ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/AnyEvent-MP/MP.pm
(Generate patch)

Comparing AnyEvent-MP/MP.pm (file contents):
Revision 1.32 by root, Wed Aug 5 19:58:46 2009 UTC vs.
Revision 1.48 by root, Thu Aug 13 02:59:42 2009 UTC

8 8
9 $NODE # contains this node's noderef 9 $NODE # contains this node's noderef
10 NODE # returns this node's noderef 10 NODE # returns this node's noderef
11 NODE $port # returns the noderef of the port 11 NODE $port # returns the noderef of the port
12 12
13 $SELF # receiving/own port id in rcv callbacks
14
15 # initialise the node so it can send/receive messages
16 initialise_node; # -OR-
17 initialise_node "localhost:4040"; # -OR-
18 initialise_node "slave/", "localhost:4040"
19
20 # ports are message endpoints
21
22 # sending messages
13 snd $port, type => data...; 23 snd $port, type => data...;
24 snd $port, @msg;
25 snd @msg_with_first_element_being_a_port;
14 26
15 $SELF # receiving/own port id in rcv callbacks 27 # creating/using miniports
28 my $miniport = port { my @msg = @_; 0 };
16 29
30 # creating/using full ports
31 my $port = port;
17 rcv $port, smartmatch => $cb->($port, @msg); 32 rcv $port, smartmatch => $cb->(@msg);
18
19 # examples:
20 rcv $port2, ping => sub { snd $_[0], "pong"; 0 }; 33 rcv $port, ping => sub { snd $_[0], "pong"; 0 };
21 rcv $port1, pong => sub { warn "pong received\n" }; 34 rcv $port, pong => sub { warn "pong received\n"; 0 };
22 snd $port2, ping => $port1;
23 35
24 # more, smarter, matches (_any_ is exported by this module) 36 # more, smarter, matches (_any_ is exported by this module)
25 rcv $port, [child_died => $pid] => sub { ... 37 rcv $port, [child_died => $pid] => sub { ...
26 rcv $port, [_any_, _any_, 3] => sub { .. $_[2] is 3 38 rcv $port, [_any_, _any_, 3] => sub { .. $_[2] is 3
27 39
40 # create a port on another node
41 my $port = spawn $node, $initfunc, @initdata;
42
43 # monitoring
44 mon $port, $cb->(@msg) # callback is invoked on death
45 mon $port, $otherport # kill otherport on abnormal death
46 mon $port, $otherport, @msg # send message on death
47
48=head1 CURRENT STATUS
49
50 AnyEvent::MP - stable API, should work
51 AnyEvent::MP::Intro - outdated
52 AnyEvent::MP::Kernel - WIP
53 AnyEvent::MP::Transport - mostly stable
54
55 stay tuned.
56
28=head1 DESCRIPTION 57=head1 DESCRIPTION
29 58
30This module (-family) implements a simple message passing framework. 59This module (-family) implements a simple message passing framework.
31 60
32Despite its simplicity, you can securely message other processes running 61Despite its simplicity, you can securely message other processes running
35For an introduction to this module family, see the L<AnyEvent::MP::Intro> 64For an introduction to this module family, see the L<AnyEvent::MP::Intro>
36manual page. 65manual page.
37 66
38At the moment, this module family is severly broken and underdocumented, 67At the moment, this module family is severly broken and underdocumented,
39so do not use. This was uploaded mainly to reserve the CPAN namespace - 68so do not use. This was uploaded mainly to reserve the CPAN namespace -
40stay tuned! The basic API should be finished, however. 69stay tuned!
41 70
42=head1 CONCEPTS 71=head1 CONCEPTS
43 72
44=over 4 73=over 4
45 74
90 119
91=cut 120=cut
92 121
93package AnyEvent::MP; 122package AnyEvent::MP;
94 123
95use AnyEvent::MP::Base; 124use AnyEvent::MP::Kernel;
96 125
97use common::sense; 126use common::sense;
98 127
99use Carp (); 128use Carp ();
100 129
101use AE (); 130use AE ();
102 131
103use base "Exporter"; 132use base "Exporter";
104 133
105our $VERSION = '0.1'; 134our $VERSION = $AnyEvent::MP::Kernel::VERSION;
135
106our @EXPORT = qw( 136our @EXPORT = qw(
107 NODE $NODE *SELF node_of _any_ 137 NODE $NODE *SELF node_of _any_
108 resolve_node initialise_node 138 resolve_node initialise_node
109 snd rcv mon kil reg psub 139 snd rcv mon kil reg psub spawn
110 port 140 port
111); 141);
112 142
113our $SELF; 143our $SELF;
114 144
123The C<NODE> function returns, and the C<$NODE> variable contains 153The C<NODE> function returns, and the C<$NODE> variable contains
124the noderef of the local node. The value is initialised by a call 154the noderef of the local node. The value is initialised by a call
125to C<become_public> or C<become_slave>, after which all local port 155to C<become_public> or C<become_slave>, after which all local port
126identifiers become invalid. 156identifiers become invalid.
127 157
128=item $noderef = node_of $portid 158=item $noderef = node_of $port
129 159
130Extracts and returns the noderef from a portid or a noderef. 160Extracts and returns the noderef from a portid or a noderef.
161
162=item initialise_node $noderef, $seednode, $seednode...
163
164=item initialise_node "slave/", $master, $master...
165
166Before a node can talk to other nodes on the network it has to initialise
167itself - the minimum a node needs to know is it's own name, and optionally
168it should know the noderefs of some other nodes in the network.
169
170This function initialises a node - it must be called exactly once (or
171never) before calling other AnyEvent::MP functions.
172
173All arguments are noderefs, which can be either resolved or unresolved.
174
175There are two types of networked nodes, public nodes and slave nodes:
176
177=over 4
178
179=item public nodes
180
181For public nodes, C<$noderef> must either be a (possibly unresolved)
182noderef, in which case it will be resolved, or C<undef> (or missing), in
183which case the noderef will be guessed.
184
185Afterwards, the node will bind itself on all endpoints and try to connect
186to all additional C<$seednodes> that are specified. Seednodes are optional
187and can be used to quickly bootstrap the node into an existing network.
188
189=item slave nodes
190
191When the C<$noderef> is the special string C<slave/>, then the node will
192become a slave node. Slave nodes cannot be contacted from outside and will
193route most of their traffic to the master node that they attach to.
194
195At least one additional noderef is required: The node will try to connect
196to all of them and will become a slave attached to the first node it can
197successfully connect to.
198
199=back
200
201This function will block until all nodes have been resolved and, for slave
202nodes, until it has successfully established a connection to a master
203server.
204
205Example: become a public node listening on the default node.
206
207 initialise_node;
208
209Example: become a public node, and try to contact some well-known master
210servers to become part of the network.
211
212 initialise_node undef, "master1", "master2";
213
214Example: become a public node listening on port C<4041>.
215
216 initialise_node 4041;
217
218Example: become a public node, only visible on localhost port 4044.
219
220 initialise_node "locahost:4044";
221
222Example: become a slave node to any of the specified master servers.
223
224 initialise_node "slave/", "master1", "192.168.13.17", "mp.example.net";
131 225
132=item $cv = resolve_node $noderef 226=item $cv = resolve_node $noderef
133 227
134Takes an unresolved node reference that may contain hostnames and 228Takes an unresolved node reference that may contain hostnames and
135abbreviated IDs, resolves all of them and returns a resolved node 229abbreviated IDs, resolves all of them and returns a resolved node
167 261
168Due to some quirks in how perl exports variables, it is impossible to 262Due to some quirks in how perl exports variables, it is impossible to
169just export C<$SELF>, all the symbols called C<SELF> are exported by this 263just export C<$SELF>, all the symbols called C<SELF> are exported by this
170module, but only C<$SELF> is currently used. 264module, but only C<$SELF> is currently used.
171 265
172=item snd $portid, type => @data 266=item snd $port, type => @data
173 267
174=item snd $portid, @msg 268=item snd $port, @msg
175 269
176Send the given message to the given port ID, which can identify either 270Send the given message to the given port ID, which can identify either
177a local or a remote port, and can be either a string or soemthignt hat 271a local or a remote port, and can be either a string or soemthignt hat
178stringifies a sa port ID (such as a port object :). 272stringifies a sa port ID (such as a port object :).
179 273
195 289
196Create a new local port object that can be used either as a pattern 290Create a new local port object that can be used either as a pattern
197matching port ("full port") or a single-callback port ("miniport"), 291matching port ("full port") or a single-callback port ("miniport"),
198depending on how C<rcv> callbacks are bound to the object. 292depending on how C<rcv> callbacks are bound to the object.
199 293
200=item $portid = port { my @msg = @_; $finished } 294=item $port = port { my @msg = @_; $finished }
201 295
202Creates a "mini port", that is, a very lightweight port without any 296Creates a "miniport", that is, a very lightweight port without any pattern
203pattern matching behind it, and returns its ID. 297matching behind it, and returns its ID. Semantically the same as creating
298a port and calling C<rcv $port, $callback> on it.
204 299
205The block will be called for every message received on the port. When the 300The block will be called for every message received on the port. When the
206callback returns a true value its job is considered "done" and the port 301callback returns a true value its job is considered "done" and the port
207will be destroyed. Otherwise it will stay alive. 302will be destroyed. Otherwise it will stay alive.
208 303
215 snd $otherport, reply => $port; 310 snd $otherport, reply => $port;
216 }; 311 };
217 312
218=cut 313=cut
219 314
315sub rcv($@);
316
220sub port(;&) { 317sub port(;&) {
221 my $id = "$UNIQ." . $ID++; 318 my $id = "$UNIQ." . $ID++;
222 my $port = "$NODE#$id"; 319 my $port = "$NODE#$id";
223 320
224 if (@_) { 321 if (@_) {
322 rcv $port, shift;
323 } else {
324 $PORT{$id} = sub { }; # nop
325 }
326
327 $port
328}
329
330=item reg $port, $name
331
332=item reg $name
333
334Registers the given port (or C<$SELF><<< if missing) under the name
335C<$name>. If the name already exists it is replaced.
336
337A port can only be registered under one well known name.
338
339A port automatically becomes unregistered when it is killed.
340
341=cut
342
343sub reg(@) {
344 my $port = @_ > 1 ? shift : $SELF || Carp::croak 'reg: called with one argument only, but $SELF not set,';
345
346 $REG{$_[0]} = $port;
347}
348
349=item rcv $port, $callback->(@msg)
350
351Replaces the callback on the specified miniport (after converting it to
352one if required).
353
354=item rcv $port, tagstring => $callback->(@msg), ...
355
356=item rcv $port, $smartmatch => $callback->(@msg), ...
357
358=item rcv $port, [$smartmatch...] => $callback->(@msg), ...
359
360Register callbacks to be called on matching messages on the given full
361port (after converting it to one if required) and return the port.
362
363The callback has to return a true value when its work is done, after
364which is will be removed, or a false value in which case it will stay
365registered.
366
367The global C<$SELF> (exported by this module) contains C<$port> while
368executing the callback.
369
370Runtime errors during callback execution will result in the port being
371C<kil>ed.
372
373If the match is an array reference, then it will be matched against the
374first elements of the message, otherwise only the first element is being
375matched.
376
377Any element in the match that is specified as C<_any_> (a function
378exported by this module) matches any single element of the message.
379
380While not required, it is highly recommended that the first matching
381element is a string identifying the message. The one-string-only match is
382also the most efficient match (by far).
383
384Example: create a port and bind receivers on it in one go.
385
386 my $port = rcv port,
387 msg1 => sub { ...; 0 },
388 msg2 => sub { ...; 0 },
389 ;
390
391Example: create a port, bind receivers and send it in a message elsewhere
392in one go:
393
394 snd $otherport, reply =>
395 rcv port,
396 msg1 => sub { ...; 0 },
397 ...
398 ;
399
400=cut
401
402sub rcv($@) {
403 my $port = shift;
404 my ($noderef, $portid) = split /#/, $port, 2;
405
406 ($NODE{$noderef} || add_node $noderef) == $NODE{""}
407 or Carp::croak "$port: rcv can only be called on local ports, caught";
408
409 if (@_ == 1) {
225 my $cb = shift; 410 my $cb = shift;
411 delete $PORT_DATA{$portid};
226 $PORT{$id} = sub { 412 $PORT{$portid} = sub {
227 local $SELF = $port; 413 local $SELF = $port;
228 eval { 414 eval {
229 &$cb 415 &$cb
230 and kil $id; 416 and kil $port;
231 }; 417 };
232 _self_die if $@; 418 _self_die if $@;
233 }; 419 };
234 } else { 420 } else {
421 my $self = $PORT_DATA{$portid} ||= do {
235 my $self = bless { 422 my $self = bless {
236 id => "$NODE#$id", 423 id => $port,
237 }, "AnyEvent::MP::Port"; 424 }, "AnyEvent::MP::Port";
238 425
239 $PORT_DATA{$id} = $self;
240 $PORT{$id} = sub { 426 $PORT{$portid} = sub {
241 local $SELF = $port; 427 local $SELF = $port;
242 428
243 eval { 429 eval {
244 for (@{ $self->{rc0}{$_[0]} }) { 430 for (@{ $self->{rc0}{$_[0]} }) {
245 $_ && &{$_->[0]} 431 $_ && &{$_->[0]}
246 && undef $_; 432 && undef $_;
247 } 433 }
248 434
249 for (@{ $self->{rcv}{$_[0]} }) { 435 for (@{ $self->{rcv}{$_[0]} }) {
250 $_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1] 436 $_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1]
251 && &{$_->[0]} 437 && &{$_->[0]}
252 && undef $_; 438 && undef $_;
253 } 439 }
254 440
255 for (@{ $self->{any} }) { 441 for (@{ $self->{any} }) {
256 $_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1] 442 $_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1]
257 && &{$_->[0]} 443 && &{$_->[0]}
258 && undef $_; 444 && undef $_;
445 }
259 } 446 };
447 _self_die if $@;
260 }; 448 };
261 _self_die if $@; 449
450 $self
262 }; 451 };
263 }
264 452
265 $port
266}
267
268=item reg $portid, $name
269
270Registers the given port under the name C<$name>. If the name already
271exists it is replaced.
272
273A port can only be registered under one well known name.
274
275A port automatically becomes unregistered when it is killed.
276
277=cut
278
279sub reg(@) {
280 my ($portid, $name) = @_;
281
282 $REG{$name} = $portid;
283}
284
285=item rcv $portid, $callback->(@msg)
286
287Replaces the callback on the specified miniport (or newly created port
288object, see C<port>). Full ports are configured with the following calls:
289
290=item rcv $portid, tagstring => $callback->(@msg), ...
291
292=item rcv $portid, $smartmatch => $callback->(@msg), ...
293
294=item rcv $portid, [$smartmatch...] => $callback->(@msg), ...
295
296Register callbacks to be called on matching messages on the given full
297port (or newly created port).
298
299The callback has to return a true value when its work is done, after
300which is will be removed, or a false value in which case it will stay
301registered.
302
303The global C<$SELF> (exported by this module) contains C<$portid> while
304executing the callback.
305
306Runtime errors wdurign callback execution will result in the port being
307C<kil>ed.
308
309If the match is an array reference, then it will be matched against the
310first elements of the message, otherwise only the first element is being
311matched.
312
313Any element in the match that is specified as C<_any_> (a function
314exported by this module) matches any single element of the message.
315
316While not required, it is highly recommended that the first matching
317element is a string identifying the message. The one-string-only match is
318also the most efficient match (by far).
