| … | |
… | |
| 363 | might affect timers and time-outs. |
363 | might affect timers and time-outs. |
| 364 | |
364 | |
| 365 | When this is the case, you can call this method, which will update the |
365 | When this is the case, you can call this method, which will update the |
| 366 | event loop's idea of "current time". |
366 | event loop's idea of "current time". |
| 367 | |
367 | |
|
|
368 | A typical example would be a script in a web server (e.g. C<mod_perl>) - |
|
|
369 | when mod_perl executes the script, then the event loop will have the wrong |
|
|
370 | idea about the "current time" (being potentially far in the past, when the |
|
|
371 | script ran the last time). In that case you should arrange a call to C<< |
|
|
372 | AnyEvent->now_update >> each time the web server process wakes up again |
|
|
373 | (e.g. at the start of your script, or in a handler). |
|
|
374 | |
| 368 | Note that updating the time I<might> cause some events to be handled. |
375 | Note that updating the time I<might> cause some events to be handled. |
| 369 | |
376 | |
| 370 | =back |
377 | =back |
| 371 | |
378 | |
| 372 | =head2 SIGNAL WATCHERS |
379 | =head2 SIGNAL WATCHERS |
| … | |
… | |
| 395 | correctly. |
402 | correctly. |
| 396 | |
403 | |
| 397 | Example: exit on SIGINT |
404 | Example: exit on SIGINT |
| 398 | |
405 | |
| 399 | my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 }); |
406 | my $w = AnyEvent->signal (signal => "INT", cb => sub { exit 1 }); |
|
|
407 | |
|
|
408 | =head3 Restart Behaviour |
|
|
409 | |
|
|
410 | While restart behaviour is up to the event loop implementation, most will |
|
|
411 | not restart syscalls (that includes L<Async::Interrupt> and AnyEvent's |
|
|
412 | pure perl implementation). |
|
|
413 | |
|
|
414 | =head3 Safe/Unsafe Signals |
|
|
415 | |
|
|
416 | Perl signals can be either "safe" (synchronous to opcode handling) or |
|
|
417 | "unsafe" (asynchronous) - the former might get delayed indefinitely, the |
|
|
418 | latter might corrupt your memory. |
|
|
419 | |
|
|
420 | AnyEvent signal handlers are, in addition, synchronous to the event loop, |
|
|
421 | i.e. they will not interrupt your running perl program but will only be |
|
|
422 | called as part of the normal event handling (just like timer, I/O etc. |
|
|
423 | callbacks, too). |
| 400 | |
424 | |
| 401 | =head3 Signal Races, Delays and Workarounds |
425 | =head3 Signal Races, Delays and Workarounds |
| 402 | |
426 | |
| 403 | Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support attaching |
427 | Many event loops (e.g. Glib, Tk, Qt, IO::Async) do not support attaching |
| 404 | callbacks to signals in a generic way, which is a pity, as you cannot |
428 | callbacks to signals in a generic way, which is a pity, as you cannot |
| … | |
… | |
| 479 | |
503 | |
| 480 | =head2 IDLE WATCHERS |
504 | =head2 IDLE WATCHERS |
| 481 | |
505 | |
| 482 | $w = AnyEvent->idle (cb => <callback>); |
506 | $w = AnyEvent->idle (cb => <callback>); |
| 483 | |
507 | |
| 484 | Sometimes there is a need to do something, but it is not so important |
508 | Repeatedly invoke the callback after the process becomes idle, until |
| 485 | to do it instantly, but only when there is nothing better to do. This |
509 | either the watcher is destroyed or new events have been detected. |
| 486 | "nothing better to do" is usually defined to be "no other events need |
|
|
| 487 | attention by the event loop". |
|
|
| 488 | |
510 | |
| 489 | Idle watchers ideally get invoked when the event loop has nothing |
511 | Idle watchers are useful when there is a need to do something, but it |
| 490 | better to do, just before it would block the process to wait for new |
512 | is not so important (or wise) to do it instantly. The callback will be |
| 491 | events. Instead of blocking, the idle watcher is invoked. |
513 | invoked only when there is "nothing better to do", which is usually |
|
|
514 | defined as "all outstanding events have been handled and no new events |
|
|
515 | have been detected". That means that idle watchers ideally get invoked |
|
|
516 | when the event loop has just polled for new events but none have been |
|
|
517 | detected. Instead of blocking to wait for more events, the idle watchers |
|
|
518 | will be invoked. |
| 492 | |
519 | |
| 493 | Most event loops unfortunately do not really support idle watchers (only |
520 | Unfortunately, most event loops do not really support idle watchers (only |
| 494 | EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent |
521 | EV, Event and Glib do it in a usable fashion) - for the rest, AnyEvent |
| 495 | will simply call the callback "from time to time". |
522 | will simply call the callback "from time to time". |
| 496 | |
523 | |
| 497 | Example: read lines from STDIN, but only process them when the |
524 | Example: read lines from STDIN, but only process them when the |
| 498 | program is otherwise idle: |
525 | program is otherwise idle: |
| … | |
… | |
| 592 | after => 1, |
619 | after => 1, |
| 593 | cb => sub { $result_ready->send }, |
620 | cb => sub { $result_ready->send }, |
| 594 | ); |
621 | ); |
| 595 | |
622 | |
| 596 | # this "blocks" (while handling events) till the callback |
623 | # this "blocks" (while handling events) till the callback |
| 597 | # calls -<send |
624 | # calls ->send |
| 598 | $result_ready->recv; |
625 | $result_ready->recv; |
| 599 | |
626 | |
| 600 | Example: wait for a timer, but take advantage of the fact that condition |
627 | Example: wait for a timer, but take advantage of the fact that condition |
| 601 | variables are also callable directly. |
628 | variables are also callable directly. |
| 602 | |
629 | |
| … | |
… | |
| 666 | one. For example, a function that pings many hosts in parallel might want |
693 | one. For example, a function that pings many hosts in parallel might want |
| 667 | to use a condition variable for the whole process. |
694 | to use a condition variable for the whole process. |
| 668 | |
695 | |
| 669 | Every call to C<< ->begin >> will increment a counter, and every call to |
696 | Every call to C<< ->begin >> will increment a counter, and every call to |
| 670 | C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end |
697 | C<< ->end >> will decrement it. If the counter reaches C<0> in C<< ->end |
| 671 | >>, the (last) callback passed to C<begin> will be executed. That callback |
698 | >>, the (last) callback passed to C<begin> will be executed, passing the |
| 672 | is I<supposed> to call C<< ->send >>, but that is not required. If no |
699 | condvar as first argument. That callback is I<supposed> to call C<< ->send |
| 673 | callback was set, C<send> will be called without any arguments. |
700 | >>, but that is not required. If no group callback was set, C<send> will |
|
|
701 | be called without any arguments. |
| 674 | |
702 | |
| 675 | You can think of C<< $cv->send >> giving you an OR condition (one call |
703 | You can think of C<< $cv->send >> giving you an OR condition (one call |
| 676 | sends), while C<< $cv->begin >> and C<< $cv->end >> giving you an AND |
704 | sends), while C<< $cv->begin >> and C<< $cv->end >> giving you an AND |
| 677 | condition (all C<begin> calls must be C<end>'ed before the condvar sends). |
705 | condition (all C<begin> calls must be C<end>'ed before the condvar sends). |
| 678 | |
706 | |
| … | |
… | |
| 705 | begung can potentially be zero: |
733 | begung can potentially be zero: |
| 706 | |
734 | |
| 707 | my $cv = AnyEvent->condvar; |
735 | my $cv = AnyEvent->condvar; |
| 708 | |
736 | |
| 709 | my %result; |
737 | my %result; |
| 710 | $cv->begin (sub { $cv->send (\%result) }); |
738 | $cv->begin (sub { shift->send (\%result) }); |
| 711 | |
739 | |
| 712 | for my $host (@list_of_hosts) { |
740 | for my $host (@list_of_hosts) { |
| 713 | $cv->begin; |
741 | $cv->begin; |
| 714 | ping_host_then_call_callback $host, sub { |
742 | ping_host_then_call_callback $host, sub { |
| 715 | $result{$host} = ...; |
743 | $result{$host} = ...; |
| … | |
… | |
| 806 | =over 4 |
834 | =over 4 |
| 807 | |
835 | |
| 808 | =item Backends that are autoprobed when no other event loop can be found. |
836 | =item Backends that are autoprobed when no other event loop can be found. |
| 809 | |
837 | |
| 810 | EV is the preferred backend when no other event loop seems to be in |
838 | EV is the preferred backend when no other event loop seems to be in |
| 811 | use. If EV is not installed, then AnyEvent will try Event, and, failing |
839 | use. If EV is not installed, then AnyEvent will fall back to its own |
| 812 | that, will fall back to its own pure-perl implementation, which is |
840 | pure-perl implementation, which is available everywhere as it comes with |
| 813 | available everywhere as it comes with AnyEvent itself. |
841 | AnyEvent itself. |
| 814 | |
842 | |
| 815 | AnyEvent::Impl::EV based on EV (interface to libev, best choice). |
843 | AnyEvent::Impl::EV based on EV (interface to libev, best choice). |
| 816 | AnyEvent::Impl::Event based on Event, very stable, few glitches. |
|
|
| 817 | AnyEvent::Impl::Perl pure-perl implementation, fast and portable. |
844 | AnyEvent::Impl::Perl pure-perl implementation, fast and portable. |
| 818 | |
845 | |
| 819 | =item Backends that are transparently being picked up when they are used. |
846 | =item Backends that are transparently being picked up when they are used. |
| 820 | |
847 | |
| 821 | These will be used when they are currently loaded when the first watcher |
848 | These will be used when they are currently loaded when the first watcher |
| 822 | is created, in which case it is assumed that the application is using |
849 | is created, in which case it is assumed that the application is using |
