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 1.23 by root, Tue Aug 4 18:46:16 2009 UTC vs.
Revision 1.57 by root, Sat Aug 15 04:34:34 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 ports, the simple way
28 my $simple_port = port { my @msg = @_; 0 };
16 29
17 rcv $port, smartmatch => $cb->($port, @msg); 30 # creating/using ports, tagged message matching
18 31 my $port = port;
19 # examples:
20 rcv $port2, ping => sub { snd $_[0], "pong"; 0 }; 32 rcv $port, ping => sub { snd $_[0], "pong"; 0 };
21 rcv $port1, pong => sub { warn "pong received\n" }; 33 rcv $port, pong => sub { warn "pong received\n"; 0 };
22 snd $port2, ping => $port1;
23 34
24 # more, smarter, matches (_any_ is exported by this module) 35 # create a port on another node
25 rcv $port, [child_died => $pid] => sub { ... 36 my $port = spawn $node, $initfunc, @initdata;
26 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46 AnyEvent::MP::Intro - outdated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.
27 51
28=head1 DESCRIPTION 52=head1 DESCRIPTION
29 53
30This module (-family) implements a simple message passing framework. 54This module (-family) implements a simple message passing framework.
31 55
35For an introduction to this module family, see the L<AnyEvent::MP::Intro> 59For an introduction to this module family, see the L<AnyEvent::MP::Intro>
36manual page. 60manual page.
37 61
38At the moment, this module family is severly broken and underdocumented, 62At the moment, this module family is severly broken and underdocumented,
39so do not use. This was uploaded mainly to reserve the CPAN namespace - 63so do not use. This was uploaded mainly to reserve the CPAN namespace -
40stay tuned! The basic API should be finished, however. 64stay tuned!
41 65
42=head1 CONCEPTS 66=head1 CONCEPTS
43 67
44=over 4 68=over 4
45 69
46=item port 70=item port
47 71
48A port is something you can send messages to with the C<snd> function, and 72A port is something you can send messages to (with the C<snd> function).
49you can register C<rcv> handlers with. All C<rcv> handlers will receive 73
74Ports allow you to register C<rcv> handlers that can match all or just
50messages they match, messages will not be queued. 75some messages. Messages will not be queued.
51 76
52=item port id - C<noderef#portname> 77=item port id - C<noderef#portname>
53 78
54A port id is always the noderef, a hash-mark (C<#>) as separator, followed 79A port ID is the concatenation of a noderef, a hash-mark (C<#>) as
55by a port name (a printable string of unspecified format). 80separator, and a port name (a printable string of unspecified format). An
81exception is the the node port, whose ID is identical to its node
82reference.
56 83
57=item node 84=item node
58 85
59A node is a single process containing at least one port - the node 86A node is a single process containing at least one port - the node port,
60port. You can send messages to node ports to let them create new ports, 87which provides nodes to manage each other remotely, and to create new
61among other things. 88ports.
62 89
63Initially, nodes are either private (single-process only) or hidden 90Nodes are either private (single-process only), slaves (connected to a
64(connected to a master node only). Only when they epxlicitly "become 91master node only) or public nodes (connectable from unrelated nodes).
65public" can you send them messages from unrelated other nodes.
66 92
67=item noderef - C<host:port,host:port...>, C<id@noderef>, C<id> 93=item noderef - C<host:port,host:port...>, C<id@noderef>, C<id>
68 94
69A noderef is a string that either uniquely identifies a given node (for 95A node reference is a string that either simply identifies the node (for
70private and hidden nodes), or contains a recipe on how to reach a given 96private and slave nodes), or contains a recipe on how to reach a given
71node (for public nodes). 97node (for public nodes).
72 98
99This recipe is simply a comma-separated list of C<address:port> pairs (for
100TCP/IP, other protocols might look different).
101
102Node references come in two flavours: resolved (containing only numerical
103addresses) or unresolved (where hostnames are used instead of addresses).
104
105Before using an unresolved node reference in a message you first have to
106resolve it.
