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

Comparing AnyEvent-MP/Changes (file contents):
Revision 1.87 by root, Sat Nov 28 12:40:21 2009 UTC vs.
Revision 1.124 by root, Sat Mar 3 13:07:19 2012 UTC

1Revision history for AnyEvent::MP 1Revision history for AnyEvent::MP
2 2
3TODO: testsuite 3TODO: testsuite
4TODO: intro: maybe simple job pool example? 4TODO: intro: maybe simple job pool example?
5TODO: intro: mention watchdog... 5TODO: intro: mention watchdog...
6TODO: use gvpe method of making contact, if possible. 6TODO: maybe disbale warnings by default?
7TODO: listener-scopes (10.0.0.1:4040@vpn) and connect-scopes ("vpn,public")
7 8
8TODO: explain service sin aemp 9faq: can't se anything
10faq: all is asynch
11faq: how to interface to non-perl nodes?
12
13TODO: update docs
14TODO: secure - do not allow remote code execution
15TODO: switch to AnyEvent::Log
16TODO: Global loaded on demand only, adds grp database and node address database
17TODO: transport connects use any global node to find addresses (maybe itself)
18TODO: maybe check g_proto version, if sent? how about node protocol?
19TODO: connect instantly in node->connect, not with timer-0
20TODO: maybe switch to md5+sha1+sha256 or so...
21TODO: up_nodes for aemp-trace replace by rpc call?
222
23 - INCOMPATIBLE CHANGE: the node protocol used by this version is not
24 compatible to versions <2 - you have to upgrade all nodes.
25 - INCOMPATIBLE CHANGE: many APIs have changed, see the section
26 "PORTING FROM VERSION 1" in the AnyEvent::MP manpage for details.
27 - AnyEvent::MP::Global is now optional - only seed nodes autoload it
28 and create a fully-meshed network with each other, normal nodes only
29 have to connect to seed nodes, other connections are on-demand.
30 - new "norc" and "force" options for MP::configure to ignore
31 or override the rc file.
32 - move seed code to Kernel.
33 - fixed small typo in MP::Transport.
34 - fix (unused) provider-kv (version sometimes missing).
35 - reduce the default export list of MP::Kernel.
36 - reduced various random delays to zero, as they
37 are hopefully no longer needed with the new global
38 protocol.
39 - do not use GMP for generating alphanumeric strings even when
40 available, use a more tricky algorithm instead.
41 - anonymous node names and remote port names have been shortened
42 considerably.
43 - aemp shell now supports "package" selection and offers an
44 $ECHO port you cna send messages to.
45 - rely on perl rand() instead of using /dev/urandom when available, as
46 hopefully it is random enough.
47 - aemp shell no longer leaks one port per command.
48 - known_nodes is gone, it has little value, use all_nodes as
49 replacement.
50
511.30 Thu Jun 30 11:30:39 CEST 2011
52 - connection errors at the right time would kill node
53 connections in a bad way (patch by Malcolm Studd).
54
551.29 Fri May 7 20:13:39 CEST 2010
56 - codename "Sadrak".
57 - fix error in callback at AnyEvent/MP/Global.pm line 339
58 (found by Sadrak).
59 - listener-less nodes had trouble sending keepalive
60 messages on write timeouts (found by Sadrak).
61 - the monitor guard could cause a memleak due
62 to autovivification if a mon was cleared after
63 the port was gone (analysed by Sadrak).
64 - do not overwrite the config file if we couldn't read it
65 for some reason (lesson demonstrated by Sadrak).
66
671.28 Thu Apr 1 21:23:54 CEST 2010
68 - accepted connections didn't correctly set up the SRCNODE
69 in some cases, leading to intra-node messages (such as monitoring)
70 to fail silently (great testcase by Sadrak).
71
721.27 Sat Mar 20 21:11:16 CET 2010
73 - avoid exceptions in the transport code if the connection
74 fails during TLS negotiation.
75 - delay new address broadcasts in seednodes, to reduce the
76 likelyhood of nodes connecting at the same time, leading to
77 message loss.
78 - improve general message passing efficiency by 8%, and special
79 case json framing for a 50% speed increase.
80 - add hints how to interface to aemp nodes from other programming
81 languages using the aemp protocol in AnyEvent::MP::Transport
82 (it can really be quite simple).
83 - handle unsupported remote framings more gracefully.
84
851.26 Wed Dec 30 16:48:37 CET 2009
86 - INCOMPATIBLE CHANGE: the high-level node protocol changed,
87 all nodes need to upgrade.
88 - remote kil's did not work (reported by Eugene Grigoriev).
89 - documented monitor_timeout, connect_interval, auth_offer,
90 auth_accept, framing_format, autocork and nodelay options
91 and make them configurable via bin/aemp.
92 - some doc updates.
93 - remote kil's were not synchronous to message sends which affected
94 nobody, as remote kils were broken.
95 - remote mon/unmonitor requests were not synchronous to message
96 sends either sometimes, although it likely didnt matter.
97
981.24 Wed Dec 9 14:59:52 CET 2009
9 - codename "the foul egg". 99 - codename "the foul egg".
10 - increased "unmonitored local port died" message level to 2, so 100 - increased "unmonitored local port died" message level to 2, so
11 it is displayed by default now. 101 it is displayed by default now.
102 - generate random seed at configure time, not module loading time,
103 so fork&configure works as expected.
104 - decreased "seed"-related message priority in Global from 0 (d'oh)
105 to 9, as they are not signifcant.
12 - new experimental module AnyEvent::MP::DataConn to establish 106 - new experimental module AnyEvent::MP::DataConn to establish
13 additional data connections between nodes. 107 additional data connections between nodes.
14 - bin/aemp now loads AnyEvent::Watchdog::Util before trying 108 - bin/aemp now loads AnyEvent::Watchdog::Util before trying
15 to call restart. 109 to call restart.
16 - bin/aemp setcert didn't properly handle the filename argument. 110 - bin/aemp setcert didn't properly handle the filename argument.
111 - removed empty and unused timeout= parameter from protocol greeting.
17 112
181.23 Fri Nov 6 18:46:26 CET 2009 1131.23 Fri Nov 6 18:46:26 CET 2009
19 - listener-less nodes were misinformed by their masters about 114 - listener-less nodes were misinformed by their masters about
20 the location of new nodes, and therefore didn't connect 115 the location of new nodes, and therefore didn't connect
21 properly in all cases. 116 properly in all cases.

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines