1 | =head1 NAME |
1 | =head1 NAME |
2 | |
2 | |
3 | AnyEvent::MP - multi-processing/message-passing framework |
3 | AnyEvent::MP - erlang-style multi-processing/message-passing framework |
4 | |
4 | |
5 | =head1 SYNOPSIS |
5 | =head1 SYNOPSIS |
6 | |
6 | |
7 | use AnyEvent::MP; |
7 | use AnyEvent::MP; |
8 | |
8 | |
9 | $NODE # contains this node's noderef |
9 | $NODE # contains this node's node ID |
10 | NODE # returns this node's noderef |
10 | NODE # returns this node's node ID |
11 | NODE $port # returns the noderef of the port |
|
|
12 | |
11 | |
13 | $SELF # receiving/own port id in rcv callbacks |
12 | $SELF # receiving/own port id in rcv callbacks |
14 | |
13 | |
|
|
14 | # initialise the node so it can send/receive messages |
|
|
15 | configure; |
|
|
16 | |
15 | # ports are message endpoints |
17 | # ports are message destinations |
16 | |
18 | |
17 | # sending messages |
19 | # sending messages |
18 | snd $port, type => data...; |
20 | snd $port, type => data...; |
19 | snd $port, @msg; |
21 | snd $port, @msg; |
20 | snd @msg_with_first_element_being_a_port; |
22 | snd @msg_with_first_element_being_a_port; |
21 | |
23 | |
22 | # miniports |
24 | # creating/using ports, the simple way |
23 | my $miniport = port { my @msg = @_; 0 }; |
25 | my $simple_port = port { my @msg = @_ }; |
24 | |
26 | |
25 | # full ports |
27 | # creating/using ports, tagged message matching |
26 | my $port = port; |
28 | my $port = port; |
27 | rcv $port, smartmatch => $cb->(@msg); |
|
|
28 | rcv $port, ping => sub { snd $_[0], "pong"; 0 }; |
29 | rcv $port, ping => sub { snd $_[0], "pong" }; |
29 | rcv $port, pong => sub { warn "pong received\n"; 0 }; |
30 | rcv $port, pong => sub { warn "pong received\n" }; |
30 | |
31 | |
31 | # remote ports |
32 | # create a port on another node |
32 | my $port = spawn $node, $initfunc, @initdata; |
33 | my $port = spawn $node, $initfunc, @initdata; |
33 | |
34 | |
34 | # more, smarter, matches (_any_ is exported by this module) |
35 | # destroy a port again |
35 | rcv $port, [child_died => $pid] => sub { ... |
36 | kil $port; # "normal" kill |
36 | rcv $port, [_any_, _any_, 3] => sub { .. $_[2] is 3 |
37 | kil $port, my_error => "everything is broken"; # error kill |
37 | |
38 | |
38 | # monitoring |
39 | # monitoring |
39 | mon $port, $cb->(@msg) # callback is invoked on death |
40 | mon $port, $cb->(@msg) # callback is invoked on death |
40 | mon $port, $otherport # kill otherport on abnormal death |
41 | mon $port, $localport # kill localport on abnormal death |
41 | mon $port, $otherport, @msg # send message on death |
42 | mon $port, $localport, @msg # send message on death |
|
|
43 | |
|
|
44 | # temporarily execute code in port context |
|
|
45 | peval $port, sub { die "kill the port!" }; |
|
|
46 | |
|
|
47 | # execute callbacks in $SELF port context |
|
|
48 | my $timer = AE::timer 1, 0, psub { |
|
|
49 | die "kill the port, delayed"; |
|
|
50 | }; |
|
|
51 | |
|
|
52 | =head1 CURRENT STATUS |
|
|
53 | |
|
|
54 | bin/aemp - stable. |
|
|
55 | AnyEvent::MP - stable API, should work. |
|
|
56 | AnyEvent::MP::Intro - explains most concepts. |
|
|
57 | AnyEvent::MP::Kernel - mostly stable API. |
|
|
58 | AnyEvent::MP::Global - stable API. |
42 | |
59 | |
43 | =head1 DESCRIPTION |
60 | =head1 DESCRIPTION |
44 | |
61 | |
45 | This module (-family) implements a simple message passing framework. |
62 | This module (-family) implements a simple message passing framework. |
46 | |
63 | |
47 | Despite its simplicity, you can securely message other processes running |
64 | Despite its simplicity, you can securely message other processes running |
48 | on the same or other hosts. |
65 | on the same or other hosts, and you can supervise entities remotely. |
49 | |
66 | |
50 | For an introduction to this module family, see the L<AnyEvent::MP::Intro> |
67 | For an introduction to this module family, see the L<AnyEvent::MP::Intro> |
51 | manual page. |
68 | manual page and the examples under F<eg/>. |
52 | |
|
|
53 | At the moment, this module family is severly broken and underdocumented, |
|
|
54 | so do not use. This was uploaded mainly to reserve the CPAN namespace - |
|
|
55 | stay tuned! The basic API should be finished, however. |
|
|
56 | |
69 | |
57 | =head1 CONCEPTS |
70 | =head1 CONCEPTS |
58 | |
71 | |
59 | =over 4 |
72 | =over 4 |
60 | |
73 | |
61 | =item port |
74 | =item port |
62 | |
75 | |
63 | A port is something you can send messages to (with the C<snd> function). |
76 | Not to be confused with a TCP port, a "port" is something you can send |
|
|
77 | messages to (with the C<snd> function). |
64 | |
78 | |
65 | Some ports allow you to register C<rcv> handlers that can match specific |
79 | Ports allow you to register C<rcv> handlers that can match all or just |
66 | messages. All C<rcv> handlers will receive messages they match, messages |
80 | some messages. Messages send to ports will not be queued, regardless of |
67 | will not be queued. |
81 | anything was listening for them or not. |
68 | |
82 | |
|
|
83 | Ports are represented by (printable) strings called "port IDs". |
|
|
84 | |
69 | =item port id - C<noderef#portname> |
85 | =item port ID - C<nodeid#portname> |
70 | |
86 | |
71 | A port id is normaly the concatenation of a noderef, a hash-mark (C<#>) as |
87 | A port ID is the concatenation of a node ID, a hash-mark (C<#>) |
72 | separator, and a port name (a printable string of unspecified format). An |
88 | as separator, and a port name (a printable string of unspecified |
73 | exception is the the node port, whose ID is identical to its node |
89 | format created by AnyEvent::MP). |
74 | reference. |
|
|
75 | |
90 | |
76 | =item node |
91 | =item node |
77 | |
92 | |
78 | A node is a single process containing at least one port - the node |
93 | A node is a single process containing at least one port - the node port, |
79 | port. You can send messages to node ports to find existing ports or to |
94 | which enables nodes to manage each other remotely, and to create new |
80 | create new ports, among other things. |
95 | ports. |
81 | |
96 | |
82 | Nodes are either private (single-process only), slaves (connected to a |
97 | Nodes are either public (have one or more listening ports) or private |
83 | master node only) or public nodes (connectable from unrelated nodes). |
98 | (no listening ports). Private nodes cannot talk to other private nodes |
|
|
99 | currently, but all nodes can talk to public nodes. |
84 | |
100 | |
85 | =item noderef - C<host:port,host:port...>, C<id@noderef>, C<id> |
101 | Nodes is represented by (printable) strings called "node IDs". |
86 | |
102 | |
87 | A node reference is a string that either simply identifies the node (for |
103 | =item node ID - C<[A-Za-z0-9_\-.:]*> |
88 | private and slave nodes), or contains a recipe on how to reach a given |
|
|
89 | node (for public nodes). |
|
|
90 | |
104 | |
91 | This recipe is simply a comma-separated list of C<address:port> pairs (for |
105 | A node ID is a string that uniquely identifies the node within a |
92 | TCP/IP, other protocols might look different). |
106 | network. Depending on the configuration used, node IDs can look like a |
|
|
107 | hostname, a hostname and a port, or a random string. AnyEvent::MP itself |
|
|
108 | doesn't interpret node IDs in any way except to uniquely identify a node. |
93 | |
109 | |
94 | Node references come in two flavours: resolved (containing only numerical |
110 | =item binds - C<ip:port> |
95 | addresses) or unresolved (where hostnames are used instead of addresses). |
|
|
96 | |
111 | |
97 | Before using an unresolved node reference in a message you first have to |
112 | Nodes can only talk to each other by creating some kind of connection to |
98 | resolve it. |
113 | each other. To do this, nodes should listen on one or more local transport |
|
|
114 | endpoints - binds. |
|
|
115 | |
|
|
116 | Currently, only standard C<ip:port> specifications can be used, which |
|
|
117 | specify TCP ports to listen on. So a bind is basically just a tcp socket |
|
|
118 | in listening mode thta accepts conenctions form other nodes. |
|
|
119 | |
|
|
120 | =item seed nodes |
|
|
121 | |
|
|
122 | When a node starts, it knows nothing about the network it is in - it |
|
|
123 | needs to connect to at least one other node that is already in the |
|
|
124 | network. These other nodes are called "seed nodes". |
|
|
125 | |
|
|
126 | Seed nodes themselves are not special - they are seed nodes only because |
|
|
127 | some other node I<uses> them as such, but any node can be used as seed |
|
|
128 | node for other nodes, and eahc node cna use a different set of seed nodes. |
|
|
129 | |
|
|
130 | In addition to discovering the network, seed nodes are also used to |
|
|
131 | maintain the network - all nodes using the same seed node form are part of |
|
|
132 | the same network. If a network is split into multiple subnets because e.g. |
|
|
133 | the network link between the parts goes down, then using the same seed |
|
|
134 | nodes for all nodes ensures that eventually the subnets get merged again. |
|
|
135 | |
|
|
136 | Seed nodes are expected to be long-running, and at least one seed node |
|
|
137 | should always be available. They should also be relatively responsive - a |
|
|
138 | seed node that blocks for long periods will slow down everybody else. |
|
|
139 | |
|
|
140 | For small networks, it's best if every node uses the same set of seed |
|
|
141 | nodes. For large networks, it can be useful to specify "regional" seed |
|
|
142 | nodes for most nodes in an area, and use all seed nodes as seed nodes for |
|
|
143 | each other. What's important is that all seed nodes connections form a |