319
320=cut
321
322sub rcv($@) {
323 my $portid = shift;
324 my ($noderef, $port) = split /#/, $port, 2;
325
326 ($NODE{$noderef} || add_node $noderef) == $NODE{""}
327 or Carp::croak "$noderef#$port: rcv can only be called on local ports, caught";
328
329 my $self = $PORT_DATA{$port}
330 or Carp::croak "$noderef#$port: rcv can only be called on message matching ports, caught";
331
332 "AnyEvent::MP::Port" eq ref $self 453 "AnyEvent::MP::Port" eq ref $self
333 or Carp::croak "$noderef#$port: rcv can only be called on message matching ports, caught"; 454 or Carp::croak "$port: rcv can only be called on message matching ports, caught";
334 455
335 while (@_) { 456 while (@_) {
336 my ($match, $cb) = splice @_, 0, 2; 457 my ($match, $cb) = splice @_, 0, 2;
337 458
338 if (!ref $match) { 459 if (!ref $match) {
339 push @{ $self->{rc0}{$match} }, [$cb]; 460 push @{ $self->{rc0}{$match} }, [$cb];
340 } elsif (("ARRAY" eq ref $match && !ref $match->[0])) { 461 } elsif (("ARRAY" eq ref $match && !ref $match->[0])) {
341 my ($type, @match) = @$match; 462 my ($type, @match) = @$match;
342 @match 463 @match
343 ? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match] 464 ? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match]
344 : push @{ $self->{rc0}{$match->[0]} }, [$cb]; 465 : push @{ $self->{rc0}{$match->[0]} }, [$cb];
345 } else { 466 } else {
346 push @{ $self->{any} }, [$cb, $match]; 467 push @{ $self->{any} }, [$cb, $match];
468 }
347 } 469 }
348 } 470 }
349 471
350 $portid 472 $port
351} 473}
352 474
353=item $closure = psub { BLOCK } 475=item $closure = psub { BLOCK }
354 476
355Remembers C<$SELF> and creates a closure out of the BLOCK. When the 477Remembers C<$SELF> and creates a closure out of the BLOCK. When the
386 $res 508 $res
387 } 509 }
388 } 510 }
389} 511}
390 512
391=item $guard = mon $portid, $cb->(@reason) 513=item $guard = mon $port, $cb->(@reason)
392 514
393=item $guard = mon $portid, $otherport 515=item $guard = mon $port, $rcvport
394 516
517=item $guard = mon $port
518
395=item $guard = mon $portid, $otherport, @msg 519=item $guard = mon $port, $rcvport, @msg
396 520
397Monitor the given port and do something when the port is killed. 521Monitor the given port and do something when the port is killed or
522messages to it were lost, and optionally return a guard that can be used
523to stop monitoring again.
398 524
525C<mon> effectively guarantees that, in the absence of hardware failures,
526that after starting the monitor, either all messages sent to the port
527will arrive, or the monitoring action will be invoked after possible
528message loss has been detected. No messages will be lost "in between"
529(after the first lost message no further messages will be received by the
530port). After the monitoring action was invoked, further messages might get
531delivered again.
532
399In the first form, the callback is simply called with any number 533In the first form (callback), the callback is simply called with any
400of C<@reason> elements (no @reason means that the port was deleted 534number of C<@reason> elements (no @reason means that the port was deleted
401"normally"). Note also that I<< the callback B<must> never die >>, so use 535"normally"). Note also that I<< the callback B<must> never die >>, so use
402C<eval> if unsure. 536C<eval> if unsure.
403 537
404In the second form, the other port will be C<kil>'ed with C<@reason>, iff 538In the second form (another port given), the other port (C<$rcvport>)
405a @reason was specified, i.e. on "normal" kils nothing happens, while 539will be C<kil>'ed with C<@reason>, iff a @reason was specified, i.e. on
406under all other conditions, the other port is killed with the same reason. 540"normal" kils nothing happens, while under all other conditions, the other
541port is killed with the same reason.
407 542
543The third form (kill self) is the same as the second form, except that
544C<$rvport> defaults to C<$SELF>.
545
408In the last form, a message of the form C<@msg, @reason> will be C<snd>. 546In the last form (message), a message of the form C<@msg, @reason> will be
547C<snd>.
548
549As a rule of thumb, monitoring requests should always monitor a port from
550a local port (or callback). The reason is that kill messages might get
551lost, just like any other message. Another less obvious reason is that
552even monitoring requests can get lost (for exmaple, when the connection
553to the other node goes down permanently). When monitoring a port locally
554these problems do not exist.
409 555
410Example: call a given callback when C<$port> is killed. 556Example: call a given callback when C<$port> is killed.
411 557
412 mon $port, sub { warn "port died because of <@_>\n" }; 558 mon $port, sub { warn "port died because of <@_>\n" };
413 559
414Example: kill ourselves when C<$port> is killed abnormally. 560Example: kill ourselves when C<$port> is killed abnormally.
415 561
416 mon $port, $self; 562 mon $port;
417 563
418Example: send us a restart message another C<$port> is killed. 564Example: send us a restart message when another C<$port> is killed.
419 565
420 mon $port, $self => "restart"; 566 mon $port, $self => "restart";
421 567
422=cut 568=cut
423 569
424sub mon { 570sub mon {
425 my ($noderef, $port) = split /#/, shift, 2; 571 my ($noderef, $port) = split /#/, shift, 2;
426 572
427 my $node = $NODE{$noderef} || add_node $noderef; 573 my $node = $NODE{$noderef} || add_node $noderef;
428 574
429 my $cb = shift; 575 my $cb = @_ ? shift : $SELF || Carp::croak 'mon: called with one argument only, but $SELF not set,';
430 576
431 unless (ref $cb) { 577 unless (ref $cb) {
432 if (@_) { 578 if (@_) {
433 # send a kill info message 579 # send a kill info message
434 my (@msg) = ($cb, @_); 580 my (@msg) = ($cb, @_);
465=cut 611=cut
466 612
467sub mon_guard { 613sub mon_guard {
468 my ($port, @refs) = @_; 614 my ($port, @refs) = @_;
469 615
616 #TODO: mon-less form?
617
470 mon $port, sub { 0 && @refs } 618 mon $port, sub { 0 && @refs }
471} 619}
472 620
473=item lnk $port1, $port2
474
475Link two ports. This is simply a shorthand for:
476
477 mon $port1, $port2;
478 mon $port2, $port1;
479
480It means that if either one is killed abnormally, the other one gets
481killed as well.
482
483=item kil $portid[, @reason] 621=item kil $port[, @reason]
484 622
485Kill the specified port with the given C<@reason>. 623Kill the specified port with the given C<@reason>.
486 624
487If no C<@reason> is specified, then the port is killed "normally" (linked 625If no C<@reason> is specified, then the port is killed "normally" (linked
488ports will not be kileld, or even notified). 626ports will not be kileld, or even notified).
494will be reported as reason C<< die => $@ >>. 632will be reported as reason C<< die => $@ >>.
495 633
496Transport/communication errors are reported as C<< transport_error => 634Transport/communication errors are reported as C<< transport_error =>
497$message >>. 635$message >>.
498 636
499=back
500
501=head1 FUNCTIONS FOR NODES
502
503=over 4
504
505=item become_public $noderef
506
507Tells the node to become a public node, i.e. reachable from other nodes.
508
509The first argument is the (unresolved) node reference of the local node
510(if missing then the empty string is used).
511
512It is quite common to not specify anything, in which case the local node
513tries to listen on the default port, or to only specify a port number, in
514which case AnyEvent::MP tries to guess the local addresses.
515
516=cut 637=cut
638
639=item $port = spawn $node, $initfunc[, @initdata]
640
641Creates a port on the node C<$node> (which can also be a port ID, in which
642case it's the node where that port resides).
643
644The port ID of the newly created port is return immediately, and it is
645permissible to immediately start sending messages or monitor the port.
646
647After the port has been created, the init function is
648called. This function must be a fully-qualified function name
649(e.g. C<MyApp::Chat::Server::init>). To specify a function in the main
650program, use C<::name>.
651
652If the function doesn't exist, then the node tries to C<require>
653the package, then the package above the package and so on (e.g.
654C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function
655exists or it runs out of package names.
656
657The init function is then called with the newly-created port as context
658object (C<$SELF>) and the C<@initdata> values as arguments.
659
660A common idiom is to pass your own port, monitor the spawned port, and
661in the init function, monitor the original port. This two-way monitoring
662ensures that both ports get cleaned up when there is a problem.
663
664Example: spawn a chat server port on C<$othernode>.
665
666 # this node, executed from within a port context:
667 my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF;
668 mon $server;
669
670 # init function on C<$othernode>
671 sub connect {
672 my ($srcport) = @_;
673
674 mon $srcport;
675
676 rcv $SELF, sub {
677 ...
678 };
679 }
680
681=cut
682
683sub _spawn {
684 my $port = shift;
685 my $init = shift;
686
687 local $SELF = "$NODE#$port";
688 eval {
689 &{ load_func $init }
690 };
691 _self_die if $@;
692}
693
694sub spawn(@) {
695 my ($noderef, undef) = split /#/, shift, 2;
696
697 my $id = "$RUNIQ." . $ID++;
698
699 $_[0] =~ /::/
700 or Carp::croak "spawn init function must be a fully-qualified name, caught";
701
702 ($NODE{$noderef} || add_node $noderef)
703 ->send (["", "AnyEvent::MP::_spawn" => $id, @_]);
704
705 "$noderef#$id"
706}
517 707
518=back 708=back
519 709
520=head1 NODE MESSAGES 710=head1 NODE MESSAGES
521 711
563 753
564=back 754=back
565 755
566=head1 AnyEvent::MP vs. Distributed Erlang 756=head1 AnyEvent::MP vs. Distributed Erlang
567 757
568AnyEvent::MP got lots of its ideas from distributed erlang (erlang node 758AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node
569== aemp node, erlang process == aemp port), so many of the documents and 759== aemp node, Erlang process == aemp port), so many of the documents and
570programming techniques employed by erlang apply to AnyEvent::MP. Here is a 760programming techniques employed by Erlang apply to AnyEvent::MP. Here is a
571sample: 761sample:
572 762
573 http://www.erlang.se/doc/programming_rules.shtml 763 http://www.Erlang.se/doc/programming_rules.shtml
574 http://erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4 764 http://Erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4
575 http://erlang.org/download/erlang-book-part1.pdf # chapters 5 and 6 765 http://Erlang.org/download/Erlang-book-part1.pdf # chapters 5 and 6
576 http://erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5 766 http://Erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5
577 767
578Despite the similarities, there are also some important differences: 768Despite the similarities, there are also some important differences:
579 769
580=over 4 770=over 4
581 771
592 782
593Erlang uses processes that selctively receive messages, and therefore 783Erlang uses processes that selctively receive messages, and therefore
594needs a queue. AEMP is event based, queuing messages would serve no useful 784needs a queue. AEMP is event based, queuing messages would serve no useful
595purpose. 785purpose.
596 786
597(But see L<Coro::MP> for a more erlang-like process model on top of AEMP). 787(But see L<Coro::MP> for a more Erlang-like process model on top of AEMP).
598 788
599=item * Erlang sends are synchronous, AEMP sends are asynchronous. 789=item * Erlang sends are synchronous, AEMP sends are asynchronous.
600 790
601Sending messages in erlang is synchronous and blocks the process. AEMP 791Sending messages in Erlang is synchronous and blocks the process. AEMP
602sends are immediate, connection establishment is handled in the 792sends are immediate, connection establishment is handled in the
603background. 793background.
604 794
605=item * Erlang can silently lose messages, AEMP cannot. 795=item * Erlang can silently lose messages, AEMP cannot.
606 796
609and c, and the other side only receives messages a and c). 799and c, and the other side only receives messages a and c).
610 800
611AEMP guarantees correct ordering, and the guarantee that there are no 801AEMP guarantees correct ordering, and the guarantee that there are no
612holes in the message sequence. 802holes in the message sequence.
613 803
614=item * In erlang, processes can be declared dead and later be found to be 804=item * In Erlang, processes can be declared dead and later be found to be
615alive. 805alive.
616 806
617In erlang it can happen that a monitored process is declared dead and 807In Erlang it can happen that a monitored process is declared dead and
618linked processes get killed, but later it turns out that the process is 808linked processes get killed, but later it turns out that the process is
619still alive - and can receive messages. 809still alive - and can receive messages.
620 810
621In AEMP, when port monitoring detects a port as dead, then that port will 811In AEMP, when port monitoring detects a port as dead, then that port will
622eventually be killed - it cannot happen that a node detects a port as dead 812eventually be killed - it cannot happen that a node detects a port as dead
623and then later sends messages to it, finding it is still alive. 813and then later sends messages to it, finding it is still alive.
624 814
625=item * Erlang can send messages to the wrong port, AEMP does not. 815=item * Erlang can send messages to the wrong port, AEMP does not.
626 816
627In erlang it is quite possible that a node that restarts reuses a process 817In Erlang it is quite possible that a node that restarts reuses a process
628ID known to other nodes for a completely different process, causing 818ID known to other nodes for a completely different process, causing
629messages destined for that process to end up in an unrelated process. 819messages destined for that process to end up in an unrelated process.
630 820
631AEMP never reuses port IDs, so old messages or old port IDs floating 821AEMP never reuses port IDs, so old messages or old port IDs floating
632around in the network will not be sent to an unrelated port. 822around in the network will not be sent to an unrelated port.
638securely authenticate nodes. 828securely authenticate nodes.
639 829
640=item * The AEMP protocol is optimised for both text-based and binary 830=item * The AEMP protocol is optimised for both text-based and binary
641communications. 831communications.
642 832
643The AEMP protocol, unlike the erlang protocol, supports both 833The AEMP protocol, unlike the Erlang protocol, supports both
644language-independent text-only protocols (good for debugging) and binary, 834language-independent text-only protocols (good for debugging) and binary,
645language-specific serialisers (e.g. Storable). 835language-specific serialisers (e.g. Storable).
646 836
647It has also been carefully designed to be implementable in other languages 837It has also been carefully designed to be implementable in other languages
648with a minimum of work while gracefully degrading fucntionality to make the 838with a minimum of work while gracefully degrading fucntionality to make the
649protocol simple. 839protocol simple.
650 840
841=item * AEMP has more flexible monitoring options than Erlang.
842
843In Erlang, you can chose to receive I<all> exit signals as messages
844or I<none>, there is no in-between, so monitoring single processes is
845difficult to implement. Monitoring in AEMP is more flexible than in
846Erlang, as one can choose between automatic kill, exit message or callback
847on a per-process basis.
848
849=item * Erlang tries to hide remote/local connections, AEMP does not.
850
851Monitoring in Erlang is not an indicator of process death/crashes,
852as linking is (except linking is unreliable in Erlang).
853
854In AEMP, you don't "look up" registered port names or send to named ports
855that might or might not be persistent. Instead, you normally spawn a port
856on the remote node. The init function monitors the you, and you monitor
857the remote port. Since both monitors are local to the node, they are much
858more reliable.
859
860This also saves round-trips and avoids sending messages to the wrong port
861(hard to do in Erlang).
862
863=back
864
865=head1 RATIONALE
866
867=over 4
868
869=item Why strings for ports and noderefs, why not objects?
870
871We considered "objects", but found that the actual number of methods
872thatc an be called are very low. Since port IDs and noderefs travel over
873the network frequently, the serialising/deserialising would add lots of
874overhead, as well as having to keep a proxy object.
875
876Strings can easily be printed, easily serialised etc. and need no special
877procedures to be "valid".
878
879And a a miniport consists of a single closure stored in a global hash - it
880can't become much cheaper.
881
882=item Why favour JSON, why not real serialising format such as Storable?
883
884In fact, any AnyEvent::MP node will happily accept Storable as framing
885format, but currently there is no way to make a node use Storable by
886default.
887
888The default framing protocol is JSON because a) JSON::XS is many times
889faster for small messages and b) most importantly, after years of
890experience we found that object serialisation is causing more problems
891than it gains: Just like function calls, objects simply do not travel
892easily over the network, mostly because they will always be a copy, so you
893always have to re-think your design.
894
895Keeping your messages simple, concentrating on data structures rather than
896objects, will keep your messages clean, tidy and efficient.
897
651=back 898=back
652 899
653=head1 SEE ALSO 900=head1 SEE ALSO
654 901
655L<AnyEvent>. 902L<AnyEvent>.

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines