1 |
=head1 NAME |
2 |
|
3 |
AnyEvent::MP - multi-processing/message-passing framework |
4 |
|
5 |
=head1 SYNOPSIS |
6 |
|
7 |
use AnyEvent::MP; |
8 |
|
9 |
$NODE # contains this node's noderef |
10 |
NODE # returns this node's noderef |
11 |
NODE $port # returns the noderef of the port |
12 |
|
13 |
$SELF # receiving/own port id in rcv callbacks |
14 |
|
15 |
# ports are message endpoints |
16 |
|
17 |
# sending messages |
18 |
snd $port, type => data...; |
19 |
snd $port, @msg; |
20 |
snd @msg_with_first_element_being_a_port; |
21 |
|
22 |
# miniports |
23 |
my $miniport = port { my @msg = @_; 0 }; |
24 |
|
25 |
# full ports |
26 |
my $port = port; |
27 |
rcv $port, smartmatch => $cb->(@msg); |
28 |
rcv $port, ping => sub { snd $_[0], "pong"; 0 }; |
29 |
rcv $port, pong => sub { warn "pong received\n"; 0 }; |
30 |
|
31 |
# remote ports |
32 |
my $port = spawn $node, $initfunc, @initdata; |
33 |
|
34 |
# more, smarter, matches (_any_ is exported by this module) |
35 |
rcv $port, [child_died => $pid] => sub { ... |
36 |
rcv $port, [_any_, _any_, 3] => sub { .. $_[2] is 3 |
37 |
|
38 |
# monitoring |
39 |
mon $port, $cb->(@msg) # callback is invoked on death |
40 |
mon $port, $otherport # kill otherport on abnormal death |
41 |
mon $port, $otherport, @msg # send message on death |
42 |
|
43 |
=head1 CURRENT STATUS |
44 |
|
45 |
AnyEvent::MP - stable API, should work |
46 |
AnyEvent::MP::Intro - outdated |
47 |
AnyEvent::MP::Kernel - WIP |
48 |
AnyEvent::MP::Transport - mostly stable |
49 |
|
50 |
stay tuned. |
51 |
|
52 |
=head1 DESCRIPTION |
53 |
|
54 |
This module (-family) implements a simple message passing framework. |
55 |
|
56 |
Despite its simplicity, you can securely message other processes running |
57 |
on the same or other hosts. |
58 |
|
59 |
For an introduction to this module family, see the L<AnyEvent::MP::Intro> |
60 |
manual page. |
61 |
|
62 |
At the moment, this module family is severly broken and underdocumented, |
63 |
so do not use. This was uploaded mainly to reserve the CPAN namespace - |
64 |
stay tuned! |
65 |
|
66 |
=head1 CONCEPTS |
67 |
|
68 |
=over 4 |
69 |
|
70 |
=item port |
71 |
|
72 |
A port is something you can send messages to (with the C<snd> function). |
73 |
|
74 |
Some ports allow you to register C<rcv> handlers that can match specific |
75 |
messages. All C<rcv> handlers will receive messages they match, messages |
76 |
will not be queued. |
77 |
|
78 |
=item port id - C<noderef#portname> |
79 |
|
80 |
A port id is normaly the concatenation of a noderef, a hash-mark (C<#>) as |
81 |
separator, and a port name (a printable string of unspecified format). An |
82 |
exception is the the node port, whose ID is identical to its node |
83 |
reference. |
84 |
|
85 |
=item node |
86 |
|
87 |
A node is a single process containing at least one port - the node |
88 |
port. You can send messages to node ports to find existing ports or to |
89 |
create new ports, among other things. |
90 |
|
91 |
Nodes are either private (single-process only), slaves (connected to a |
92 |
master node only) or public nodes (connectable from unrelated nodes). |
93 |
|
94 |
=item noderef - C<host:port,host:port...>, C<id@noderef>, C<id> |
95 |
|
96 |
A node reference is a string that either simply identifies the node (for |
97 |
private and slave nodes), or contains a recipe on how to reach a given |
98 |
node (for public nodes). |
99 |
|
100 |
This recipe is simply a comma-separated list of C<address:port> pairs (for |
101 |
TCP/IP, other protocols might look different). |
102 |
|
103 |
Node references come in two flavours: resolved (containing only numerical |
104 |
addresses) or unresolved (where hostnames are used instead of addresses). |
105 |
|
106 |
Before using an unresolved node reference in a message you first have to |
107 |
resolve it. |
108 |
|
109 |
=back |
110 |
|
111 |
=head1 VARIABLES/FUNCTIONS |
112 |
|
113 |
=over 4 |
114 |
|
115 |
=cut |
116 |
|
117 |
package AnyEvent::MP; |
118 |
|
119 |
use AnyEvent::MP::Kernel; |
120 |
|
121 |
use common::sense; |
122 |
|
123 |
use Carp (); |
124 |
|
125 |
use AE (); |
126 |
|
127 |
use base "Exporter"; |
128 |
|
129 |
our $VERSION = $AnyEvent::MP::Kernel::VERSION; |
130 |
|
131 |
our @EXPORT = qw( |
132 |
NODE $NODE *SELF node_of _any_ |
133 |
resolve_node initialise_node |
134 |
snd rcv mon kil reg psub spawn |
135 |
port |
136 |
); |
137 |
|
138 |
our $SELF; |
139 |
|
140 |
sub _self_die() { |
141 |
my $msg = $@; |
142 |
$msg =~ s/\n+$// unless ref $msg; |
143 |
kil $SELF, die => $msg; |
144 |
} |
145 |
|
146 |
=item $thisnode = NODE / $NODE |
147 |
|
148 |
The C<NODE> function returns, and the C<$NODE> variable contains |
149 |
the noderef of the local node. The value is initialised by a call |
150 |
to C<become_public> or C<become_slave>, after which all local port |
151 |
identifiers become invalid. |
152 |
|
153 |
=item $noderef = node_of $port |
154 |
|
155 |
Extracts and returns the noderef from a portid or a noderef. |
156 |
|
157 |
=item initialise_node $noderef, $seednode, $seednode... |
158 |
|
159 |
=item initialise_node "slave/", $master, $master... |
160 |
|
161 |
Before a node can talk to other nodes on the network it has to initialise |
162 |
itself - the minimum a node needs to know is it's own name, and optionally |
163 |
it should know the noderefs of some other nodes in the network. |
164 |
|
165 |
This function initialises a node - it must be called exactly once (or |
166 |
never) before calling other AnyEvent::MP functions. |
167 |
|
168 |
All arguments are noderefs, which can be either resolved or unresolved. |
169 |
|
170 |
There are two types of networked nodes, public nodes and slave nodes: |
171 |
|
172 |
=over 4 |
173 |
|
174 |
=item public nodes |
175 |
|
176 |
For public nodes, C<$noderef> must either be a (possibly unresolved) |
177 |
noderef, in which case it will be resolved, or C<undef> (or missing), in |
178 |
which case the noderef will be guessed. |
179 |
|
180 |
Afterwards, the node will bind itself on all endpoints and try to connect |
181 |
to all additional C<$seednodes> that are specified. Seednodes are optional |
182 |
and can be used to quickly bootstrap the node into an existing network. |
183 |
|
184 |
=item slave nodes |
185 |
|
186 |
When the C<$noderef> is the special string C<slave/>, then the node will |
187 |
become a slave node. Slave nodes cannot be contacted from outside and will |
188 |
route most of their traffic to the master node that they attach to. |
189 |
|
190 |
At least one additional noderef is required: The node will try to connect |
191 |
to all of them and will become a slave attached to the first node it can |
192 |
successfully connect to. |
193 |
|
194 |
=back |
195 |
|
196 |
This function will block until all nodes have been resolved and, for slave |
197 |
nodes, until it has successfully established a connection to a master |
198 |
server. |
199 |
|
200 |
Example: become a public node listening on the default node. |
201 |
|
202 |
initialise_node; |
203 |
|
204 |
Example: become a public node, and try to contact some well-known master |
205 |
servers to become part of the network. |
206 |
|
207 |
initialise_node undef, "master1", "master2"; |
208 |
|
209 |
Example: become a public node listening on port C<4041>. |
210 |
|
211 |
initialise_node 4041; |
212 |
|
213 |
Example: become a public node, only visible on localhost port 4044. |
214 |
|
215 |
initialise_node "locahost:4044"; |
216 |
|
217 |
Example: become a slave node to any of the specified master servers. |
218 |
|
219 |
initialise_node "slave/", "master1", "192.168.13.17", "mp.example.net"; |
220 |
|
221 |
=item $cv = resolve_node $noderef |
222 |
|
223 |
Takes an unresolved node reference that may contain hostnames and |
224 |
abbreviated IDs, resolves all of them and returns a resolved node |
225 |
reference. |
226 |
|
227 |
In addition to C<address:port> pairs allowed in resolved noderefs, the |
228 |
following forms are supported: |
229 |
|
230 |
=over 4 |
231 |
|
232 |
=item the empty string |
233 |
|
234 |
An empty-string component gets resolved as if the default port (4040) was |
235 |
specified. |
236 |
|
237 |
=item naked port numbers (e.g. C<1234>) |
238 |
|
239 |
These are resolved by prepending the local nodename and a colon, to be |
240 |
further resolved. |
241 |
|
242 |
=item hostnames (e.g. C<localhost:1234>, C<localhost>) |
243 |
|
244 |
These are resolved by using AnyEvent::DNS to resolve them, optionally |
245 |
looking up SRV records for the C<aemp=4040> port, if no port was |
246 |
specified. |
247 |
|
248 |
=back |
249 |
|
250 |
=item $SELF |
251 |
|
252 |
Contains the current port id while executing C<rcv> callbacks or C<psub> |
253 |
blocks. |
254 |
|
255 |
=item SELF, %SELF, @SELF... |
256 |
|
257 |
Due to some quirks in how perl exports variables, it is impossible to |
258 |
just export C<$SELF>, all the symbols called C<SELF> are exported by this |
259 |
module, but only C<$SELF> is currently used. |
260 |
|
261 |
=item snd $port, type => @data |
262 |
|
263 |
=item snd $port, @msg |
264 |
|
265 |
Send the given message to the given port ID, which can identify either |
266 |
a local or a remote port, and can be either a string or soemthignt hat |
267 |
stringifies a sa port ID (such as a port object :). |
268 |
|
269 |
While the message can be about anything, it is highly recommended to use a |
270 |
string as first element (a portid, or some word that indicates a request |
271 |
type etc.). |
272 |
|
273 |
The message data effectively becomes read-only after a call to this |
274 |
function: modifying any argument is not allowed and can cause many |
275 |
problems. |
276 |
|
277 |
The type of data you can transfer depends on the transport protocol: when |
278 |
JSON is used, then only strings, numbers and arrays and hashes consisting |
279 |
of those are allowed (no objects). When Storable is used, then anything |
280 |
that Storable can serialise and deserialise is allowed, and for the local |
281 |
node, anything can be passed. |
282 |
|
283 |
=item $local_port = port |
284 |
|
285 |
Create a new local port object that can be used either as a pattern |
286 |
matching port ("full port") or a single-callback port ("miniport"), |
287 |
depending on how C<rcv> callbacks are bound to the object. |
288 |
|
289 |
=item $port = port { my @msg = @_; $finished } |
290 |
|
291 |
Creates a "miniport", that is, a very lightweight port without any pattern |
292 |
matching behind it, and returns its ID. Semantically the same as creating |
293 |
a port and calling C<rcv $port, $callback> on it. |
294 |
|
295 |
The block will be called for every message received on the port. When the |
296 |
callback returns a true value its job is considered "done" and the port |
297 |
will be destroyed. Otherwise it will stay alive. |
298 |
|
299 |
The message will be passed as-is, no extra argument (i.e. no port id) will |
300 |
be passed to the callback. |
301 |
|
302 |
If you need the local port id in the callback, this works nicely: |
303 |
|
304 |
my $port; $port = port { |
305 |
snd $otherport, reply => $port; |
306 |
}; |
307 |
|
308 |
=cut |
309 |
|
310 |
sub rcv($@); |
311 |
|
312 |
sub port(;&) { |
313 |
my $id = "$UNIQ." . $ID++; |
314 |
my $port = "$NODE#$id"; |
315 |
|
316 |
if (@_) { |
317 |
rcv $port, shift; |
318 |
} else { |
319 |
$PORT{$id} = sub { }; # nop |
320 |
} |
321 |
|
322 |
$port |
323 |
} |
324 |
|
325 |
=item reg $port, $name |
326 |
|
327 |
=item reg $name |
328 |
|
329 |