| 823 | them. This means that AnyEvent will automatically pick the right backend |
850 | them. This means that AnyEvent will automatically pick the right backend |
| 824 | when the main program loads an event module before anything starts to |
851 | when the main program loads an event module before anything starts to |
| 825 | create watchers. Nothing special needs to be done by the main program. |
852 | create watchers. Nothing special needs to be done by the main program. |
| 826 | |
853 | |
|
|
854 | AnyEvent::Impl::Event based on Event, very stable, few glitches. |
| 827 | AnyEvent::Impl::Glib based on Glib, slow but very stable. |
855 | AnyEvent::Impl::Glib based on Glib, slow but very stable. |
| 828 | AnyEvent::Impl::Tk based on Tk, very broken. |
856 | AnyEvent::Impl::Tk based on Tk, very broken. |
| 829 | AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse. |
857 | AnyEvent::Impl::EventLib based on Event::Lib, leaks memory and worse. |
| 830 | AnyEvent::Impl::POE based on POE, very slow, some limitations. |
858 | AnyEvent::Impl::POE based on POE, very slow, some limitations. |
| 831 | AnyEvent::Impl::Irssi used when running within irssi. |
859 | AnyEvent::Impl::Irssi used when running within irssi. |
| … | |
… | |
| 941 | You should check C<$AnyEvent::MODEL> before adding to this array, though: |
969 | You should check C<$AnyEvent::MODEL> before adding to this array, though: |
| 942 | if it is defined then the event loop has already been detected, and the |
970 | if it is defined then the event loop has already been detected, and the |
| 943 | array will be ignored. |
971 | array will be ignored. |
| 944 | |
972 | |
| 945 | Best use C<AnyEvent::post_detect { BLOCK }> when your application allows |
973 | Best use C<AnyEvent::post_detect { BLOCK }> when your application allows |
| 946 | it,as it takes care of these details. |
974 | it, as it takes care of these details. |
| 947 | |
975 | |
| 948 | This variable is mainly useful for modules that can do something useful |
976 | This variable is mainly useful for modules that can do something useful |
| 949 | when AnyEvent is used and thus want to know when it is initialised, but do |
977 | when AnyEvent is used and thus want to know when it is initialised, but do |
| 950 | not need to even load it by default. This array provides the means to hook |
978 | not need to even load it by default. This array provides the means to hook |
| 951 | into AnyEvent passively, without loading it. |
979 | into AnyEvent passively, without loading it. |
|
|
980 | |
|
|
981 | Example: To load Coro::AnyEvent whenever Coro and AnyEvent are used |
|
|
982 | together, you could put this into Coro (this is the actual code used by |
|
|
983 | Coro to accomplish this): |
|
|
984 | |
|
|
985 | if (defined $AnyEvent::MODEL) { |
|
|
986 | # AnyEvent already initialised, so load Coro::AnyEvent |
|
|
987 | require Coro::AnyEvent; |
|
|
988 | } else { |
|
|
989 | # AnyEvent not yet initialised, so make sure to load Coro::AnyEvent |
|
|
990 | # as soon as it is |
|
|
991 | push @AnyEvent::post_detect, sub { require Coro::AnyEvent }; |
|
|
992 | } |
| 952 | |
993 | |
| 953 | =back |
994 | =back |
| 954 | |
995 | |
| 955 | =head1 WHAT TO DO IN A MODULE |
996 | =head1 WHAT TO DO IN A MODULE |
| 956 | |
997 | |
| … | |
… | |
| 1105 | |
1146 | |
| 1106 | package AnyEvent; |
1147 | package AnyEvent; |
| 1107 | |
1148 | |
| 1108 | # basically a tuned-down version of common::sense |
1149 | # basically a tuned-down version of common::sense |
| 1109 | sub common_sense { |
1150 | sub common_sense { |
| 1110 | # no warnings |
1151 | # from common:.sense 1.0 |
| 1111 | ${^WARNING_BITS} ^= ${^WARNING_BITS}; |
1152 | ${^WARNING_BITS} = "\xfc\x3f\x33\x00\x0f\xf3\xcf\xc0\xf3\xfc\x33\x00"; |
| 1112 | # use strict vars subs |
1153 | # use strict vars subs - NO UTF-8, as Util.pm doesn't like this atm. (uts46data.pl) |
| 1113 | $^H |= 0x00000600; |
1154 | $^H |= 0x00000600; |
| 1114 | } |
1155 | } |
| 1115 | |
1156 | |
| 1116 | BEGIN { AnyEvent::common_sense } |
1157 | BEGIN { AnyEvent::common_sense } |
| 1117 | |
1158 | |
| 1118 | use Carp (); |
1159 | use Carp (); |
| 1119 | |
1160 | |
| 1120 | our $VERSION = 4.91; |
1161 | our $VERSION = '5.24'; |
| 1121 | our $MODEL; |
1162 | our $MODEL; |
| 1122 | |
1163 | |
| 1123 | our $AUTOLOAD; |
1164 | our $AUTOLOAD; |
| 1124 | our @ISA; |
1165 | our @ISA; |
| 1125 | |
1166 | |
| 1126 | our @REGISTRY; |
1167 | our @REGISTRY; |
| 1127 | |
1168 | |
| 1128 | our $WIN32; |
|
|
| 1129 | |
|
|
| 1130 | our $VERBOSE; |
1169 | our $VERBOSE; |
| 1131 | |
1170 | |
| 1132 | BEGIN { |
1171 | BEGIN { |
|
|
1172 | eval "sub CYGWIN(){" . (($^O =~ /cygwin/i) *1) . "}"; |
| 1133 | eval "sub WIN32(){ " . (($^O =~ /mswin32/i)*1) ." }"; |
1173 | eval "sub WIN32 (){" . (($^O =~ /mswin32/i)*1) . "}"; |
| 1134 | eval "sub TAINT(){ " . (${^TAINT}*1) . " }"; |
1174 | eval "sub TAINT (){" . (${^TAINT} *1) . "}"; |
| 1135 | |
1175 | |
| 1136 | delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV} |
1176 | delete @ENV{grep /^PERL_ANYEVENT_/, keys %ENV} |
| 1137 | if ${^TAINT}; |
1177 | if ${^TAINT}; |
| 1138 | |
1178 | |
| 1139 | $VERBOSE = $ENV{PERL_ANYEVENT_VERBOSE}*1; |
1179 | $VERBOSE = $ENV{PERL_ANYEVENT_VERBOSE}*1; |
| … | |
… | |
| 1151 | $ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6"; |
1191 | $ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6"; |
| 1152 | } |
1192 | } |
| 1153 | |
1193 | |
| 1154 | my @models = ( |
1194 | my @models = ( |
| 1155 | [EV:: => AnyEvent::Impl::EV:: , 1], |
1195 | [EV:: => AnyEvent::Impl::EV:: , 1], |
| 1156 | [Event:: => AnyEvent::Impl::Event::, 1], |
|
|
| 1157 | [AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl:: , 1], |
1196 | [AnyEvent::Impl::Perl:: => AnyEvent::Impl::Perl:: , 1], |
| 1158 | # everything below here will not (normally) be autoprobed |
1197 | # everything below here will not (normally) be autoprobed |
| 1159 | # as the pureperl backend should work everywhere |
1198 | # as the pureperl backend should work everywhere |
| 1160 | # and is usually faster |
1199 | # and is usually faster |
|
|
1200 | [Event:: => AnyEvent::Impl::Event::, 1], |
| 1161 | [Glib:: => AnyEvent::Impl::Glib:: , 1], # becomes extremely slow with many watchers |
1201 | [Glib:: => AnyEvent::Impl::Glib:: , 1], # becomes extremely slow with many watchers |
| 1162 | [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy |
1202 | [Event::Lib:: => AnyEvent::Impl::EventLib::], # too buggy |
| 1163 | [Irssi:: => AnyEvent::Impl::Irssi::], # Irssi has a bogus "Event" package |
1203 | [Irssi:: => AnyEvent::Impl::Irssi::], # Irssi has a bogus "Event" package |
| 1164 | [Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles |
1204 | [Tk:: => AnyEvent::Impl::Tk::], # crashes with many handles |
| 1165 | [Qt:: => AnyEvent::Impl::Qt::], # requires special main program |
1205 | [Qt:: => AnyEvent::Impl::Qt::], # requires special main program |
| … | |
… | |
| 1168 | [Prima:: => AnyEvent::Impl::POE::], |
1208 | [Prima:: => AnyEvent::Impl::POE::], |
| 1169 | # IO::Async is just too broken - we would need workarounds for its |
1209 | # IO::Async is just too broken - we would need workarounds for its |
| 1170 | # byzantine signal and broken child handling, among others. |
1210 | # byzantine signal and broken child handling, among others. |
| 1171 | # IO::Async is rather hard to detect, as it doesn't have any |
1211 | # IO::Async is rather hard to detect, as it doesn't have any |
| 1172 | # obvious default class. |
1212 | # obvious default class. |
| 1173 | # [0, IO::Async:: => AnyEvent::Impl::IOAsync::], # requires special main program |
1213 | [IO::Async:: => AnyEvent::Impl::IOAsync::], # requires special main program |
| 1174 | # [0, IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # requires special main program |
1214 | [IO::Async::Loop:: => AnyEvent::Impl::IOAsync::], # requires special main program |
| 1175 | # [0, IO::Async::Notifier:: => AnyEvent::Impl::IOAsync::], # requires special main program |
1215 | [IO::Async::Notifier:: => AnyEvent::Impl::IOAsync::], # requires special main program |
|
|
1216 | [AnyEvent::Impl::IOAsync:: => AnyEvent::Impl::IOAsync::], # requires special main program |
| 1176 | ); |
1217 | ); |
| 1177 | |
1218 | |
| 1178 | our %method = map +($_ => 1), |
1219 | our %method = map +($_ => 1), |
| 1179 | qw(io timer time now now_update signal child idle condvar one_event DESTROY); |
1220 | qw(io timer time now now_update signal child idle condvar one_event DESTROY); |
| 1180 | |
1221 | |
| … | |
… | |
| 1287 | # we assume CLOEXEC is already set by perl in all important cases |
1328 | # we assume CLOEXEC is already set by perl in all important cases |
| 1288 | |
1329 | |
| 1289 | ($fh2, $rw) |
1330 | ($fh2, $rw) |
| 1290 | } |
1331 | } |
| 1291 | |
1332 | |
| 1292 | ############################################################################# |
1333 | =head1 SIMPLIFIED AE API |
| 1293 | # "new" API, currently only emulation of it |
1334 | |
| 1294 | ############################################################################# |
1335 | Starting with version 5.0, AnyEvent officially supports a second, much |
|
|
1336 | simpler, API that is designed to reduce the calling, typing and memory |
|
|
1337 | overhead. |
|
|
1338 | |
|
|
1339 | See the L<AE> manpage for details. |
|
|
1340 | |
|
|
1341 | =cut |
| 1295 | |
1342 | |
| 1296 | package AE; |
1343 | package AE; |
|
|
1344 | |
|
|
1345 | our $VERSION = $AnyEvent::VERSION; |
| 1297 | |
1346 | |
| 1298 | sub io($$$) { |
1347 | sub io($$$) { |
| 1299 | AnyEvent->io (fh => $_[0], poll => $_[1] ? "w" : "r", cb => $_[2]) |
1348 | AnyEvent->io (fh => $_[0], poll => $_[1] ? "w" : "r", cb => $_[2]) |
| 1300 | } |
1349 | } |
| 1301 | |
1350 | |
| 1302 | sub timer($$$) { |
1351 | sub timer($$$) { |
| 1303 | AnyEvent->timer (after => $_[0], interval => $_[1], cb => $_[2]); |
1352 | AnyEvent->timer (after => $_[0], interval => $_[1], cb => $_[2]) |
| 1304 | } |
1353 | } |
| 1305 | |
1354 | |
| 1306 | sub signal($$) { |
1355 | sub signal($$) { |
| 1307 | AnyEvent->signal (signal => $_[0], cb => $_[1]); |
1356 | AnyEvent->signal (signal => $_[0], cb => $_[1]) |
| 1308 | } |
1357 | } |
| 1309 | |
1358 | |
| 1310 | sub child($$) { |
1359 | sub child($$) { |
| 1311 | AnyEvent->child (pid => $_[0], cb => $_[1]); |
1360 | AnyEvent->child (pid => $_[0], cb => $_[1]) |
| 1312 | } |
1361 | } |
| 1313 | |
1362 | |
| 1314 | sub idle($) { |
1363 | sub idle($) { |
| 1315 | AnyEvent->idle (cb => $_[0]); |
1364 | AnyEvent->idle (cb => $_[0]) |
| 1316 | } |
1365 | } |
| 1317 | |
1366 | |
| 1318 | sub cv(;&) { |
1367 | sub cv(;&) { |
| 1319 | AnyEvent->condvar (@_ ? (cb => $_[0]) : ()) |
1368 | AnyEvent->condvar (@_ ? (cb => $_[0]) : ()) |
| 1320 | } |
1369 | } |
| … | |
… | |
| 1333 | |
1382 | |
| 1334 | package AnyEvent::Base; |
1383 | package AnyEvent::Base; |
| 1335 | |
1384 | |
| 1336 | # default implementations for many methods |
1385 | # default implementations for many methods |
| 1337 | |
1386 | |
| 1338 | sub _time { |
1387 | sub _time() { |
| 1339 | # probe for availability of Time::HiRes |
1388 | # probe for availability of Time::HiRes |
| 1340 | if (eval "use Time::HiRes (); Time::HiRes::time (); 1") { |
1389 | if (eval "use Time::HiRes (); Time::HiRes::time (); 1") { |
| 1341 | warn "AnyEvent: using Time::HiRes for sub-second timing accuracy.\n" if $VERBOSE >= 8; |
1390 | warn "AnyEvent: using Time::HiRes for sub-second timing accuracy.\n" if $VERBOSE >= 8; |
| 1342 | *_time = \&Time::HiRes::time; |
1391 | *_time = \&Time::HiRes::time; |
| 1343 | # if (eval "use POSIX (); (POSIX::times())... |
1392 | # if (eval "use POSIX (); (POSIX::times())... |
| … | |
… | |
| 1363 | |
1412 | |
| 1364 | our $HAVE_ASYNC_INTERRUPT; |
1413 | our $HAVE_ASYNC_INTERRUPT; |
| 1365 | |
1414 | |
| 1366 | sub _have_async_interrupt() { |
1415 | sub _have_async_interrupt() { |
| 1367 | $HAVE_ASYNC_INTERRUPT = 1*(!$ENV{PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT} |
1416 | $HAVE_ASYNC_INTERRUPT = 1*(!$ENV{PERL_ANYEVENT_AVOID_ASYNC_INTERRUPT} |
| 1368 | && eval "use Async::Interrupt 1.0 (); 1") |
1417 | && eval "use Async::Interrupt 1.02 (); 1") |
| 1369 | unless defined $HAVE_ASYNC_INTERRUPT; |
1418 | unless defined $HAVE_ASYNC_INTERRUPT; |
| 1370 | |
1419 | |
| 1371 | $HAVE_ASYNC_INTERRUPT |
1420 | $HAVE_ASYNC_INTERRUPT |
| 1372 | } |
1421 | } |
| 1373 | |
1422 | |
| … | |
… | |
| 1376 | our ($SIG_COUNT, $SIG_TW); |
1425 | our ($SIG_COUNT, $SIG_TW); |
| 1377 | |
1426 | |
| 1378 | sub _signal_exec { |
1427 | sub _signal_exec { |
| 1379 | $HAVE_ASYNC_INTERRUPT |
1428 | $HAVE_ASYNC_INTERRUPT |
| 1380 | ? $SIGPIPE_R->drain |
1429 | ? $SIGPIPE_R->drain |
| 1381 | : sysread $SIGPIPE_R, my $dummy, 9; |
1430 | : sysread $SIGPIPE_R, (my $dummy), 9; |
| 1382 | |
1431 | |
| 1383 | while (%SIG_EV) { |
1432 | while (%SIG_EV) { |
| 1384 | for (keys %SIG_EV) { |
1433 | for (keys %SIG_EV) { |
| 1385 | delete $SIG_EV{$_}; |
1434 | delete $SIG_EV{$_}; |
| 1386 | $_->() for values %{ $SIG_CB{$_} || {} }; |
1435 | $_->() for values %{ $SIG_CB{$_} || {} }; |
| … | |
… | |
| 1902 | warn "read: $input\n"; # output what has been read |
1951 | warn "read: $input\n"; # output what has been read |
| 1903 | $cv->send if $input =~ /^q/i; # quit program if /^q/i |
1952 | $cv->send if $input =~ /^q/i; # quit program if /^q/i |
| 1904 | }, |
1953 | }, |
| 1905 | ); |
1954 | ); |
| 1906 | |
1955 | |
| 1907 | my $time_watcher; # can only be used once |
|
|
| 1908 | |
|
|
| 1909 | sub new_timer { |
|
|
| 1910 | $timer = AnyEvent->timer (after => 1, cb => sub { |
1956 | my $time_watcher = AnyEvent->timer (after => 1, interval => 1, cb => sub { |
| 1911 | warn "timeout\n"; # print 'timeout' about every second |
1957 | warn "timeout\n"; # print 'timeout' at most every second |
| 1912 | &new_timer; # and restart the time |
|
|
| 1913 | }); |
1958 | }); |
| 1914 | } |
|
|
| 1915 | |
|
|
| 1916 | new_timer; # create first timer |
|
|
| 1917 | |
1959 | |
| 1918 | $cv->recv; # wait until user enters /^q/i |
1960 | $cv->recv; # wait until user enters /^q/i |
| 1919 | |
1961 | |
| 1920 | =head1 REAL-WORLD EXAMPLE |
1962 | =head1 REAL-WORLD EXAMPLE |
| 1921 | |
1963 | |
| … | |
… | |
| 2052 | through AnyEvent. The benchmark creates a lot of timers (with a zero |
2094 | through AnyEvent. The benchmark creates a lot of timers (with a zero |
| 2053 | timeout) and I/O watchers (watching STDOUT, a pty, to become writable, |
2095 | timeout) and I/O watchers (watching STDOUT, a pty, to become writable, |
| 2054 | which it is), lets them fire exactly once and destroys them again. |
2096 | which it is), lets them fire exactly once and destroys them again. |
| 2055 | |
2097 | |
| 2056 | Source code for this benchmark is found as F<eg/bench> in the AnyEvent |
2098 | Source code for this benchmark is found as F<eg/bench> in the AnyEvent |
| 2057 | distribution. |
2099 | distribution. It uses the L<AE> interface, which makes a real difference |
|
|
2100 | for the EV and Perl backends only. |
| 2058 | |
2101 | |
| 2059 | =head3 Explanation of the columns |
2102 | =head3 Explanation of the columns |
| 2060 | |
2103 | |
| 2061 | I<watcher> is the number of event watchers created/destroyed. Since |
2104 | I<watcher> is the number of event watchers created/destroyed. Since |
| 2062 | different event models feature vastly different performances, each event |
2105 | different event models feature vastly different performances, each event |
| … | |
… | |
| 2083 | watcher. |
2126 | watcher. |
| 2084 | |
2127 | |
| 2085 | =head3 Results |
2128 | =head3 Results |
| 2086 | |
2129 | |
| 2087 | name watchers bytes create invoke destroy comment |
2130 | name watchers bytes create invoke destroy comment |
| 2088 | EV/EV 400000 224 0.47 0.35 0.27 EV native interface |
2131 | EV/EV 100000 223 0.47 0.43 0.27 EV native interface |
| 2089 | EV/Any 100000 224 2.88 0.34 0.27 EV + AnyEvent watchers |
2132 | EV/Any 100000 223 0.48 0.42 0.26 EV + AnyEvent watchers |
| 2090 | CoroEV/Any 100000 224 2.85 0.35 0.28 coroutines + Coro::Signal |
2133 | Coro::EV/Any 100000 223 0.47 0.42 0.26 coroutines + Coro::Signal |
| 2091 | Perl/Any 100000 452 4.13 0.73 0.95 pure perl implementation |
2134 | Perl/Any 100000 431 2.70 0.74 0.92 pure perl implementation |
| 2092 | Event/Event 16000 517 32.20 31.80 0.81 Event native interface |
2135 | Event/Event 16000 516 31.16 31.84 0.82 Event native interface |
| 2093 | Event/Any 16000 590 35.85 31.55 1.06 Event + AnyEvent watchers |
2136 | Event/Any 16000 1203 42.61 34.79 1.80 Event + AnyEvent watchers |
| 2094 | IOAsync/Any 16000 989 38.10 32.77 11.13 via IO::Async::Loop::IO_Poll |
2137 | IOAsync/Any 16000 1911 41.92 27.45 16.81 via IO::Async::Loop::IO_Poll |
| 2095 | IOAsync/Any 16000 990 37.59 29.50 10.61 via IO::Async::Loop::Epoll |
2138 | IOAsync/Any 16000 1726 40.69 26.37 15.25 via IO::Async::Loop::Epoll |
| 2096 | Glib/Any 16000 1357 102.33 12.31 51.00 quadratic behaviour |
2139 | Glib/Any 16000 1118 89.00 12.57 51.17 quadratic behaviour |
| 2097 | Tk/Any 2000 1860 27.20 66.31 14.00 SEGV with >> 2000 watchers |
2140 | Tk/Any 2000 1346 20.96 10.75 8.00 SEGV with >> 2000 watchers |
| 2098 | POE/Event 2000 6328 109.99 751.67 14.02 via POE::Loop::Event |
2141 | POE/Any 2000 6951 108.97 795.32 14.24 via POE::Loop::Event |
| 2099 | POE/Select 2000 6027 94.54 809.13 579.80 via POE::Loop::Select |
2142 | POE/Any 2000 6648 94.79 774.40 575.51 via POE::Loop::Select |
| 2100 | |
2143 | |
| 2101 | =head3 Discussion |
2144 | =head3 Discussion |
| 2102 | |
2145 | |
| 2103 | The benchmark does I<not> measure scalability of the event loop very |
2146 | The benchmark does I<not> measure scalability of the event loop very |
| 2104 | well. For example, a select-based event loop (such as the pure perl one) |
2147 | well. For example, a select-based event loop (such as the pure perl one) |
| … | |
… | |
| 2116 | benchmark machine, handling an event takes roughly 1600 CPU cycles with |
2159 | benchmark machine, handling an event takes roughly 1600 CPU cycles with |
| 2117 | EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU |
2160 | EV, 3100 CPU cycles with AnyEvent's pure perl loop and almost 3000000 CPU |
| 2118 | cycles with POE. |
2161 | cycles with POE. |
| 2119 | |
2162 | |
| 2120 | C<EV> is the sole leader regarding speed and memory use, which are both |
2163 | C<EV> is the sole leader regarding speed and memory use, which are both |
| 2121 | maximal/minimal, respectively. Even when going through AnyEvent, it uses |
2164 | maximal/minimal, respectively. When using the L<AE> API there is zero |
|
|
2165 | overhead (when going through the AnyEvent API create is about 5-6 times |
|
|
2166 | slower, with other times being equal, so still uses far less memory than |
| 2122 | far less memory than any other event loop and is still faster than Event |
2167 | any other event loop and is still faster than Event natively). |
| 2123 | natively. |
|
|
| 2124 | |
2168 | |
| 2125 | The pure perl implementation is hit in a few sweet spots (both the |
2169 | The pure perl implementation is hit in a few sweet spots (both the |
| 2126 | constant timeout and the use of a single fd hit optimisations in the perl |
2170 | constant timeout and the use of a single fd hit optimisations in the perl |
| 2127 | interpreter and the backend itself). Nevertheless this shows that it |
2171 | interpreter and the backend itself). Nevertheless this shows that it |
| 2128 | adds very little overhead in itself. Like any select-based backend its |
2172 | adds very little overhead in itself. Like any select-based backend its |
| … | |
… | |
| 2202 | In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100 |
2246 | In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100 |
| 2203 | (1%) are active. This mirrors the activity of large servers with many |
2247 | (1%) are active. This mirrors the activity of large servers with many |
| 2204 | connections, most of which are idle at any one point in time. |
2248 | connections, most of which are idle at any one point in time. |
| 2205 | |
2249 | |
| 2206 | Source code for this benchmark is found as F<eg/bench2> in the AnyEvent |
2250 | Source code for this benchmark is found as F<eg/bench2> in the AnyEvent |
| 2207 | distribution. |
2251 | distribution. It uses the L<AE> interface, which makes a real difference |
|
|
2252 | for the EV and Perl backends only. |
| 2208 | |
2253 | |
| 2209 | =head3 Explanation of the columns |
2254 | =head3 Explanation of the columns |
| 2210 | |
2255 | |
| 2211 | I<sockets> is the number of sockets, and twice the number of "servers" (as |
2256 | I<sockets> is the number of sockets, and twice the number of "servers" (as |
| 2212 | each server has a read and write socket end). |
2257 | each server has a read and write socket end). |
| … | |
… | |
| 2220 | a new one that moves the timeout into the future. |
2265 | a new one that moves the timeout into the future. |
| 2221 | |
2266 | |
| 2222 | =head3 Results |
2267 | =head3 Results |
| 2223 | |
2268 | |
| 2224 | name sockets create request |
2269 | name sockets create request |
| 2225 | EV 20000 69.01 11.16 |
2270 | EV 20000 62.66 7.99 |
| 2226 | Perl 20000 73.32 35.87 |
2271 | Perl 20000 68.32 32.64 |
| 2227 | IOAsync 20000 157.00 98.14 epoll |
2272 | IOAsync 20000 174.06 101.15 epoll |
| 2228 | IOAsync 20000 159.31 616.06 poll |
2273 | IOAsync 20000 174.67 610.84 poll |
| 2229 | Event 20000 212.62 257.32 |
2274 | Event 20000 202.69 242.91 |
| 2230 | Glib 20000 651.16 1896.30 |
2275 | Glib 20000 557.01 1689.52 |
| 2231 | POE 20000 349.67 12317.24 uses POE::Loop::Event |
2276 | POE 20000 341.54 12086.32 uses POE::Loop::Event |
| 2232 | |
2277 | |
| 2233 | =head3 Discussion |
2278 | =head3 Discussion |
| 2234 | |
2279 | |
| 2235 | This benchmark I<does> measure scalability and overall performance of the |
2280 | This benchmark I<does> measure scalability and overall performance of the |
| 2236 | particular event loop. |
2281 | particular event loop. |
| … | |
… | |
| 2362 | As you can see, the AnyEvent + EV combination even beats the |
2407 | As you can see, the AnyEvent + EV combination even beats the |
| 2363 | hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl |
2408 | hand-optimised "raw sockets benchmark", while AnyEvent + its pure perl |
| 2364 | backend easily beats IO::Lambda and POE. |
2409 | backend easily beats IO::Lambda and POE. |
| 2365 | |
2410 | |
| 2366 | And even the 100% non-blocking version written using the high-level (and |
2411 | And even the 100% non-blocking version written using the high-level (and |
| 2367 | slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda by a |
2412 | slow :) L<AnyEvent::Handle> abstraction beats both POE and IO::Lambda |
| 2368 | large margin, even though it does all of DNS, tcp-connect and socket I/O |
2413 | higher level ("unoptimised") abstractions by a large margin, even though |
| 2369 | in a non-blocking way. |
2414 | it does all of DNS, tcp-connect and socket I/O in a non-blocking way. |
| 2370 | |
2415 | |
| 2371 | The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and |
2416 | The two AnyEvent benchmarks programs can be found as F<eg/ae0.pl> and |
| 2372 | F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are |
2417 | F<eg/ae2.pl> in the AnyEvent distribution, the remaining benchmarks are |
| 2373 | part of the IO::lambda distribution and were used without any changes. |
2418 | part of the IO::Lambda distribution and were used without any changes. |
| 2374 | |
2419 | |
| 2375 | |
2420 | |
| 2376 | =head1 SIGNALS |
2421 | =head1 SIGNALS |
| 2377 | |
2422 | |
| 2378 | AnyEvent currently installs handlers for these signals: |
2423 | AnyEvent currently installs handlers for these signals: |
| … | |
… | |
| 2420 | it's built-in modules) are required to use it. |
2465 | it's built-in modules) are required to use it. |
| 2421 | |
2466 | |
| 2422 | That does not mean that AnyEvent won't take advantage of some additional |
2467 | That does not mean that AnyEvent won't take advantage of some additional |
| 2423 | modules if they are installed. |
2468 | modules if they are installed. |
| 2424 | |
2469 | |
| 2425 | This section epxlains which additional modules will be used, and how they |
2470 | This section explains which additional modules will be used, and how they |
| 2426 | affect AnyEvent's operetion. |
2471 | affect AnyEvent's operation. |
| 2427 | |
2472 | |
| 2428 | =over 4 |
2473 | =over 4 |
| 2429 | |
2474 | |
| 2430 | =item L<Async::Interrupt> |
2475 | =item L<Async::Interrupt> |
| 2431 | |
2476 | |
| … | |
… | |
| 2436 | catch the signals) with some delay (default is 10 seconds, look for |
2481 | catch the signals) with some delay (default is 10 seconds, look for |
| 2437 | C<$AnyEvent::MAX_SIGNAL_LATENCY>). |
2482 | C<$AnyEvent::MAX_SIGNAL_LATENCY>). |
| 2438 | |
2483 | |
| 2439 | If this module is available, then it will be used to implement signal |
2484 | If this module is available, then it will be used to implement signal |
| 2440 | catching, which means that signals will not be delayed, and the event loop |
2485 | catching, which means that signals will not be delayed, and the event loop |
| 2441 | will not be interrupted regularly, which is more efficient (And good for |
2486 | will not be interrupted regularly, which is more efficient (and good for |
| 2442 | battery life on laptops). |
2487 | battery life on laptops). |
| 2443 | |
2488 | |
| 2444 | This affects not just the pure-perl event loop, but also other event loops |
2489 | This affects not just the pure-perl event loop, but also other event loops |
| 2445 | that have no signal handling on their own (e.g. Glib, Tk, Qt). |
2490 | that have no signal handling on their own (e.g. Glib, Tk, Qt). |
| 2446 | |
2491 | |
| … | |
… | |
| 2467 | lot less memory), but otherwise doesn't affect guard operation much. It is |
2512 | lot less memory), but otherwise doesn't affect guard operation much. It is |
| 2468 | purely used for performance. |
2513 | purely used for performance. |
| 2469 | |
2514 | |
| 2470 | =item L<JSON> and L<JSON::XS> |
2515 | =item L<JSON> and L<JSON::XS> |
| 2471 | |
2516 | |
| 2472 | This module is required when you want to read or write JSON data via |
2517 | One of these modules is required when you want to read or write JSON data |
| 2473 | L<AnyEvent::Handle>. It is also written in pure-perl, but can take |
2518 | via L<AnyEvent::Handle>. It is also written in pure-perl, but can take |
| 2474 | advantage of the ultra-high-speed L<JSON::XS> module when it is installed. |
2519 | advantage of the ultra-high-speed L<JSON::XS> module when it is installed. |
| 2475 | |
2520 | |
| 2476 | In fact, L<AnyEvent::Handle> will use L<JSON::XS> by default if it is |
2521 | In fact, L<AnyEvent::Handle> will use L<JSON::XS> by default if it is |
| 2477 | installed. |
2522 | installed. |
| 2478 | |
2523 | |
| … | |
… | |
| 2493 | |
2538 | |
| 2494 | |
2539 | |
| 2495 | =head1 FORK |
2540 | =head1 FORK |
| 2496 | |
2541 | |
| 2497 | Most event libraries are not fork-safe. The ones who are usually are |
2542 | Most event libraries are not fork-safe. The ones who are usually are |
| 2498 | because they rely on inefficient but fork-safe C<select> or C<poll> |
2543 | because they rely on inefficient but fork-safe C<select> or C<poll> calls |
| 2499 | calls. Only L<EV> is fully fork-aware. |
2544 | - higher performance APIs such as BSD's kqueue or the dreaded Linux epoll |
|
|
2545 | are usually badly thought-out hacks that are incompatible with fork in |
|
|
2546 | one way or another. Only L<EV> is fully fork-aware and ensures that you |
|
|
2547 | continue event-processing in both parent and child (or both, if you know |
|
|
2548 | what you are doing). |
|
|
2549 | |
|
|
2550 | This means that, in general, you cannot fork and do event processing in |
|
|
2551 | the child if the event library was initialised before the fork (which |
|
|
2552 | usually happens when the first AnyEvent watcher is created, or the library |
|
|
2553 | is loaded). |
| 2500 | |
2554 | |
| 2501 | If you have to fork, you must either do so I<before> creating your first |
2555 | If you have to fork, you must either do so I<before> creating your first |
| 2502 | watcher OR you must not use AnyEvent at all in the child OR you must do |
2556 | watcher OR you must not use AnyEvent at all in the child OR you must do |
| 2503 | something completely out of the scope of AnyEvent. |
2557 | something completely out of the scope of AnyEvent. |
|
|
2558 | |
|
|
2559 | The problem of doing event processing in the parent I<and> the child |
|
|
2560 | is much more complicated: even for backends that I<are> fork-aware or |
|
|
2561 | fork-safe, their behaviour is not usually what you want: fork clones all |
|
|
2562 | watchers, that means all timers, I/O watchers etc. are active in both |
|
|
2563 | parent and child, which is almost never what you want. USing C<exec> |
|
|
2564 | to start worker children from some kind of manage rprocess is usually |
|
|
2565 | preferred, because it is much easier and cleaner, at the expense of having |
|
|
2566 | to have another binary. |
| 2504 | |
2567 | |
| 2505 | |
2568 | |
| 2506 | =head1 SECURITY CONSIDERATIONS |
2569 | =head1 SECURITY CONSIDERATIONS |
| 2507 | |
2570 | |
| 2508 | AnyEvent can be forced to load any event model via |
2571 | AnyEvent can be forced to load any event model via |