107
73=back 108=back
74 109
75=head1 VARIABLES/FUNCTIONS 110=head1 VARIABLES/FUNCTIONS
76 111
77=over 4 112=over 4
78 113
79=cut 114=cut
80 115
81package AnyEvent::MP; 116package AnyEvent::MP;
82 117
83use AnyEvent::MP::Base; 118use AnyEvent::MP::Kernel;
84 119
85use common::sense; 120use common::sense;
86 121
87use Carp (); 122use Carp ();
88 123
89use AE (); 124use AE ();
90 125
91use base "Exporter"; 126use base "Exporter";
92 127
93our $VERSION = '0.02'; 128our $VERSION = $AnyEvent::MP::Kernel::VERSION;
129
94our @EXPORT = qw( 130our @EXPORT = qw(
95 NODE $NODE *SELF node_of _any_ 131 NODE $NODE *SELF node_of _any_
96 become_slave become_public 132 resolve_node initialise_node
97 snd rcv mon kil reg psub 133 snd rcv mon kil reg psub spawn
98 port 134 port
99); 135);
100 136
101our $SELF; 137our $SELF;
102 138
106 kil $SELF, die => $msg; 142 kil $SELF, die => $msg;
107} 143}
108 144
109=item $thisnode = NODE / $NODE 145=item $thisnode = NODE / $NODE
110 146
111The C<NODE> function returns, and the C<$NODE> variable contains 147The C<NODE> function returns, and the C<$NODE> variable contains the
112the noderef of the local node. The value is initialised by a call 148noderef of the local node. The value is initialised by a call to
113to C<become_public> or C<become_slave>, after which all local port 149C<initialise_node>.
114identifiers become invalid.
115 150
116=item $noderef = node_of $portid 151=item $noderef = node_of $port
117 152
118Extracts and returns the noderef from a portid or a noderef. 153Extracts and returns the noderef from a port ID or a noderef.
154
155=item initialise_node $noderef, $seednode, $seednode...
156
157=item initialise_node "slave/", $master, $master...
158
159Before a node can talk to other nodes on the network it has to initialise
160itself - the minimum a node needs to know is it's own name, and optionally
161it should know the noderefs of some other nodes in the network.
162
163This function initialises a node - it must be called exactly once (or
164never) before calling other AnyEvent::MP functions.
165
166All arguments (optionally except for the first) are noderefs, which can be
167either resolved or unresolved.
168
169The first argument will be looked up in the configuration database first
170(if it is C<undef> then the current nodename will be used instead) to find
171the relevant configuration profile (see L<aemp>). If none is found then
172the default configuration is used. The configuration supplies additional
173seed/master nodes and can override the actual noderef.
174
175There are two types of networked nodes, public nodes and slave nodes:
176
177=over 4
178
179=item public nodes
180
181For public nodes, C<$noderef> (supplied either directly to
182C<initialise_node> or indirectly via a profile or the nodename) must be a
183noderef (possibly unresolved, in which case it will be resolved).
184
185After resolving, the node will bind itself on all endpoints and try to
186connect to all additional C<$seednodes> that are specified. Seednodes are
187optional and can be used to quickly bootstrap the node into an existing
188network.
189
190=item slave nodes
191
192When the C<$noderef> (either as given or overriden by the config file)
193is the special string C<slave/>, then the node will become a slave
194node. Slave nodes cannot be contacted from outside and will route most of
195their traffic to the master node that they attach to.
196
197At least one additional noderef is required (either by specifying it
198directly or because it is part of the configuration profile): The node
199will try to connect to all of them and will become a slave attached to the
200first node it can successfully connect to.
201
202Note that slave nodes cannot change their name, and consequently, their
203master, so if the master goes down, the slave node will not function well
204anymore until it can re-establish conenciton to its master. This makes
205slave nodes unsuitable for long-term nodes or fault-tolerant networks.
206
207=back
208
209This function will block until all nodes have been resolved and, for slave
210nodes, until it has successfully established a connection to a master
211server.
212
213All the seednodes will also be specially marked to automatically retry
214connecting to them infinitely.
215
216Example: become a public node listening on the guessed noderef, or the one
217specified via C<aemp> for the current node. This should be the most common
218form of invocation for "daemon"-type nodes.
219
220 initialise_node;
221
222Example: become a slave node to any of the the seednodes specified via
223C<aemp>. This form is often used for commandline clients.
224
225 initialise_node "slave/";
226
227Example: become a slave node to any of the specified master servers. This
228form is also often used for commandline clients.
229
230 initialise_node "slave/", "master1", "192.168.13.17", "mp.example.net";
231
232Example: become a public node, and try to contact some well-known master
233servers to become part of the network.
234
235 initialise_node undef, "master1", "master2";
236
237Example: become a public node listening on port C<4041>.
238
239 initialise_node 4041;
240
241Example: become a public node, only visible on localhost port 4044.
242
243 initialise_node "localhost:4044";
244
245=item $cv = resolve_node $noderef
246
247Takes an unresolved node reference that may contain hostnames and
248abbreviated IDs, resolves all of them and returns a resolved node
249reference.
250
251In addition to C<address:port> pairs allowed in resolved noderefs, the
252following forms are supported:
253
254=over 4
255
256=item the empty string
257
258An empty-string component gets resolved as if the default port (4040) was
259specified.
260
261=item naked port numbers (e.g. C<1234>)
262
263These are resolved by prepending the local nodename and a colon, to be
264further resolved.
265
266=item hostnames (e.g. C<localhost:1234>, C<localhost>)
267
268These are resolved by using AnyEvent::DNS to resolve them, optionally
269looking up SRV records for the C<aemp=4040> port, if no port was
270specified.
271
272=back
119 273
120=item $SELF 274=item $SELF
121 275
122Contains the current port id while executing C<rcv> callbacks or C<psub> 276Contains the current port id while executing C<rcv> callbacks or C<psub>
123blocks. 277blocks.
126 280
127Due to some quirks in how perl exports variables, it is impossible to 281Due to some quirks in how perl exports variables, it is impossible to
128just export C<$SELF>, all the symbols called C<SELF> are exported by this 282just export C<$SELF>, all the symbols called C<SELF> are exported by this
129module, but only C<$SELF> is currently used. 283module, but only C<$SELF> is currently used.
130 284
131=item snd $portid, type => @data 285=item snd $port, type => @data
132 286
133=item snd $portid, @msg 287=item snd $port, @msg
134 288
135Send the given message to the given port ID, which can identify either 289Send the given message to the given port ID, which can identify either
136a local or a remote port, and can be either a string or soemthignt hat 290a local or a remote port, and must be a port ID.
137stringifies a sa port ID (such as a port object :).
138 291
139While the message can be about anything, it is highly recommended to use a 292While the message can be about anything, it is highly recommended to use a
140string as first element (a portid, or some word that indicates a request 293string as first element (a port ID, or some word that indicates a request
141type etc.). 294type etc.).
142 295
143The message data effectively becomes read-only after a call to this 296The message data effectively becomes read-only after a call to this
144function: modifying any argument is not allowed and can cause many 297function: modifying any argument is not allowed and can cause many
145problems. 298problems.
148JSON is used, then only strings, numbers and arrays and hashes consisting 301JSON is used, then only strings, numbers and arrays and hashes consisting
149of those are allowed (no objects). When Storable is used, then anything 302of those are allowed (no objects). When Storable is used, then anything
150that Storable can serialise and deserialise is allowed, and for the local 303that Storable can serialise and deserialise is allowed, and for the local
151node, anything can be passed. 304node, anything can be passed.
152 305
153=item kil $portid[, @reason] 306=item $local_port = port
154 307
155Kill the specified port with the given C<@reason>. 308Create a new local port object and returns its port ID. Initially it has
309no callbacks set and will throw an error when it receives messages.
156 310
157If no C<@reason> is specified, then the port is killed "normally" (linked 311=item $local_port = port { my @msg = @_ }
158ports will not be kileld, or even notified).
159 312
160Otherwise, linked ports get killed with the same reason (second form of 313Creates a new local port, and returns its ID. Semantically the same as
161C<mon>, see below). 314creating a port and calling C<rcv $port, $callback> on it.
162 315
163Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks 316The block will be called for every message received on the port, with the
164will be reported as reason C<< die => $@ >>. 317global variable C<$SELF> set to the port ID. Runtime errors will cause the
318port to be C<kil>ed. The message will be passed as-is, no extra argument
319(i.e. no port ID) will be passed to the callback.
165 320
166Transport/communication errors are reported as C<< transport_error => 321If you want to stop/destroy the port, simply C<kil> it:
167$message >>.
168 322
169=item $guard = mon $portid, $cb->(@reason) 323 my $port = port {
170 324 my @msg = @_;
171=item $guard = mon $portid, $otherport 325 ...
172 326 kil $SELF;
173=item $guard = mon $portid, $otherport, @msg 327 };
174
175Monitor the given port and do something when the port is killed.
176
177In the first form, the callback is simply called with any number
178of C<@reason> elements (no @reason means that the port was deleted
179"normally"). Note also that I<< the callback B<must> never die >>, so use
180C<eval> if unsure.
181
182In the second form, the other port will be C<kil>'ed with C<@reason>, iff
183a @reason was specified, i.e. on "normal" kils nothing happens, while
184under all other conditions, the other port is killed with the same reason.
185
186In the last form, a message of the form C<@msg, @reason> will be C<snd>.
187
188Example: call a given callback when C<$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, $self => "restart";
199 328
200=cut 329=cut
201 330
202sub mon { 331sub rcv($@);
203 my ($noderef, $port, $cb) = ((split /#/, shift, 2), shift);
204 332
205 my $node = $NODE{$noderef} || add_node $noderef; 333sub _kilme {
206 334 die "received message on port without callback";
207 #TODO: ports must not be references
208 if (!ref $cb or "AnyEvent::MP::Port" eq ref $cb) {
209 if (@_) {
210 # send a kill info message
211 my (@msg) = ($cb, @_);
212 $cb = sub { snd @msg, @_ };
213 } else {
214 # simply kill other port
215 my $port = $cb;
216 $cb = sub { kil $port, @_ if @_ };
217 }
218 }
219
220 $node->monitor ($port, $cb);
221
222 defined wantarray
223 and AnyEvent::Util::guard { $node->unmonitor ($port, $cb) }
224} 335}
225
226=item $guard = mon_guard $port, $ref, $ref...
227
228Monitors the given C<$port> and keeps the passed references. When the port
229is killed, the references will be freed.
230
231Optionally returns a guard that will stop the monitoring.
232
233This function is useful when you create e.g. timers or other watchers and
234want to free them when the port gets killed:
235
236 $port->rcv (start => sub {
237 my $timer; $timer = mon_guard $port, AE::timer 1, 1, sub {
238 undef $timer if 0.9 < rand;
239 });
240 });
241
242=cut
243
244sub mon_guard {
245 my ($port, @refs) = @_;
246
247 mon $port, sub { 0 && @refs }
248}
249
250=item $local_port = port
251
252Create a new local port object that supports message matching.
253
254=item $portid = port { my @msg = @_; $finished }
255
256Creates a "mini port", that is, a very lightweight port without any
257pattern matching behind it, and returns its ID.
258
259The block will be called for every message received on the port. When the
260callback returns a true value its job is considered "done" and the port
261will be destroyed. Otherwise it will stay alive.
262
263The message will be passed as-is, no extra argument (i.e. no port id) will
264be passed to the callback.
265
266If you need the local port id in the callback, this works nicely:
267
268 my $port; $port = miniport {
269 snd $otherport, reply => $port;
270 };
271
272=cut
273 336
274sub port(;&) { 337sub port(;&) {
275 my $id = "$UNIQ." . $ID++; 338 my $id = "$UNIQ." . $ID++;
276 my $port = "$NODE#$id"; 339 my $port = "$NODE#$id";
277 340
278 if (@_) { 341 rcv $port, shift || \&_kilme;
279 my $cb = shift;
280 $PORT{$id} = sub {
281 local $SELF = $port;
282 eval {
283 &$cb
284 and kil $id;
285 };
286 _self_die if $@;
287 };
288 } else {
289 my $self = bless {
290 id => "$NODE#$id",
291 }, "AnyEvent::MP::Port";
292
293 $PORT_DATA{$id} = $self;
294 $PORT{$id} = sub {
295 local $SELF = $port;
296
297 eval {
298 for (@{ $self->{rc0}{$_[0]} }) {
299 $_ && &{$_->[0]}
300 && undef $_;
301 }
302
303 for (@{ $self->{rcv}{$_[0]} }) {
304 $_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1]
305 && &{$_->[0]}
306 && undef $_;
307 }
308
309 for (@{ $self->{any} }) {
310 $_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1]
311 && &{$_->[0]}
312 && undef $_;
313 }
314 };
315 _self_die if $@;
316 };
317 }
318 342
319 $port 343 $port
320} 344}
321 345
322=item reg $portid, $name 346=item rcv $local_port, $callback->(@msg)
323 347
324Registers the given port under the name C<$name>. If the name already 348Replaces the default callback on the specified port. There is no way to
325exists it is replaced. 349remove the default callback: use C<sub { }> to disable it, or better
350C<kil> the port when it is no longer needed.
326 351
327A port can only be registered under one well known name. 352The global C<$SELF> (exported by this module) contains C<$port> while
353executing the callback. Runtime errors during callback execution will
354result in the port being C<kil>ed.
328 355
329A port automatically becomes unregistered when it is killed. 356The default callback received all messages not matched by a more specific
357C<tag> match.
358
359=item rcv $local_port, tag => $callback->(@msg_without_tag), ...
360
361Register (or replace) callbacks to be called on messages starting with the
362given tag on the given port (and return the port), or unregister it (when
363C<$callback> is C<$undef> or missing). There can only be one callback
364registered for each tag.
365
366The original message will be passed to the callback, after the first
367element (the tag) has been removed. The callback will use the same
368environment as the default callback (see above).
369
370Example: create a port and bind receivers on it in one go.
371
372 my $port = rcv port,
373 msg1 => sub { ... },
374 msg2 => sub { ... },
375 ;
376
377Example: create a port, bind receivers and send it in a message elsewhere
378in one go:
379
380 snd $otherport, reply =>
381 rcv port,
382 msg1 => sub { ... },
383 ...
384 ;
385
386Example: temporarily register a rcv callback for a tag matching some port
387(e.g. for a rpc reply) and unregister it after a message was received.
388
389 rcv $port, $otherport => sub {
390 my @reply = @_;
391
392 rcv $SELF, $otherport;
393 };
330 394
331=cut 395=cut
332 396
333sub reg(@) {
334 my ($portid, $name) = @_;
335
336 $REG{$name} = $portid;
337}
338
339=item rcv $portid, tagstring => $callback->(@msg), ...
340
341=item rcv $portid, $smartmatch => $callback->(@msg), ...
342
343=item rcv $portid, [$smartmatch...] => $callback->(@msg), ...
344
345Register callbacks to be called on matching messages on the given port.
346
347The callback has to return a true value when its work is done, after
348which is will be removed, or a false value in which case it will stay
349registered.
350
351The global C<$SELF> (exported by this module) contains C<$portid> while
352executing the callback.
353
354Runtime errors wdurign callback execution will result in the port being
355C<kil>ed.
356
357If the match is an array reference, then it will be matched against the
358first elements of the message, otherwise only the first element is being
359matched.
360
361Any element in the match that is specified as C<_any_> (a function
362exported by this module) matches any single element of the message.
363
364While not required, it is highly recommended that the first matching
365element is a string identifying the message. The one-string-only match is
366also the most efficient match (by far).
367
368=cut
369
370sub rcv($@) { 397sub rcv($@) {
398 my $port = shift;
371 my ($noderef, $port) = split /#/, shift, 2; 399 my ($noderef, $portid) = split /#/, $port, 2;
372 400
373 ($NODE{$noderef} || add_node $noderef) == $NODE{""} 401 ($NODE{$noderef} || add_node $noderef) == $NODE{""}
374 or Carp::croak "$noderef#$port: rcv can only be called on local ports, caught"; 402 or Carp::croak "$port: rcv can only be called on local ports, caught";
375
376 my $self = $PORT_DATA{$port}
377 or Carp::croak "$noderef#$port: rcv can only be called on message matching ports, caught";
378
379 "AnyEvent::MP::Port" eq ref $self
380 or Carp::croak "$noderef#$port: rcv can only be called on message matching ports, caught";
381 403
382 while (@_) { 404 while (@_) {
383 my ($match, $cb) = splice @_, 0, 2;
384
385 if (!ref $match) { 405 if (ref $_[0]) {
386 push @{ $self->{rc0}{$match} }, [$cb]; 406 if (my $self = $PORT_DATA{$portid}) {
387 } elsif (("ARRAY" eq ref $match && !ref $match->[0])) { 407 "AnyEvent::MP::Port" eq ref $self
388 my ($type, @match) = @$match; 408 or Carp::croak "$port: rcv can only be called on message matching ports, caught";
389 @match 409
390 ? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match] 410 $self->[2] = shift;
391 : push @{ $self->{rc0}{$match->[0]} }, [$cb];
392 } else { 411 } else {
393 push @{ $self->{any} }, [$cb, $match]; 412 my $cb = shift;
413 $PORT{$portid} = sub {
414 local $SELF = $port;
415 eval { &$cb }; _self_die if $@;
416 };
417 }
418 } elsif (defined $_[0]) {
419 my $self = $PORT_DATA{$portid} ||= do {
420 my $self = bless [$PORT{$port} || sub { }, { }, $port], "AnyEvent::MP::Port";
421
422 $PORT{$portid} = sub {
423 local $SELF = $port;
424
425 if (my $cb = $self->[1]{$_[0]}) {
426 shift;
427 eval { &$cb }; _self_die if $@;
428 } else {
429 &{ $self->[0] };
430 }
431 };
432
433 $self
434 };
435
436 "AnyEvent::MP::Port" eq ref $self
437 or Carp::croak "$port: rcv can only be called on message matching ports, caught";
438
439 my ($tag, $cb) = splice @_, 0, 2;
440
441 if (defined $cb) {
442 $self->[1]{$tag} = $cb;
443 } else {
444 delete $self->[1]{$tag};
445 }
394 } 446 }
395 } 447 }
448
449 $port
396} 450}
397 451
398=item $closure = psub { BLOCK } 452=item $closure = psub { BLOCK }
399 453
400Remembers C<$SELF> and creates a closure out of the BLOCK. When the 454Remembers C<$SELF> and creates a closure out of the BLOCK. When the
431 $res 485 $res
432 } 486 }
433 } 487 }
434} 488}
435 489
490=item $guard = mon $port, $cb->(@reason)
491
492=item $guard = mon $port, $rcvport
493
494=item $guard = mon $port
495
496=item $guard = mon $port, $rcvport, @msg
497
498Monitor the given port and do something when the port is killed or
499messages to it were lost, and optionally return a guard that can be used
500to stop monitoring again.
501
502C<mon> effectively guarantees that, in the absence of hardware failures,
503that after starting the monitor, either all messages sent to the port
504will arrive, or the monitoring action will be invoked after possible
505message loss has been detected. No messages will be lost "in between"
506(after the first lost message no further messages will be received by the
507port). After the monitoring action was invoked, further messages might get
508delivered again.
509
510In the first form (callback), the callback is simply called with any
511number of C<@reason> elements (no @reason means that the port was deleted
512"normally"). Note also that I<< the callback B<must> never die >>, so use
513C<eval> if unsure.
514
515In the second form (another port given), the other port (C<$rcvport>)
516will be C<kil>'ed with C<@reason>, iff a @reason was specified, i.e. on
517"normal" kils nothing happens, while under all other conditions, the other
518port is killed with the same reason.
519
520The third form (kill self) is the same as the second form, except that
521C<$rvport> defaults to C<$SELF>.
522
523In the last form (message), a message of the form C<@msg, @reason> will be
524C<snd>.
525
526As a rule of thumb, monitoring requests should always monitor a port from
527a local port (or callback). The reason is that kill messages might get
528lost, just like any other message. Another less obvious reason is that
529even monitoring requests can get lost (for exmaple, when the connection
530to the other node goes down permanently). When monitoring a port locally
531these problems do not exist.
532
533Example: call a given callback when C<$port> is killed.
534
535 mon $port, sub { warn "port died because of <@_>\n" };
536
537Example: kill ourselves when C<$port> is killed abnormally.
538
539 mon $port;
540
541Example: send us a restart message when another C<$port> is killed.
542
543 mon $port, $self => "restart";
544
545=cut
546
547sub mon {
548 my ($noderef, $port) = split /#/, shift, 2;
549
550 my $node = $NODE{$noderef} || add_node $noderef;
551
552 my $cb = @_ ? shift : $SELF || Carp::croak 'mon: called with one argument only, but $SELF not set,';
553
554 unless (ref $cb) {
555 if (@_) {
556 # send a kill info message
557 my (@msg) = ($cb, @_);
558 $cb = sub { snd @msg, @_ };
559 } else {
560 # simply kill other port
561 my $port = $cb;
562 $cb = sub { kil $port, @_ if @_ };
563 }
564 }
565
566 $node->monitor ($port, $cb);
567
568 defined wantarray
569 and AnyEvent::Util::guard { $node->unmonitor ($port, $cb) }
570}
571
572=item $guard = mon_guard $port, $ref, $ref...
573
574Monitors the given C<$port> and keeps the passed references. When the port
575is killed, the references will be freed.
576
577Optionally returns a guard that will stop the monitoring.
578
579This function is useful when you create e.g. timers or other watchers and
580want to free them when the port gets killed:
581
582 $port->rcv (start => sub {
583 my $timer; $timer = mon_guard $port, AE::timer 1, 1, sub {
584 undef $timer if 0.9 < rand;
585 });
586 });
587
588=cut
589
590sub mon_guard {
591 my ($port, @refs) = @_;
592
593 #TODO: mon-less form?
594
595 mon $port, sub { 0 && @refs }
596}
597
598=item kil $port[, @reason]
599
600Kill the specified port with the given C<@reason>.
601
602If no C<@reason> is specified, then the port is killed "normally" (linked
603ports will not be kileld, or even notified).
604
605Otherwise, linked ports get killed with the same reason (second form of
606C<mon>, see below).
607
608Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks
609will be reported as reason C<< die => $@ >>.
610
611Transport/communication errors are reported as C<< transport_error =>
612$message >>.
613
614=cut
615
616=item $port = spawn $node, $initfunc[, @initdata]
617
618Creates a port on the node C<$node> (which can also be a port ID, in which
619case it's the node where that port resides).
620
621The port ID of the newly created port is return immediately, and it is
622permissible to immediately start sending messages or monitor the port.
623
624After the port has been created, the init function is
625called. This function must be a fully-qualified function name
626(e.g. C<MyApp::Chat::Server::init>). To specify a function in the main
627program, use C<::name>.