|
|
144 | complete graph, so that the network cannot split into separate subnets |
|
|
145 | forever. |
|
|
146 | |
|
|
147 | Seed nodes are represented by seed IDs. |
|
|
148 | |
|
|
149 | =item seed IDs - C<host:port> |
|
|
150 | |
|
|
151 | Seed IDs are transport endpoint(s) (usually a hostname/IP address and a |
|
|
152 | TCP port) of nodes that should be used as seed nodes. |
|
|
153 | |
|
|
154 | =item global nodes |
|
|
155 | |
|
|
156 | An AEMP network needs a discovery service - nodes need to know how to |
|
|
157 | connect to other nodes they only know by name. In addition, AEMP offers a |
|
|
158 | distributed "group database", which maps group names to a list of strings |
|
|
159 | - for example, to register worker ports. |
|
|
160 | |
|
|
161 | A network needs at least one global node to work, and allows every node to |
|
|
162 | be a global node. |
|
|
163 | |
|
|
164 | Any node that loads the L<AnyEvent::MP::Global> module becomes a global |
|
|
165 | node and tries to keep connections to all other nodes. So while it can |
|
|
166 | make sense to make every node "global" in small networks, it usually makes |
|
|
167 | sense to only make seed nodes into global nodes in large networks (nodes |
|
|
168 | keep connections to seed nodes and global nodes, so makign them the same |
|
|
169 | reduces overhead). |
99 | |
170 | |
100 | =back |
171 | =back |
101 | |
172 | |
102 | =head1 VARIABLES/FUNCTIONS |
173 | =head1 VARIABLES/FUNCTIONS |
103 | |
174 | |
… | |
… | |
105 | |
176 | |
106 | =cut |
177 | =cut |
107 | |
178 | |
108 | package AnyEvent::MP; |
179 | package AnyEvent::MP; |
109 | |
180 | |
|
|
181 | use AnyEvent::MP::Config (); |
110 | use AnyEvent::MP::Base; |
182 | use AnyEvent::MP::Kernel; |
|
|
183 | use AnyEvent::MP::Kernel qw(%NODE %PORT %PORT_DATA $UNIQ $RUNIQ $ID); |
111 | |
184 | |
112 | use common::sense; |
185 | use common::sense; |
113 | |
186 | |
114 | use Carp (); |
187 | use Carp (); |
115 | |
188 | |
116 | use AE (); |
189 | use AE (); |
|
|
190 | use Guard (); |
117 | |
191 | |
118 | use base "Exporter"; |
192 | use base "Exporter"; |
119 | |
193 | |
120 | our $VERSION = '0.1'; |
194 | our $VERSION = $AnyEvent::MP::Config::VERSION; |
|
|
195 | |
121 | our @EXPORT = qw( |
196 | our @EXPORT = qw( |
122 | NODE $NODE *SELF node_of _any_ |
197 | NODE $NODE *SELF node_of after |
123 | resolve_node initialise_node |
198 | configure |
124 | snd rcv mon kil reg psub spawn |
199 | snd rcv mon mon_guard kil psub peval spawn cal |
125 | port |
200 | port |
|
|
201 | db_set db_del db_reg |
|
|
202 | db_mon db_family db_keys db_values |
126 | ); |
203 | ); |
127 | |
204 | |
128 | our $SELF; |
205 | our $SELF; |
129 | |
206 | |
130 | sub _self_die() { |
207 | sub _self_die() { |
… | |
… | |
133 | kil $SELF, die => $msg; |
210 | kil $SELF, die => $msg; |
134 | } |
211 | } |
135 | |
212 | |
136 | =item $thisnode = NODE / $NODE |
213 | =item $thisnode = NODE / $NODE |
137 | |
214 | |
138 | The C<NODE> function returns, and the C<$NODE> variable contains |
215 | The C<NODE> function returns, and the C<$NODE> variable contains, the node |
139 | the noderef of the local node. The value is initialised by a call |
216 | ID of the node running in the current process. This value is initialised by |
140 | to C<become_public> or C<become_slave>, after which all local port |
217 | a call to C<configure>. |
141 | identifiers become invalid. |
|
|
142 | |
218 | |
143 | =item $noderef = node_of $port |
219 | =item $nodeid = node_of $port |
144 | |
220 | |
145 | Extracts and returns the noderef from a portid or a noderef. |
221 | Extracts and returns the node ID from a port ID or a node ID. |
146 | |
222 | |
147 | =item initialise_node $noderef, $seednode, $seednode... |
223 | =item configure $profile, key => value... |
148 | |
224 | |
149 | =item initialise_node "slave/", $master, $master... |
225 | =item configure key => value... |
150 | |
226 | |
151 | Before a node can talk to other nodes on the network it has to initialise |
227 | Before a node can talk to other nodes on the network (i.e. enter |
152 | itself - the minimum a node needs to know is it's own name, and optionally |
228 | "distributed mode") it has to configure itself - the minimum a node needs |
153 | it should know the noderefs of some other nodes in the network. |
229 | to know is its own name, and optionally it should know the addresses of |
|
|
230 | some other nodes in the network to discover other nodes. |
154 | |
231 | |
155 | This function initialises a node - it must be called exactly once (or |
232 | This function configures a node - it must be called exactly once (or |
156 | never) before calling other AnyEvent::MP functions. |
233 | never) before calling other AnyEvent::MP functions. |
157 | |
234 | |
158 | All arguments are noderefs, which can be either resolved or unresolved. |
235 | The key/value pairs are basically the same ones as documented for the |
159 | |
236 | F<aemp> command line utility (sans the set/del prefix), with these additions: |
160 | There are two types of networked nodes, public nodes and slave nodes: |
|
|
161 | |
237 | |
162 | =over 4 |
238 | =over 4 |
163 | |
239 | |
164 | =item public nodes |
240 | =item norc => $boolean (default false) |
165 | |
241 | |
166 | For public nodes, C<$noderef> must either be a (possibly unresolved) |
242 | If true, then the rc file (e.g. F<~/.perl-anyevent-mp>) will I<not> |
167 | noderef, in which case it will be resolved, or C<undef> (or missing), in |
243 | be consulted - all configuraiton options must be specified in the |
168 | which case the noderef will be guessed. |
244 | C<configure> call. |
169 | |
245 | |
170 | Afterwards, the node will bind itself on all endpoints and try to connect |
246 | =item force => $boolean (default false) |
171 | to all additional C<$seednodes> that are specified. Seednodes are optional |
|
|
172 | and can be used to quickly bootstrap the node into an existing network. |
|
|
173 | |
247 | |
174 | =item slave nodes |
248 | IF true, then the values specified in the C<configure> will take |
|
|
249 | precedence over any values configured via the rc file. The default is for |
|
|
250 | the rc file to override any options specified in the program. |
175 | |
251 | |
176 | When the C<$noderef> is the special string C<slave/>, then the node will |
252 | =item secure => $pass->($nodeid) |
177 | become a slave node. Slave nodes cannot be contacted from outside and will |
|
|
178 | route most of their traffic to the master node that they attach to. |
|
|
179 | |
253 | |
180 | At least one additional noderef is required: The node will try to connect |
254 | In addition to specifying a boolean, you can specify a code reference that |
181 | to all of them and will become a slave attached to the first node it can |
255 | is called for every remote execution attempt - the execution request is |
182 | successfully connect to. |
256 | granted iff the callback returns a true value. |
|
|
257 | |
|
|
258 | See F<semp setsecure> for more info. |
183 | |
259 | |
184 | =back |
260 | =back |
185 | |
261 | |
186 | This function will block until all nodes have been resolved and, for slave |
|
|
187 | nodes, until it has successfully established a connection to a master |
|
|
188 | server. |
|
|
189 | |
|
|
190 | Example: become a public node listening on the default node. |
|
|
191 | |
|
|
192 | initialise_node; |
|
|
193 | |
|
|
194 | Example: become a public node, and try to contact some well-known master |
|
|
195 | servers to become part of the network. |
|
|
196 | |
|
|
197 | initialise_node undef, "master1", "master2"; |
|
|
198 | |
|
|
199 | Example: become a public node listening on port C<4041>. |
|
|
200 | |
|
|
201 | initialise_node 4041; |
|
|
202 | |
|
|
203 | Example: become a public node, only visible on localhost port 4044. |
|
|
204 | |
|
|
205 | initialise_node "locahost:4044"; |
|
|
206 | |
|
|
207 | Example: become a slave node to any of the specified master servers. |
|
|
208 | |
|
|
209 | initialise_node "slave/", "master1", "192.168.13.17", "mp.example.net"; |
|
|
210 | |
|
|
211 | =item $cv = resolve_node $noderef |
|
|
212 | |
|
|
213 | Takes an unresolved node reference that may contain hostnames and |
|
|
214 | abbreviated IDs, resolves all of them and returns a resolved node |
|
|
215 | reference. |
|
|
216 | |
|
|
217 | In addition to C<address:port> pairs allowed in resolved noderefs, the |
|
|
218 | following forms are supported: |
|
|
219 | |
|
|
220 | =over 4 |
262 | =over 4 |
221 | |
263 | |
222 | =item the empty string |
264 | =item step 1, gathering configuration from profiles |
223 | |
265 | |
224 | An empty-string component gets resolved as if the default port (4040) was |
266 | The function first looks up a profile in the aemp configuration (see the |
225 | specified. |
267 | L<aemp> commandline utility). The profile name can be specified via the |
|
|
268 | named C<profile> parameter or can simply be the first parameter). If it is |
|
|
269 | missing, then the nodename (F<uname -n>) will be used as profile name. |
226 | |
270 | |
227 | =item naked port numbers (e.g. C<1234>) |
271 | The profile data is then gathered as follows: |
228 | |
272 | |
229 | These are resolved by prepending the local nodename and a colon, to be |
273 | First, all remaining key => value pairs (all of which are conveniently |
230 | further resolved. |
274 | undocumented at the moment) will be interpreted as configuration |
|
|
275 | data. Then they will be overwritten by any values specified in the global |
|
|
276 | default configuration (see the F<aemp> utility), then the chain of |
|
|
277 | profiles chosen by the profile name (and any C<parent> attributes). |
231 | |
278 | |
232 | =item hostnames (e.g. C<localhost:1234>, C<localhost>) |
279 | That means that the values specified in the profile have highest priority |
|
|
280 | and the values specified directly via C<configure> have lowest priority, |
|
|
281 | and can only be used to specify defaults. |
233 | |
282 | |
234 | These are resolved by using AnyEvent::DNS to resolve them, optionally |
283 | If the profile specifies a node ID, then this will become the node ID of |
235 | looking up SRV records for the C<aemp=4040> port, if no port was |
284 | this process. If not, then the profile name will be used as node ID, with |
236 | specified. |
285 | a unique randoms tring (C</%u>) appended. |
|
|
286 | |
|
|
287 | The node ID can contain some C<%> sequences that are expanded: C<%n> |
|
|
288 | is expanded to the local nodename, C<%u> is replaced by a random |
|
|
289 | strign to make the node unique. For example, the F<aemp> commandline |
|
|
290 | utility uses C<aemp/%n/%u> as nodename, which might expand to |
|
|
291 | C<aemp/cerebro/ZQDGSIkRhEZQDGSIkRhE>. |
|
|
292 | |
|
|
293 | =item step 2, bind listener sockets |
|
|
294 | |
|
|
295 | The next step is to look up the binds in the profile, followed by binding |
|
|
296 | aemp protocol listeners on all binds specified (it is possible and valid |
|
|
297 | to have no binds, meaning that the node cannot be contacted form the |
|
|
298 | outside. This means the node cannot talk to other nodes that also have no |
|
|
299 | binds, but it can still talk to all "normal" nodes). |
|
|
300 | |
|
|
301 | If the profile does not specify a binds list, then a default of C<*> is |
|
|
302 | used, meaning the node will bind on a dynamically-assigned port on every |
|
|
303 | local IP address it finds. |
|
|
304 | |
|
|
305 | =item step 3, connect to seed nodes |
|
|
306 | |
|
|
307 | As the last step, the seed ID list from the profile is passed to the |
|
|
308 | L<AnyEvent::MP::Global> module, which will then use it to keep |
|
|
309 | connectivity with at least one node at any point in time. |
237 | |
310 | |
238 | =back |
311 | =back |
|
|
312 | |
|
|
313 | Example: become a distributed node using the local node name as profile. |
|
|
314 | This should be the most common form of invocation for "daemon"-type nodes. |
|
|
315 | |
|
|
316 | configure |
|
|
317 | |
|
|
318 | Example: become a semi-anonymous node. This form is often used for |
|
|
319 | commandline clients. |
|
|
320 | |
|
|
321 | configure nodeid => "myscript/%n/%u"; |
|
|
322 | |
|
|
323 | Example: configure a node using a profile called seed, which is suitable |
|
|
324 | for a seed node as it binds on all local addresses on a fixed port (4040, |
|
|
325 | customary for aemp). |
|
|
326 | |
|
|
327 | # use the aemp commandline utility |
|
|
328 | # aemp profile seed binds '*:4040' |
|
|
329 | |
|
|
330 | # then use it |
|
|
331 | configure profile => "seed"; |
|
|
332 | |
|
|
333 | # or simply use aemp from the shell again: |
|
|
334 | # aemp run profile seed |
|
|
335 | |
|
|
336 | # or provide a nicer-to-remember nodeid |
|
|
337 | # aemp run profile seed nodeid "$(hostname)" |
239 | |
338 | |
240 | =item $SELF |
339 | =item $SELF |
241 | |
340 | |
242 | Contains the current port id while executing C<rcv> callbacks or C<psub> |
341 | Contains the current port id while executing C<rcv> callbacks or C<psub> |
243 | blocks. |
342 | blocks. |
244 | |
343 | |
245 | =item SELF, %SELF, @SELF... |
344 | =item *SELF, SELF, %SELF, @SELF... |
246 | |
345 | |
247 | Due to some quirks in how perl exports variables, it is impossible to |
346 | Due to some quirks in how perl exports variables, it is impossible to |
248 | just export C<$SELF>, all the symbols called C<SELF> are exported by this |
347 | just export C<$SELF>, all the symbols named C<SELF> are exported by this |
249 | module, but only C<$SELF> is currently used. |
348 | module, but only C<$SELF> is currently used. |
250 | |
349 | |
251 | =item snd $port, type => @data |
350 | =item snd $port, type => @data |
252 | |
351 | |
253 | =item snd $port, @msg |
352 | =item snd $port, @msg |
254 | |
353 | |
255 | Send the given message to the given port ID, which can identify either |
354 | Send the given message to the given port, which can identify either a |
256 | a local or a remote port, and can be either a string or soemthignt hat |
355 | local or a remote port, and must be a port ID. |
257 | stringifies a sa port ID (such as a port object :). |
|
|
258 | |
356 | |
259 | While the message can be about anything, it is highly recommended to use a |
357 | While the message can be almost anything, it is highly recommended to |
260 | string as first element (a portid, or some word that indicates a request |
358 | use a string as first element (a port ID, or some word that indicates a |
261 | type etc.). |
359 | request type etc.) and to consist if only simple perl values (scalars, |
|
|
360 | arrays, hashes) - if you think you need to pass an object, think again. |
262 | |
361 | |
263 | The message data effectively becomes read-only after a call to this |
362 | The message data logically becomes read-only after a call to this |
264 | function: modifying any argument is not allowed and can cause many |
363 | function: modifying any argument (or values referenced by them) is |
265 | problems. |
364 | forbidden, as there can be considerable time between the call to C<snd> |
|
|
365 | and the time the message is actually being serialised - in fact, it might |
|
|
366 | never be copied as within the same process it is simply handed to the |
|
|
367 | receiving port. |
266 | |
368 | |
267 | The type of data you can transfer depends on the transport protocol: when |
369 | The type of data you can transfer depends on the transport protocol: when |
268 | JSON is used, then only strings, numbers and arrays and hashes consisting |
370 | JSON is used, then only strings, numbers and arrays and hashes consisting |
269 | of those are allowed (no objects). When Storable is used, then anything |
371 | of those are allowed (no objects). When Storable is used, then anything |
270 | that Storable can serialise and deserialise is allowed, and for the local |
372 | that Storable can serialise and deserialise is allowed, and for the local |
271 | node, anything can be passed. |
373 | node, anything can be passed. Best rely only on the common denominator of |
|
|
374 | these. |
272 | |
375 | |
273 | =item $local_port = port |
376 | =item $local_port = port |
274 | |
377 | |
275 | Create a new local port object that can be used either as a pattern |
378 | Create a new local port object and returns its port ID. Initially it has |
276 | matching port ("full port") or a single-callback port ("miniport"), |
379 | no callbacks set and will throw an error when it receives messages. |
277 | depending on how C<rcv> callbacks are bound to the object. |
|
|
278 | |
380 | |
279 | =item $port = port { my @msg = @_; $finished } |
381 | =item $local_port = port { my @msg = @_ } |
280 | |
382 | |
281 | Creates a "miniport", that is, a very lightweight port without any pattern |
383 | Creates a new local port, and returns its ID. Semantically the same as |
282 | matching behind it, and returns its ID. Semantically the same as creating |
|
|
283 | a port and calling C<rcv $port, $callback> on it. |
384 | creating a port and calling C<rcv $port, $callback> on it. |
284 | |
385 | |
285 | The block will be called for every message received on the port. When the |
386 | The block will be called for every message received on the port, with the |
286 | callback returns a true value its job is considered "done" and the port |
387 | global variable C<$SELF> set to the port ID. Runtime errors will cause the |
287 | will be destroyed. Otherwise it will stay alive. |
388 | port to be C<kil>ed. The message will be passed as-is, no extra argument |
|
|
389 | (i.e. no port ID) will be passed to the callback. |
288 | |
390 | |
289 | The message will be passed as-is, no extra argument (i.e. no port id) will |
391 | If you want to stop/destroy the port, simply C<kil> it: |
290 | be passed to the callback. |
|
|
291 | |
392 | |
292 | If you need the local port id in the callback, this works nicely: |
393 | my $port = port { |
293 | |
394 | my @msg = @_; |
294 | my $port; $port = port { |
395 | ... |
295 | snd $otherport, reply => $port; |
396 | kil $SELF; |
296 | }; |
397 | }; |
297 | |
398 | |
298 | =cut |
399 | =cut |
299 | |
400 | |
300 | sub rcv($@); |
401 | sub rcv($@); |
301 | |
402 | |
|
|
403 | sub _kilme { |
|
|
404 | die "received message on port without callback"; |
|
|
405 | } |
|
|
406 | |
302 | sub port(;&) { |
407 | sub port(;&) { |
303 | my $id = "$UNIQ." . $ID++; |
408 | my $id = $UNIQ . ++$ID; |
304 | my $port = "$NODE#$id"; |
409 | my $port = "$NODE#$id"; |
305 | |
410 | |
306 | if (@_) { |
411 | rcv $port, shift || \&_kilme; |
307 | rcv $port, shift; |
|
|
308 | } else { |
|
|
309 | $PORT{$id} = sub { }; # nop |
|
|
310 | } |
|
|
311 | |
412 | |
312 | $port |
413 | $port |
313 | } |
414 | } |
314 | |
415 | |
315 | =item reg $port, $name |
|
|
316 | |
|
|
317 | =item reg $name |
|
|
318 | |
|
|
319 | Registers the given port (or C<$SELF><<< if missing) under the name |
|
|
320 | C<$name>. If the name already exists it is replaced. |
|
|
321 | |
|
|
322 | A port can only be registered under one well known name. |
|
|
323 | |
|
|
324 | A port automatically becomes unregistered when it is killed. |
|
|
325 | |
|
|
326 | =cut |
|
|
327 | |
|
|
328 | sub reg(@) { |
|
|
329 | my $port = @_ > 1 ? shift : $SELF || Carp::croak 'reg: called with one argument only, but $SELF not set,'; |
|
|
330 | |
|
|
331 | $REG{$_[0]} = $port; |
|
|
332 | } |
|
|
333 | |
|
|
334 | =item rcv $port, $callback->(@msg) |
416 | =item rcv $local_port, $callback->(@msg) |
335 | |
417 | |
336 | Replaces the callback on the specified miniport (after converting it to |
418 | Replaces the default callback on the specified port. There is no way to |
337 | one if required). |
419 | remove the default callback: use C<sub { }> to disable it, or better |
338 | |
420 | C<kil> the port when it is no longer needed. |
339 | =item rcv $port, tagstring => $callback->(@msg), ... |
|
|
340 | |
|
|
341 | =item rcv $port, $smartmatch => $callback->(@msg), ... |
|
|
342 | |
|
|
343 | =item rcv $port, [$smartmatch...] => $callback->(@msg), ... |
|
|
344 | |
|
|
345 | Register callbacks to be called on matching messages on the given full |
|
|
346 | port (after converting it to one if required) and return the port. |
|
|
347 | |
|
|
348 | The callback has to return a true value when its work is done, after |
|
|
349 | which is will be removed, or a false value in which case it will stay |
|
|
350 | registered. |
|
|
351 | |
421 | |
352 | The global C<$SELF> (exported by this module) contains C<$port> while |
422 | The global C<$SELF> (exported by this module) contains C<$port> while |
353 | executing the callback. |
423 | executing the callback. Runtime errors during callback execution will |
|
|
424 | result in the port being C<kil>ed. |
354 | |
425 | |
355 | Runtime errors during callback execution will result in the port being |
426 | The default callback received all messages not matched by a more specific |
356 | C<kil>ed. |
427 | C<tag> match. |
357 | |
428 | |
358 | If the match is an array reference, then it will be matched against the |
429 | =item rcv $local_port, tag => $callback->(@msg_without_tag), ... |
359 | first elements of the message, otherwise only the first element is being |
|
|
360 | matched. |
|
|
361 | |
430 | |
362 | Any element in the match that is specified as C<_any_> (a function |
431 | Register (or replace) callbacks to be called on messages starting with the |
363 | exported by this module) matches any single element of the message. |
432 | given tag on the given port (and return the port), or unregister it (when |
|
|
433 | C<$callback> is C<$undef> or missing). There can only be one callback |
|
|
434 | registered for each tag. |
364 | |
435 | |
365 | While not required, it is highly recommended that the first matching |
436 | The original message will be passed to the callback, after the first |
366 | element is a string identifying the message. The one-string-only match is |
437 | element (the tag) has been removed. The callback will use the same |
367 | also the most efficient match (by far). |
438 | environment as the default callback (see above). |
368 | |
439 | |
369 | Example: create a port and bind receivers on it in one go. |
440 | Example: create a port and bind receivers on it in one go. |
370 | |
441 | |
371 | my $port = rcv port, |
442 | my $port = rcv port, |
372 | msg1 => sub { ...; 0 }, |
443 | msg1 => sub { ... }, |
373 | msg2 => sub { ...; 0 }, |
444 | msg2 => sub { ... }, |
374 | ; |
445 | ; |
375 | |
446 | |
376 | Example: create a port, bind receivers and send it in a message elsewhere |
447 | Example: create a port, bind receivers and send it in a message elsewhere |
377 | in one go: |
448 | in one go: |
378 | |
449 | |
379 | snd $otherport, reply => |
450 | snd $otherport, reply => |
380 | rcv port, |
451 | rcv port, |
381 | msg1 => sub { ...; 0 }, |
452 | msg1 => sub { ... }, |
382 | ... |
453 | ... |
383 | ; |
454 | ; |
384 | |
455 | |
|
|
456 | Example: temporarily register a rcv callback for a tag matching some port |
|
|
457 | (e.g. for an rpc reply) and unregister it after a message was received. |
|
|
458 | |
|
|
459 | rcv $port, $otherport => sub { |
|
|
460 | my @reply = @_; |
|
|
461 | |
|
|
462 | rcv $SELF, $otherport; |
|
|
463 | }; |
|
|
464 | |
385 | =cut |
465 | =cut |
386 | |
466 | |
387 | sub rcv($@) { |
467 | sub rcv($@) { |
388 | my $port = shift; |
468 | my $port = shift; |
389 | my ($noderef, $portid) = split /#/, $port, 2; |
469 | my ($nodeid, $portid) = split /#/, $port, 2; |
390 | |
470 | |
391 | ($NODE{$noderef} || add_node $noderef) == $NODE{""} |
471 | $NODE{$nodeid} == $NODE{""} |
392 | or Carp::croak "$port: rcv can only be called on local ports, caught"; |
472 | or Carp::croak "$port: rcv can only be called on local ports, caught"; |
393 | |
473 | |
394 | if (@_ == 1) { |
474 | while (@_) { |
|
|
475 | if (ref $_[0]) { |
|
|
476 | if (my $self = $PORT_DATA{$portid}) { |
|
|
477 | "AnyEvent::MP::Port" eq ref $self |
|
|
478 | or Carp::croak "$port: rcv can only be called on message matching ports, caught"; |
|
|
479 | |
|
|
480 | $self->[0] = shift; |
|
|
481 | } else { |
395 | my $cb = shift; |
482 | my $cb = shift; |
396 | delete $PORT_DATA{$portid}; |
|
|
397 | $PORT{$portid} = sub { |
483 | $PORT{$portid} = sub { |
398 | local $SELF = $port; |
484 | local $SELF = $port; |
399 | eval { |
485 | eval { &$cb }; _self_die if $@; |
400 | &$cb |
486 | }; |
401 | and kil $port; |
|
|
402 | }; |
487 | } |
403 | _self_die if $@; |
488 | } elsif (defined $_[0]) { |
404 | }; |
|
|
405 | } else { |
|
|
406 | my $self = $PORT_DATA{$portid} ||= do { |
489 | my $self = $PORT_DATA{$portid} ||= do { |
407 | my $self = bless { |
490 | my $self = bless [$PORT{$portid} || sub { }, { }, $port], "AnyEvent::MP::Port"; |
408 | id => $port, |
|
|
409 | }, "AnyEvent::MP::Port"; |
|
|
410 | |
491 | |
411 | $PORT{$portid} = sub { |
492 | $PORT{$portid} = sub { |
412 | local $SELF = $port; |
493 | local $SELF = $port; |
413 | |
494 | |
414 | eval { |
|
|
415 | for (@{ $self->{rc0}{$_[0]} }) { |
495 | if (my $cb = $self->[1]{$_[0]}) { |
416 | $_ && &{$_->[0]} |
496 | shift; |
417 | && undef $_; |
497 | eval { &$cb }; _self_die if $@; |
418 | } |
498 | } else { |
419 | |
|
|
420 | for (@{ $self->{rcv}{$_[0]} }) { |
|
|
421 | $_ && [@_[1 .. @{$_->[1]}]] ~~ $_->[1] |
|
|
422 | && &{$_->[0]} |
499 | &{ $self->[0] }; |
423 | && undef $_; |
|
|
424 | } |
|
|
425 | |
|
|
426 | for (@{ $self->{any} }) { |
|
|
427 | $_ && [@_[0 .. $#{$_->[1]}]] ~~ $_->[1] |
|
|
428 | && &{$_->[0]} |
|
|
429 | && undef $_; |
|
|
430 | } |
500 | } |
431 | }; |
501 | }; |
432 | _self_die if $@; |
502 | |
|
|
503 | $self |
433 | }; |
504 | }; |
434 | |
505 | |
435 | $self |
|
|
436 | }; |
|
|
437 | |
|
|
438 | "AnyEvent::MP::Port" eq ref $self |
506 | "AnyEvent::MP::Port" eq ref $self |
439 | or Carp::croak "$port: rcv can only be called on message matching ports, caught"; |
507 | or Carp::croak "$port: rcv can only be called on message matching ports, caught"; |
440 | |
508 | |
441 | while (@_) { |
|
|
442 | my ($match, $cb) = splice @_, 0, 2; |
509 | my ($tag, $cb) = splice @_, 0, 2; |
443 | |
510 | |
444 | if (!ref $match) { |
511 | if (defined $cb) { |
445 | push @{ $self->{rc0}{$match} }, [$cb]; |
512 | $self->[1]{$tag} = $cb; |
446 | } elsif (("ARRAY" eq ref $match && !ref $match->[0])) { |
|
|
447 | my ($type, @match) = @$match; |
|
|
448 | @match |
|
|
449 | ? push @{ $self->{rcv}{$match->[0]} }, [$cb, \@match] |
|
|
450 | : push @{ $self->{rc0}{$match->[0]} }, [$cb]; |
|
|
451 | } else { |
513 | } else { |
452 | push @{ $self->{any} }, [$cb, $match]; |
514 | delete $self->[1]{$tag}; |
453 | } |
515 | } |
454 | } |
516 | } |
455 | } |
517 | } |
456 | |
518 | |
457 | $port |
519 | $port |
458 | } |
520 | } |
459 | |
521 | |
|
|
522 | =item peval $port, $coderef[, @args] |
|
|
523 | |
|
|
524 | Evaluates the given C<$codref> within the contetx of C<$port>, that is, |
|
|
525 | when the code throews an exception the C<$port> will be killed. |
|
|
526 | |
|
|
527 | Any remaining args will be passed to the callback. Any return values will |
|
|
528 | be returned to the caller. |
|
|
529 | |
|
|
530 | This is useful when you temporarily want to execute code in the context of |
|
|
531 | a port. |
|
|
532 | |
|
|
533 | Example: create a port and run some initialisation code in it's context. |
|
|
534 | |
|
|
535 | my $port = port { ... }; |
|
|
536 | |
|
|
537 | peval $port, sub { |
|
|
538 | init |
|
|
539 | or die "unable to init"; |
|
|
540 | }; |
|
|
541 | |
|
|
542 | =cut |
|
|
543 | |
|
|
544 | sub peval($$) { |
|
|
545 | local $SELF = shift; |
|
|
546 | my $cb = shift; |
|
|
547 | |
|
|
548 | if (wantarray) { |
|
|
549 | my @res = eval { &$cb }; |
|
|
550 | _self_die if $@; |
|
|
551 | @res |
|
|
552 | } else { |
|
|
553 | my $res = eval { &$cb }; |
|
|
554 | _self_die if $@; |
|
|
555 | $res |
|
|
556 | } |
|
|
557 | } |
|
|
558 | |
460 | =item $closure = psub { BLOCK } |
559 | =item $closure = psub { BLOCK } |
461 | |
560 | |
462 | Remembers C<$SELF> and creates a closure out of the BLOCK. When the |
561 | Remembers C<$SELF> and creates a closure out of the BLOCK. When the |
463 | closure is executed, sets up the environment in the same way as in C<rcv> |
562 | closure is executed, sets up the environment in the same way as in C<rcv> |
464 | callbacks, i.e. runtime errors will cause the port to get C<kil>ed. |
563 | callbacks, i.e. runtime errors will cause the port to get C<kil>ed. |
|
|
564 | |
|
|
565 | The effect is basically as if it returned C<< sub { peval $SELF, sub { |
|
|
566 | BLOCK }, @_ } >>. |
465 | |
567 | |
466 | This is useful when you register callbacks from C<rcv> callbacks: |
568 | This is useful when you register callbacks from C<rcv> callbacks: |
467 | |
569 | |
468 | rcv delayed_reply => sub { |
570 | rcv delayed_reply => sub { |
469 | my ($delay, @reply) = @_; |
571 | my ($delay, @reply) = @_; |
… | |
… | |
493 | $res |
595 | $res |
494 | } |
596 | } |
495 | } |
597 | } |
496 | } |
598 | } |
497 | |
599 | |
498 | =item $guard = mon $port, $cb->(@reason) |
600 | =item $guard = mon $port, $cb->(@reason) # call $cb when $port dies |
499 | |
601 | |
500 | =item $guard = mon $port, $rcvport |
602 | =item $guard = mon $port, $rcvport # kill $rcvport when $port dies |
501 | |
603 | |
502 | =item $guard = mon $port |
604 | =item $guard = mon $port # kill $SELF when $port dies |
503 | |
605 | |
504 | =item $guard = mon $port, $rcvport, @msg |
606 | =item $guard = mon $port, $rcvport, @msg # send a message when $port dies |
505 | |
607 | |
506 | Monitor the given port and do something when the port is killed, and |
608 | Monitor the given port and do something when the port is killed or |
507 | optionally return a guard that can be used to stop monitoring again. |
609 | messages to it were lost, and optionally return a guard that can be used |
|
|
610 | to stop monitoring again. |
508 | |
611 | |
509 | In the first form (callback), the callback is simply called with any |
612 | In the first form (callback), the callback is simply called with any |
510 | number of C<@reason> elements (no @reason means that the port was deleted |
613 | number of C<@reason> elements (no @reason means that the port was deleted |
511 | "normally"). Note also that I<< the callback B<must> never die >>, so use |
614 | "normally"). Note also that I<< the callback B<must> never die >>, so use |
512 | C<eval> if unsure. |
615 | C<eval> if unsure. |
513 | |
616 | |
514 | In the second form (another port given), the other port (C<$rcvport) |
617 | In the second form (another port given), the other port (C<$rcvport>) |
515 | will be C<kil>'ed with C<@reason>, iff a @reason was specified, i.e. on |
618 | will be C<kil>'ed with C<@reason>, if a @reason was specified, i.e. on |
516 | "normal" kils nothing happens, while under all other conditions, the other |
619 | "normal" kils nothing happens, while under all other conditions, the other |
517 | port is killed with the same reason. |
620 | port is killed with the same reason. |
518 | |
621 | |
519 | The third form (kill self) is the same as the second form, except that |
622 | The third form (kill self) is the same as the second form, except that |
520 | C<$rvport> defaults to C<$SELF>. |
623 | C<$rvport> defaults to C<$SELF>. |
521 | |
624 | |
522 | In the last form (message), a message of the form C<@msg, @reason> will be |
625 | In the last form (message), a message of the form C<@msg, @reason> will be |
523 | C<snd>. |
626 | C<snd>. |
|
|
627 | |
|
|
628 | Monitoring-actions are one-shot: once messages are lost (and a monitoring |
|
|
629 | alert was raised), they are removed and will not trigger again. |
524 | |
630 | |
525 | As a rule of thumb, monitoring requests should always monitor a port from |
631 | As a rule of thumb, monitoring requests should always monitor a port from |
526 | a local port (or callback). The reason is that kill messages might get |
632 | a local port (or callback). The reason is that kill messages might get |
527 | lost, just like any other message. Another less obvious reason is that |
633 | lost, just like any other message. Another less obvious reason is that |
528 | even monitoring requests can get lost (for exmaple, when the connection |
634 | even monitoring requests can get lost (for example, when the connection |
529 | to the other node goes down permanently). When monitoring a port locally |
635 | to the other node goes down permanently). When monitoring a port locally |
530 | these problems do not exist. |
636 | these problems do not exist. |
531 | |
637 | |
|
|
638 | C<mon> effectively guarantees that, in the absence of hardware failures, |
|
|
639 | after starting the monitor, either all messages sent to the port will |
|
|
640 | arrive, or the monitoring action will be invoked after possible message |
|
|
641 | loss has been detected. No messages will be lost "in between" (after |
|
|
642 | the first lost message no further messages will be received by the |
|
|
643 | port). After the monitoring action was invoked, further messages might get |
|
|
644 | delivered again. |
|
|
645 | |
|
|
646 | Inter-host-connection timeouts and monitoring depend on the transport |
|
|
647 | used. The only transport currently implemented is TCP, and AnyEvent::MP |
|
|
648 | relies on TCP to detect node-downs (this can take 10-15 minutes on a |
|
|
649 | non-idle connection, and usually around two hours for idle connections). |
|
|
650 | |
|
|
651 | This means that monitoring is good for program errors and cleaning up |
|
|
652 | stuff eventually, but they are no replacement for a timeout when you need |
|
|
653 | to ensure some maximum latency. |
|
|
654 | |
532 | Example: call a given callback when C<$port> is killed. |
655 | Example: call a given callback when C<$port> is killed. |
533 | |
656 | |
534 | mon $port, sub { warn "port died because of <@_>\n" }; |
657 | mon $port, sub { warn "port died because of <@_>\n" }; |
535 | |
658 | |
536 | Example: kill ourselves when C<$port> is killed abnormally. |
659 | Example: kill ourselves when C<$port> is killed abnormally. |
… | |
… | |
542 | mon $port, $self => "restart"; |
665 | mon $port, $self => "restart"; |
543 | |
666 | |
544 | =cut |
667 | =cut |
545 | |
668 | |
546 | sub mon { |
669 | sub mon { |
547 | my ($noderef, $port) = split /#/, shift, 2; |
670 | my ($nodeid, $port) = split /#/, shift, 2; |
548 | |
671 | |
549 | my $node = $NODE{$noderef} || add_node $noderef; |
672 | my $node = $NODE{$nodeid} || add_node $nodeid; |
550 | |
673 | |
551 | my $cb = @_ ? $_[0] : $SELF || Carp::croak 'mon: called with one argument only, but $SELF not set,'; |
674 | my $cb = @_ ? shift : $SELF || Carp::croak 'mon: called with one argument only, but $SELF not set,'; |
552 | |
675 | |
553 | unless (ref $cb) { |
676 | unless (ref $cb) { |
554 | if (@_) { |
677 | if (@_) { |
555 | # send a kill info message |
678 | # send a kill info message |
556 | my (@msg) = @_; |
679 | my (@msg) = ($cb, @_); |
557 | $cb = sub { snd @msg, @_ }; |
680 | $cb = sub { snd @msg, @_ }; |
558 | } else { |
681 | } else { |
559 | # simply kill other port |
682 | # simply kill other port |
560 | my $port = $cb; |
683 | my $port = $cb; |
561 | $cb = sub { kil $port, @_ if @_ }; |
684 | $cb = sub { kil $port, @_ if @_ }; |
… | |
… | |
563 | } |
686 | } |
564 | |
687 | |
565 | $node->monitor ($port, $cb); |
688 | $node->monitor ($port, $cb); |
566 | |
689 | |
567 | defined wantarray |
690 | defined wantarray |
568 | and AnyEvent::Util::guard { $node->unmonitor ($port, $cb) } |
691 | and ($cb += 0, Guard::guard { $node->unmonitor ($port, $cb) }) |
569 | } |
692 | } |
570 | |
693 | |
571 | =item $guard = mon_guard $port, $ref, $ref... |
694 | =item $guard = mon_guard $port, $ref, $ref... |
572 | |
695 | |
573 | Monitors the given C<$port> and keeps the passed references. When the port |
696 | Monitors the given C<$port> and keeps the passed references. When the port |
574 | is killed, the references will be freed. |
697 | is killed, the references will be freed. |
575 | |
698 | |
576 | Optionally returns a guard that will stop the monitoring. |
699 | Optionally returns a guard that will stop the monitoring. |
577 | |
700 | |
578 | This function is useful when you create e.g. timers or other watchers and |
701 | This function is useful when you create e.g. timers or other watchers and |
579 | want to free them when the port gets killed: |
702 | want to free them when the port gets killed (note the use of C<psub>): |
580 | |
703 | |
581 | $port->rcv (start => sub { |
704 | $port->rcv (start => sub { |
582 | my $timer; $timer = mon_guard $port, AE::timer 1, 1, sub { |
705 | my $timer; $timer = mon_guard $port, AE::timer 1, 1, psub { |
583 | undef $timer if 0.9 < rand; |
706 | undef $timer if 0.9 < rand; |
584 | }); |
707 | }); |
585 | }); |
708 | }); |
586 | |
709 | |
587 | =cut |
710 | =cut |
… | |
… | |
596 | |
719 | |
597 | =item kil $port[, @reason] |
720 | =item kil $port[, @reason] |
598 | |
721 | |
599 | Kill the specified port with the given C<@reason>. |
722 | Kill the specified port with the given C<@reason>. |
600 | |
723 | |
601 | If no C<@reason> is specified, then the port is killed "normally" (linked |
724 | If no C<@reason> is specified, then the port is killed "normally" - |
602 | ports will not be kileld, or even notified). |
725 | monitor callback will be invoked, but the kil will not cause linked ports |
|
|
726 | (C<mon $mport, $lport> form) to get killed. |
603 | |
727 | |
604 | Otherwise, linked ports get killed with the same reason (second form of |
728 | If a C<@reason> is specified, then linked ports (C<mon $mport, $lport> |
605 | C<mon>, see below). |
729 | form) get killed with the same reason. |
606 | |
730 | |
607 | Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks |
731 | Runtime errors while evaluating C<rcv> callbacks or inside C<psub> blocks |
608 | will be reported as reason C<< die => $@ >>. |
732 | will be reported as reason C<< die => $@ >>. |
609 | |
733 | |
610 | Transport/communication errors are reported as C<< transport_error => |
734 | Transport/communication errors are reported as C<< transport_error => |
… | |
… | |
615 | =item $port = spawn $node, $initfunc[, @initdata] |
739 | =item $port = spawn $node, $initfunc[, @initdata] |
616 | |
740 | |
617 | Creates a port on the node C<$node> (which can also be a port ID, in which |
741 | Creates a port on the node C<$node> (which can also be a port ID, in which |
618 | case it's the node where that port resides). |
742 | case it's the node where that port resides). |
619 | |
743 | |
620 | The port ID of the newly created port is return immediately, and it is |
744 | The port ID of the newly created port is returned immediately, and it is |
621 | permissible to immediately start sending messages or monitor the port. |
745 | possible to immediately start sending messages or to monitor the port. |
622 | |
746 | |
623 | After the port has been created, the init function is |
747 | After the port has been created, the init function is called on the remote |
624 | called. This function must be a fully-qualified function name |
748 | node, in the same context as a C<rcv> callback. This function must be a |
625 | (e.g. C<MyApp::Chat::Server::init>). |
749 | fully-qualified function name (e.g. C<MyApp::Chat::Server::init>). To |
|
|
750 | specify a function in the main program, use C<::name>. |
626 | |
751 | |
627 | If the function doesn't exist, then the node tries to C<require> |
752 | If the function doesn't exist, then the node tries to C<require> |
628 | the package, then the package above the package and so on (e.g. |
753 | the package, then the package above the package and so on (e.g. |
629 | C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function |
754 | C<MyApp::Chat::Server>, C<MyApp::Chat>, C<MyApp>) until the function |
630 | exists or it runs out of package names. |
755 | exists or it runs out of package names. |
631 | |
756 | |
632 | The init function is then called with the newly-created port as context |
757 | The init function is then called with the newly-created port as context |
633 | object (C<$SELF>) and the C<@initdata> values as arguments. |
758 | object (C<$SELF>) and the C<@initdata> values as arguments. It I<must> |
|
|
759 | call one of the C<rcv> functions to set callbacks on C<$SELF>, otherwise |
|
|
760 | the port might not get created. |
634 | |
761 | |
635 | A common idiom is to pass your own port, monitor the spawned port, and |
762 | A common idiom is to pass a local port, immediately monitor the spawned |
636 | in the init function, monitor the original port. This two-way monitoring |
763 | port, and in the remote init function, immediately monitor the passed |
637 | ensures that both ports get cleaned up when there is a problem. |
764 | local port. This two-way monitoring ensures that both ports get cleaned up |
|
|
765 | when there is a problem. |
|
|
766 | |
|
|
767 | C<spawn> guarantees that the C<$initfunc> has no visible effects on the |
|
|
768 | caller before C<spawn> returns (by delaying invocation when spawn is |
|
|
769 | called for the local node). |
638 | |
770 | |
639 | Example: spawn a chat server port on C<$othernode>. |
771 | Example: spawn a chat server port on C<$othernode>. |
640 | |
772 | |
641 | # this node, executed from within a port context: |
773 | # this node, executed from within a port context: |
642 | my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF; |
774 | my $server = spawn $othernode, "MyApp::Chat::Server::connect", $SELF; |
… | |
… | |
657 | |
789 | |
658 | sub _spawn { |
790 | sub _spawn { |
659 | my $port = shift; |
791 | my $port = shift; |
660 | my $init = shift; |
792 | my $init = shift; |
661 | |
793 | |
|
|
794 | # rcv will create the actual port |
662 | local $SELF = "$NODE#$port"; |
795 | local $SELF = "$NODE#$port"; |
663 | eval { |
796 | eval { |
664 | &{ load_func $init } |
797 | &{ load_func $init } |
665 | }; |
798 | }; |
666 | _self_die if $@; |
799 | _self_die if $@; |
667 | } |
800 | } |
668 | |
801 | |
669 | sub spawn(@) { |
802 | sub spawn(@) { |
670 | my ($noderef, undef) = split /#/, shift, 2; |
803 | my ($nodeid, undef) = split /#/, shift, 2; |
671 | |
804 | |
672 | my $id = "$RUNIQ." . $ID++; |
805 | my $id = $RUNIQ . ++$ID; |
673 | |
806 | |
674 | $_[0] =~ /::/ |
807 | $_[0] =~ /::/ |
675 | or Carp::croak "spawn init function must be a fully-qualified name, caught"; |
808 | or Carp::croak "spawn init function must be a fully-qualified name, caught"; |
676 | |
809 | |
677 | ($NODE{$noderef} || add_node $noderef) |
810 | snd_to_func $nodeid, "AnyEvent::MP::_spawn" => $id, @_; |
678 | ->send (["", "AnyEvent::MP::_spawn" => $id, @_]); |
|
|
679 | |
811 | |
680 | "$noderef#$id" |
812 | "$nodeid#$id" |
681 | } |
813 | } |
682 | |
814 | |
|
|
815 | |
|
|
816 | =item after $timeout, @msg |
|
|
817 | |
|
|
818 | =item after $timeout, $callback |
|
|
819 | |
|
|
820 | Either sends the given message, or call the given callback, after the |
|
|
821 | specified number of seconds. |
|
|
822 | |
|
|
823 | This is simply a utility function that comes in handy at times - the |
|
|
824 | AnyEvent::MP author is not convinced of the wisdom of having it, though, |
|
|
825 | so it may go away in the future. |
|
|
826 | |
|
|
827 | =cut |
|
|
828 | |
|
|
829 | sub after($@) { |
|
|
830 | my ($timeout, @action) = @_; |
|
|
831 | |
|
|
832 | my $t; $t = AE::timer $timeout, 0, sub { |
|
|
833 | undef $t; |
|
|
834 | ref $action[0] |
|
|
835 | ? $action[0]() |
|
|
836 | : snd @action; |
|
|
837 | }; |
|
|
838 | } |
|
|
839 | |
|
|
840 | #=item $cb2 = timeout $seconds, $cb[, @args] |
|
|
841 | |
|
|
842 | =item cal $port, @msg, $callback[, $timeout] |
|
|
843 | |
|
|
844 | A simple form of RPC - sends a message to the given C<$port> with the |
|
|
845 | given contents (C<@msg>), but adds a reply port to the message. |
|
|
846 | |
|
|
847 | The reply port is created temporarily just for the purpose of receiving |
|
|
848 | the reply, and will be C<kil>ed when no longer needed. |
|
|
849 | |
|
|
850 | A reply message sent to the port is passed to the C<$callback> as-is. |
|
|
851 | |
|
|
852 | If an optional time-out (in seconds) is given and it is not C<undef>, |
|
|
853 | then the callback will be called without any arguments after the time-out |
|
|
854 | elapsed and the port is C<kil>ed. |
|
|
855 | |
|
|
856 | If no time-out is given (or it is C<undef>), then the local port will |
|
|
857 | monitor the remote port instead, so it eventually gets cleaned-up. |
|
|
858 | |
|
|
859 | Currently this function returns the temporary port, but this "feature" |
|
|
860 | might go in future versions unless you can make a convincing case that |
|
|
861 | this is indeed useful for something. |
|
|
862 | |
|
|
863 | =cut |
|
|
864 | |
|
|
865 | sub cal(@) { |
|
|
866 | my $timeout = ref $_[-1] ? undef : pop; |
|
|
867 | my $cb = pop; |
|
|
868 | |
|
|
869 | my $port = port { |
|
|
870 | undef $timeout; |
|
|
871 | kil $SELF; |
|
|
872 | &$cb; |
|
|
873 | }; |
|
|
874 | |
|
|
875 | if (defined $timeout) { |
|
|
876 | $timeout = AE::timer $timeout, 0, sub { |
|
|
877 | undef $timeout; |
|
|
878 | kil $port; |
|
|
879 | $cb->(); |
|
|
880 | }; |
|
|
881 | } else { |
|
|
882 | mon $_[0], sub { |
|
|
883 | kil $port; |
|
|
884 | $cb->(); |
|
|
885 | }; |
|
|
886 | } |
|
|
887 | |
|
|
888 | push @_, $port; |
|
|
889 | &snd; |
|
|
890 | |
|
|
891 | $port |
|
|
892 | } |
|
|
893 | |
683 | =back |
894 | =back |
684 | |
895 | |
685 | =head1 NODE MESSAGES |
896 | =head1 DISTRIBUTED DATABASE |
686 | |
897 | |
687 | Nodes understand the following messages sent to them. Many of them take |
898 | AnyEvent::MP comes with a simple distributed database. The database will |
688 | arguments called C<@reply>, which will simply be used to compose a reply |
899 | be mirrored asynchronously at all global nodes. Other nodes bind to one of |
689 | message - C<$reply[0]> is the port to reply to, C<$reply[1]> the type and |
900 | the global nodes for their needs. |
690 | the remaining arguments are simply the message data. |
|
|
691 | |
901 | |
692 | While other messages exist, they are not public and subject to change. |
902 | The database consists of a two-level hash - a hash contains a hash which |
|
|
903 | contains values. |
693 | |
904 | |
|
|
905 | The top level hash key is called "family", and the second-level hash key |
|
|
906 | is called "subkey" or simply "key". |
|
|
907 | |
|
|
908 | The family must be alphanumeric, i.e. start with a letter and consist |
|
|
909 | of letters, digits, underscores and colons (C<[A-Za-z][A-Za-z0-9_:]*>, |
|
|
910 | pretty much like Perl module names. |
|
|
911 | |
|
|
912 | As the family namespace is global, it is recommended to prefix family names |
|
|
913 | with the name of the application or module using it. |
|
|
914 | |
|
|
915 | The subkeys must be non-empty strings, with no further restrictions. |
|
|
916 | |
|
|
917 | The values should preferably be strings, but other perl scalars should |
|
|
918 | work as well (such as undef, arrays and hashes). |
|
|
919 | |
|
|
920 | Every database entry is owned by one node - adding the same family/subkey |
|
|
921 | combination on multiple nodes will not cause discomfort for AnyEvent::MP, |
|
|
922 | but the result might be nondeterministic, i.e. the key might have |
|
|
923 | different values on different nodes. |
|
|
924 | |
|
|
925 | Different subkeys in the same family can be owned by different nodes |
|
|
926 | without problems, and in fact, this is the common method to create worker |
|
|
927 | pools. For example, a worker port for image scaling might do this: |
|
|
928 | |
|
|
929 | db_set my_image_scalers => $port; |
|
|
930 | |
|
|
931 | And clients looking for an image scaler will want to get the |
|
|
932 | C<my_image_scalers> keys from time to time: |
|
|
933 | |
|
|
934 | db_keys my_image_scalers => sub { |
|
|
935 | @ports = @{ $_[0] }; |
|
|
936 | }; |
|
|
937 | |
|
|
938 | Or better yet, they want to monitor the database family, so they always |
|
|
939 | have a reasonable up-to-date copy: |
|
|
940 | |
|
|
941 | db_mon my_image_scalers => sub { |
|
|
942 | @ports = keys %{ $_[0] }; |
|
|
943 | }; |
|
|
944 | |
|
|
945 | In general, you can set or delete single subkeys, but query and monitor |
|
|
946 | whole families only. |
|
|
947 | |
|
|
948 | If you feel the need to monitor or query a single subkey, try giving it |
|
|
949 | it's own family. |
|
|
950 | |
694 | =over 4 |
951 | =over |
|
|
952 | |
|
|
953 | =item db_set $family => $subkey [=> $value] |
|
|
954 | |
|
|
955 | Sets (or replaces) a key to the database - if C<$value> is omitted, |
|
|
956 | C<undef> is used instead. |
|
|
957 | |
|
|
958 | =item db_del $family => $subkey... |
|
|
959 | |
|
|
960 | Deletes one or more subkeys from the database family. |
|
|
961 | |
|
|
962 | =item $guard = db_reg $family => $subkey [=> $value] |
|
|
963 | |
|
|
964 | Sets the key on the database and returns a guard. When the guard is |
|
|
965 | destroyed, the key is deleted from the database. If C<$value> is missing, |
|
|
966 | then C<undef> is used. |
|
|
967 | |
|
|
968 | =item db_family $family => $cb->(\%familyhash) |
|
|
969 | |
|
|
970 | Queries the named database C<$family> and call the callback with the |
|
|
971 | family represented as a hash. You can keep and freely modify the hash. |
|
|
972 | |
|
|
973 | =item db_keys $family => $cb->(\@keys) |
|
|
974 | |
|
|
975 | Same as C<db_family>, except it only queries the family I<subkeys> and passes |
|
|
976 | them as array reference to the callback. |
|
|
977 | |
|
|
978 | =item db_values $family => $cb->(\@values) |
|
|
979 | |
|
|
980 | Same as C<db_family>, except it only queries the family I<values> and passes them |
|
|
981 | as array reference to the callback. |
|
|
982 | |
|
|
983 | =item $guard = db_mon $family => $cb->($familyhash, \@added, \@changed, \@deleted) |
|
|
984 | |
|
|
985 | Creates a monitor on the given database family. Each time a key is set |
|
|
986 | or or is deleted the callback is called with a hash containing the |
|
|
987 | database family and three lists of added, changed and deleted subkeys, |
|
|
988 | respectively. If no keys have changed then the array reference might be |
|
|
989 | C<undef> or even missing. |
|
|
990 | |
|
|
991 | The family hash reference and the key arrays belong to AnyEvent::MP and |
|
|
992 | B<must not be modified or stored> by the callback. When in doubt, make a |
|
|
993 | copy. |
|
|
994 | |
|
|
995 | As soon as possible after the monitoring starts, the callback will be |
|
|
996 | called with the intiial contents of the family, even if it is empty, |
|
|
997 | i.e. there will always be a timely call to the callback with the current |
|
|
998 | contents. |
|
|
999 | |
|
|
1000 | It is possible that the callback is called with a change event even though |
|
|
1001 | the subkey is already present and the value has not changed. |
|
|
1002 | |
|
|
1003 | The monitoring stops when the guard object is destroyed. |
|
|
1004 | |
|
|
1005 | Example: on every change to the family "mygroup", print out all keys. |
|
|
1006 | |
|
|
1007 | my $guard = db_mon mygroup => sub { |
|
|
1008 | my ($family, $a, $c, $d) = @_; |
|
|
1009 | print "mygroup members: ", (join " ", keys %$family), "\n"; |
|
|
1010 | }; |
|
|
1011 | |
|
|
1012 | Exmaple: wait until the family "My::Module::workers" is non-empty. |
|
|
1013 | |
|
|
1014 | my $guard; $guard = db_mon My::Module::workers => sub { |
|
|
1015 | my ($family, $a, $c, $d) = @_; |
|
|
1016 | return unless %$family; |
|
|
1017 | undef $guard; |
|
|
1018 | print "My::Module::workers now nonempty\n"; |
|
|
1019 | }; |
|
|
1020 | |
|
|
1021 | Example: print all changes to the family "AnyRvent::Fantasy::Module". |
|
|
1022 | |
|
|
1023 | my $guard = db_mon AnyRvent::Fantasy::Module => sub { |
|
|
1024 | my ($family, $a, $c, $d) = @_; |
|
|
1025 | |
|
|
1026 | print "+$_=$family->{$_}\n" for @$a; |
|
|
1027 | print "*$_=$family->{$_}\n" for @$c; |
|
|
1028 | print "-$_=$family->{$_}\n" for @$d; |
|
|
1029 | }; |
695 | |
1030 | |
696 | =cut |
1031 | =cut |
697 | |
|
|
698 | =item lookup => $name, @reply |
|
|
699 | |
|
|
700 | Replies with the port ID of the specified well-known port, or C<undef>. |
|
|
701 | |
|
|
702 | =item devnull => ... |
|
|
703 | |
|
|
704 | Generic data sink/CPU heat conversion. |
|
|
705 | |
|
|
706 | =item relay => $port, @msg |
|
|
707 | |
|
|
708 | Simply forwards the message to the given port. |
|
|
709 | |
|
|
710 | =item eval => $string[ @reply] |
|
|
711 | |
|
|
712 | Evaluates the given string. If C<@reply> is given, then a message of the |
|
|
713 | form C<@reply, $@, @evalres> is sent. |
|
|
714 | |
|
|
715 | Example: crash another node. |
|
|
716 | |
|
|
717 | snd $othernode, eval => "exit"; |
|
|
718 | |
|
|
719 | =item time => @reply |
|
|
720 | |
|
|
721 | Replies the the current node time to C<@reply>. |
|
|
722 | |
|
|
723 | Example: tell the current node to send the current time to C<$myport> in a |
|
|
724 | C<timereply> message. |
|
|
725 | |
|
|
726 | snd $NODE, time => $myport, timereply => 1, 2; |
|
|
727 | # => snd $myport, timereply => 1, 2, <time> |
|
|
728 | |
1032 | |
729 | =back |
1033 | =back |
730 | |
1034 | |
731 | =head1 AnyEvent::MP vs. Distributed Erlang |
1035 | =head1 AnyEvent::MP vs. Distributed Erlang |
732 | |
1036 | |
733 | AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node |
1037 | AnyEvent::MP got lots of its ideas from distributed Erlang (Erlang node |
734 | == aemp node, Erlang process == aemp port), so many of the documents and |
1038 | == aemp node, Erlang process == aemp port), so many of the documents and |
735 | programming techniques employed by Erlang apply to AnyEvent::MP. Here is a |
1039 | programming techniques employed by Erlang apply to AnyEvent::MP. Here is a |
736 | sample: |
1040 | sample: |
737 | |
1041 | |
738 | http://www.Erlang.se/doc/programming_rules.shtml |
1042 | http://www.erlang.se/doc/programming_rules.shtml |
739 | http://Erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4 |
1043 | http://erlang.org/doc/getting_started/part_frame.html # chapters 3 and 4 |
740 | http://Erlang.org/download/Erlang-book-part1.pdf # chapters 5 and 6 |
1044 | http://erlang.org/download/erlang-book-part1.pdf # chapters 5 and 6 |
741 | http://Erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5 |
1045 | http://erlang.org/download/armstrong_thesis_2003.pdf # chapters 4 and 5 |
742 | |
1046 | |
743 | Despite the similarities, there are also some important differences: |
1047 | Despite the similarities, there are also some important differences: |
744 | |
1048 | |
745 | =over 4 |
1049 | =over 4 |
746 | |
1050 | |
747 | =item * Node references contain the recipe on how to contact them. |
1051 | =item * Node IDs are arbitrary strings in AEMP. |
748 | |
1052 | |
749 | Erlang relies on special naming and DNS to work everywhere in the |
1053 | Erlang relies on special naming and DNS to work everywhere in the same |
750 | same way. AEMP relies on each node knowing it's own address(es), with |
1054 | way. AEMP relies on each node somehow knowing its own address(es) (e.g. by |
751 | convenience functionality. |
1055 | configuration or DNS), and possibly the addresses of some seed nodes, but |
|
|
1056 | will otherwise discover other nodes (and their IDs) itself. |
752 | |
1057 | |
753 | This means that AEMP requires a less tightly controlled environment at the |
1058 | =item * Erlang has a "remote ports are like local ports" philosophy, AEMP |
754 | cost of longer node references and a slightly higher management overhead. |
1059 | uses "local ports are like remote ports". |
|
|
1060 | |
|
|
1061 | The failure modes for local ports are quite different (runtime errors |
|
|
1062 | only) then for remote ports - when a local port dies, you I<know> it dies, |
|
|
1063 | when a connection to another node dies, you know nothing about the other |
|
|
1064 | port. |
|
|
1065 | |
|
|
1066 | Erlang pretends remote ports are as reliable as local ports, even when |
|
|
1067 | they are not. |
|
|
1068 | |
|
|
1069 | AEMP encourages a "treat remote ports differently" philosophy, with local |
|
|
1070 | ports being the special case/exception, where transport errors cannot |
|
|
1071 | occur. |
755 | |
1072 | |
756 | =item * Erlang uses processes and a mailbox, AEMP does not queue. |
1073 | =item * Erlang uses processes and a mailbox, AEMP does not queue. |
757 | |
1074 | |
758 | Erlang uses processes that selctively receive messages, and therefore |
1075 | Erlang uses processes that selectively receive messages out of order, and |
759 | needs a queue. AEMP is event based, queuing messages would serve no useful |
1076 | therefore needs a queue. AEMP is event based, queuing messages would serve |
760 | purpose. |
1077 | no useful purpose. For the same reason the pattern-matching abilities |
|
|
1078 | of AnyEvent::MP are more limited, as there is little need to be able to |
|
|
1079 | filter messages without dequeuing them. |
761 | |
1080 | |
762 | (But see L<Coro::MP> for a more Erlang-like process model on top of AEMP). |
1081 | This is not a philosophical difference, but simply stems from AnyEvent::MP |
|
|
1082 | being event-based, while Erlang is process-based. |
|
|
1083 | |
|
|
1084 | You cna have a look at L<Coro::MP> for a more Erlang-like process model on |
|
|
1085 | top of AEMP and Coro threads. |
763 | |
1086 | |
764 | =item * Erlang sends are synchronous, AEMP sends are asynchronous. |
1087 | =item * Erlang sends are synchronous, AEMP sends are asynchronous. |
765 | |
1088 | |
766 | Sending messages in Erlang is synchronous and blocks the process. AEMP |
1089 | Sending messages in Erlang is synchronous and blocks the process until |
767 | sends are immediate, connection establishment is handled in the |
1090 | a conenction has been established and the message sent (and so does not |
768 | background. |
1091 | need a queue that can overflow). AEMP sends return immediately, connection |
|
|
1092 | establishment is handled in the background. |
769 | |
1093 | |
770 | =item * Erlang can silently lose messages, AEMP cannot. |
1094 | =item * Erlang suffers from silent message loss, AEMP does not. |
771 | |
1095 | |
772 | Erlang makes few guarantees on messages delivery - messages can get lost |
1096 | Erlang implements few guarantees on messages delivery - messages can get |
773 | without any of the processes realising it (i.e. you send messages a, b, |
1097 | lost without any of the processes realising it (i.e. you send messages a, |
774 | and c, and the other side only receives messages a and c). |
1098 | b, and c, and the other side only receives messages a and c). |
775 | |
1099 | |
776 | AEMP guarantees correct ordering, and the guarantee that there are no |
1100 | AEMP guarantees (modulo hardware errors) correct ordering, and the |
|
|
1101 | guarantee that after one message is lost, all following ones sent to the |
|
|
1102 | same port are lost as well, until monitoring raises an error, so there are |
777 | holes in the message sequence. |
1103 | no silent "holes" in the message sequence. |
778 | |
1104 | |
779 | =item * In Erlang, processes can be declared dead and later be found to be |
1105 | If you want your software to be very reliable, you have to cope with |
780 | alive. |
1106 | corrupted and even out-of-order messages in both Erlang and AEMP. AEMP |
781 | |
1107 | simply tries to work better in common error cases, such as when a network |
782 | In Erlang it can happen that a monitored process is declared dead and |
1108 | link goes down. |
783 | linked processes get killed, but later it turns out that the process is |
|
|
784 | still alive - and can receive messages. |
|
|
785 | |
|
|
786 | In AEMP, when port monitoring detects a port as dead, then that port will |
|
|
787 | eventually be killed - it cannot happen that a node detects a port as dead |
|
|
788 | and then later sends messages to it, finding it is still alive. |
|
|
789 | |
1109 | |
790 | =item * Erlang can send messages to the wrong port, AEMP does not. |
1110 | =item * Erlang can send messages to the wrong port, AEMP does not. |
791 | |
1111 | |
792 | In Erlang it is quite possible that a node that restarts reuses a process |
1112 | In Erlang it is quite likely that a node that restarts reuses an Erlang |
793 | ID known to other nodes for a completely different process, causing |
1113 | process ID known to other nodes for a completely different process, |
794 | messages destined for that process to end up in an unrelated process. |
1114 | causing messages destined for that process to end up in an unrelated |
|
|
1115 | process. |
795 | |
1116 | |
796 | AEMP never reuses port IDs, so old messages or old port IDs floating |
1117 | AEMP does not reuse port IDs, so old messages or old port IDs floating |
797 | around in the network will not be sent to an unrelated port. |
1118 | around in the network will not be sent to an unrelated port. |
798 | |
1119 | |
799 | =item * Erlang uses unprotected connections, AEMP uses secure |
1120 | =item * Erlang uses unprotected connections, AEMP uses secure |
800 | authentication and can use TLS. |
1121 | authentication and can use TLS. |
801 | |
1122 | |
802 | AEMP can use a proven protocol - SSL/TLS - to protect connections and |
1123 | AEMP can use a proven protocol - TLS - to protect connections and |
803 | securely authenticate nodes. |
1124 | securely authenticate nodes. |
804 | |
1125 | |
805 | =item * The AEMP protocol is optimised for both text-based and binary |
1126 | =item * The AEMP protocol is optimised for both text-based and binary |
806 | communications. |
1127 | communications. |
807 | |
1128 | |
808 | The AEMP protocol, unlike the Erlang protocol, supports both |
1129 | The AEMP protocol, unlike the Erlang protocol, supports both programming |
809 | language-independent text-only protocols (good for debugging) and binary, |
1130 | language independent text-only protocols (good for debugging), and binary, |
810 | language-specific serialisers (e.g. Storable). |
1131 | language-specific serialisers (e.g. Storable). By default, unless TLS is |
|
|
1132 | used, the protocol is actually completely text-based. |
811 | |
1133 | |
812 | It has also been carefully designed to be implementable in other languages |
1134 | It has also been carefully designed to be implementable in other languages |
813 | with a minimum of work while gracefully degrading fucntionality to make the |
1135 | with a minimum of work while gracefully degrading functionality to make the |
814 | protocol simple. |
1136 | protocol simple. |
815 | |
1137 | |
816 | =item * AEMP has more flexible monitoring options than Erlang. |
1138 | =item * AEMP has more flexible monitoring options than Erlang. |
817 | |
1139 | |
818 | In Erlang, you can chose to receive I<all> exit signals as messages |
1140 | In Erlang, you can chose to receive I<all> exit signals as messages or |
819 | or I<none>, there is no in-between, so monitoring single processes is |
1141 | I<none>, there is no in-between, so monitoring single Erlang processes is |
820 | difficult to implement. Monitoring in AEMP is more flexible than in |
1142 | difficult to implement. |
821 | Erlang, as one can choose between automatic kill, exit message or callback |
1143 | |
822 | on a per-process basis. |
1144 | Monitoring in AEMP is more flexible than in Erlang, as one can choose |
|
|
1145 | between automatic kill, exit message or callback on a per-port basis. |
823 | |
1146 | |
824 | =item * Erlang tries to hide remote/local connections, AEMP does not. |
1147 | =item * Erlang tries to hide remote/local connections, AEMP does not. |
825 | |
1148 | |
826 | Monitoring in Erlang is not an indicator of process death/crashes, |
1149 | Monitoring in Erlang is not an indicator of process death/crashes, in the |
827 | as linking is (except linking is unreliable in Erlang). |
1150 | same way as linking is (except linking is unreliable in Erlang). |
828 | |
1151 | |
829 | In AEMP, you don't "look up" registered port names or send to named ports |
1152 | In AEMP, you don't "look up" registered port names or send to named ports |
830 | that might or might not be persistent. Instead, you normally spawn a port |
1153 | that might or might not be persistent. Instead, you normally spawn a port |
831 | on the remote node. The init function monitors the you, and you monitor |
1154 | on the remote node. The init function monitors you, and you monitor the |
832 | the remote port. Since both monitors are local to the node, they are much |
1155 | remote port. Since both monitors are local to the node, they are much more |
833 | more reliable. |
1156 | reliable (no need for C<spawn_link>). |
834 | |
1157 | |
835 | This also saves round-trips and avoids sending messages to the wrong port |
1158 | This also saves round-trips and avoids sending messages to the wrong port |
836 | (hard to do in Erlang). |
1159 | (hard to do in Erlang). |
837 | |
1160 | |
838 | =back |
1161 | =back |
839 | |
1162 | |
|
|
1163 | =head1 RATIONALE |
|
|
1164 | |
|
|
1165 | =over 4 |
|
|
1166 | |
|
|
1167 | =item Why strings for port and node IDs, why not objects? |
|
|
1168 | |
|
|
1169 | We considered "objects", but found that the actual number of methods |
|
|
1170 | that can be called are quite low. Since port and node IDs travel over |
|
|
1171 | the network frequently, the serialising/deserialising would add lots of |
|
|
1172 | overhead, as well as having to keep a proxy object everywhere. |
|
|
1173 | |
|
|
1174 | Strings can easily be printed, easily serialised etc. and need no special |
|
|
1175 | procedures to be "valid". |
|
|
1176 | |
|
|
1177 | And as a result, a port with just a default receiver consists of a single |
|
|
1178 | code reference stored in a global hash - it can't become much cheaper. |
|
|
1179 | |
|
|
1180 | =item Why favour JSON, why not a real serialising format such as Storable? |
|
|
1181 | |
|
|
1182 | In fact, any AnyEvent::MP node will happily accept Storable as framing |
|
|
1183 | format, but currently there is no way to make a node use Storable by |
|
|
1184 | default (although all nodes will accept it). |
|
|
1185 | |
|
|
1186 | The default framing protocol is JSON because a) JSON::XS is many times |
|
|
1187 | faster for small messages and b) most importantly, after years of |
|
|
1188 | experience we found that object serialisation is causing more problems |
|
|
1189 | than it solves: Just like function calls, objects simply do not travel |
|
|
1190 | easily over the network, mostly because they will always be a copy, so you |
|
|
1191 | always have to re-think your design. |
|
|
1192 | |
|
|
1193 | Keeping your messages simple, concentrating on data structures rather than |
|
|
1194 | objects, will keep your messages clean, tidy and efficient. |
|
|
1195 | |
|
|
1196 | =back |
|
|
1197 | |
840 | =head1 SEE ALSO |
1198 | =head1 SEE ALSO |
|
|
1199 | |
|
|
1200 | L<AnyEvent::MP::Intro> - a gentle introduction. |
|
|
1201 | |
|
|
1202 | L<AnyEvent::MP::Kernel> - more, lower-level, stuff. |
|
|
1203 | |
|
|
1204 | L<AnyEvent::MP::Global> - network maintenance and port groups, to find |
|
|
1205 | your applications. |
|
|
1206 | |
|
|
1207 | L<AnyEvent::MP::DataConn> - establish data connections between nodes. |
|
|
1208 | |
|
|
1209 | L<AnyEvent::MP::LogCatcher> - simple service to display log messages from |
|
|
1210 | all nodes. |
841 | |
1211 | |
842 | L<AnyEvent>. |
1212 | L<AnyEvent>. |
843 | |
1213 | |
844 | =head1 AUTHOR |
1214 | =head1 AUTHOR |
845 | |
1215 | |