… | |
… | |
57 | as those use one of the supported event loops. It is trivial to add new |
57 | as those use one of the supported event loops. It is trivial to add new |
58 | event loops to AnyEvent, too, so it is future-proof). |
58 | event loops to AnyEvent, too, so it is future-proof). |
59 | |
59 | |
60 | In addition to being free of having to use I<the one and only true event |
60 | In addition to being free of having to use I<the one and only true event |
61 | model>, AnyEvent also is free of bloat and policy: with POE or similar |
61 | model>, AnyEvent also is free of bloat and policy: with POE or similar |
62 | modules, you get an enourmous amount of code and strict rules you have to |
62 | modules, you get an enormous amount of code and strict rules you have to |
63 | follow. AnyEvent, on the other hand, is lean and up to the point, by only |
63 | follow. AnyEvent, on the other hand, is lean and up to the point, by only |
64 | offering the functionality that is necessary, in as thin as a wrapper as |
64 | offering the functionality that is necessary, in as thin as a wrapper as |
65 | technically possible. |
65 | technically possible. |
66 | |
66 | |
67 | Of course, if you want lots of policy (this can arguably be somewhat |
67 | Of course, if you want lots of policy (this can arguably be somewhat |
… | |
… | |
108 | |
108 | |
109 | =head1 WATCHERS |
109 | =head1 WATCHERS |
110 | |
110 | |
111 | AnyEvent has the central concept of a I<watcher>, which is an object that |
111 | AnyEvent has the central concept of a I<watcher>, which is an object that |
112 | stores relevant data for each kind of event you are waiting for, such as |
112 | stores relevant data for each kind of event you are waiting for, such as |
113 | the callback to call, the filehandle to watch, etc. |
113 | the callback to call, the file handle to watch, etc. |
114 | |
114 | |
115 | These watchers are normal Perl objects with normal Perl lifetime. After |
115 | These watchers are normal Perl objects with normal Perl lifetime. After |
116 | creating a watcher it will immediately "watch" for events and invoke the |
116 | creating a watcher it will immediately "watch" for events and invoke the |
117 | callback when the event occurs (of course, only when the event model |
117 | callback when the event occurs (of course, only when the event model |
118 | is in control). |
118 | is in control). |
… | |
… | |
237 | |
237 | |
238 | Although the callback might get passed parameters, their value and |
238 | Although the callback might get passed parameters, their value and |
239 | presence is undefined and you cannot rely on them. Portable AnyEvent |
239 | presence is undefined and you cannot rely on them. Portable AnyEvent |
240 | callbacks cannot use arguments passed to signal watcher callbacks. |
240 | callbacks cannot use arguments passed to signal watcher callbacks. |
241 | |
241 | |
242 | Multiple signal occurances can be clumped together into one callback |
242 | Multiple signal occurrences can be clumped together into one callback |
243 | invocation, and callback invocation will be synchronous. synchronous means |
243 | invocation, and callback invocation will be synchronous. Synchronous means |
244 | that it might take a while until the signal gets handled by the process, |
244 | that it might take a while until the signal gets handled by the process, |
245 | but it is guarenteed not to interrupt any other callbacks. |
245 | but it is guaranteed not to interrupt any other callbacks. |
246 | |
246 | |
247 | The main advantage of using these watchers is that you can share a signal |
247 | The main advantage of using these watchers is that you can share a signal |
248 | between multiple watchers. |
248 | between multiple watchers. |
249 | |
249 | |
250 | This watcher might use C<%SIG>, so programs overwriting those signals |
250 | This watcher might use C<%SIG>, so programs overwriting those signals |
… | |
… | |
310 | Condition variables can be created by calling the C<< AnyEvent->condvar |
310 | Condition variables can be created by calling the C<< AnyEvent->condvar |
311 | >> method, usually without arguments. The only argument pair allowed is |
311 | >> method, usually without arguments. The only argument pair allowed is |
312 | C<cb>, which specifies a callback to be called when the condition variable |
312 | C<cb>, which specifies a callback to be called when the condition variable |
313 | becomes true. |