628
629If the function doesn't exist, then the node tries to C<require>
630the package, then the package above the package and so on (e.g.
631C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function
632exists or it runs out of package names.
633
634The init function is then called with the newly-created port as context
635object (C<$SELF>) and the C<@initdata> values as arguments.
636
637A common idiom is to pass your own port, monitor the spawned port, and
638in the init function, monitor the original port. This two-way monitoring
639ensures that both ports get cleaned up when there is a problem.
640
641Example: spawn a chat server port on C<$othernode>.
642
643 # this node, executed from within a port context:
644 my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF;
645 mon $server;
646
647 # init function on C<$othernode>
648 sub connect {
649 my ($srcport) = @_;
650
651 mon $srcport;
652
653 rcv $SELF, sub {
654 ...
655 };
656 }
657
658=cut
659
660sub _spawn {
661 my $port = shift;
662 my $init = shift;
663
664 local $SELF = "$NODE#$port";
665 eval {
666 &{ load_func $init }
667 };
668 _self_die if $@;
669}
670
671sub spawn(@) {
672 my ($noderef, undef) = split /#/, shift, 2;
673
674 my $id = "$RUNIQ." . $ID++;
675
676 $_[0] =~ /::/
677 or Carp::croak "spawn init function must be a fully-qualified name, caught";
678
679 snd_to_func $noderef, "AnyEvent::MP::_spawn" => $id, @_;
680
681 "$noderef#$id"
682}
683
436=back 684=back
437 685
438=head1 FUNCTIONS FOR NODES 686=head1 AnyEvent::MP vs. Distributed Erlang
687
688AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node
689== aemp node, Erlang process == aemp port), so many of the documents and
690programming techniques employed by Erlang apply to AnyEvent::MP. Here is a
691sample:
692
693 http://www.Erlang.se/doc/programming_rules.shtml
694 http://Erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4
695 http://Erlang.org/download/Erlang-book-part1.pdf # chapters 5 and 6
696 http://Erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5
697
698Despite the similarities, there are also some important differences:
439 699
440=over 4 700=over 4
441 701
442=item become_public endpoint... 702=item * Node references contain the recipe on how to contact them.
443 703
444Tells the node to become a public node, i.e. reachable from other nodes. 704Erlang relies on special naming and DNS to work everywhere in the
705same way. AEMP relies on each node knowing it's own address(es), with
706convenience functionality.
445 707
446If no arguments are given, or the first argument is C<undef>, then 708This means that AEMP requires a less tightly controlled environment at the
447AnyEvent::MP tries to bind on port C<4040> on all IP addresses that the 709cost of longer node references and a slightly higher management overhead.
448local nodename resolves to.
449 710
450Otherwise the first argument must be an array-reference with transport 711=item * Erlang has a "remote ports are like local ports" philosophy, AEMP
451endpoints ("ip:port", "hostname:port") or port numbers (in which case the 712uses "local ports are like remote ports".
452local nodename is used as hostname). The endpoints are all resolved and
453will become the node reference.
454 713
455=cut 714The failure modes for local ports are quite different (runtime errors
715only) then for remote ports - when a local port dies, you I<know> it dies,
716when a connection to another node dies, you know nothing about the other
717port.
718
719Erlang pretends remote ports are as reliable as local ports, even when
720they are not.
721
722AEMP encourages a "treat remote ports differently" philosophy, with local
723ports being the special case/exception, where transport errors cannot
724occur.
725
726=item * Erlang uses processes and a mailbox, AEMP does not queue.
727
728Erlang uses processes that selectively receive messages, and therefore
729needs a queue. AEMP is event based, queuing messages would serve no
730useful purpose. For the same reason the pattern-matching abilities of
731AnyEvent::MP are more limited, as there is little need to be able to
732filter messages without dequeing them.
733
734(But see L<Coro::MP> for a more Erlang-like process model on top of AEMP).
735
736=item * Erlang sends are synchronous, AEMP sends are asynchronous.
737
738Sending messages in Erlang is synchronous and blocks the process (and
739so does not need a queue that can overflow). AEMP sends are immediate,
740connection establishment is handled in the background.
741
742=item * Erlang suffers from silent message loss, AEMP does not.
743
744Erlang makes few guarantees on messages delivery - messages can get lost
745without any of the processes realising it (i.e. you send messages a, b,
746and c, and the other side only receives messages a and c).
747
748AEMP guarantees correct ordering, and the guarantee that there are no
749holes in the message sequence.
750
751=item * In Erlang, processes can be declared dead and later be found to be
752alive.
753
754In Erlang it can happen that a monitored process is declared dead and
755linked processes get killed, but later it turns out that the process is
756still alive - and can receive messages.
757
758In AEMP, when port monitoring detects a port as dead, then that port will
759eventually be killed - it cannot happen that a node detects a port as dead
760and then later sends messages to it, finding it is still alive.
761
762=item * Erlang can send messages to the wrong port, AEMP does not.
763
764In Erlang it is quite likely that a node that restarts reuses a process ID
765known to other nodes for a completely different process, causing messages
766destined for that process to end up in an unrelated process.
767
768AEMP never reuses port IDs, so old messages or old port IDs floating
769around in the network will not be sent to an unrelated port.
770
771=item * Erlang uses unprotected connections, AEMP uses secure
772authentication and can use TLS.
773
774AEMP can use a proven protocol - SSL/TLS - to protect connections and
775securely authenticate nodes.
776
777=item * The AEMP protocol is optimised for both text-based and binary
778communications.
779
780The AEMP protocol, unlike the Erlang protocol, supports both
781language-independent text-only protocols (good for debugging) and binary,
782language-specific serialisers (e.g. Storable).
783
784It has also been carefully designed to be implementable in other languages
785with a minimum of work while gracefully degrading fucntionality to make the
786protocol simple.
787
788=item * AEMP has more flexible monitoring options than Erlang.
789
790In Erlang, you can chose to receive I<all> exit signals as messages
791or I<none>, there is no in-between, so monitoring single processes is
792difficult to implement. Monitoring in AEMP is more flexible than in
793Erlang, as one can choose between automatic kill, exit message or callback
794on a per-process basis.
795
796=item * Erlang tries to hide remote/local connections, AEMP does not.
797
798Monitoring in Erlang is not an indicator of process death/crashes,
799as linking is (except linking is unreliable in Erlang).
800
801In AEMP, you don't "look up" registered port names or send to named ports
802that might or might not be persistent. Instead, you normally spawn a port
803on the remote node. The init function monitors the you, and you monitor
804the remote port. Since both monitors are local to the node, they are much
805more reliable.
806
807This also saves round-trips and avoids sending messages to the wrong port
808(hard to do in Erlang).
456 809
457=back 810=back
458 811
459=head1 NODE MESSAGES 812=head1 RATIONALE
460
461Nodes understand the following messages sent to them. Many of them take
462arguments called C<@reply>, which will simply be used to compose a reply
463message - C<$reply[0]> is the port to reply to, C<$reply[1]> the type and
464the remaining arguments are simply the message data.
465 813
466=over 4 814=over 4
467 815
468=cut 816=item Why strings for ports and noderefs, why not objects?
469 817
470=item lookup => $name, @reply 818We considered "objects", but found that the actual number of methods
819thatc an be called are very low. Since port IDs and noderefs travel over
820the network frequently, the serialising/deserialising would add lots of
821overhead, as well as having to keep a proxy object.
471 822
472Replies with the port ID of the specified well-known port, or C<undef>. 823Strings can easily be printed, easily serialised etc. and need no special
824procedures to be "valid".
473 825
474=item devnull => ... 826And a a miniport consists of a single closure stored in a global hash - it
827can't become much cheaper.
475 828
476Generic data sink/CPU heat conversion. 829=item Why favour JSON, why not real serialising format such as Storable?
477 830
478=item relay => $port, @msg 831In fact, any AnyEvent::MP node will happily accept Storable as framing
832format, but currently there is no way to make a node use Storable by
833default.
479 834
480Simply forwards the message to the given port. 835The default framing protocol is JSON because a) JSON::XS is many times
836faster for small messages and b) most importantly, after years of
837experience we found that object serialisation is causing more problems
838than it gains: Just like function calls, objects simply do not travel
839easily over the network, mostly because they will always be a copy, so you
840always have to re-think your design.
481 841
482=item eval => $string[ @reply] 842Keeping your messages simple, concentrating on data structures rather than
483 843objects, will keep your messages clean, tidy and efficient.
484Evaluates the given string. If C<@reply> is given, then a message of the
485form C<@reply, $@, @evalres> is sent.
486
487Example: crash another node.
488
489 snd $othernode, eval => "exit";
490
491=item time => @reply
492
493Replies the the current node time to C<@reply>.
494
495Example: tell the current node to send the current time to C<$myport> in a
496C<timereply> message.
497
498 snd $NODE, time => $myport, timereply => 1, 2;
499 # => snd $myport, timereply => 1, 2, <time>
500 844
501=back 845=back
502 846
503=head1 SEE ALSO 847=head1 SEE ALSO
504 848

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines