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

Comparing AnyEvent-MP/Changes (file contents):
Revision 1.52 by root, Sat Sep 5 22:48:27 2009 UTC vs.
Revision 1.118 by root, Tue Feb 28 18:37:24 2012 UTC

1Revision history for AnyEvent::MP 1Revision history for AnyEvent::MP
2 2
3TODO: global protocol fix?
4TODO: testsuite 3TODO: testsuite
5TODO: make more parameters configurable: framing, tls, auth, timeouts
6TODO: intro: maybe simple job pool example? 4TODO: intro: maybe simple job pool example?
7TODO: intro: mention watchdog... 5TODO: intro: mention watchdog...
8TODO: aemp manpage more examples 6TODO: maybe disbale warnings by default?
9TODO: group monitoring 7TODO: networkdomains
8TODO: group proxy nodes - store group membership for other nodes
9TODO: switch to AnyEvent::Log
10 10
11TODO: cal, cal_func 11faq: can't se anything
12faq: all is asynch
13faq: how to interface to non-perl nodes?
14
15TODO: AnyEvent::MP gets grp_reg, grp_get
16TODO: AnyEvent::MP keeps list of all "global" nodes
17TODO: Global loaded on demand only, adds grp database and node address database
18TODO: transport connects use any global node to find addresses (maybe itself)
19TODO: not really
20
212
22 - INCOMPATIBLE CHANGE: the node protocol used by this version is not
23 compatible to versions <2 - you have to upgrade all node
24 - new "norc" and "force" options for MP::configure to ignore
25 or override the rc file.
26 - move seed code to Kernel.
27 - fixed small typo in MP::Transport.
28 - fix (unused) provider-kv (version sometimes missing).
29 - reduce the default export list of MP::Kernel.
30
311.30 Thu Jun 30 11:30:39 CEST 2011
32 - connection errors at the right time would kill node
33 connections in a bad way (patch by Malcolm Studd).
34
351.29 Fri May 7 20:13:39 CEST 2010
36 - codename "Sadrak".
37 - fix error in callback at AnyEvent/MP/Global.pm line 339
38 (found by Sadrak).
39 - listener-less nodes had trouble sending keepalive
40 messages on write timeouts (found by Sadrak).
41 - the monitor guard could cause a memleak due
42 to autovivification if a mon was cleared after
43 the port was gone (analysed by Sadrak).
44 - do not overwrite the config file if we couldn't read it
45 for some reason (lesson demonstrated by Sadrak).
46
471.28 Thu Apr 1 21:23:54 CEST 2010
48 - accepted connections didn't correctly set up the SRCNODE
49 in some cases, leading to intra-node messages (such as monitoring)
50 to fail silently (great testcase by Sadrak).
51
521.27 Sat Mar 20 21:11:16 CET 2010
53 - avoid exceptions in the transport code if the connection
54 fails during TLS negotiation.
55 - delay new address broadcasts in seednodes, to reduce the
56 likelyhood of nodes connecting at the same time, leading to
57 message loss.
58 - improve general message passing efficiency by 8%, and special
59 case json framing for a 50% speed increase.
60 - add hints how to interface to aemp nodes from other programming
61 languages using the aemp protocol in AnyEvent::MP::Transport
62 (it can really be quite simple).
63 - handle unsupported remote framings more gracefully.
64
651.26 Wed Dec 30 16:48:37 CET 2009
66 - INCOMPATIBLE CHANGE: the high-level node protocol changed,
67 all nodes need to upgrade.
68 - remote kil's did not work (reported by Eugene Grigoriev).
69 - documented monitor_timeout, connect_interval, auth_offer,
70 auth_accept, framing_format, autocork and nodelay options
71 and make them configurable via bin/aemp.
72 - some doc updates.
73 - remote kil's were not synchronous to message sends which affected
74 nobody, as remote kils were broken.
75 - remote mon/unmonitor requests were not synchronous to message
76 sends either sometimes, although it likely didnt matter.
77
781.24 Wed Dec 9 14:59:52 CET 2009
79 - codename "the foul egg".
80 - increased "unmonitored local port died" message level to 2, so
81 it is displayed by default now.
82 - generate random seed at configure time, not module loading time,
83 so fork&configure works as expected.
84 - decreased "seed"-related message priority in Global from 0 (d'oh)
85 to 9, as they are not signifcant.
86 - new experimental module AnyEvent::MP::DataConn to establish
87 additional data connections between nodes.
88 - bin/aemp now loads AnyEvent::Watchdog::Util before trying
89 to call restart.
90 - bin/aemp setcert didn't properly handle the filename argument.
91 - removed empty and unused timeout= parameter from protocol greeting.
92
931.23 Fri Nov 6 18:46:26 CET 2009
94 - listener-less nodes were misinformed by their masters about
95 the location of new nodes, and therefore didn't connect
96 properly in all cases.
97 - messages send during node-up processing could sometimes get lost,
98 which would hamper global's ability to mesh the network.
99 - fixed AnyEvent::MP::Global::grp_reg to return a guard also
100 in scalar context.
101 - fixed AnyEvent::MP::Kernel::mon_nodes to return a guard also
102 in scalar context.
103 - try to improve error reporting when automatically loading
104 a function - being unable to load a module will now
105 stop the process and report the error.
106
1071.22 Sat Oct 17 03:41:47 CEST 2009
108 - fix two rcv-bugs: after adding a tagged rcv, the default
109 callback would not be used or changed anymore (reported
110 by Felix Antonius Wilhelm Ostmann, with apologies for
111 taking so long to releae the fix).
112 - added Coro::MP introduction to AnyEvent::MP::Intro.
113 - add peval function.
114
1151.21 Mon Oct 5 21:58:36 CEST 2009
116 - always delay kil messages.
117 - remote monitor requests were never cleaned up (found by
118 Sadrak and analysed by elmex).
119 - move $VERSION to AnyEvent::MP and AnyEvent::MP only.
120
1211.2 Thu Oct 1 22:59:35 CEST 2009
122 - fix a rather big bug where message processing could be
123 inhibited completely.
124 - grp_mon could not return a guard due to a typo.
125 - implement primitive aemp shell.
126 - do not use bound address as listen length (reported by
127 Felix Antonius Wilhelm Ostman).
128 - no longer export an undefined "reg" function from AE::MP.
129 - the mon_guard return value no longer keeps an additional
130 reference to the passed refs.
131
1321.1 Fri Sep 11 04:34:03 CEST 2009
133 - bumped the transport protocol to version 1 - sorry, but there
134 were too many bugfixes/changes.
135 - new function AnyEvent::MP::cal for simple rpc.
12 - renamed AnyEvent::MP::Global functions to grp_reg|get|mon. 136 - renamed AnyEvent::MP::Global functions to grp_reg|get|mon.
137 - implemented monitoring for Global groups.
13 - removed all userspace time-outs from the transport, instead 138 - removed all userspace time-outs from the transport, instead
14 rely on tcp retransmit timeouts and tcp keepalive. 139 rely on tcp retransmit timeouts and tcp keepalive.
15 - spawn now delays spawns on the local node artificially, so 140 - spawn now delays spawns on the local node artificially, so
16 it can return before invoking the init function. 141 it can return before invoking the init function.
142 - break endless recursion between ping-pong ports on the local
143 node after ~50 iterations.
17 - support JSON-encoded aemp arguments. 144 - support JSON-encoded aemp arguments.
18 - added aemp restart. 145 - added aemp restart.
19 - support for starting init functions with parameters added. 146 - support for starting init functions with parameters added.
147 - data_format, auth_offer and auth_accept are now configurable, albeit
148 not documented.
149 - new service: AnyEvent::MP::LogCatcher.
150 - rely on perl srand'ing better than we could.
151 - try to backport to 5.8.2, unfortunately, recursive dependencies
152 rely on Module::Build, which fails at 5.8.7, so YMMV.
153 - use a vastly different reseeding strategy.
154 - use seed nodes to notify nodes about new nodes.
155 - in the case where a ndoe learns new addresses after it tried
156 connecting, prefer the new addrseses over the old ones.
20 157
211.0 Wed Sep 2 20:40:43 CEST 2009 1581.0 Wed Sep 2 20:40:43 CEST 2009
22 - randomise more connection times, but reduce delay to 0.2s (seeds) 159 - randomise more connection times, but reduce delay to 0.2s (seeds)
23 and 0.05s (normal connects). 160 and 0.05s (normal connects).
24 - upgrade to and require AnyEvent::Watchdog 1.0. 161 - upgrade to and require AnyEvent::Watchdog 1.0.

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines