specific Functions and Data Members"
864.IP "ev_signal_init (ev_signal *, callback, int signum)" 4 1565.IP "ev_signal_init (ev_signal *, callback, int signum)" 4
865.IX Item "ev_signal_init (ev_signal *, callback, int signum)" 1566.IX Item "ev_signal_init (ev_signal *, callback, int signum)"
866.PD 0 1567.PD 0
867.IP "ev_signal_set (ev_signal *, int signum)" 4 1568.IP "ev_signal_set (ev_signal *, int signum)" 4
868.IX Item "ev_signal_set (ev_signal *, int signum)" 1569.IX Item "ev_signal_set (ev_signal *, int signum)"
869.PD 1570.PD
870Configures the watcher to trigger on the given signal number (usually one 1571Configures the watcher to trigger on the given signal number (usually one
871of the \f(CW\*(C`SIGxxx\*(C'\fR constants). 1572of the \f(CW\*(C`SIGxxx\*(C'\fR constants).
1573.IP "int signum [read\-only]" 4
1574.IX Item "int signum [read-only]"
1575The signal the watcher watches out for.
872.ie n .Sh """ev_child"" \- wait for pid status changes" 1576.ie n .Sh """ev_child"" \- watch out for process status changes"
873.el .Sh "\f(CWev_child\fP \- wait for pid status changes" 1577.el .Sh "\f(CWev_child\fP \- watch out for process status changes"
874.IX Subsection "ev_child - wait for pid status changes" 1578.IX Subsection "ev_child - watch out for process status changes"
875Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to 1579Child watchers trigger when your process receives a \s-1SIGCHLD\s0 in response to
876some child status changes (most typically when a child of yours dies). 1580some child status changes (most typically when a child of yours dies).
1581.PP
1582\fIWatcher-Specific Functions and Data Members\fR
1583.IX Subsection "Watcher-Specific Functions and Data Members"
877.IP "ev_child_init (ev_child *, callback, int pid)" 4 1584.IP "ev_child_init (ev_child *, callback, int pid)" 4
878.IX Item "ev_child_init (ev_child *, callback, int pid)" 1585.IX Item "ev_child_init (ev_child *, callback, int pid)"
879.PD 0 1586.PD 0
880.IP "ev_child_set (ev_child *, int pid)" 4 1587.IP "ev_child_set (ev_child *, int pid)" 4
881.IX Item "ev_child_set (ev_child *, int pid)" 1588.IX Item "ev_child_set (ev_child *, int pid)"
… …
884\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look 1591\&\fIany\fR process if \f(CW\*(C`pid\*(C'\fR is specified as \f(CW0\fR). The callback can look
885at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see 1592at the \f(CW\*(C`rstatus\*(C'\fR member of the \f(CW\*(C`ev_child\*(C'\fR watcher structure to see
886the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems 1593the status word (use the macros from \f(CW\*(C`sys/wait.h\*(C'\fR and see your systems
887\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the 1594\&\f(CW\*(C`waitpid\*(C'\fR documentation). The \f(CW\*(C`rpid\*(C'\fR member contains the pid of the
888process causing the status change. 1595process causing the status change.
1596.IP "int pid [read\-only]" 4
1597.IX Item "int pid [read-only]"
1598The process id this watcher watches out for, or \f(CW0\fR, meaning any process id.
1599.IP "int rpid [read\-write]" 4
1600.IX Item "int rpid [read-write]"
1601The process id that detected a status change.
1602.IP "int rstatus [read\-write]" 4
1603.IX Item "int rstatus [read-write]"
1604The process exit/trace status caused by \f(CW\*(C`rpid\*(C'\fR (see your systems
1605\&\f(CW\*(C`waitpid\*(C'\fR and \f(CW\*(C`sys/wait.h\*(C'\fR documentation for details).
1606.PP
1607Example: Try to exit cleanly on \s-1SIGINT\s0 and \s-1SIGTERM\s0.
1608.PP
1609.Vb 5
1610\& static void
1611\& sigint_cb (struct ev_loop *loop, struct ev_signal *w, int revents)
1612\& {
1613\& ev_unloop (loop, EVUNLOOP_ALL);
1614\& }
1615.Ve
1616.PP
1617.Vb 3
1618\& struct ev_signal signal_watcher;
1619\& ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
1620\& ev_signal_start (loop, &sigint_cb);
1621.Ve
1622.ie n .Sh """ev_stat"" \- did the file attributes just change?"
1623.el .Sh "\f(CWev_stat\fP \- did the file attributes just change?"
1624.IX Subsection "ev_stat - did the file attributes just change?"
1625This watches a filesystem path for attribute changes. That is, it calls
1626\&\f(CW\*(C`stat\*(C'\fR regularly (or when the \s-1OS\s0 says it changed) and sees if it changed
1627compared to the last time, invoking the callback if it did.
1628.PP
1629The path does not need to exist: changing from \*(L"path exists\*(R" to \*(L"path does
1630not exist\*(R" is a status change like any other. The condition \*(L"path does
1631not exist\*(R" is signified by the \f(CW\*(C`st_nlink\*(C'\fR field being zero (which is
1632otherwise always forced to be at least one) and all the other fields of
1633the stat buffer having unspecified contents.
1634.PP
1635The path \fIshould\fR be absolute and \fImust not\fR end in a slash. If it is
1636relative and your working directory changes, the behaviour is undefined.
1637.PP
1638Since there is no standard to do this, the portable implementation simply
1639calls \f(CW\*(C`stat (2)\*(C'\fR regularly on the path to see if it changed somehow. You
1640can specify a recommended polling interval for this case. If you specify
1641a polling interval of \f(CW0\fR (highly recommended!) then a \fIsuitable,
1642unspecified default\fR value will be used (which you can expect to be around
1643five seconds, although this might change dynamically). Libev will also
1644impose a minimum interval which is currently around \f(CW0.1\fR, but thats
1645usually overkill.
1646.PP
1647This watcher type is not meant for massive numbers of stat watchers,
1648as even with OS-supported change notifications, this can be
1649resource\-intensive.
1650.PP
1651At the time of this writing, only the Linux inotify interface is
1652implemented (implementing kqueue support is left as an exercise for the
1653reader). Inotify will be used to give hints only and should not change the
1654semantics of \f(CW\*(C`ev_stat\*(C'\fR watchers, which means that libev sometimes needs
1655to fall back to regular polling again even with inotify, but changes are
1656usually detected immediately, and if the file exists there will be no
1657polling.
1658.PP
1659\fIInotify\fR
1660.IX Subsection "Inotify"
1661.PP
1662When \f(CW\*(C`inotify (7)\*(C'\fR support has been compiled into libev (generally only
1663available on Linux) and present at runtime, it will be used to speed up
1664change detection where possible. The inotify descriptor will be created lazily
1665when the first \f(CW\*(C`ev_stat\*(C'\fR watcher is being started.
1666.PP
1667Inotify presense does not change the semantics of \f(CW\*(C`ev_stat\*(C'\fR watchers
1668except that changes might be detected earlier, and in some cases, to avoid
1669making regular \f(CW\*(C`stat\*(C'\fR calls. Even in the presense of inotify support
1670there are many cases where libev has to resort to regular \f(CW\*(C`stat\*(C'\fR polling.
1671.PP
1672(There is no support for kqueue, as apparently it cannot be used to
1673implement this functionality, due to the requirement of having a file
1674descriptor open on the object at all times).
1675.PP
1676\fIThe special problem of stat time resolution\fR
1677.IX Subsection "The special problem of stat time resolution"
1678.PP
1679The \f(CW\*(C`stat ()\*(C'\fR syscall only supports full-second resolution portably, and
1680even on systems where the resolution is higher, many filesystems still
1681only support whole seconds.
1682.PP
1683That means that, if the time is the only thing that changes, you might
1684miss updates: on the first update, \f(CW\*(C`ev_stat\*(C'\fR detects a change and calls
1685your callback, which does something. When there is another update within
1686the same second, \f(CW\*(C`ev_stat\*(C'\fR will be unable to detect it.
1687.PP
1688The solution to this is to delay acting on a change for a second (or till
1689the next second boundary), using a roughly one-second delay \f(CW\*(C`ev_timer\*(C'\fR
1690(\f(CW\*(C`ev_timer_set (w, 0., 1.01); ev_timer_again (loop, w)\*(C'\fR). The \f(CW.01\fR
1691is added to work around small timing inconsistencies of some operating
1692systems.
1693.PP
1694\fIWatcher-Specific Functions and Data Members\fR
1695.IX Subsection "Watcher-Specific Functions and Data Members"
1696.IP "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)" 4
1697.IX Item "ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)"
1698.PD 0
1699.IP "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)" 4
1700.IX Item "ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)"
1701.PD
1702Configures the watcher to wait for status changes of the given
1703\&\f(CW\*(C`path\*(C'\fR. The \f(CW\*(C`interval\*(C'\fR is a hint on how quickly a change is expected to
1704be detected and should normally be specified as \f(CW0\fR to let libev choose
1705a suitable value. The memory pointed to by \f(CW\*(C`path\*(C'\fR must point to the same
1706path for as long as the watcher is active.
1707.Sp
1708The callback will be receive \f(CW\*(C`EV_STAT\*(C'\fR when a change was detected,
1709relative to the attributes at the time the watcher was started (or the
1710last change was detected).
1711.IP "ev_stat_stat (ev_stat *)" 4
1712.IX Item "ev_stat_stat (ev_stat *)"
1713Updates the stat buffer immediately with new values. If you change the
1714watched path in your callback, you could call this fucntion to avoid
1715detecting this change (while introducing a race condition). Can also be
1716useful simply to find out the new values.
1717.IP "ev_statdata attr [read\-only]" 4
1718.IX Item "ev_statdata attr [read-only]"
1719The most-recently detected attributes of the file. Although the type is of
1720\&\f(CW\*(C`ev_statdata\*(C'\fR, this is usually the (or one of the) \f(CW\*(C`struct stat\*(C'\fR types
1721suitable for your system. If the \f(CW\*(C`st_nlink\*(C'\fR member is \f(CW0\fR, then there
1722was some error while \f(CW\*(C`stat\*(C'\fRing the file.
1723.IP "ev_statdata prev [read\-only]" 4
1724.IX Item "ev_statdata prev [read-only]"
1725The previous attributes of the file. The callback gets invoked whenever
1726\&\f(CW\*(C`prev\*(C'\fR != \f(CW\*(C`attr\*(C'\fR.
1727.IP "ev_tstamp interval [read\-only]" 4
1728.IX Item "ev_tstamp interval [read-only]"
1729The specified interval.
1730.IP "const char *path [read\-only]" 4
1731.IX Item "const char *path [read-only]"
1732The filesystem path that is being watched.
1733.PP
1734\fIExamples\fR
1735.IX Subsection "Examples"
1736.PP
1737Example: Watch \f(CW\*(C`/etc/passwd\*(C'\fR for attribute changes.
1738.PP
1739.Vb 15
1740\& static void
1741\& passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
1742\& {
1743\& /* /etc/passwd changed in some way */
1744\& if (w->attr.st_nlink)
1745\& {
1746\& printf ("passwd current size %ld\en", (long)w->attr.st_size);
1747\& printf ("passwd current atime %ld\en", (long)w->attr.st_mtime);
1748\& printf ("passwd current mtime %ld\en", (long)w->attr.st_mtime);
1749\& }
1750\& else
1751\& /* you shalt not abuse printf for puts */
1752\& puts ("wow, /etc/passwd is not there, expect problems. "
1753\& "if this is windows, they already arrived\en");
1754\& }
1755.Ve
1756.PP
1757.Vb 2
1758\& ...
1759\& ev_stat passwd;
1760.Ve
1761.PP
1762.Vb 2
1763\& ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
1764\& ev_stat_start (loop, &passwd);
1765.Ve
1766.PP
1767Example: Like above, but additionally use a one-second delay so we do not
1768miss updates (however, frequent updates will delay processing, too, so
1769one might do the work both on \f(CW\*(C`ev_stat\*(C'\fR callback invocation \fIand\fR on
1770\&\f(CW\*(C`ev_timer\*(C'\fR callback invocation).
1771.PP
1772.Vb 2
1773\& static ev_stat passwd;
1774\& static ev_timer timer;
1775.Ve
1776.PP
1777.Vb 4
1778\& static void
1779\& timer_cb (EV_P_ ev_timer *w, int revents)
1780\& {
1781\& ev_timer_stop (EV_A_ w);
1782.Ve
1783.PP
1784.Vb 2
1785\& /* now it's one second after the most recent passwd change */
1786\& }
1787.Ve
1788.PP
1789.Vb 6
1790\& static void
1791\& stat_cb (EV_P_ ev_stat *w, int revents)
1792\& {
1793\& /* reset the one-second timer */
1794\& ev_timer_again (EV_A_ &timer);
1795\& }
1796.Ve
1797.PP
1798.Vb 4
1799\& ...
1800\& ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
1801\& ev_stat_start (loop, &passwd);
1802\& ev_timer_init (&timer, timer_cb, 0., 1.01);
1803.Ve
889.ie n .Sh """ev_idle"" \- when you've got nothing better to do" 1804.ie n .Sh """ev_idle"" \- when you've got nothing better to do..."
890.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do" 1805.el .Sh "\f(CWev_idle\fP \- when you've got nothing better to do..."
891.IX Subsection "ev_idle - when you've got nothing better to do" 1806.IX Subsection "ev_idle - when you've got nothing better to do..."
892Idle watchers trigger events when there are no other events are pending 1807Idle watchers trigger events when no other events of the same or higher
893(prepare, check and other idle watchers do not count). That is, as long 1808priority are pending (prepare, check and other idle watchers do not
894as your process is busy handling sockets or timeouts (or even signals, 1809count).
895imagine) it will not be triggered. But when your process is idle all idle 1810.PP
896watchers are being called again and again, once per event loop iteration \- 1811That is, as long as your process is busy handling sockets or timeouts
1812(or even signals, imagine) of the same or higher priority it will not be
1813triggered. But when your process is idle (or only lower-priority watchers
1814are pending), the idle watchers are being called once per event loop
897until stopped, that is, or your process receives more events and becomes 1815iteration \- until stopped, that is, or your process receives more events
898busy. 1816and becomes busy again with higher priority stuff.
899.PP 1817.PP
900The most noteworthy effect is that as long as any idle watchers are 1818The most noteworthy effect is that as long as any idle watchers are
901active, the process will not block when waiting for new events. 1819active, the process will not block when waiting for new events.
902.PP 1820.PP
903Apart from keeping your process non-blocking (which is a useful 1821Apart from keeping your process non-blocking (which is a useful
904effect on its own sometimes), idle watchers are a good place to do 1822effect on its own sometimes), idle watchers are a good place to do
905\&\*(L"pseudo\-background processing\*(R", or delay processing stuff to after the 1823\&\*(L"pseudo\-background processing\*(R", or delay processing stuff to after the
906event loop has handled all outstanding events. 1824event loop has handled all outstanding events.
1825.PP
1826\fIWatcher-Specific Functions and Data Members\fR
1827.IX Subsection "Watcher-Specific Functions and Data Members"
907.IP "ev_idle_init (ev_signal *, callback)" 4 1828.IP "ev_idle_init (ev_signal *, callback)" 4
908.IX Item "ev_idle_init (ev_signal *, callback)" 1829.IX Item "ev_idle_init (ev_signal *, callback)"
909Initialises and configures the idle watcher \- it has no parameters of any 1830Initialises and configures the idle watcher \- it has no parameters of any
910kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless, 1831kind. There is a \f(CW\*(C`ev_idle_set\*(C'\fR macro, but using it is utterly pointless,
911believe me. 1832believe me.
1833.PP
1834Example: Dynamically allocate an \f(CW\*(C`ev_idle\*(C'\fR watcher, start it, and in the
1835callback, free it. Also, use no error checking, as usual.
1836.PP
1837.Vb 7
1838\& static void
1839\& idle_cb (struct ev_loop *loop, struct ev_idle *w, int revents)
1840\& {
1841\& free (w);
1842\& // now do something you wanted to do when the program has
1843\& // no longer asnything immediate to do.
1844\& }
1845.Ve
1846.PP
1847.Vb 3
1848\& struct ev_idle *idle_watcher = malloc (sizeof (struct ev_idle));
1849\& ev_idle_init (idle_watcher, idle_cb);
1850\& ev_idle_start (loop, idle_cb);
1851.Ve
912.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop" 1852.ie n .Sh """ev_prepare""\fP and \f(CW""ev_check"" \- customise your event loop!"
913.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop" 1853.el .Sh "\f(CWev_prepare\fP and \f(CWev_check\fP \- customise your event loop!"
914.IX Subsection "ev_prepare and ev_check - customise your event loop" 1854.IX Subsection "ev_prepare and ev_check - customise your event loop!"
915Prepare and check watchers are usually (but not always) used in tandem: 1855Prepare and check watchers are usually (but not always) used in tandem:
916prepare watchers get invoked before the process blocks and check watchers 1856prepare watchers get invoked before the process blocks and check watchers
917afterwards. 1857afterwards.
918.PP 1858.PP
1859You \fImust not\fR call \f(CW\*(C`ev_loop\*(C'\fR or similar functions that enter
1860the current event loop from either \f(CW\*(C`ev_prepare\*(C'\fR or \f(CW\*(C`ev_check\*(C'\fR
1861watchers. Other loops than the current one are fine, however. The
1862rationale behind this is that you do not need to check for recursion in
1863those watchers, i.e. the sequence will always be \f(CW\*(C`ev_prepare\*(C'\fR, blocking,
1864\&\f(CW\*(C`ev_check\*(C'\fR so if you have one watcher of each kind they will always be
1865called in pairs bracketing the blocking call.
1866.PP
919Their main purpose is to integrate other event mechanisms into libev. This 1867Their main purpose is to integrate other event mechanisms into libev and
920could be used, for example, to track variable changes, implement your own 1868their use is somewhat advanced. This could be used, for example, to track
921watchers, integrate net-snmp or a coroutine library and lots more. 1869variable changes, implement your own watchers, integrate net-snmp or a
1870coroutine library and lots more. They are also occasionally useful if
1871you cache some data and want to flush it before blocking (for example,
1872in X programs you might want to do an \f(CW\*(C`XFlush ()\*(C'\fR in an \f(CW\*(C`ev_prepare\*(C'\fR
1873watcher).
922.PP 1874.PP
923This is done by examining in each prepare call which file descriptors need 1875This is done by examining in each prepare call which file descriptors need
924to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for 1876to be watched by the other library, registering \f(CW\*(C`ev_io\*(C'\fR watchers for
925them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries 1877them and starting an \f(CW\*(C`ev_timer\*(C'\fR watcher for any timeouts (many libraries
926provide just this functionality). Then, in the check watcher you check for 1878provide just this functionality). Then, in the check watcher you check for
… …
935are ready to run (it's actually more complicated: it only runs coroutines 1887are ready to run (it's actually more complicated: it only runs coroutines
936with priority higher than or equal to the event loop and one coroutine 1888with priority higher than or equal to the event loop and one coroutine
937of lower priority, but only once, using idle watchers to keep the event 1889of lower priority, but only once, using idle watchers to keep the event
938loop from blocking if lower-priority coroutines are active, thus mapping 1890loop from blocking if lower-priority coroutines are active, thus mapping
939low-priority coroutines to idle/background tasks). 1891low-priority coroutines to idle/background tasks).
1892.PP
1893It is recommended to give \f(CW\*(C`ev_check\*(C'\fR watchers highest (\f(CW\*(C`EV_MAXPRI\*(C'\fR)
1894priority, to ensure that they are being run before any other watchers
1895after the poll. Also, \f(CW\*(C`ev_check\*(C'\fR watchers (and \f(CW\*(C`ev_prepare\*(C'\fR watchers,
1896too) should not activate (\*(L"feed\*(R") events into libev. While libev fully
1897supports this, they will be called before other \f(CW\*(C`ev_check\*(C'\fR watchers
1898did their job. As \f(CW\*(C`ev_check\*(C'\fR watchers are often used to embed other
1899(non\-libev) event loops those other event loops might be in an unusable
1900state until their \f(CW\*(C`ev_check\*(C'\fR watcher ran (always remind yourself to
1901coexist peacefully with others).
1902.PP
1903\fIWatcher-Specific Functions and Data Members\fR
1904.IX Subsection "Watcher-Specific Functions and Data Members"
940.IP "ev_prepare_init (ev_prepare *, callback)" 4 1905.IP "ev_prepare_init (ev_prepare *, callback)" 4
941.IX Item "ev_prepare_init (ev_prepare *, callback)" 1906.IX Item "ev_prepare_init (ev_prepare *, callback)"
942.PD 0 1907.PD 0
943.IP "ev_check_init (ev_check *, callback)" 4 1908.IP "ev_check_init (ev_check *, callback)" 4
944.IX Item "ev_check_init (ev_check *, callback)" 1909.IX Item "ev_check_init (ev_check *, callback)"
945.PD 1910.PD
946Initialises and configures the prepare or check watcher \- they have no 1911Initialises and configures the prepare or check watcher \- they have no
947parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR 1912parameters of any kind. There are \f(CW\*(C`ev_prepare_set\*(C'\fR and \f(CW\*(C`ev_check_set\*(C'\fR
948macros, but using them is utterly, utterly and completely pointless. 1913macros, but using them is utterly, utterly and completely pointless.
1914.PP
1915There are a number of principal ways to embed other event loops or modules
1916into libev. Here are some ideas on how to include libadns into libev
1917(there is a Perl module named \f(CW\*(C`EV::ADNS\*(C'\fR that does this, which you could
1918use for an actually working example. Another Perl module named \f(CW\*(C`EV::Glib\*(C'\fR
1919embeds a Glib main context into libev, and finally, \f(CW\*(C`Glib::EV\*(C'\fR embeds \s-1EV\s0
1920into the Glib event loop).
1921.PP
1922Method 1: Add \s-1IO\s0 watchers and a timeout watcher in a prepare handler,
1923and in a check watcher, destroy them and call into libadns. What follows
1924is pseudo-code only of course. This requires you to either use a low
1925priority for the check watcher or use \f(CW\*(C`ev_clear_pending\*(C'\fR explicitly, as
1926the callbacks for the IO/timeout watchers might not have been called yet.
1927.PP
1928.Vb 2
1929\& static ev_io iow [nfd];
1930\& static ev_timer tw;
1931.Ve
1932.PP
1933.Vb 4
1934\& static void
1935\& io_cb (ev_loop *loop, ev_io *w, int revents)
1936\& {
1937\& }
1938.Ve
1939.PP
1940.Vb 8
1941\& // create io watchers for each fd and a timer before blocking
1942\& static void
1943\& adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents)
1944\& {
1945\& int timeout = 3600000;
1946\& struct pollfd fds [nfd];
1947\& // actual code will need to loop here and realloc etc.
1948\& adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
1949.Ve
1950.PP
1951.Vb 3
1952\& /* the callback is illegal, but won't be called as we stop during check */
1953\& ev_timer_init (&tw, 0, timeout * 1e-3);
1954\& ev_timer_start (loop, &tw);
1955.Ve
1956.PP
1957.Vb 6
1958\& // create one ev_io per pollfd
1959\& for (int i = 0; i < nfd; ++i)
1960\& {
1961\& ev_io_init (iow + i, io_cb, fds [i].fd,
1962\& ((fds [i].events & POLLIN ? EV_READ : 0)
1963\& | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
1964.Ve
1965.PP
1966.Vb 4
1967\& fds [i].revents = 0;
1968\& ev_io_start (loop, iow + i);
1969\& }
1970\& }
1971.Ve
1972.PP
1973.Vb 5
1974\& // stop all watchers after blocking
1975\& static void
1976\& adns_check_cb (ev_loop *loop, ev_check *w, int revents)
1977\& {
1978\& ev_timer_stop (loop, &tw);
1979.Ve
1980.PP
1981.Vb 8
1982\& for (int i = 0; i < nfd; ++i)
1983\& {
1984\& // set the relevant poll flags
1985\& // could also call adns_processreadable etc. here
1986\& struct pollfd *fd = fds + i;
1987\& int revents = ev_clear_pending (iow + i);
1988\& if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
1989\& if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
1990.Ve
1991.PP
1992.Vb 3
1993\& // now stop the watcher
1994\& ev_io_stop (loop, iow + i);
1995\& }
1996.Ve
1997.PP
1998.Vb 2
1999\& adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
2000\& }
2001.Ve
2002.PP
2003Method 2: This would be just like method 1, but you run \f(CW\*(C`adns_afterpoll\*(C'\fR
2004in the prepare watcher and would dispose of the check watcher.
2005.PP
2006Method 3: If the module to be embedded supports explicit event
2007notification (adns does), you can also make use of the actual watcher
2008callbacks, and only destroy/create the watchers in the prepare watcher.
2009.PP
2010.Vb 5
2011\& static void
2012\& timer_cb (EV_P_ ev_timer *w, int revents)
2013\& {
2014\& adns_state ads = (adns_state)w->data;
2015\& update_now (EV_A);
2016.Ve
2017.PP
2018.Vb 2
2019\& adns_processtimeouts (ads, &tv_now);
2020\& }
2021.Ve
2022.PP
2023.Vb 5
2024\& static void
2025\& io_cb (EV_P_ ev_io *w, int revents)
2026\& {
2027\& adns_state ads = (adns_state)w->data;
2028\& update_now (EV_A);
2029.Ve
2030.PP
2031.Vb 3
2032\& if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
2033\& if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
2034\& }
2035.Ve
2036.PP
2037.Vb 1
2038\& // do not ever call adns_afterpoll
2039.Ve
2040.PP
2041Method 4: Do not use a prepare or check watcher because the module you
2042want to embed is too inflexible to support it. Instead, youc na override
2043their poll function. The drawback with this solution is that the main
2044loop is now no longer controllable by \s-1EV\s0. The \f(CW\*(C`Glib::EV\*(C'\fR module does
2045this.
2046.PP
2047.Vb 4
2048\& static gint
2049\& event_poll_func (GPollFD *fds, guint nfds, gint timeout)
2050\& {
2051\& int got_events = 0;
2052.Ve
2053.PP
2054.Vb 2
2055\& for (n = 0; n < nfds; ++n)
2056\& // create/start io watcher that sets the relevant bits in fds[n] and increment got_events
2057.Ve
2058.PP
2059.Vb 2
2060\& if (timeout >= 0)
2061\& // create/start timer
2062.Ve
2063.PP
2064.Vb 2
2065\& // poll
2066\& ev_loop (EV_A_ 0);
2067.Ve
2068.PP
2069.Vb 3
2070\& // stop timer again
2071\& if (timeout >= 0)
2072\& ev_timer_stop (EV_A_ &to);
2073.Ve
2074.PP
2075.Vb 3
2076\& // stop io watchers again - their callbacks should have set
2077\& for (n = 0; n < nfds; ++n)
2078\& ev_io_stop (EV_A_ iow [n]);
2079.Ve
2080.PP
2081.Vb 2
2082\& return got_events;
2083\& }
2084.Ve
2085.ie n .Sh """ev_embed"" \- when one backend isn't enough..."
2086.el .Sh "\f(CWev_embed\fP \- when one backend isn't enough..."
2087.IX Subsection "ev_embed - when one backend isn't enough..."
2088This is a rather advanced watcher type that lets you embed one event loop
2089into another (currently only \f(CW\*(C`ev_io\*(C'\fR events are supported in the embedded
2090loop, other types of watchers might be handled in a delayed or incorrect
2091fashion and must not be used).
2092.PP
2093There are primarily two reasons you would want that: work around bugs and
2094prioritise I/O.
2095.PP
2096As an example for a bug workaround, the kqueue backend might only support
2097sockets on some platform, so it is unusable as generic backend, but you
2098still want to make use of it because you have many sockets and it scales
2099so nicely. In this case, you would create a kqueue-based loop and embed it
2100into your default loop (which might use e.g. poll). Overall operation will
2101be a bit slower because first libev has to poll and then call kevent, but
2102at least you can use both at what they are best.
2103.PP
2104As for prioritising I/O: rarely you have the case where some fds have
2105to be watched and handled very quickly (with low latency), and even
2106priorities and idle watchers might have too much overhead. In this case
2107you would put all the high priority stuff in one loop and all the rest in
2108a second one, and embed the second one in the first.
2109.PP
2110As long as the watcher is active, the callback will be invoked every time
2111there might be events pending in the embedded loop. The callback must then
2112call \f(CW\*(C`ev_embed_sweep (mainloop, watcher)\*(C'\fR to make a single sweep and invoke
2113their callbacks (you could also start an idle watcher to give the embedded
2114loop strictly lower priority for example). You can also set the callback
2115to \f(CW0\fR, in which case the embed watcher will automatically execute the
2116embedded loop sweep.
2117.PP
2118As long as the watcher is started it will automatically handle events. The
2119callback will be invoked whenever some events have been handled. You can
2120set the callback to \f(CW0\fR to avoid having to specify one if you are not
2121interested in that.
2122.PP
2123Also, there have not currently been made special provisions for forking:
2124when you fork, you not only have to call \f(CW\*(C`ev_loop_fork\*(C'\fR on both loops,
2125but you will also have to stop and restart any \f(CW\*(C`ev_embed\*(C'\fR watchers
2126yourself.
2127.PP
2128Unfortunately, not all backends are embeddable, only the ones returned by
2129\&\f(CW\*(C`ev_embeddable_backends\*(C'\fR are, which, unfortunately, does not include any
2130portable one.
2131.PP
2132So when you want to use this feature you will always have to be prepared
2133that you cannot get an embeddable loop. The recommended way to get around
2134this is to have a separate variables for your embeddable loop, try to
2135create it, and if that fails, use the normal loop for everything:
2136.PP
2137.Vb 3
2138\& struct ev_loop *loop_hi = ev_default_init (0);
2139\& struct ev_loop *loop_lo = 0;
2140\& struct ev_embed embed;
2141.Ve
2142.PP
2143.Vb 5
2144\& // see if there is a chance of getting one that works
2145\& // (remember that a flags value of 0 means autodetection)
2146\& loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
2147\& ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
2148\& : 0;
2149.Ve
2150.PP
2151.Vb 8
2152\& // if we got one, then embed it, otherwise default to loop_hi
2153\& if (loop_lo)
2154\& {
2155\& ev_embed_init (&embed, 0, loop_lo);
2156\& ev_embed_start (loop_hi, &embed);
2157\& }
2158\& else
2159\& loop_lo = loop_hi;
2160.Ve
2161.PP
2162\fIWatcher-Specific Functions and Data Members\fR
2163.IX Subsection "Watcher-Specific Functions and Data Members"
2164.IP "ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)" 4
2165.IX Item "ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)"
2166.PD 0
2167.IP "ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)" 4
2168.IX Item "ev_embed_set (ev_embed *, callback, struct ev_loop *embedded_loop)"
2169.PD
2170Configures the watcher to embed the given loop, which must be
2171embeddable. If the callback is \f(CW0\fR, then \f(CW\*(C`ev_embed_sweep\*(C'\fR will be
2172invoked automatically, otherwise it is the responsibility of the callback
2173to invoke it (it will continue to be called until the sweep has been done,
2174if you do not want thta, you need to temporarily stop the embed watcher).
2175.IP "ev_embed_sweep (loop, ev_embed *)" 4
2176.IX Item "ev_embed_sweep (loop, ev_embed *)"
2177Make a single, non-blocking sweep over the embedded loop. This works
2178similarly to \f(CW\*(C`ev_loop (embedded_loop, EVLOOP_NONBLOCK)\*(C'\fR, but in the most
2179apropriate way for embedded loops.
2180.IP "struct ev_loop *other [read\-only]" 4
2181.IX Item "struct ev_loop *other [read-only]"
2182The embedded event loop.
2183.ie n .Sh """ev_fork"" \- the audacity to resume the event loop after a fork"
2184.el .Sh "\f(CWev_fork\fP \- the audacity to resume the event loop after a fork"
2185.IX Subsection "ev_fork - the audacity to resume the event loop after a fork"
2186Fork watchers are called when a \f(CW\*(C`fork ()\*(C'\fR was detected (usually because
2187whoever is a good citizen cared to tell libev about it by calling
2188\&\f(CW\*(C`ev_default_fork\*(C'\fR or \f(CW\*(C`ev_loop_fork\*(C'\fR). The invocation is done before the
2189event loop blocks next and before \f(CW\*(C`ev_check\*(C'\fR watchers are being called,
2190and only in the child after the fork. If whoever good citizen calling
2191\&\f(CW\*(C`ev_default_fork\*(C'\fR cheats and calls it in the wrong process, the fork
2192handlers will be invoked, too, of course.
2193.PP
2194\fIWatcher-Specific Functions and Data Members\fR
2195.IX Subsection "Watcher-Specific Functions and Data Members"
2196.IP "ev_fork_init (ev_signal *, callback)" 4
2197.IX Item "ev_fork_init (ev_signal *, callback)"
2198Initialises and configures the fork watcher \- it has no parameters of any
2199kind. There is a \f(CW\*(C`ev_fork_set\*(C'\fR macro, but using it is utterly pointless,
2200believe me.
949.SH "OTHER FUNCTIONS" 2201.SH "OTHER FUNCTIONS"
950.IX Header "OTHER FUNCTIONS" 2202.IX Header "OTHER FUNCTIONS"
951There are some other functions of possible interest. Described. Here. Now. 2203There are some other functions of possible interest. Described. Here. Now.
952.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4 2204.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4
953.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 2205.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)"
… …
982.Ve 2234.Ve
983.Sp 2235.Sp
984.Vb 1 2236.Vb 1
985\& ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0); 2237\& ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
986.Ve 2238.Ve
987.IP "ev_feed_event (loop, watcher, int events)" 4 2239.IP "ev_feed_event (ev_loop *, watcher *, int revents)" 4
988.IX Item "ev_feed_event (loop, watcher, int events)" 2240.IX Item "ev_feed_event (ev_loop *, watcher *, int revents)"
989Feeds the given event set into the event loop, as if the specified event 2241Feeds the given event set into the event loop, as if the specified event
990had happened for the specified watcher (which must be a pointer to an 2242had happened for the specified watcher (which must be a pointer to an
991initialised but not necessarily started event watcher). 2243initialised but not necessarily started event watcher).
992.IP "ev_feed_fd_event (loop, int fd, int revents)" 4 2244.IP "ev_feed_fd_event (ev_loop *, int fd, int revents)" 4
993.IX Item "ev_feed_fd_event (loop, int fd, int revents)" 2245.IX Item "ev_feed_fd_event (ev_loop *, int fd, int revents)"
994Feed an event on the given fd, as if a file descriptor backend detected 2246Feed an event on the given fd, as if a file descriptor backend detected
995the given events it. 2247the given events it.
996.IP "ev_feed_signal_event (loop, int signum)" 4 2248.IP "ev_feed_signal_event (ev_loop *loop, int signum)" 4
997.IX Item "ev_feed_signal_event (loop, int signum)" 2249.IX Item "ev_feed_signal_event (ev_loop *loop, int signum)"
998Feed an event as if the given signal occured (loop must be the default loop!). 2250Feed an event as if the given signal occured (\f(CW\*(C`loop\*(C'\fR must be the default
2251loop!).
999.SH "LIBEVENT EMULATION" 2252.SH "LIBEVENT EMULATION"
1000.IX Header "LIBEVENT EMULATION" 2253.IX Header "LIBEVENT EMULATION"
1001Libev offers a compatibility emulation layer for libevent. It cannot 2254Libev offers a compatibility emulation layer for libevent. It cannot
1002emulate the internals of libevent, so here are some usage hints: 2255emulate the internals of libevent, so here are some usage hints:
1003.IP "* Use it by including <event.h>, as usual." 4 2256.IP "* Use it by including <event.h>, as usual." 4
… …
1014.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4 2267.IP "* The libev emulation is \fInot\fR \s-1ABI\s0 compatible to libevent, you need to use the libev header file and library." 4
1015.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library." 2268.IX Item "The libev emulation is not ABI compatible to libevent, you need to use the libev header file and library."
1016.PD 2269.PD
1017.SH "\*(C+ SUPPORT" 2270.SH "\*(C+ SUPPORT"
1018.IX Header " SUPPORT" 2271.IX Header " SUPPORT"
1019\&\s-1TBD\s0. 2272Libev comes with some simplistic wrapper classes for \*(C+ that mainly allow
2273you to use some convinience methods to start/stop watchers and also change
2274the callback model to a model using method callbacks on objects.
2275.PP
2276To use it,
2277.PP
2278.Vb 1
2279\& #include <ev++.h>
2280.Ve
2281.PP
2282This automatically includes \fIev.h\fR and puts all of its definitions (many
2283of them macros) into the global namespace. All \*(C+ specific things are
2284put into the \f(CW\*(C`ev\*(C'\fR namespace. It should support all the same embedding
2285options as \fIev.h\fR, most notably \f(CW\*(C`EV_MULTIPLICITY\*(C'\fR.
2286.PP
2287Care has been taken to keep the overhead low. The only data member the \*(C+
2288classes add (compared to plain C\-style watchers) is the event loop pointer
2289that the watcher is associated with (or no additional members at all if
2290you disable \f(CW\*(C`EV_MULTIPLICITY\*(C'\fR when embedding libev).
2291.PP
2292Currently, functions, and static and non-static member functions can be
2293used as callbacks. Other types should be easy to add as long as they only
2294need one additional pointer for context. If you need support for other
2295types of functors please contact the author (preferably after implementing
2296it).
2297.PP
2298Here is a list of things available in the \f(CW\*(C`ev\*(C'\fR namespace:
2299.ie n .IP """ev::READ""\fR, \f(CW""ev::WRITE"" etc." 4
2300.el .IP "\f(CWev::READ\fR, \f(CWev::WRITE\fR etc." 4
2301.IX Item "ev::READ, ev::WRITE etc."
2302These are just enum values with the same values as the \f(CW\*(C`EV_READ\*(C'\fR etc.
2303macros from \fIev.h\fR.
2304.ie n .IP """ev::tstamp""\fR, \f(CW""ev::now""" 4
2305.el .IP "\f(CWev::tstamp\fR, \f(CWev::now\fR" 4
2306.IX Item "ev::tstamp, ev::now"
2307Aliases to the same types/functions as with the \f(CW\*(C`ev_\*(C'\fR prefix.
2308.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
2309.el .IP "\f(CWev::io\fR, \f(CWev::timer\fR, \f(CWev::periodic\fR, \f(CWev::idle\fR, \f(CWev::sig\fR etc." 4
2310.IX Item "ev::io, ev::timer, ev::periodic, ev::idle, ev::sig etc."
2311For each \f(CW\*(C`ev_TYPE\*(C'\fR watcher in \fIev.h\fR there is a corresponding class of
2312the same name in the \f(CW\*(C`ev\*(C'\fR namespace, with the exception of \f(CW\*(C`ev_signal\*(C'\fR
2313which is called \f(CW\*(C`ev::sig\*(C'\fR to avoid clashes with the \f(CW\*(C`signal\*(C'\fR macro
2314defines by many implementations.
2315.Sp
2316All of those classes have these methods:
2317.RS 4
2318.IP "ev::TYPE::TYPE ()" 4
2319.IX Item "ev::TYPE::TYPE ()"
2320.PD 0
2321.IP "ev::TYPE::TYPE (struct ev_loop *)" 4
2322.IX Item "ev::TYPE::TYPE (struct ev_loop *)"
2323.IP "ev::TYPE::~TYPE" 4
2324.IX Item "ev::TYPE::~TYPE"
2325.PD
2326The constructor (optionally) takes an event loop to associate the watcher
2327with. If it is omitted, it will use \f(CW\*(C`EV_DEFAULT\*(C'\fR.
2328.Sp
2329The constructor calls \f(CW\*(C`ev_init\*(C'\fR for you, which means you have to call the
2330\&\f(CW\*(C`set\*(C'\fR method before starting it.
2331.Sp
2332It will not set a callback, however: You have to call the templated \f(CW\*(C`set\*(C'\fR
2333method to set a callback before you can start the watcher.
2334.Sp
2335(The reason why you have to use a method is a limitation in \*(C+ which does
2336not allow explicit template arguments for constructors).
2337.Sp
2338The destructor automatically stops the watcher if it is active.
2339.IP "w\->set<class, &class::method> (object *)" 4
2340.IX Item "w->set<class, &class::method> (object *)"
2341This method sets the callback method to call. The method has to have a
2342signature of \f(CW\*(C`void (*)(ev_TYPE &, int)\*(C'\fR, it receives the watcher as
2343first argument and the \f(CW\*(C`revents\*(C'\fR as second. The object must be given as
2344parameter and is stored in the \f(CW\*(C`data\*(C'\fR member of the watcher.
2345.Sp
2346This method synthesizes efficient thunking code to call your method from
2347the C callback that libev requires. If your compiler can inline your
2348callback (i.e. it is visible to it at the place of the \f(CW\*(C`set\*(C'\fR call and
2349your compiler is good :), then the method will be fully inlined into the
2350thunking function, making it as fast as a direct C callback.
2351.Sp
2352Example: simple class declaration and watcher initialisation
2353.Sp
2354.Vb 4
2355\& struct myclass
2356\& {
2357\& void io_cb (ev::io &w, int revents) { }
2358\& }
2359.Ve
2360.Sp
2361.Vb 3
2362\& myclass obj;
2363\& ev::io iow;
2364\& iow.set <myclass, &myclass::io_cb> (&obj);
2365.Ve
2366.IP "w\->set<function> (void *data = 0)" 4
2367.IX Item "w->set<function> (void *data = 0)"
2368Also sets a callback, but uses a static method or plain function as
2369callback. The optional \f(CW\*(C`data\*(C'\fR argument will be stored in the watcher's
2370\&\f(CW\*(C`data\*(C'\fR member and is free for you to use.
2371.Sp
2372The prototype of the \f(CW\*(C`function\*(C'\fR must be \f(CW\*(C`void (*)(ev::TYPE &w, int)\*(C'\fR.
2373.Sp
2374See the method\-\f(CW\*(C`set\*(C'\fR above for more details.
2375.Sp
2376Example:
2377.Sp
2378.Vb 2
2379\& static void io_cb (ev::io &w, int revents) { }
2380\& iow.set <io_cb> ();
2381.Ve
2382.IP "w\->set (struct ev_loop *)" 4
2383.IX Item "w->set (struct ev_loop *)"
2384Associates a different \f(CW\*(C`struct ev_loop\*(C'\fR with this watcher. You can only
2385do this when the watcher is inactive (and not pending either).
2386.IP "w\->set ([args])" 4
2387.IX Item "w->set ([args])"
2388Basically the same as \f(CW\*(C`ev_TYPE_set\*(C'\fR, with the same args. Must be
2389called at least once. Unlike the C counterpart, an active watcher gets
2390automatically stopped and restarted when reconfiguring it with this
2391method.
2392.IP "w\->start ()" 4
2393.IX Item "w->start ()"
2394Starts the watcher. Note that there is no \f(CW\*(C`loop\*(C'\fR argument, as the
2395constructor already stores the event loop.
2396.IP "w\->stop ()" 4
2397.IX Item "w->stop ()"
2398Stops the watcher if it is active. Again, no \f(CW\*(C`loop\*(C'\fR argument.
2399.ie n .IP "w\->again () (""ev::timer""\fR, \f(CW""ev::periodic"" only)" 4
2400.el .IP "w\->again () (\f(CWev::timer\fR, \f(CWev::periodic\fR only)" 4
2401.IX Item "w->again () (ev::timer, ev::periodic only)"
2402For \f(CW\*(C`ev::timer\*(C'\fR and \f(CW\*(C`ev::periodic\*(C'\fR, this invokes the corresponding
2403\&\f(CW\*(C`ev_TYPE_again\*(C'\fR function.
2404.ie n .IP "w\->sweep () (""ev::embed"" only)" 4
2405.el .IP "w\->sweep () (\f(CWev::embed\fR only)" 4
2406.IX Item "w->sweep () (ev::embed only)"
2407Invokes \f(CW\*(C`ev_embed_sweep\*(C'\fR.
2408.ie n .IP "w\->update () (""ev::stat"" only)" 4
2409.el .IP "w\->update () (\f(CWev::stat\fR only)" 4
2410.IX Item "w->update () (ev::stat only)"
2411Invokes \f(CW\*(C`ev_stat_stat\*(C'\fR.
2412.RE
2413.RS 4
2414.RE
2415.PP
2416Example: Define a class with an \s-1IO\s0 and idle watcher, start one of them in
2417the constructor.
2418.PP
2419.Vb 4
2420\& class myclass
2421\& {
2422\& ev_io io; void io_cb (ev::io &w, int revents);
2423\& ev_idle idle void idle_cb (ev::idle &w, int revents);
2424.Ve
2425.PP
2426.Vb 2
2427\& myclass ();
2428\& }
2429.Ve
2430.PP
2431.Vb 4
2432\& myclass::myclass (int fd)
2433\& {
2434\& io .set <myclass, &myclass::io_cb > (this);
2435\& idle.set <myclass, &myclass::idle_cb> (this);
2436.Ve
2437.PP
2438.Vb 2
2439\& io.start (fd, ev::READ);
2440\& }
2441.Ve
2442.SH "MACRO MAGIC"
2443.IX Header "MACRO MAGIC"
2444Libev can be compiled with a variety of options, the most fundamantal
2445of which is \f(CW\*(C`EV_MULTIPLICITY\*(C'\fR. This option determines whether (most)
2446functions and callbacks have an initial \f(CW\*(C`struct ev_loop *\*(C'\fR argument.
2447.PP
2448To make it easier to write programs that cope with either variant, the
2449following macros are defined:
2450.ie n .IP """EV_A""\fR, \f(CW""EV_A_""" 4
2451.el .IP "\f(CWEV_A\fR, \f(CWEV_A_\fR" 4
2452.IX Item "EV_A, EV_A_"
2453This provides the loop \fIargument\fR for functions, if one is required (\*(L"ev
2454loop argument\*(R"). The \f(CW\*(C`EV_A\*(C'\fR form is used when this is the sole argument,
2455\&\f(CW\*(C`EV_A_\*(C'\fR is used when other arguments are following. Example:
2456.Sp
2457.Vb 3
2458\& ev_unref (EV_A);
2459\& ev_timer_add (EV_A_ watcher);
2460\& ev_loop (EV_A_ 0);
2461.Ve
2462.Sp
2463It assumes the variable \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR is in scope,
2464which is often provided by the following macro.
2465.ie n .IP """EV_P""\fR, \f(CW""EV_P_""" 4
2466.el .IP "\f(CWEV_P\fR, \f(CWEV_P_\fR" 4
2467.IX Item "EV_P, EV_P_"
2468This provides the loop \fIparameter\fR for functions, if one is required (\*(L"ev
2469loop parameter\*(R"). The \f(CW\*(C`EV_P\*(C'\fR form is used when this is the sole parameter,
2470\&\f(CW\*(C`EV_P_\*(C'\fR is used when other parameters are following. Example:
2471.Sp
2472.Vb 2
2473\& // this is how ev_unref is being declared
2474\& static void ev_unref (EV_P);
2475.Ve
2476.Sp
2477.Vb 2
2478\& // this is how you can declare your typical callback
2479\& static void cb (EV_P_ ev_timer *w, int revents)
2480.Ve
2481.Sp
2482It declares a parameter \f(CW\*(C`loop\*(C'\fR of type \f(CW\*(C`struct ev_loop *\*(C'\fR, quite
2483suitable for use with \f(CW\*(C`EV_A\*(C'\fR.
2484.ie n .IP """EV_DEFAULT""\fR, \f(CW""EV_DEFAULT_""" 4
2485.el .IP "\f(CWEV_DEFAULT\fR, \f(CWEV_DEFAULT_\fR" 4
2486.IX Item "EV_DEFAULT, EV_DEFAULT_"
2487Similar to the other two macros, this gives you the value of the default
2488loop, if multiple loops are supported (\*(L"ev loop default\*(R").
2489.PP
2490Example: Declare and initialise a check watcher, utilising the above
2491macros so it will work regardless of whether multiple loops are supported
2492or not.
2493.PP
2494.Vb 5
2495\& static void
2496\& check_cb (EV_P_ ev_timer *w, int revents)
2497\& {
2498\& ev_check_stop (EV_A_ w);
2499\& }
2500.Ve
2501.PP
2502.Vb 4
2503\& ev_check check;
2504\& ev_check_init (&check, check_cb);
2505\& ev_check_start (EV_DEFAULT_ &check);
2506\& ev_loop (EV_DEFAULT_ 0);
2507.Ve
2508.SH "EMBEDDING"
2509.IX Header "EMBEDDING"
2510Libev can (and often is) directly embedded into host
2511applications. Examples of applications that embed it include the Deliantra
2512Game Server, the \s-1EV\s0 perl module, the \s-1GNU\s0 Virtual Private Ethernet (gvpe)
2513and rxvt\-unicode.
2514.PP
2515The goal is to enable you to just copy the necessary files into your
2516source directory without having to change even a single line in them, so
2517you can easily upgrade by simply copying (or having a checked-out copy of
2518libev somewhere in your source tree).
2519.Sh "\s-1FILESETS\s0"
2520.IX Subsection "FILESETS"
2521Depending on what features you need you need to include one or more sets of files
2522in your app.
2523.PP
2524\fI\s-1CORE\s0 \s-1EVENT\s0 \s-1LOOP\s0\fR
2525.IX Subsection "CORE EVENT LOOP"
2526.PP
2527To include only the libev core (all the \f(CW\*(C`ev_*\*(C'\fR functions), with manual
2528configuration (no autoconf):
2529.PP
2530.Vb 2
2531\& #define EV_STANDALONE 1
2532\& #include "ev.c"
2533.Ve
2534.PP
2535This will automatically include \fIev.h\fR, too, and should be done in a
2536single C source file only to provide the function implementations. To use
2537it, do the same for \fIev.h\fR in all files wishing to use this \s-1API\s0 (best
2538done by writing a wrapper around \fIev.h\fR that you can include instead and
2539where you can put other configuration options):
2540.PP
2541.Vb 2
2542\& #define EV_STANDALONE 1
2543\& #include "ev.h"
2544.Ve
2545.PP
2546Both header files and implementation files can be compiled with a \*(C+
2547compiler (at least, thats a stated goal, and breakage will be treated
2548as a bug).
2549.PP
2550You need the following files in your source tree, or in a directory
2551in your include path (e.g. in libev/ when using \-Ilibev):
2552.PP
2553.Vb 4
2554\& ev.h
2555\& ev.c
2556\& ev_vars.h
2557\& ev_wrap.h
2558.Ve
2559.PP
2560.Vb 1
2561\& ev_win32.c required on win32 platforms only
2562.Ve
2563.PP
2564.Vb 5
2565\& ev_select.c only when select backend is enabled (which is enabled by default)
2566\& ev_poll.c only when poll backend is enabled (disabled by default)
2567\& ev_epoll.c only when the epoll backend is enabled (disabled by default)
2568\& ev_kqueue.c only when the kqueue backend is enabled (disabled by default)
2569\& ev_port.c only when the solaris port backend is enabled (disabled by default)
2570.Ve
2571.PP
2572\&\fIev.c\fR includes the backend files directly when enabled, so you only need
2573to compile this single file.
2574.PP
2575\fI\s-1LIBEVENT\s0 \s-1COMPATIBILITY\s0 \s-1API\s0\fR
2576.IX Subsection "LIBEVENT COMPATIBILITY API"
2577.PP
2578To include the libevent compatibility \s-1API\s0, also include:
2579.PP
2580.Vb 1
2581\& #include "event.c"
2582.Ve
2583.PP
2584in the file including \fIev.c\fR, and:
2585.PP
2586.Vb 1
2587\& #include "event.h"
2588.Ve
2589.PP
2590in the files that want to use the libevent \s-1API\s0. This also includes \fIev.h\fR.
2591.PP
2592You need the following additional files for this:
2593.PP
2594.Vb 2
2595\& event.h
2596\& event.c
2597.Ve
2598.PP
2599\fI\s-1AUTOCONF\s0 \s-1SUPPORT\s0\fR
2600.IX Subsection "AUTOCONF SUPPORT"
2601.PP
2602Instead of using \f(CW\*(C`EV_STANDALONE=1\*(C'\fR and providing your config in
2603whatever way you want, you can also \f(CW\*(C`m4_include([libev.m4])\*(C'\fR in your
2604\&\fIconfigure.ac\fR and leave \f(CW\*(C`EV_STANDALONE\*(C'\fR undefined. \fIev.c\fR will then
2605include \fIconfig.h\fR and configure itself accordingly.
2606.PP
2607For this of course you need the m4 file:
2608.PP
2609.Vb 1
2610\& libev.m4
2611.Ve
2612.Sh "\s-1PREPROCESSOR\s0 \s-1SYMBOLS/MACROS\s0"
2613.IX Subsection "PREPROCESSOR SYMBOLS/MACROS"
2614Libev can be configured via a variety of preprocessor symbols you have to define
2615before including any of its files. The default is not to build for multiplicity
2616and only include the select backend.
2617.IP "\s-1EV_STANDALONE\s0" 4
2618.IX Item "EV_STANDALONE"
2619Must always be \f(CW1\fR if you do not use autoconf configuration, which
2620keeps libev from including \fIconfig.h\fR, and it also defines dummy
2621implementations for some libevent functions (such as logging, which is not
2622supported). It will also not define any of the structs usually found in
2623\&\fIevent.h\fR that are not directly supported by the libev core alone.
2624.IP "\s-1EV_USE_MONOTONIC\s0" 4
2625.IX Item "EV_USE_MONOTONIC"
2626If defined to be \f(CW1\fR, libev will try to detect the availability of the
2627monotonic clock option at both compiletime and runtime. Otherwise no use
2628of the monotonic clock option will be attempted. If you enable this, you
2629usually have to link against librt or something similar. Enabling it when
2630the functionality isn't available is safe, though, although you have
2631to make sure you link against any libraries where the \f(CW\*(C`clock_gettime\*(C'\fR
2632function is hiding in (often \fI\-lrt\fR).
2633.IP "\s-1EV_USE_REALTIME\s0" 4
2634.IX Item "EV_USE_REALTIME"
2635If defined to be \f(CW1\fR, libev will try to detect the availability of the
2636realtime clock option at compiletime (and assume its availability at
2637runtime if successful). Otherwise no use of the realtime clock option will
2638be attempted. This effectively replaces \f(CW\*(C`gettimeofday\*(C'\fR by \f(CW\*(C`clock_get
2639(CLOCK_REALTIME, ...)\*(C'\fR and will not normally affect correctness. See the
2640note about libraries in the description of \f(CW\*(C`EV_USE_MONOTONIC\*(C'\fR, though.
2641.IP "\s-1EV_USE_NANOSLEEP\s0" 4
2642.IX Item "EV_USE_NANOSLEEP"
2643If defined to be \f(CW1\fR, libev will assume that \f(CW\*(C`nanosleep ()\*(C'\fR is available
2644and will use it for delays. Otherwise it will use \f(CW\*(C`select ()\*(C'\fR.
2645.IP "\s-1EV_USE_SELECT\s0" 4
2646.IX Item "EV_USE_SELECT"
2647If undefined or defined to be \f(CW1\fR, libev will compile in support for the
2648\&\f(CW\*(C`select\*(C'\fR(2) backend. No attempt at autodetection will be done: if no
2649other method takes over, select will be it. Otherwise the select backend
2650will not be compiled in.
2651.IP "\s-1EV_SELECT_USE_FD_SET\s0" 4
2652.IX Item "EV_SELECT_USE_FD_SET"
2653If defined to \f(CW1\fR, then the select backend will use the system \f(CW\*(C`fd_set\*(C'\fR
2654structure. This is useful if libev doesn't compile due to a missing
2655\&\f(CW\*(C`NFDBITS\*(C'\fR or \f(CW\*(C`fd_mask\*(C'\fR definition or it misguesses the bitset layout on
2656exotic systems. This usually limits the range of file descriptors to some
2657low limit such as 1024 or might have other limitations (winsocket only
2658allows 64 sockets). The \f(CW\*(C`FD_SETSIZE\*(C'\fR macro, set before compilation, might
2659influence the size of the \f(CW\*(C`fd_set\*(C'\fR used.
2660.IP "\s-1EV_SELECT_IS_WINSOCKET\s0" 4
2661.IX Item "EV_SELECT_IS_WINSOCKET"
2662When defined to \f(CW1\fR, the select backend will assume that
2663select/socket/connect etc. don't understand file descriptors but
2664wants osf handles on win32 (this is the case when the select to
2665be used is the winsock select). This means that it will call
2666\&\f(CW\*(C`_get_osfhandle\*(C'\fR on the fd to convert it to an \s-1OS\s0 handle. Otherwise,
2667it is assumed that all these functions actually work on fds, even
2668on win32. Should not be defined on non\-win32 platforms.
2669.IP "\s-1EV_USE_POLL\s0" 4
2670.IX Item "EV_USE_POLL"
2671If defined to be \f(CW1\fR, libev will compile in support for the \f(CW\*(C`poll\*(C'\fR(2)
2672backend. Otherwise it will be enabled on non\-win32 platforms. It
2673takes precedence over select.
2674.IP "\s-1EV_USE_EPOLL\s0" 4
2675.IX Item "EV_USE_EPOLL"
2676If defined to be \f(CW1\fR, libev will compile in support for the Linux
2677\&\f(CW\*(C`epoll\*(C'\fR(7) backend. Its availability will be detected at runtime,
2678otherwise another method will be used as fallback. This is the
2679preferred backend for GNU/Linux systems.
2680.IP "\s-1EV_USE_KQUEUE\s0" 4
2681.IX Item "EV_USE_KQUEUE"
2682If defined to be \f(CW1\fR, libev will compile in support for the \s-1BSD\s0 style
2683\&\f(CW\*(C`kqueue\*(C'\fR(2) backend. Its actual availability will be detected at runtime,
2684otherwise another method will be used as fallback. This is the preferred
2685backend for \s-1BSD\s0 and BSD-like systems, although on most BSDs kqueue only
2686supports some types of fds correctly (the only platform we found that
2687supports ptys for example was NetBSD), so kqueue might be compiled in, but
2688not be used unless explicitly requested. The best way to use it is to find
2689out whether kqueue supports your type of fd properly and use an embedded
2690kqueue loop.
2691.IP "\s-1EV_USE_PORT\s0" 4
2692.IX Item "EV_USE_PORT"
2693If defined to be \f(CW1\fR, libev will compile in support for the Solaris
269410 port style backend. Its availability will be detected at runtime,
2695otherwise another method will be used as fallback. This is the preferred
2696backend for Solaris 10 systems.
2697.IP "\s-1EV_USE_DEVPOLL\s0" 4
2698.IX Item "EV_USE_DEVPOLL"
2699reserved for future expansion, works like the \s-1USE\s0 symbols above.
2700.IP "\s-1EV_USE_INOTIFY\s0" 4
2701.IX Item "EV_USE_INOTIFY"
2702If defined to be \f(CW1\fR, libev will compile in support for the Linux inotify
2703interface to speed up \f(CW\*(C`ev_stat\*(C'\fR watchers. Its actual availability will
2704be detected at runtime.
2705.IP "\s-1EV_H\s0" 4
2706.IX Item "EV_H"
2707The name of the \fIev.h\fR header file used to include it. The default if
2708undefined is \f(CW"ev.h"\fR in \fIevent.h\fR and \fIev.c\fR. This can be used to
2709virtually rename the \fIev.h\fR header file in case of conflicts.
2710.IP "\s-1EV_CONFIG_H\s0" 4
2711.IX Item "EV_CONFIG_H"
2712If \f(CW\*(C`EV_STANDALONE\*(C'\fR isn't \f(CW1\fR, this variable can be used to override
2713\&\fIev.c\fR's idea of where to find the \fIconfig.h\fR file, similarly to
2714\&\f(CW\*(C`EV_H\*(C'\fR, above.
2715.IP "\s-1EV_EVENT_H\s0" 4
2716.IX Item "EV_EVENT_H"
2717Similarly to \f(CW\*(C`EV_H\*(C'\fR, this macro can be used to override \fIevent.c\fR's idea
2718of how the \fIevent.h\fR header can be found, the dfeault is \f(CW"event.h"\fR.
2719.IP "\s-1EV_PROTOTYPES\s0" 4
2720.IX Item "EV_PROTOTYPES"
2721If defined to be \f(CW0\fR, then \fIev.h\fR will not define any function
2722prototypes, but still define all the structs and other symbols. This is
2723occasionally useful if you want to provide your own wrapper functions
2724around libev functions.
2725.IP "\s-1EV_MULTIPLICITY\s0" 4
2726.IX Item "EV_MULTIPLICITY"
2727If undefined or defined to \f(CW1\fR, then all event-loop-specific functions
2728will have the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument, and you can create
2729additional independent event loops. Otherwise there will be no support
2730for multiple event loops and there is no first event loop pointer
2731argument. Instead, all functions act on the single default loop.
2732.IP "\s-1EV_MINPRI\s0" 4
2733.IX Item "EV_MINPRI"
2734.PD 0
2735.IP "\s-1EV_MAXPRI\s0" 4
2736.IX Item "EV_MAXPRI"
2737.PD
2738The range of allowed priorities. \f(CW\*(C`EV_MINPRI\*(C'\fR must be smaller or equal to
2739\&\f(CW\*(C`EV_MAXPRI\*(C'\fR, but otherwise there are no non-obvious limitations. You can
2740provide for more priorities by overriding those symbols (usually defined
2741to be \f(CW\*(C`\-2\*(C'\fR and \f(CW2\fR, respectively).
2742.Sp
2743When doing priority-based operations, libev usually has to linearly search
2744all the priorities, so having many of them (hundreds) uses a lot of space
2745and time, so using the defaults of five priorities (\-2 .. +2) is usually
2746fine.
2747.Sp
2748If your embedding app does not need any priorities, defining these both to
2749\&\f(CW0\fR will save some memory and cpu.
2750.IP "\s-1EV_PERIODIC_ENABLE\s0" 4
2751.IX Item "EV_PERIODIC_ENABLE"
2752If undefined or defined to be \f(CW1\fR, then periodic timers are supported. If
2753defined to be \f(CW0\fR, then they are not. Disabling them saves a few kB of
2754code.
2755.IP "\s-1EV_IDLE_ENABLE\s0" 4
2756.IX Item "EV_IDLE_ENABLE"
2757If undefined or defined to be \f(CW1\fR, then idle watchers are supported. If
2758defined to be \f(CW0\fR, then they are not. Disabling them saves a few kB of
2759code.
2760.IP "\s-1EV_EMBED_ENABLE\s0" 4
2761.IX Item "EV_EMBED_ENABLE"
2762If undefined or defined to be \f(CW1\fR, then embed watchers are supported. If
2763defined to be \f(CW0\fR, then they are not.
2764.IP "\s-1EV_STAT_ENABLE\s0" 4
2765.IX Item "EV_STAT_ENABLE"
2766If undefined or defined to be \f(CW1\fR, then stat watchers are supported. If
2767defined to be \f(CW0\fR, then they are not.
2768.IP "\s-1EV_FORK_ENABLE\s0" 4
2769.IX Item "EV_FORK_ENABLE"
2770If undefined or defined to be \f(CW1\fR, then fork watchers are supported. If
2771defined to be \f(CW0\fR, then they are not.
2772.IP "\s-1EV_MINIMAL\s0" 4
2773.IX Item "EV_MINIMAL"
2774If you need to shave off some kilobytes of code at the expense of some
2775speed, define this symbol to \f(CW1\fR. Currently only used for gcc to override
2776some inlining decisions, saves roughly 30% codesize of amd64.
2777.IP "\s-1EV_PID_HASHSIZE\s0" 4
2778.IX Item "EV_PID_HASHSIZE"
2779\&\f(CW\*(C`ev_child\*(C'\fR watchers use a small hash table to distribute workload by
2780pid. The default size is \f(CW16\fR (or \f(CW1\fR with \f(CW\*(C`EV_MINIMAL\*(C'\fR), usually more
2781than enough. If you need to manage thousands of children you might want to
2782increase this value (\fImust\fR be a power of two).
2783.IP "\s-1EV_INOTIFY_HASHSIZE\s0" 4
2784.IX Item "EV_INOTIFY_HASHSIZE"
2785\&\f(CW\*(C`ev_stat\*(C'\fR watchers use a small hash table to distribute workload by
2786inotify watch id. The default size is \f(CW16\fR (or \f(CW1\fR with \f(CW\*(C`EV_MINIMAL\*(C'\fR),
2787usually more than enough. If you need to manage thousands of \f(CW\*(C`ev_stat\*(C'\fR
2788watchers you might want to increase this value (\fImust\fR be a power of
2789two).
2790.IP "\s-1EV_COMMON\s0" 4
2791.IX Item "EV_COMMON"
2792By default, all watchers have a \f(CW\*(C`void *data\*(C'\fR member. By redefining
2793this macro to a something else you can include more and other types of
2794members. You have to define it each time you include one of the files,
2795though, and it must be identical each time.
2796.Sp
2797For example, the perl \s-1EV\s0 module uses something like this:
2798.Sp
2799.Vb 3
2800\& #define EV_COMMON \e
2801\& SV *self; /* contains this struct */ \e
2802\& SV *cb_sv, *fh /* note no trailing ";" */
2803.Ve
2804.IP "\s-1EV_CB_DECLARE\s0 (type)" 4
2805.IX Item "EV_CB_DECLARE (type)"
2806.PD 0
2807.IP "\s-1EV_CB_INVOKE\s0 (watcher, revents)" 4
2808.IX Item "EV_CB_INVOKE (watcher, revents)"
2809.IP "ev_set_cb (ev, cb)" 4
2810.IX Item "ev_set_cb (ev, cb)"
2811.PD
2812Can be used to change the callback member declaration in each watcher,
2813and the way callbacks are invoked and set. Must expand to a struct member
2814definition and a statement, respectively. See the \fIev.h\fR header file for
2815their default definitions. One possible use for overriding these is to
2816avoid the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument in all cases, or to use
2817method calls instead of plain function calls in \*(C+.
2818.Sh "\s-1EXPORTED\s0 \s-1API\s0 \s-1SYMBOLS\s0"
2819.IX Subsection "EXPORTED API SYMBOLS"
2820If you need to re-export the \s-1API\s0 (e.g. via a dll) and you need a list of
2821exported symbols, you can use the provided \fISymbol.*\fR files which list
2822all public symbols, one per line:
2823.Sp
2824.Vb 2
2825\& Symbols.ev for libev proper
2826\& Symbols.event for the libevent emulation
2827.Ve
2828.Sp
2829This can also be used to rename all public symbols to avoid clashes with
2830multiple versions of libev linked together (which is obviously bad in
2831itself, but sometimes it is inconvinient to avoid this).
2832.Sp
2833A sed command like this will create wrapper \f(CW\*(C`#define\*(C'\fR's that you need to
2834include before including \fIev.h\fR:
2835.Sp
2836.Vb 1
2837\& <Symbols.ev sed -e "s/.*/#define & myprefix_&/" >wrap.h
2838.Ve
2839.Sp
2840This would create a file \fIwrap.h\fR which essentially looks like this:
2841.Sp
2842.Vb 4
2843\& #define ev_backend myprefix_ev_backend
2844\& #define ev_check_start myprefix_ev_check_start
2845\& #define ev_check_stop myprefix_ev_check_stop
2846\& ...
2847.Ve
2848.Sh "\s-1EXAMPLES\s0"
2849.IX Subsection "EXAMPLES"
2850For a real-world example of a program the includes libev
2851verbatim, you can have a look at the \s-1EV\s0 perl module
2852(<http://software.schmorp.de/pkg/EV.html>). It has the libev files in
2853the \fIlibev/\fR subdirectory and includes them in the \fI\s-1EV/EVAPI\s0.h\fR (public
2854interface) and \fI\s-1EV\s0.xs\fR (implementation) files. Only the \fI\s-1EV\s0.xs\fR file
2855will be compiled. It is pretty complex because it provides its own header
2856file.
2857.Sp
2858The usage in rxvt-unicode is simpler. It has a \fIev_cpp.h\fR header file
2859that everybody includes and which overrides some configure choices:
2860.Sp
2861.Vb 9
2862\& #define EV_MINIMAL 1
2863\& #define EV_USE_POLL 0
2864\& #define EV_MULTIPLICITY 0
2865\& #define EV_PERIODIC_ENABLE 0
2866\& #define EV_STAT_ENABLE 0
2867\& #define EV_FORK_ENABLE 0
2868\& #define EV_CONFIG_H <config.h>
2869\& #define EV_MINPRI 0
2870\& #define EV_MAXPRI 0
2871.Ve
2872.Sp
2873.Vb 1
2874\& #include "ev++.h"
2875.Ve
2876.Sp
2877And a \fIev_cpp.C\fR implementation file that contains libev proper and is compiled:
2878.Sp
2879.Vb 2
2880\& #include "ev_cpp.h"
2881\& #include "ev.c"
2882.Ve
2883.SH "COMPLEXITIES"
2884.IX Header "COMPLEXITIES"
2885In this section the complexities of (many of) the algorithms used inside
2886libev will be explained. For complexity discussions about backends see the
2887documentation for \f(CW\*(C`ev_default_init\*(C'\fR.
2888.Sp
2889All of the following are about amortised time: If an array needs to be
2890extended, libev needs to realloc and move the whole array, but this
2891happens asymptotically never with higher number of elements, so O(1) might
2892mean it might do a lengthy realloc operation in rare cases, but on average
2893it is much faster and asymptotically approaches constant time.
2894.RS 4
2895.IP "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)" 4
2896.IX Item "Starting and stopping timer/periodic watchers: O(log skipped_other_timers)"
2897This means that, when you have a watcher that triggers in one hour and
2898there are 100 watchers that would trigger before that then inserting will
2899have to skip roughly seven (\f(CW\*(C`ld 100\*(C'\fR) of these watchers.
2900.IP "Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)" 4
2901.IX Item "Changing timer/periodic watchers (by autorepeat or calling again): O(log skipped_other_timers)"
2902That means that changing a timer costs less than removing/adding them
2903as only the relative motion in the event queue has to be paid for.
2904.IP "Starting io/check/prepare/idle/signal/child watchers: O(1)" 4
2905.IX Item "Starting io/check/prepare/idle/signal/child watchers: O(1)"
2906These just add the watcher into an array or at the head of a list.
2907.IP "Stopping check/prepare/idle watchers: O(1)" 4
2908.IX Item "Stopping check/prepare/idle watchers: O(1)"
2909.PD 0
2910.IP "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % \s-1EV_PID_HASHSIZE\s0))" 4
2911.IX Item "Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE))"
2912.PD
2913These watchers are stored in lists then need to be walked to find the
2914correct watcher to remove. The lists are usually short (you don't usually
2915have many watchers waiting for the same fd or signal).
2916.IP "Finding the next timer in each loop iteration: O(1)" 4
2917.IX Item "Finding the next timer in each loop iteration: O(1)"
2918By virtue of using a binary heap, the next timer is always found at the
2919beginning of the storage array.
2920.IP "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)" 4
2921.IX Item "Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd)"
2922A change means an I/O watcher gets started or stopped, which requires
2923libev to recalculate its status (and possibly tell the kernel, depending
2924on backend and wether \f(CW\*(C`ev_io_set\*(C'\fR was used).
2925.IP "Activating one watcher (putting it into the pending state): O(1)" 4
2926.IX Item "Activating one watcher (putting it into the pending state): O(1)"
2927.PD 0
2928.IP "Priority handling: O(number_of_priorities)" 4
2929.IX Item "Priority handling: O(number_of_priorities)"
2930.PD
2931Priorities are implemented by allocating some space for each
2932priority. When doing priority-based operations, libev usually has to
2933linearly search all the priorities, but starting/stopping and activating
2934watchers becomes O(1) w.r.t. prioritiy handling.
2935.RE
2936.RS 4
1020.SH "AUTHOR" 2937.SH "AUTHOR"
1021.IX Header "AUTHOR" 2938.IX Header "AUTHOR"
1022Marc Lehmann <libev@schmorp.de>. 2939Marc Lehmann <libev@schmorp.de>.

Diff Legend

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