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 1.21 by root, Tue Aug 4 14:10:51 2009 UTC vs.
Revision 1.28 by root, Tue Aug 4 22:16:54 2009 UTC

4 4
5=head1 SYNOPSIS 5=head1 SYNOPSIS
6 6
7 use AnyEvent::MP; 7 use AnyEvent::MP;
8 8
9 NODE # returns this node identifier
10 $NODE # contains this node identifier 9 $NODE # contains this node's noderef
10 NODE # returns this node's noderef
11 NODE $port # returns the noderef of the port
11 12
12 snd $port, type => data...; 13 snd $port, type => data...;
14
15 $SELF # receiving/own port id in rcv callbacks
13 16
14 rcv $port, smartmatch => $cb->($port, @msg); 17 rcv $port, smartmatch => $cb->($port, @msg);
15 18
16 # examples: 19 # examples:
17 rcv $port2, ping => sub { snd $_[0], "pong"; 0 }; 20 rcv $port2, ping => sub { snd $_[0], "pong"; 0 };
27This module (-family) implements a simple message passing framework. 30This module (-family) implements a simple message passing framework.
28 31
29Despite its simplicity, you can securely message other processes running 32Despite its simplicity, you can securely message other processes running
30on the same or other hosts. 33on the same or other hosts.
31 34
35For an introduction to this module family, see the L<AnyEvent::MP::Intro>
36manual page.
37
32At the moment, this module family is severly brokena nd underdocumented, 38At the moment, this module family is severly broken and underdocumented,
33so do not use. This was uploaded mainly to reserve the CPAN namespace - 39so do not use. This was uploaded mainly to reserve the CPAN namespace -
34stay tuned! 40stay tuned! The basic API should be finished, however.
35 41
36=head1 CONCEPTS 42=head1 CONCEPTS
37 43
38=over 4 44=over 4
39 45
82 88
83use AE (); 89use AE ();
84 90
85use base "Exporter"; 91use base "Exporter";
86 92
87our $VERSION = '0.02'; 93our $VERSION = '0.1';
88our @EXPORT = qw( 94our @EXPORT = qw(
89 NODE $NODE $PORT snd rcv mon kil _any_ 95 NODE $NODE *SELF node_of _any_
90 create_port create_port_on
91 miniport
92 become_slave become_public 96 become_slave become_public
97 snd rcv mon kil reg psub
98 port
93); 99);
94 100
101our $SELF;
102
103sub _self_die() {
104 my $msg = $@;
105 $msg =~ s/\n+$// unless ref $msg;
106 kil $SELF, die => $msg;
107}
108
95=item NODE / $NODE 109=item $thisnode = NODE / $NODE
96 110
97The C<NODE ()> function and the C<$NODE> variable contain the noderef of 111The C<NODE> function returns, and the C<$NODE> variable contains
98the local node. The value is initialised by a call to C<become_public> or 112the noderef of the local node. The value is initialised by a call
99C<become_slave>, after which all local port identifiers become invalid. 113to C<become_public> or C<become_slave>, after which all local port
114identifiers become invalid.
115
116=item $noderef = node_of $portid
117
118Extracts and returns the noderef from a portid or a noderef.
119
120=item $SELF
121
122Contains the current port id while executing C<rcv> callbacks or C<psub>
123blocks.
124
125=item SELF, %SELF, @SELF...
126
127Due to some quirks in how perl exports variables, it is impossible to
128just export C<$SELF>, all the symbols called C<SELF> are exported by this
129module, but only C<$SELF> is currently used.
100 130
101=item snd $portid, type => @data 131=item snd $portid, type => @data
102 132
103=item snd $portid, @msg 133=item snd $portid, @msg
104 134
118JSON is used, then only strings, numbers and arrays and hashes consisting 148JSON is used, then only strings, numbers and arrays and hashes consisting
119of those are allowed (no objects). When Storable is used, then anything 149of those are allowed (no objects). When Storable is used, then anything
120that Storable can serialise and deserialise is allowed, and for the local 150that Storable can serialise and deserialise is allowed, and for the local
121node, anything can be passed. 151node, anything can be passed.
122 152
153=item kil $portid[, @reason]
154
155Kill the specified port with the given C<@reason>.
156
157If no C<@reason> is specified, then the port is killed "normally" (linked
158ports will not be kileld, or even notified).
159
160Otherwise, linked ports get killed with the same reason (second form of
161C<mon>, see below).
162
163Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks
164will be reported as reason C<< die => $@ >>.
165
166Transport/communication errors are reported as C<< transport_error =>
167$message >>.
168
123=item $guard = mon $portid, $cb->() 169=item $guard = mon $portid, $cb->(@reason)
124 170
125=item $guard = mon $portid, $otherport 171=item $guard = mon $portid, $otherport
126 172
127=item $guard = mon $portid, $otherport, @msg 173=item $guard = mon $portid, $otherport, @msg
128 174
129Monitor the given port and call the given callback when the port is 175Monitor the given port and do something when the port is killed.
130destroyed or connection to it's node is lost.
131 176
132#TODO 177In the first form, the callback is simply called with any number
178of C<@reason> elements (no @reason means that the port was deleted
179"normally"). Note also that I<< the callback B<must> never die >>, so use
180C<eval> if unsure.
181
182In the second form, the other port will be C<kil>'ed with C<@reason>, iff
183a @reason was specified, i.e. on "normal" kils nothing happens, while
184under all other conditions, the other port is killed with the same reason.
185
186In the last form, a message of the form C<@msg, @reason> will be C<snd>.
187
188Example: call a given callback when C<$port> is killed.
189
190 mon $port, sub { warn "port died because of <@_>\n" };
191
192Example: kill ourselves when C<$port> is killed abnormally.
193
194 mon $port, $self;
195
196Example: send us a restart message another C<$port> is killed.
197
198 mon $port, $self => "restart";
133 199
134=cut 200=cut
135 201
136sub mon { 202sub mon {
137 my ($noderef, $port, $cb) = ((split /#/, shift, 2), shift); 203 my ($noderef, $port, $cb) = ((split /#/, shift, 2), shift);
138 204
139 my $node = AnyEvent::MP::Base::add_node $noderef; 205 my $node = $NODE{$noderef} || add_node $noderef;
140 206
141 #TODO: ports must not be references 207 #TODO: ports must not be references
142 if (!ref $cb or "AnyEvent::MP::Port" eq ref $cb) { 208 if (!ref $cb or "AnyEvent::MP::Port" eq ref $cb) {
143 if (@_) { 209 if (@_) {
144 # send a kill info message 210 # send a kill info message
145 my (@msg) = ($cb, @_); 211 my (@msg) = ($cb, @_);
146 $cb = sub { snd @msg, @_ }; 212 $cb = sub { snd @msg, @_ };
147 } else { 213 } else {
148 # simply kill other port 214 # simply kill other port
149 my $port = $cb; 215 my $port = $cb;
150 $cb = sub { kil $port, @_ }; 216 $cb = sub { kil $port, @_ if @_ };
151 } 217 }
152 } 218 }
153 219
154 $node->monitor ($port, $cb); 220 $node->monitor ($port, $cb);
155 221
179 my ($port, @refs) = @_; 245 my ($port, @refs) = @_;
180 246
181 mon $port, sub { 0 && @refs } 247 mon $port, sub { 0 && @refs }
182} 248}
183 249
250=item lnk $port1, $port2
251
252Link two ports. This is simply a shorthand for:
253
254 mon $port1, $port2;
255 mon $port2, $port1;
256
257It means that if either one is killed abnormally, the other one gets
258killed as well.
259
184=item $local_port = create_port 260=item $local_port = port
185 261
186Create a new local port object. See the next section for allowed methods. 262Create a new local port object that supports message matching.
187 263
188=cut
189
190sub create_port {
191 my $id = "$AnyEvent::MP::Base::UNIQ." . $AnyEvent::MP::Base::ID++;
192
193 my $self = bless {
194 id => "$NODE#$id",
195 }, "AnyEvent::MP::Port";
196
197 $AnyEvent::MP::Base::PORT{$id} = sub {
198 unshift @_, $self;
199
200 for (@{ $self->{rc0}{$_[1]} }) {
201 $_ && &{$_->[0]}
202 && undef $_;
203 }
204
205 for (@{ $self->{rcv}{$_[1]} }) {
206 $_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1]
207 && &{$_->[0]}
208 && undef $_;
209 }
210
211 for (@{ $self->{any} }) {
212 $_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1]
213 && &{$_->[0]}
214 && undef $_;
215 }
216 };
217
218 $self
219}
220
221=item $portid = miniport { my @msg = @_; $finished } 264=item $portid = port { my @msg = @_; $finished }
222 265
223Creates a "mini port", that is, a very lightweight port without any 266Creates a "mini port", that is, a very lightweight port without any
224pattern matching behind it, and returns its ID. 267pattern matching behind it, and returns its ID.
225 268
226The block will be called for every message received on the port. When the 269The block will be called for every message received on the port. When the
236 snd $otherport, reply => $port; 279 snd $otherport, reply => $port;
237 }; 280 };
238 281
239=cut 282=cut
240 283
241sub miniport(&) { 284sub port(;&) {
285 my $id = "$UNIQ." . $ID++;
286 my $port = "$NODE#$id";
287
288 if (@_) {
242 my $cb = shift; 289 my $cb = shift;
243 my $id = "$AnyEvent::MP::Base::UNIQ." . $AnyEvent::MP::Base::ID++; 290 $PORT{$id} = sub {
244 291 local $SELF = $port;
245 $AnyEvent::MP::Base::PORT{$id} = sub { 292 eval {
246 &$cb 293 &$cb
247 and kil $id; 294 and kil $id;
295 };
296 _self_die if $@;
297 };
298 } else {
299 my $self = bless {
300 id => "$NODE#$id",
301 }, "AnyEvent::MP::Port";
302
303 $PORT_DATA{$id} = $self;
304 $PORT{$id} = sub {
305 local $SELF = $port;
306
307 eval {
308 for (@{ $self->{rc0}{$_[0]} }) {
309 $_ && &{$_->[0]}
310 && undef $_;
311 }
312
313 for (@{ $self->{rcv}{$_[0]} }) {
314 $_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1]
315 && &{$_->[0]}
316 && undef $_;
317 }
318
319 for (@{ $self->{any} }) {
320 $_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1]
321 && &{$_->[0]}
322 && undef $_;
323 }
324 };
325 _self_die if $@;
326 };
248 }; 327 }
249 328
250 "$NODE#$id" 329 $port
251} 330}
252 331
253package AnyEvent::MP::Port; 332=item reg $portid, $name
254 333
255=back 334Registers the given port under the name C<$name>. If the name already
335exists it is replaced.
256 336
257=head1 METHODS FOR PORT OBJECTS 337A port can only be registered under one well known name.
258 338
259=over 4 339A port automatically becomes unregistered when it is killed.
260 340
261=item "$port"
262
263A port object stringifies to its port ID, so can be used directly for
264C<snd> operations.
265
266=cut 341=cut
267 342
268use overload 343sub reg(@) {
269 '""' => sub { $_[0]{id} }, 344 my ($portid, $name) = @_;
270 fallback => 1;
271 345
272sub TO_JSON { $_[0]{id} } 346 $REG{$name} = $portid;
347}
273 348
274=item $port->rcv (type => $callback->($port, @msg)) 349=item rcv $portid, tagstring => $callback->(@msg), ...
275 350
276=item $port->rcv ($smartmatch => $callback->($port, @msg)) 351=item rcv $portid, $smartmatch => $callback->(@msg), ...
277 352
278=item $port->rcv ([$smartmatch...] => $callback->($port, @msg)) 353=item rcv $portid, [$smartmatch...] => $callback->(@msg), ...
279 354
280Register a callback on the given port. 355Register callbacks to be called on matching messages on the given port.
281 356
282The callback has to return a true value when its work is done, after 357The callback has to return a true value when its work is done, after
283which is will be removed, or a false value in which case it will stay 358which is will be removed, or a false value in which case it will stay
284registered. 359registered.
285 360
361The global C<$SELF> (exported by this module) contains C<$portid> while
362executing the callback.
363
364Runtime errors wdurign callback execution will result in the port being
365C<kil>ed.
366
286If the match is an array reference, then it will be matched against the 367If the match is an array reference, then it will be matched against the
287first elements of the message, otherwise only the first element is being 368first elements of the message, otherwise only the first element is being
288matched. 369matched.
289 370
290Any element in the match that is specified as C<_any_> (a function 371Any element in the match that is specified as C<_any_> (a function
295also the most efficient match (by far). 376also the most efficient match (by far).
296 377
297=cut 378=cut
298 379
299sub rcv($@) { 380sub rcv($@) {
300 my ($self, $match, $cb) = @_; 381 my ($noderef, $port) = split /#/, shift, 2;
301 382
383 ($NODE{$noderef} || add_node $noderef) == $NODE{""}
384 or Carp::croak "$noderef#$port: rcv can only be called on local ports, caught";
385
386 my $self = $PORT_DATA{$port}
387 or Carp::croak "$noderef#$port: rcv can only be called on message matching ports, caught";
388
389 "AnyEvent::MP::Port" eq ref $self
390 or Carp::croak "$noderef#$port: rcv can only be called on message matching ports, caught";
391
392 while (@_) {
393 my ($match, $cb) = splice @_, 0, 2;
394
302 if (!ref $match) { 395 if (!ref $match) {
303 push @{ $self->{rc0}{$match} }, [$cb]; 396 push @{ $self->{rc0}{$match} }, [$cb];
304 } elsif (("ARRAY" eq ref $match && !ref $match->[0])) { 397 } elsif (("ARRAY" eq ref $match && !ref $match->[0])) {
305 my ($type, @match) = @$match; 398 my ($type, @match) = @$match;
306 @match 399 @match
307 ? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match] 400 ? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match]
308 : push @{ $self->{rc0}{$match->[0]} }, [$cb]; 401 : push @{ $self->{rc0}{$match->[0]} }, [$cb];
309 } else { 402 } else {
310 push @{ $self->{any} }, [$cb, $match]; 403 push @{ $self->{any} }, [$cb, $match];
404 }
311 } 405 }
312} 406}
313 407
314=item $port->register ($name) 408=item $closure = psub { BLOCK }
315 409
316Registers the given port under the well known name C<$name>. If the name 410Remembers C<$SELF> and creates a closure out of the BLOCK. When the
317already exists it is replaced. 411closure is executed, sets up the environment in the same way as in C<rcv>
412callbacks, i.e. runtime errors will cause the port to get C<kil>ed.
318 413
319A port can only be registered under one well known name. 414This is useful when you register callbacks from C<rcv> callbacks:
320 415
321=cut 416 rcv delayed_reply => sub {
417 my ($delay, @reply) = @_;
418 my $timer = AE::timer $delay, 0, psub {
419 snd @reply, $SELF;
420 };
421 };
322 422
323sub register { 423=cut
324 my ($self, $name) = @_;
325 424
326 $self->{wkname} = $name; 425sub psub(&) {
327 $AnyEvent::MP::Base::WKP{$name} = "$self"; 426 my $cb = shift;
427
428 my $port = $SELF
429 or Carp::croak "psub can only be called from within rcv or psub callbacks, not";
430
431 sub {
432 local $SELF = $port;
433
434 if (wantarray) {
435 my @res = eval { &$cb };
436 _self_die if $@;
437 @res
438 } else {
439 my $res = eval { &$cb };
440 _self_die if $@;
441 $res
442 }
443 }
328} 444}
329 445
330=item $port->destroy
331
332Explicitly destroy/remove/nuke/vaporise the port.
333
334Ports are normally kept alive by their mere existance alone, and need to
335be destroyed explicitly.
336
337=cut
338
339sub destroy {
340 my ($self) = @_;
341
342 delete $AnyEvent::MP::Base::WKP{ $self->{wkname} };
343
344 AnyEvent::MP::Base::kil $self->{id};
345}
346
347=back 446=back
348 447
349=head1 FUNCTIONS FOR NODES 448=head1 FUNCTIONS FOR NODES
350 449
351=over 4 450=over 4
352
353=item mon $noderef, $callback->($noderef, $status, $)
354
355Monitors the given noderef.
356 451
357=item become_public endpoint... 452=item become_public endpoint...
358 453
359Tells the node to become a public node, i.e. reachable from other nodes. 454Tells the node to become a public node, i.e. reachable from other nodes.
360 455
380 475
381=over 4 476=over 4
382 477
383=cut 478=cut
384 479
385=item wkp => $name, @reply 480=item lookup => $name, @reply
386 481
387Replies with the port ID of the specified well-known port, or C<undef>. 482Replies with the port ID of the specified well-known port, or C<undef>.
388 483
389=item devnull => ... 484=item devnull => ...
390 485
413 snd $NODE, time => $myport, timereply => 1, 2; 508 snd $NODE, time => $myport, timereply => 1, 2;
414 # => snd $myport, timereply => 1, 2, <time> 509 # => snd $myport, timereply => 1, 2, <time>
415 510
416=back 511=back
417 512
513=head1 AnyEvent::MP vs. Distributed Erlang
514
515AnyEvent::MP got lots of its ideas from distributed erlang (erlang node
516== aemp node, erlang process == aemp port), so many of the documents and
517programming techniques employed by erlang apply to AnyEvent::MP. Here is a
518sample:
519
520 http://www.erlang.se/doc/programming_rules.shtml
521 http://erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4
522 http://erlang.org/download/erlang-book-part1.pdf # chapters 5 and 6
523 http://erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5
524
525Despite the similarities, there are also some important differences:
526
527=over 4
528
529=item * Node references contain the recipe on how to contact them.
530
531Erlang relies on special naming and DNS to work everywhere in the
532same way. AEMP relies on each node knowing it's own address(es), with
533convenience functionality.
534
535This means that AEMP requires a less tightly controlled environment at the
536cost of longer node references and a slightly higher management overhead.
537
538=item * Erlang uses processes and a mailbox, AEMP does not queue.
539
540Erlang uses processes that selctively receive messages, and therefore
541needs a queue. AEMP is event based, queuing messages would serve no useful
542purpose.
543
544(But see L<Coro::MP> for a more erlang-like process model on top of AEMP).
545
546=item * Erlang sends are synchronous, AEMP sends are asynchronous.
547
548Sending messages in erlang is synchronous and blocks the process. AEMP
549sends are immediate, connection establishment is handled in the
550background.
551
552=item * Erlang can silently lose messages, AEMP cannot.
553
554Erlang makes few guarantees on messages delivery - messages can get lost
555without any of the processes realising it (i.e. you send messages a, b,
556and c, and the other side only receives messages a and c).
557
558AEMP guarantees correct ordering, and the guarantee that there are no
559holes in the message sequence.
560
561=item * In erlang, processes can be declared dead and later be found to be
562alive.
563
564In erlang it can happen that a monitored process is declared dead and
565linked processes get killed, but later it turns out that the process is
566still alive - and can receive messages.
567
568In AEMP, when port monitoring detects a port as dead, then that port will
569eventually be killed - it cannot happen that a node detects a port as dead
570and then later sends messages to it, finding it is still alive.
571
572=item * Erlang can send messages to the wrong port, AEMP does not.
573
574In erlang it is quite possible that a node that restarts reuses a process
575ID known to other nodes for a completely different process, causing
576messages destined for that process to end up in an unrelated process.
577
578AEMP never reuses port IDs, so old messages or old port IDs floating
579around in the network will not be sent to an unrelated port.
580
581=item * Erlang uses unprotected connections, AEMP uses secure
582authentication and can use TLS.
583
584AEMP can use a proven protocol - SSL/TLS - to protect connections and
585securely authenticate nodes.
586
587=item * The AEMP protocol is optimised for both text-based and binary
588communications.
589
590The AEMP protocol, unlike the erlang protocol, supports both
591language-independent text-only protocols (good for debugging) and binary,
592language-specific serialisers (e.g. Storable).
593
594It has also been carefully designed to be implementable in other languages
595with a minimum of work while gracefully degrading fucntionality to make the
596protocol simple.
597
598=back
599
418=head1 SEE ALSO 600=head1 SEE ALSO
419 601
420L<AnyEvent>. 602L<AnyEvent>.
421 603
422=head1 AUTHOR 604=head1 AUTHOR

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines