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

Comparing AnyEvent-MP/Changes (file contents):
Revision 1.4 by root, Mon Aug 3 15:40:53 2009 UTC vs.
Revision 1.116 by root, Sun Feb 26 10:29:59 2012 UTC

1Revision history for AnyEvent::MP 1Revision history for AnyEvent::MP
2 2
3TODO: tls as auth-method? 3TODO: testsuite
4TODO: intro: maybe simple job pool example?
5TODO: intro: mention watchdog...
6TODO: maybe disbale warnings by default?
7TODO: networkdomains
8TODO: group proxy nodes - store group membership for other nodes
9TODO: switch to AnyEvent::Log
10
11faq: can't se anything
12faq: all is asynch
13faq: how to interface to non-perl nodes?
14
15<<<<<<< Changes
16TODO: Anyevent::MP gets grp_reg, grp_get
17TODO: AnyEvent::MP keeps list of all "global" nodes
18TODO: Global loaded on demand only, adds grp database and node address database
19TODO: transport connects use any global node to find addresses (maybe itself)
20TODO: not really
21
22 - move seed code to Kernel.
23 - fixed small typo in AnyEvent::MP::Transport.
24
25>>>>>>> 1.115
261.30 Thu Jun 30 11:30:39 CEST 2011
27 - connection errors at the right time would kill node
28 connections in a bad way (patch by Malcolm Studd).
29
301.29 Fri May 7 20:13:39 CEST 2010
31 - codename "Sadrak".
32 - fix error in callback at AnyEvent/MP/Global.pm line 339
33 (found by Sadrak).
34 - listener-less nodes had trouble sending keepalive
35 messages on write timeouts (found by Sadrak).
36 - the monitor guard could cause a memleak due
37 to autovivification if a mon was cleared after
38 the port was gone (analysed by Sadrak).
39 - do not overwrite the config file if we couldn't read it
40 for some reason (lesson demonstrated by Sadrak).
41
421.28 Thu Apr 1 21:23:54 CEST 2010
43 - accepted connections didn't correctly set up the SRCNODE
44 in some cases, leading to intra-node messages (such as monitoring)
45 to fail silently (great testcase by Sadrak).
46
471.27 Sat Mar 20 21:11:16 CET 2010
48 - avoid exceptions in the transport code if the connection
49 fails during TLS negotiation.
50 - delay new address broadcasts in seednodes, to reduce the
51 likelyhood of nodes connecting at the same time, leading to
52 message loss.
53 - improve general message passing efficiency by 8%, and special
54 case json framing for a 50% speed increase.
55 - add hints how to interface to aemp nodes from other programming
56 languages using the aemp protocol in AnyEvent::MP::Transport
57 (it can really be quite simple).
58 - handle unsupported remote framings more gracefully.
59
601.26 Wed Dec 30 16:48:37 CET 2009
61 - INCOMPATIBLE CHANGE: the high-level node protocol changed,
62 all nodes need to upgrade.
63 - remote kil's did not work (reported by Eugene Grigoriev).
64 - documented monitor_timeout, connect_interval, auth_offer,
65 auth_accept, framing_format, autocork and nodelay options
66 and make them configurable via bin/aemp.
67 - some doc updates.
68 - remote kil's were not synchronous to message sends which affected
69 nobody, as remote kils were broken.
70 - remote mon/unmonitor requests were not synchronous to message
71 sends either sometimes, although it likely didnt matter.
72
731.24 Wed Dec 9 14:59:52 CET 2009
74 - codename "the foul egg".
75 - increased "unmonitored local port died" message level to 2, so
76 it is displayed by default now.
77 - generate random seed at configure time, not module loading time,
78 so fork&configure works as expected.
79 - decreased "seed"-related message priority in Global from 0 (d'oh)
80 to 9, as they are not signifcant.
81 - new experimental module AnyEvent::MP::DataConn to establish
82 additional data connections between nodes.
83 - bin/aemp now loads AnyEvent::Watchdog::Util before trying
84 to call restart.
85 - bin/aemp setcert didn't properly handle the filename argument.
86 - removed empty and unused timeout= parameter from protocol greeting.
87
881.23 Fri Nov 6 18:46:26 CET 2009
89 - listener-less nodes were misinformed by their masters about
90 the location of new nodes, and therefore didn't connect
91 properly in all cases.
92 - messages send during node-up processing could sometimes get lost,
93 which would hamper global's ability to mesh the network.
94 - fixed AnyEvent::MP::Global::grp_reg to return a guard also
95 in scalar context.
96 - fixed AnyEvent::MP::Kernel::mon_nodes to return a guard also
97 in scalar context.
98 - try to improve error reporting when automatically loading
99 a function - being unable to load a module will now
100 stop the process and report the error.
101
1021.22 Sat Oct 17 03:41:47 CEST 2009
103 - fix two rcv-bugs: after adding a tagged rcv, the default
104 callback would not be used or changed anymore (reported
105 by Felix Antonius Wilhelm Ostmann, with apologies for
106 taking so long to releae the fix).