313 | becomes true. |
314 | |
314 | |
315 | After creation, the conditon variable is "false" until it becomes "true" |
315 | After creation, the condition variable is "false" until it becomes "true" |
316 | by calling the C<send> method. |
316 | by calling the C<send> method. |
317 | |
317 | |
318 | Condition variables are similar to callbacks, except that you can |
318 | Condition variables are similar to callbacks, except that you can |
319 | optionally wait for them. They can also be called merge points - points |
319 | optionally wait for them. They can also be called merge points - points |
320 | in time where multiple outstandign events have been processed. And yet |
320 | in time where multiple outstanding events have been processed. And yet |
321 | another way to call them is transations - each condition variable can be |
321 | another way to call them is transactions - each condition variable can be |
322 | used to represent a transaction, which finishes at some point and delivers |
322 | used to represent a transaction, which finishes at some point and delivers |
323 | a result. |
323 | a result. |
324 | |
324 | |
325 | Condition variables are very useful to signal that something has finished, |
325 | Condition variables are very useful to signal that something has finished, |
326 | for example, if you write a module that does asynchronous http requests, |
326 | for example, if you write a module that does asynchronous http requests, |
… | |
… | |
332 | you can block your main program until an event occurs - for example, you |
332 | you can block your main program until an event occurs - for example, you |
333 | could C<< ->recv >> in your main program until the user clicks the Quit |
333 | could C<< ->recv >> in your main program until the user clicks the Quit |
334 | button of your app, which would C<< ->send >> the "quit" event. |
334 | button of your app, which would C<< ->send >> the "quit" event. |
335 | |
335 | |
336 | Note that condition variables recurse into the event loop - if you have |
336 | Note that condition variables recurse into the event loop - if you have |
337 | two pieces of code that call C<< ->recv >> in a round-robbin fashion, you |
337 | two pieces of code that call C<< ->recv >> in a round-robin fashion, you |
338 | lose. Therefore, condition variables are good to export to your caller, but |
338 | lose. Therefore, condition variables are good to export to your caller, but |
339 | you should avoid making a blocking wait yourself, at least in callbacks, |
339 | you should avoid making a blocking wait yourself, at least in callbacks, |
340 | as this asks for trouble. |
340 | as this asks for trouble. |
341 | |
341 | |
342 | Condition variables are represented by hash refs in perl, and the keys |
342 | Condition variables are represented by hash refs in perl, and the keys |
… | |
… | |
443 | doesn't execute once). |
443 | doesn't execute once). |
444 | |
444 | |
445 | This is the general pattern when you "fan out" into multiple subrequests: |
445 | This is the general pattern when you "fan out" into multiple subrequests: |
446 | use an outer C<begin>/C<end> pair to set the callback and ensure C<end> |
446 | use an outer C<begin>/C<end> pair to set the callback and ensure C<end> |
447 | is called at least once, and then, for each subrequest you start, call |
447 | is called at least once, and then, for each subrequest you start, call |
448 | C<begin> and for eahc subrequest you finish, call C<end>. |
448 | C<begin> and for each subrequest you finish, call C<end>. |
449 | |
449 | |
450 | =back |
450 | =back |
451 | |
451 | |
452 | =head3 METHODS FOR CONSUMERS |
452 | =head3 METHODS FOR CONSUMERS |
453 | |
453 | |
… | |
… | |
475 | (programs might want to do that to stay interactive), so I<if you are |
475 | (programs might want to do that to stay interactive), so I<if you are |
476 | using this from a module, never require a blocking wait>, but let the |
476 | using this from a module, never require a blocking wait>, but let the |
477 | caller decide whether the call will block or not (for example, by coupling |
477 | caller decide whether the call will block or not (for example, by coupling |
478 | condition variables with some kind of request results and supporting |
478 | condition variables with some kind of request results and supporting |
479 | callbacks so the caller knows that getting the result will not block, |
479 | callbacks so the caller knows that getting the result will not block, |
480 | while still suppporting blocking waits if the caller so desires). |
480 | while still supporting blocking waits if the caller so desires). |
481 | |
481 | |
482 | Another reason I<never> to C<< ->recv >> in a module is that you cannot |
482 | Another reason I<never> to C<< ->recv >> in a module is that you cannot |
483 | sensibly have two C<< ->recv >>'s in parallel, as that would require |
483 | sensibly have two C<< ->recv >>'s in parallel, as that would require |
484 | multiple interpreters or coroutines/threads, none of which C<AnyEvent> |
484 | multiple interpreters or coroutines/threads, none of which C<AnyEvent> |
485 | can supply. |
485 | can supply. |
… | |
… | |
705 | our @ISA; |
705 | our @ISA; |
706 | |
706 | |
707 | our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1; |
707 | our $verbose = $ENV{PERL_ANYEVENT_VERBOSE}*1; |
708 | |
708 | |
709 | our @REGISTRY; |
709 | our @REGISTRY; |
|
|
710 | |
|
|
711 | our %PROTOCOL; # (ipv4|ipv6) => (1|2) |
|
|
712 | |
|
|
713 | { |
|
|
714 | my $idx; |
|
|
715 | $PROTOCOL{$_} = ++$idx |
|
|
716 | for split /\s*,\s*/, $ENV{PERL_ANYEVENT_PROTOCOLS} || "ipv4,ipv6"; |
|
|
717 | } |
710 | |
718 | |
711 | my @models = ( |
719 | my @models = ( |
712 | [EV:: => AnyEvent::Impl::EV::], |
720 | [EV:: => AnyEvent::Impl::EV::], |
713 | [Event:: => AnyEvent::Impl::Event::], |
721 | [Event:: => AnyEvent::Impl::Event::], |
714 | [Tk:: => AnyEvent::Impl::Tk::], |
722 | [Tk:: => AnyEvent::Impl::Tk::], |
… | |
… | |
1021 | model it chooses. |
1029 | model it chooses. |
1022 | |
1030 | |
1023 | =item C<PERL_ANYEVENT_MODEL> |
1031 | =item C<PERL_ANYEVENT_MODEL> |
1024 | |
1032 | |
1025 | This can be used to specify the event model to be used by AnyEvent, before |
1033 | This can be used to specify the event model to be used by AnyEvent, before |
1026 | autodetection and -probing kicks in. It must be a string consisting |
1034 | auto detection and -probing kicks in. It must be a string consisting |
1027 | entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended |
1035 | entirely of ASCII letters. The string C<AnyEvent::Impl::> gets prepended |
1028 | and the resulting module name is loaded and if the load was successful, |
1036 | and the resulting module name is loaded and if the load was successful, |
1029 | used as event model. If it fails to load AnyEvent will proceed with |
1037 | used as event model. If it fails to load AnyEvent will proceed with |
1030 | autodetection and -probing. |
1038 | auto detection and -probing. |
1031 | |
1039 | |
1032 | This functionality might change in future versions. |
1040 | This functionality might change in future versions. |
1033 | |
1041 | |
1034 | For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you |
1042 | For example, to force the pure perl model (L<AnyEvent::Impl::Perl>) you |
1035 | could start your program like this: |
1043 | could start your program like this: |
… | |
… | |
1038 | |
1046 | |
1039 | =item C<PERL_ANYEVENT_PROTOCOLS> |
1047 | =item C<PERL_ANYEVENT_PROTOCOLS> |
1040 | |
1048 | |
1041 | Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences |
1049 | Used by both L<AnyEvent::DNS> and L<AnyEvent::Socket> to determine preferences |
1042 | for IPv4 or IPv6. The default is unspecified (and might change, or be the result |
1050 | for IPv4 or IPv6. The default is unspecified (and might change, or be the result |
1043 | of autoprobing). |
1051 | of auto probing). |
1044 | |
1052 | |
1045 | Must be set to a comma-separated list of protocols or address families, |
1053 | Must be set to a comma-separated list of protocols or address families, |
1046 | current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be |
1054 | current supported: C<ipv4> and C<ipv6>. Only protocols mentioned will be |
1047 | used, and preference will be given to protocols mentioned earlier in the |
1055 | used, and preference will be given to protocols mentioned earlier in the |
1048 | list. |
1056 | list. |
1049 | |
1057 | |
|
|
1058 | This variable can effectively be used for denial-of-service attacks |
|
|
1059 | against local programs (e.g. when setuid), although the impact is likely |
|
|
1060 | small, as the program has to handle connection errors already- |
|
|
1061 | |
1050 | Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6, |
1062 | Examples: C<PERL_ANYEVENT_PROTOCOLS=ipv4,ipv6> - prefer IPv4 over IPv6, |
1051 | but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4> |
1063 | but support both and try to use both. C<PERL_ANYEVENT_PROTOCOLS=ipv4> |
1052 | - only support IPv4, never try to resolve or contact IPv6 |
1064 | - only support IPv4, never try to resolve or contact IPv6 |
1053 | addressses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or |
1065 | addresses. C<PERL_ANYEVENT_PROTOCOLS=ipv6,ipv4> support either IPv4 or |
1054 | IPv6, but prefer IPv6 over IPv4. |
1066 | IPv6, but prefer IPv6 over IPv4. |
|
|
1067 | |
|
|
1068 | =item C<PERL_ANYEVENT_EDNS0> |
|
|
1069 | |
|
|
1070 | Used by L<AnyEvent::DNS> to decide whether to use the EDNS0 extension |
|
|
1071 | for DNS. This extension is generally useful to reduce DNS traffic, but |
|
|
1072 | some (broken) firewalls drop such DNS packets, which is why it is off by |
|
|
1073 | default. |
|
|
1074 | |
|
|
1075 | Setting this variable to C<1> will cause L<AnyEvent::DNS> to announce |
|
|
1076 | EDNS0 in its DNS requests. |
1055 | |
1077 | |
1056 | =back |
1078 | =back |
1057 | |
1079 | |
1058 | =head1 EXAMPLE PROGRAM |
1080 | =head1 EXAMPLE PROGRAM |
1059 | |
1081 | |
… | |
… | |
1145 | syswrite $txn->{fh}, $txn->{request} |
1167 | syswrite $txn->{fh}, $txn->{request} |
1146 | or die "connection or write error"; |
1168 | or die "connection or write error"; |
1147 | $txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r }); |
1169 | $txn->{w} = AnyEvent->io (fh => $txn->{fh}, poll => 'r', cb => sub { $txn->fh_ready_r }); |
1148 | |
1170 | |
1149 | Again, C<fh_ready_r> waits till all data has arrived, and then stores the |
1171 | Again, C<fh_ready_r> waits till all data has arrived, and then stores the |
1150 | result and signals any possible waiters that the request ahs finished: |
1172 | result and signals any possible waiters that the request has finished: |
1151 | |
1173 | |
1152 | sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf}; |
1174 | sysread $txn->{fh}, $txn->{buf}, length $txn->{$buf}; |
1153 | |
1175 | |
1154 | if (end-of-file or data complete) { |
1176 | if (end-of-file or data complete) { |
1155 | $txn->{result} = $txn->{buf}; |
1177 | $txn->{result} = $txn->{buf}; |
… | |
… | |
1163 | |
1185 | |
1164 | $txn->{finished}->recv; |
1186 | $txn->{finished}->recv; |
1165 | return $txn->{result}; |
1187 | return $txn->{result}; |
1166 | |
1188 | |
1167 | The actual code goes further and collects all errors (C<die>s, exceptions) |
1189 | The actual code goes further and collects all errors (C<die>s, exceptions) |
1168 | that occured during request processing. The C<result> method detects |
1190 | that occurred during request processing. The C<result> method detects |
1169 | whether an exception as thrown (it is stored inside the $txn object) |
1191 | whether an exception as thrown (it is stored inside the $txn object) |
1170 | and just throws the exception, which means connection errors and other |
1192 | and just throws the exception, which means connection errors and other |
1171 | problems get reported tot he code that tries to use the result, not in a |
1193 | problems get reported tot he code that tries to use the result, not in a |
1172 | random callback. |
1194 | random callback. |
1173 | |
1195 | |
… | |
… | |
1219 | of various event loops I prepared some benchmarks. |
1241 | of various event loops I prepared some benchmarks. |
1220 | |
1242 | |
1221 | =head2 BENCHMARKING ANYEVENT OVERHEAD |
1243 | =head2 BENCHMARKING ANYEVENT OVERHEAD |
1222 | |
1244 | |
1223 | Here is a benchmark of various supported event models used natively and |
1245 | Here is a benchmark of various supported event models used natively and |
1224 | through anyevent. The benchmark creates a lot of timers (with a zero |
1246 | through AnyEvent. The benchmark creates a lot of timers (with a zero |
1225 | timeout) and I/O watchers (watching STDOUT, a pty, to become writable, |
1247 | timeout) and I/O watchers (watching STDOUT, a pty, to become writable, |
1226 | which it is), lets them fire exactly once and destroys them again. |
1248 | which it is), lets them fire exactly once and destroys them again. |
1227 | |
1249 | |
1228 | Source code for this benchmark is found as F<eg/bench> in the AnyEvent |
1250 | Source code for this benchmark is found as F<eg/bench> in the AnyEvent |
1229 | distribution. |
1251 | distribution. |
… | |
… | |
1352 | |
1374 | |
1353 | =back |
1375 | =back |
1354 | |
1376 | |
1355 | =head2 BENCHMARKING THE LARGE SERVER CASE |
1377 | =head2 BENCHMARKING THE LARGE SERVER CASE |
1356 | |
1378 | |
1357 | This benchmark atcually benchmarks the event loop itself. It works by |
1379 | This benchmark actually benchmarks the event loop itself. It works by |
1358 | creating a number of "servers": each server consists of a socketpair, a |
1380 | creating a number of "servers": each server consists of a socket pair, a |
1359 | timeout watcher that gets reset on activity (but never fires), and an I/O |
1381 | timeout watcher that gets reset on activity (but never fires), and an I/O |
1360 | watcher waiting for input on one side of the socket. Each time the socket |
1382 | watcher waiting for input on one side of the socket. Each time the socket |
1361 | watcher reads a byte it will write that byte to a random other "server". |
1383 | watcher reads a byte it will write that byte to a random other "server". |
1362 | |
1384 | |
1363 | The effect is that there will be a lot of I/O watchers, only part of which |
1385 | The effect is that there will be a lot of I/O watchers, only part of which |
1364 | are active at any one point (so there is a constant number of active |
1386 | are active at any one point (so there is a constant number of active |
1365 | fds for each loop iterstaion, but which fds these are is random). The |
1387 | fds for each loop iteration, but which fds these are is random). The |
1366 | timeout is reset each time something is read because that reflects how |
1388 | timeout is reset each time something is read because that reflects how |
1367 | most timeouts work (and puts extra pressure on the event loops). |
1389 | most timeouts work (and puts extra pressure on the event loops). |
1368 | |
1390 | |
1369 | In this benchmark, we use 10000 socketpairs (20000 sockets), of which 100 |
1391 | In this benchmark, we use 10000 socket pairs (20000 sockets), of which 100 |
1370 | (1%) are active. This mirrors the activity of large servers with many |
1392 | (1%) are active. This mirrors the activity of large servers with many |
1371 | connections, most of which are idle at any one point in time. |
1393 | connections, most of which are idle at any one point in time. |
1372 | |
1394 | |
1373 | Source code for this benchmark is found as F<eg/bench2> in the AnyEvent |
1395 | Source code for this benchmark is found as F<eg/bench2> in the AnyEvent |
1374 | distribution. |
1396 | distribution. |
… | |
… | |
1376 | =head3 Explanation of the columns |
1398 | =head3 Explanation of the columns |
1377 | |
1399 | |
1378 | I<sockets> is the number of sockets, and twice the number of "servers" (as |
1400 | I<sockets> is the number of sockets, and twice the number of "servers" (as |
1379 | each server has a read and write socket end). |
1401 | each server has a read and write socket end). |
1380 | |
1402 | |
1381 | I<create> is the time it takes to create a socketpair (which is |
1403 | I<create> is the time it takes to create a socket pair (which is |
1382 | nontrivial) and two watchers: an I/O watcher and a timeout watcher. |
1404 | nontrivial) and two watchers: an I/O watcher and a timeout watcher. |
1383 | |
1405 | |
1384 | I<request>, the most important value, is the time it takes to handle a |
1406 | I<request>, the most important value, is the time it takes to handle a |
1385 | single "request", that is, reading the token from the pipe and forwarding |
1407 | single "request", that is, reading the token from the pipe and forwarding |
1386 | it to another server. This includes deleting the old timeout and creating |
1408 | it to another server. This includes deleting the old timeout and creating |
… | |
… | |
1459 | speed most when you have lots of watchers, not when you only have a few of |
1481 | speed most when you have lots of watchers, not when you only have a few of |
1460 | them). |
1482 | them). |
1461 | |
1483 | |
1462 | EV is again fastest. |
1484 | EV is again fastest. |
1463 | |
1485 | |
1464 | Perl again comes second. It is noticably faster than the C-based event |
1486 | Perl again comes second. It is noticeably faster than the C-based event |
1465 | loops Event and Glib, although the difference is too small to really |
1487 | loops Event and Glib, although the difference is too small to really |
1466 | matter. |
1488 | matter. |
1467 | |
1489 | |
1468 | POE also performs much better in this case, but is is still far behind the |
1490 | POE also performs much better in this case, but is is still far behind the |
1469 | others. |
1491 | others. |