Registers the given port (or C<$SELF><<< if missing) under the name |
330 |
C<$name>. If the name already exists it is replaced. |
331 |
|
332 |
A port can only be registered under one well known name. |
333 |
|
334 |
A port automatically becomes unregistered when it is killed. |
335 |
|
336 |
=cut |
337 |
|
338 |
sub reg(@) { |
339 |
my $port = @_ > 1 ? shift : $SELF || Carp::croak 'reg: called with one argument only, but $SELF not set,'; |
340 |
|
341 |
$REG{$_[0]} = $port; |
342 |
} |
343 |
|
344 |
=item rcv $port, $callback->(@msg) |
345 |
|
346 |
Replaces the callback on the specified miniport (after converting it to |
347 |
one if required). |
348 |
|
349 |
=item rcv $port, tagstring => $callback->(@msg), ... |
350 |
|
351 |
=item rcv $port, $smartmatch => $callback->(@msg), ... |
352 |
|
353 |
=item rcv $port, [$smartmatch...] => $callback->(@msg), ... |
354 |
|
355 |
Register callbacks to be called on matching messages on the given full |
356 |
port (after converting it to one if required) and return the port. |
357 |
|
358 |
The callback has to return a true value when its work is done, after |
359 |
which is will be removed, or a false value in which case it will stay |
360 |
registered. |
361 |
|
362 |
The global C<$SELF> (exported by this module) contains C<$port> while |
363 |
executing the callback. |
364 |
|
365 |
Runtime errors during callback execution will result in the port being |
366 |
C<kil>ed. |
367 |
|
368 |
If the match is an array reference, then it will be matched against the |
369 |
first elements of the message, otherwise only the first element is being |
370 |
matched. |
371 |
|
372 |
Any element in the match that is specified as C<_any_> (a function |
373 |
exported by this module) matches any single element of the message. |
374 |
|
375 |
While not required, it is highly recommended that the first matching |
376 |
element is a string identifying the message. The one-string-only match is |
377 |
also the most efficient match (by far). |
378 |
|
379 |
Example: create a port and bind receivers on it in one go. |
380 |
|
381 |
my $port = rcv port, |
382 |
msg1 => sub { ...; 0 }, |
383 |
msg2 => sub { ...; 0 }, |
384 |
; |
385 |
|
386 |
Example: create a port, bind receivers and send it in a message elsewhere |
387 |
in one go: |
388 |
|
389 |
snd $otherport, reply => |
390 |
rcv port, |
391 |
msg1 => sub { ...; 0 }, |
392 |
... |
393 |
; |
394 |
|
395 |
=cut |
396 |
|
397 |
sub rcv($@) { |
398 |
my $port = shift; |
399 |
my ($noderef, $portid) = split /#/, $port, 2; |
400 |
|
401 |
($NODE{$noderef} || add_node $noderef) == $NODE{""} |
402 |
or Carp::croak "$port: rcv can only be called on local ports, caught"; |
403 |
|
404 |
if (@_ == 1) { |
405 |
my $cb = shift; |
406 |
delete $PORT_DATA{$portid}; |
407 |
$PORT{$portid} = sub { |
408 |
local $SELF = $port; |
409 |
eval { |
410 |
&$cb |
411 |
and kil $port; |
412 |
}; |
413 |
_self_die if $@; |
414 |
}; |
415 |
} else { |
416 |
my $self = $PORT_DATA{$portid} ||= do { |
417 |
my $self = bless { |
418 |
id => $port, |
419 |
}, "AnyEvent::MP::Port"; |
420 |
|
421 |
$PORT{$portid} = sub { |
422 |
local $SELF = $port; |
423 |
|
424 |
eval { |
425 |
for (@{ $self->{rc0}{$_[0]} }) { |
426 |
$_ && &{$_->[0]} |
427 |
&& undef $_; |
428 |
} |
429 |
|
430 |
for (@{ $self->{rcv}{$_[0]} }) { |
431 |
$_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1] |
432 |
&& &{$_->[0]} |
433 |
&& undef $_; |
434 |
} |
435 |
|
436 |
for (@{ $self->{any} }) { |
437 |
$_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1] |
438 |
&& &{$_->[0]} |
439 |
&& undef $_; |
440 |
} |
441 |
}; |
442 |
_self_die if $@; |
443 |
}; |
444 |
|
445 |
$self |
446 |
}; |
447 |
|
448 |
"AnyEvent::MP::Port" eq ref $self |
449 |
or Carp::croak "$port: rcv can only be called on message matching ports, caught"; |
450 |
|
451 |
while (@_) { |
452 |
my ($match, $cb) = splice @_, 0, 2; |
453 |
|
454 |
if (!ref $match) { |
455 |
push @{ $self->{rc0}{$match} }, [$cb]; |
456 |
} elsif (("ARRAY" eq ref $match && !ref $match->[0])) { |
457 |
my ($type, @match) = @$match; |
458 |
@match |
459 |
? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match] |
460 |
: push @{ $self->{rc0}{$match->[0]} }, [$cb]; |
461 |
} else { |
462 |
push @{ $self->{any} }, [$cb, $match]; |
463 |
} |
464 |
} |
465 |
} |
466 |
|
467 |
$port |
468 |
} |
469 |
|
470 |
=item $closure = psub { BLOCK } |
471 |
|
472 |
Remembers C<$SELF> and creates a closure out of the BLOCK. When the |
473 |
closure is executed, sets up the environment in the same way as in C<rcv> |
474 |
callbacks, i.e. runtime errors will cause the port to get C<kil>ed. |
475 |
|
476 |
This is useful when you register callbacks from C<rcv> callbacks: |
477 |
|
478 |
rcv delayed_reply => sub { |
479 |
my ($delay, @reply) = @_; |
480 |
my $timer = AE::timer $delay, 0, psub { |
481 |
snd @reply, $SELF; |
482 |
}; |
483 |
}; |
484 |
|
485 |
=cut |
486 |
|
487 |
sub psub(&) { |
488 |
my $cb = shift; |
489 |
|
490 |
my $port = $SELF |
491 |
or Carp::croak "psub can only be called from within rcv or psub callbacks, not"; |
492 |
|
493 |
sub { |
494 |
local $SELF = $port; |
495 |
|
496 |
if (wantarray) { |
497 |
my @res = eval { &$cb }; |
498 |
_self_die if $@; |
499 |
@res |
500 |
} else { |
501 |
my $res = eval { &$cb }; |
502 |
_self_die if $@; |
503 |
$res |
504 |
} |
505 |
} |
506 |
} |
507 |
|
508 |
=item $guard = mon $port, $cb->(@reason) |
509 |
|
510 |
=item $guard = mon $port, $rcvport |
511 |
|
512 |
=item $guard = mon $port |
513 |
|
514 |
=item $guard = mon $port, $rcvport, @msg |
515 |
|
516 |
Monitor the given port and do something when the port is killed or |
517 |
messages to it were lost, and optionally return a guard that can be used |
518 |
to stop monitoring again. |
519 |
|
520 |
C<mon> effectively guarantees that, in the absence of hardware failures, |
521 |
that after starting the monitor, either all messages sent to the port |
522 |
will arrive, or the monitoring action will be invoked after possible |
523 |
message loss has been detected. No messages will be lost "in between" |
524 |
(after the first lost message no further messages will be received by the |
525 |
port). After the monitoring action was invoked, further messages might get |
526 |
delivered again. |
527 |
|
528 |
In the first form (callback), the callback is simply called with any |
529 |
number of C<@reason> elements (no @reason means that the port was deleted |
530 |
"normally"). Note also that I<< the callback B<must> never die >>, so use |
531 |
C<eval> if unsure. |
532 |
|
533 |
In the second form (another port given), the other port (C<$rcvport>) |
534 |
will be C<kil>'ed with C<@reason>, iff a @reason was specified, i.e. on |
535 |
"normal" kils nothing happens, while under all other conditions, the other |
536 |
port is killed with the same reason. |
537 |
|
538 |
The third form (kill self) is the same as the second form, except that |
539 |
C<$rvport> defaults to C<$SELF>. |
540 |
|
541 |
In the last form (message), a message of the form C<@msg, @reason> will be |
542 |
C<snd>. |
543 |
|
544 |
As a rule of thumb, monitoring requests should always monitor a port from |
545 |
a local port (or callback). The reason is that kill messages might get |
546 |
lost, just like any other message. Another less obvious reason is that |
547 |
even monitoring requests can get lost (for exmaple, when the connection |
548 |
to the other node goes down permanently). When monitoring a port locally |
549 |
these problems do not exist. |
550 |
|
551 |
Example: call a given callback when C<$port> is killed. |
552 |
|
553 |
mon $port, sub { warn "port died because of <@_>\n" }; |
554 |
|
555 |
Example: kill ourselves when C<$port> is killed abnormally. |
556 |
|
557 |
mon $port; |
558 |
|
559 |
Example: send us a restart message when another C<$port> is killed. |
560 |
|
561 |
mon $port, $self => "restart"; |
562 |
|
563 |
=cut |
564 |
|
565 |
sub mon { |
566 |
my ($noderef, $port) = split /#/, shift, 2; |
567 |
|
568 |
my $node = $NODE{$noderef} || add_node $noderef; |
569 |
|
570 |
my $cb = @_ ? shift : $SELF || Carp::croak 'mon: called with one argument only, but $SELF not set,'; |
571 |
|
572 |
unless (ref $cb) { |
573 |
if (@_) { |
574 |
# send a kill info message |
575 |
my (@msg) = ($cb, @_); |
576 |
$cb = sub { snd @msg, @_ }; |
577 |
} else { |
578 |
# simply kill other port |
579 |
my $port = $cb; |
580 |
$cb = sub { kil $port, @_ if @_ }; |
581 |
} |
582 |
} |
583 |
|
584 |
$node->monitor ($port, $cb); |
585 |
|
586 |
defined wantarray |
587 |
and AnyEvent::Util::guard { $node->unmonitor ($port, $cb) } |
588 |
} |
589 |
|
590 |
=item $guard = mon_guard $port, $ref, $ref... |
591 |
|
592 |
Monitors the given C<$port> and keeps the passed references. When the port |
593 |
is killed, the references will be freed. |
594 |
|
595 |
Optionally returns a guard that will stop the monitoring. |
596 |
|
597 |
This function is useful when you create e.g. timers or other watchers and |
598 |
want to free them when the port gets killed: |
599 |
|
600 |
$port->rcv (start => sub { |
601 |
my $timer; $timer = mon_guard $port, AE::timer 1, 1, sub { |
602 |
undef $timer if 0.9 < rand; |
603 |
}); |
604 |
}); |
605 |
|
606 |
=cut |
607 |
|
608 |
sub mon_guard { |
609 |
my ($port, @refs) = @_; |
610 |
|
611 |
#TODO: mon-less form? |
612 |
|
613 |
mon $port, sub { 0 && @refs } |
614 |
} |
615 |
|
616 |
=item kil $port[, @reason] |
617 |
|
618 |
Kill the specified port with the given C<@reason>. |
619 |
|
620 |
If no C<@reason> is specified, then the port is killed "normally" (linked |
621 |
ports will not be kileld, or even notified). |
622 |
|
623 |
Otherwise, linked ports get killed with the same reason (second form of |
624 |
C<mon>, see below). |
625 |
|
626 |
Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks |
627 |
will be reported as reason C<< die => $@ >>. |
628 |
|
629 |
Transport/communication errors are reported as C<< transport_error => |
630 |
$message >>. |
631 |
|
632 |
=cut |
633 |
|
634 |
=item $port = spawn $node, $initfunc[, @initdata] |
635 |
|
636 |
Creates a port on the node C<$node> (which can also be a port ID, in which |
637 |
case it's the node where that port resides). |
638 |
|
639 |
The port ID of the newly created port is return immediately, and it is |
640 |
permissible to immediately start sending messages or monitor the port. |
641 |
|
642 |
After the port has been created, the init function is |
643 |
called. This function must be a fully-qualified function name |
644 |
(e.g. C<MyApp::Chat::Server::init>). To specify a function in the main |
645 |
program, use C<::name>. |
646 |
|
647 |
If the function doesn't exist, then the node tries to C<require> |
648 |
the package, then the package above the package and so on (e.g. |
649 |
C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function |
650 |
exists or it runs out of package names. |
651 |
|
652 |
The init function is then called with the newly-created port as context |
653 |
object (C<$SELF>) and the C<@initdata> values as arguments. |
654 |
|
655 |
A common idiom is to pass your own port, monitor the spawned port, and |
656 |
in the init function, monitor the original port. This two-way monitoring |
657 |
ensures that both ports get cleaned up when there is a problem. |
658 |
|
659 |
Example: spawn a chat server port on C<$othernode>. |
660 |
|
661 |
# this node, executed from within a port context: |
662 |
my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF; |
663 |
mon $server; |
664 |
|
665 |
# init function on C<$othernode> |
666 |
sub connect { |
667 |
my ($srcport) = @_; |
668 |
|
669 |
mon $srcport; |
670 |
|
671 |
rcv $SELF, sub { |
672 |
... |
673 |
}; |
674 |
} |
675 |
|
676 |
=cut |
677 |
|
678 |
sub _spawn { |
679 |
my $port = shift; |
680 |
my $init = shift; |
681 |
|
682 |
local $SELF = "$NODE#$port"; |
683 |
eval { |
684 |
&{ load_func $init } |
685 |
}; |
686 |
_self_die if $@; |
687 |
} |
688 |
|
689 |
sub spawn(@) { |
690 |
my ($noderef, undef) = split /#/, shift, 2; |
691 |
|
692 |
my $id = "$RUNIQ." . $ID++; |
693 |
|
694 |
$_[0] =~ /::/ |
695 |
or Carp::croak "spawn init function must be a fully-qualified name, caught"; |
696 |
|
697 |
($NODE{$noderef} || add_node $noderef) |
698 |
->send (["", "AnyEvent::MP::_spawn" => $id, @_]); |
699 |
|
700 |
"$noderef#$id" |
701 |
} |
702 |
|
703 |
=back |
704 |
|
705 |
=head1 NODE MESSAGES |
706 |
|
707 |
Nodes understand the following messages sent to them. Many of them take |
708 |
arguments called C<@reply>, which will simply be used to compose a reply |
709 |
message - C<$reply[0]> is the port to reply to, C<$reply[1]> the type and |
710 |
the remaining arguments are simply the message data. |
711 |
|
712 |
While other messages exist, they are not public and subject to change. |
713 |
|
714 |
=over 4 |
715 |
|
716 |
=cut |
717 |
|
718 |
=item lookup => $name, @reply |
719 |
|
720 |
Replies with the port ID of the specified well-known port, or C<undef>. |
721 |
|
722 |
=item devnull => ... |
723 |
|
724 |
Generic data sink/CPU heat conversion. |
725 |
|
726 |
=item relay => $port, @msg |
727 |
|
728 |
Simply forwards the message to the given port. |
729 |
|
730 |
=item eval => $string[ @reply] |
731 |
|
732 |
Evaluates the given string. If C<@reply> is given, then a message of the |
733 |
form C<@reply, $@, @evalres> is sent. |
734 |
|
735 |
Example: crash another node. |
736 |
|
737 |
snd $othernode, eval => "exit"; |
738 |
|
739 |
=item time => @reply |
740 |
|
741 |
Replies the the current node time to C<@reply>. |
742 |
|
743 |
Example: tell the current node to send the current time to C<$myport> in a |
744 |
C<timereply> message. |
745 |
|
746 |
snd $NODE, time => $myport, timereply => 1, 2; |
747 |
# => snd $myport, timereply => 1, 2, <time> |
748 |
|
749 |
=back |
750 |
|
751 |
=head1 AnyEvent::MP vs. Distributed Erlang |
752 |
|
753 |
AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node |
754 |
== aemp node, Erlang process == aemp port), so many of the documents and |
755 |
programming techniques employed by Erlang apply to AnyEvent::MP. Here is a |
756 |
sample: |
757 |
|
758 |
http://www.Erlang.se/doc/programming_rules.shtml |
759 |
http://Erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4 |
760 |
http://Erlang.org/download/Erlang-book-part1.pdf # chapters 5 and 6 |
761 |
http://Erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5 |
762 |
|
763 |
Despite the similarities, there are also some important differences: |
764 |
|
765 |
=over 4 |
766 |
|
767 |
=item * Node references contain the recipe on how to contact them. |
768 |
|
769 |
Erlang relies on special naming and DNS to work everywhere in the |
770 |
same way. AEMP relies on each node knowing it's own address(es), with |
771 |
convenience functionality. |
772 |
|
773 |
This means that AEMP requires a less tightly controlled environment at the |
774 |
cost of longer node references and a slightly higher management overhead. |
775 |
|
776 |
=item * Erlang uses processes and a mailbox, AEMP does not queue. |
777 |
|
778 |
Erlang uses processes that selctively receive messages, and therefore |
779 |
needs a queue. AEMP is event based, queuing messages would serve no useful |
780 |
purpose. |
781 |
|
782 |
(But see L<Coro::MP> for a more Erlang-like process model on top of AEMP). |
783 |
|
784 |
=item * Erlang sends are synchronous, AEMP sends are asynchronous. |
785 |
|
786 |
Sending messages in Erlang is synchronous and blocks the process. AEMP |
787 |
sends are immediate, connection establishment is handled in the |
788 |
background. |
789 |
|
790 |
=item * Erlang can silently lose messages, AEMP cannot. |
791 |
|
792 |
Erlang makes few guarantees on messages delivery - messages can get lost |
793 |
without any of the processes realising it (i.e. you send messages a, b, |
794 |
and c, and the other side only receives messages a and c). |
795 |
|
796 |
AEMP guarantees correct ordering, and the guarantee that there are no |
797 |
holes in the message sequence. |
798 |
|
799 |
=item * In Erlang, processes can be declared dead and later be found to be |
800 |
alive. |
801 |
|
802 |
In Erlang it can happen that a monitored process is declared dead and |
803 |
linked processes get killed, but later it turns out that the process is |
804 |
still alive - and can receive messages. |
805 |
|
806 |
In AEMP, when port monitoring detects a port as dead, then that port will |
807 |
eventually be killed - it cannot happen that a node detects a port as dead |
808 |
and then later sends messages to it, finding it is still alive. |
809 |
|
810 |
=item * Erlang can send messages to the wrong port, AEMP does not. |
811 |
|
812 |
In Erlang it is quite possible that a node that restarts reuses a process |
813 |
ID known to other nodes for a completely different process, causing |
814 |
messages destined for that process to end up in an unrelated process. |
815 |
|
816 |
AEMP never reuses port IDs, so old messages or old port IDs floating |
817 |
around in the network will not be sent to an unrelated port. |
818 |
|
819 |
=item * Erlang uses unprotected connections, AEMP uses secure |
820 |
authentication and can use TLS. |
821 |
|
822 |
AEMP can use a proven protocol - SSL/TLS - to protect connections and |
823 |
securely authenticate nodes. |
824 |
|
825 |
=item * The AEMP protocol is optimised for both text-based and binary |
826 |
communications. |
827 |
|
828 |
The AEMP protocol, unlike the Erlang protocol, supports both |
829 |
language-independent text-only protocols (good for debugging) and binary, |
830 |
language-specific serialisers (e.g. Storable). |
831 |
|
832 |
It has also been carefully designed to be implementable in other languages |
833 |
with a minimum of work while gracefully degrading fucntionality to make the |
834 |
protocol simple. |
835 |
|
836 |
=item * AEMP has more flexible monitoring options than Erlang. |
837 |
|
838 |
In Erlang, you can chose to receive I<all> exit signals as messages |
839 |
or I<none>, there is no in-between, so monitoring single processes is |
840 |
difficult to implement. Monitoring in AEMP is more flexible than in |
841 |
Erlang, as one can choose between automatic kill, exit message or callback |
842 |
on a per-process basis. |
843 |
|
844 |
=item * Erlang tries to hide remote/local connections, AEMP does not. |
845 |
|
846 |
Monitoring in Erlang is not an indicator of process death/crashes, |
847 |
as linking is (except linking is unreliable in Erlang). |
848 |
|
849 |
In AEMP, you don't "look up" registered port names or send to named ports |
850 |
that might or might not be persistent. Instead, you normally spawn a port |
851 |
on the remote node. The init function monitors the you, and you monitor |
852 |
the remote port. Since both monitors are local to the node, they are much |
853 |
more reliable. |
854 |
|
855 |
This also saves round-trips and avoids sending messages to the wrong port |
856 |
(hard to do in Erlang). |
857 |
|
858 |
=back |
859 |
|
860 |
=head1 SEE ALSO |
861 |
|
862 |
L<AnyEvent>. |
863 |
|
864 |
=head1 AUTHOR |
865 |
|
866 |
Marc Lehmann <schmorp@schmorp.de> |
867 |
http://home.schmorp.de/ |
868 |
|
869 |
=cut |
870 |
|
871 |
1 |
872 |
|