107 - added Coro::MP introduction to AnyEvent::MP::Intro.
108 - add peval function.
109
1101.21 Mon Oct 5 21:58:36 CEST 2009
111 - always delay kil messages.
112 - remote monitor requests were never cleaned up (found by
113 Sadrak and analysed by elmex).
114 - move $VERSION to AnyEvent::MP and AnyEvent::MP only.
115
1161.2 Thu Oct 1 22:59:35 CEST 2009
117 - fix a rather big bug where message processing could be
118 inhibited completely.
119 - grp_mon could not return a guard due to a typo.
120 - implement primitive aemp shell.
121 - do not use bound address as listen length (reported by
122 Felix Antonius Wilhelm Ostman).
123 - no longer export an undefined "reg" function from AE::MP.
124 - the mon_guard return value no longer keeps an additional
125 reference to the passed refs.
126
1271.1 Fri Sep 11 04:34:03 CEST 2009
128 - bumped the transport protocol to version 1 - sorry, but there
129 were too many bugfixes/changes.
130 - new function AnyEvent::MP::cal for simple rpc.
131 - renamed AnyEvent::MP::Global functions to grp_reg|get|mon.
132 - implemented monitoring for Global groups.
133 - removed all userspace time-outs from the transport, instead
134 rely on tcp retransmit timeouts and tcp keepalive.
135 - spawn now delays spawns on the local node artificially, so
136 it can return before invoking the init function.
137 - break endless recursion between ping-pong ports on the local
138 node after ~50 iterations.
139 - support JSON-encoded aemp arguments.
140 - added aemp restart.
141 - support for starting init functions with parameters added.
142 - data_format, auth_offer and auth_accept are now configurable, albeit
143 not documented.
144 - new service: AnyEvent::MP::LogCatcher.
145 - rely on perl srand'ing better than we could.
146 - try to backport to 5.8.2, unfortunately, recursive dependencies
147 rely on Module::Build, which fails at 5.8.7, so YMMV.
148 - use a vastly different reseeding strategy.
149 - use seed nodes to notify nodes about new nodes.
150 - in the case where a ndoe learns new addresses after it tried
151 connecting, prefer the new addrseses over the old ones.
152
1531.0 Wed Sep 2 20:40:43 CEST 2009
154 - randomise more connection times, but reduce delay to 0.2s (seeds)
155 and 0.05s (normal connects).
156 - upgrade to and require AnyEvent::Watchdog 1.0.
157
1580.95 Mon Aug 31 22:03:55 CEST 2009
159 - tutorial finally covers all the basics, maybe not optimally, but it
160 is covered.
161 - rename initialise_node to configure and accepts key => value
162 pairs for configuration.
163 - give aemp a full manpage.
164 - bin/aemp now expects comma-separated lists and supports multiple
165 subcommands.
166 - new aemp eval subcommand.
167 - new aemp parent subcommand.
168 - nodes get freed when no longer needed.
169 - support "*" hostname to specify all local interface addresses
170 (without localhost etc.).
171 - support "*" port to specify a dynamically-assigned port (which
172 is the default now).
173 - fix per-transport memory leak.
174
1750.9 Sat Aug 29 18:47:45 CEST 2009
176 - tutorial was updated to reflect the current API, but
177 does not touch monitoring yet.
178 - got rid of noderefs and slave nodes entirely, introduce
179 free-text node identifiers and anonymous nodes.
180 - document security issues with tls_anon method and introduce
181 optional tls_md6_64_256 method.
182
1830.8 Wed Aug 19 07:54:50 CEST 2009
184 - added AnyEvent::MP::Global service.
185 - support "package::" names as service names.
186 - many, many, many bugfixes and tweaks.
187
1880.7 Sat Aug 15 01:18:51 CEST 2009
189 - sorry, another major API simplification.
190 - greatly improved aemp configuration abilities.
191 - introduced profile management.
192 - updated exmaples in eg/.
193 - added Kernel::snd_to_func.
194
1950.6 Thu Aug 13 03:15:05 CEST 2009
196 - do away with .aemp-secret and use .perl-anyevent-mp as json config
197 file.
198 - introduce bin/aemp to configure nodes.
199 - too many changes to describe.
200
2010.4 Sun Aug 9 18:05:49 CEST 2009
202 - API should now be really stable :)
203 - distinguish between resolved and unresolved noderefs.
204 - certificates in .aemp-secret are now supported.
205 - change port creation semantics.
206 - replace become_xxx by initialise_node.
207 - implement AE::MP::spawn.
208 - implement transport timeouts, monitoring, reconnecting.
209 - simplify and partially document the transport protocol.
210
2110.1 Tue Aug 4 23:07:11 CEST 2009
212 - AnyEvent::MP API should be mostly stable by now.
4 213
50.02 Sun Aug 2 17:46:43 CEST 2009 2140.02 Sun Aug 2 17:46:43 CEST 2009
6 - design has stabilised a bit, sitll unusable. 215 - design has stabilised a bit, sitll unusable.
7 216
80.01 Sat Aug 1 11:43:31 CEST 2009 2170.01 Sat Aug 1 11:43:31 CEST 2009

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines