| 1 | Revision history for AnyEvent::MP |
1 | Revision history for AnyEvent::MP |
| 2 | |
2 | |
| 3 | TODO: testsuite |
3 | TODO: testsuite |
| 4 | TODO: intro: maybe simple job pool example? |
4 | TODO: intro: maybe simple job pool example? |
| 5 | TODO: intro: mention watchdog... |
5 | TODO: intro: mention watchdog... |
| 6 | TODO: use gvpe method of making contact, if possible. |
6 | TODO: maybe disbale warnings by default? |
|
|
7 | TODO: listener-scopes (10.0.0.1:4040@vpn) and connect-scopes ("vpn,public") |
| 7 | |
8 | |
| 8 | TODO: explain service sin aemp |
9 | faq: can't se anything |
|
|
10 | faq: all is asynch |
|
|
11 | faq: how to interface to non-perl nodes? |
|
|
12 | |
|
|
13 | TODO: update docs |
|
|
14 | TODO: secure - do not allow remote code execution |
|
|
15 | TODO: switch to AnyEvent::Log |
|
|
16 | TODO: Global loaded on demand only, adds grp database and node address database |
|
|
17 | TODO: transport connects use any global node to find addresses (maybe itself) |
|
|
18 | TODO: maybe check g_proto version, if sent? how about node protocol? |
|
|
19 | TODO: connect instantly in node->connect, not with timer-0 |
|
|
20 | TODO: maybe switch to md5+sha1+sha256 or so... |
|
|
21 | 2 |
|
|
22 | - INCOMPATIBLE CHANGE: the node protocol used by this version is not |
|
|
23 | compatible to versions <2 - you have to upgrade all nodes. |
|
|
24 | - AnyEvent::MP::Global is now optional - only seed nodes autoload it |
|
|
25 | and create a fully-meshed network with each other, normal nodes only |
|
|
26 | have to connect to seed nodes, other connections are on-demand. |
|
|
27 | - new "norc" and "force" options for MP::configure to ignore |
|
|
28 | or override the rc file. |
|
|
29 | - move seed code to Kernel. |
|
|
30 | - fixed small typo in MP::Transport. |
|
|
31 | - fix (unused) provider-kv (version sometimes missing). |
|
|
32 | - reduce the default export list of MP::Kernel. |
|
|
33 | - reduced various random delays to zero, as they |
|
|
34 | are hopefully no longer needed with the new global |
|
|
35 | protocol. |
|
|
36 | - do not use GMP for generating alphanumeric strings, |
|
|
37 | use base64 with _ and - as extra chars instead. |
|
|
38 | - anonymous node names and remote port names have been shortened |
|
|
39 | considerably. |
|
|
40 | |
|
|
41 | 1.30 Thu Jun 30 11:30:39 CEST 2011 |
|
|
42 | - connection errors at the right time would kill node |
|
|
43 | connections in a bad way (patch by Malcolm Studd). |
|
|
44 | |
|
|
45 | 1.29 Fri May 7 20:13:39 CEST 2010 |
|
|
46 | - codename "Sadrak". |
|
|
47 | - fix error in callback at AnyEvent/MP/Global.pm line 339 |
|
|
48 | (found by Sadrak). |
|
|
49 | - listener-less nodes had trouble sending keepalive |
|
|
50 | messages on write timeouts (found by Sadrak). |
|
|
51 | - the monitor guard could cause a memleak due |
|
|
52 | to autovivification if a mon was cleared after |
|
|
53 | the port was gone (analysed by Sadrak). |
|
|
54 | - do not overwrite the config file if we couldn't read it |
|
|
55 | for some reason (lesson demonstrated by Sadrak). |
|
|
56 | |
|
|
57 | 1.28 Thu Apr 1 21:23:54 CEST 2010 |
|
|
58 | - accepted connections didn't correctly set up the SRCNODE |
|
|
59 | in some cases, leading to intra-node messages (such as monitoring) |
|
|
60 | to fail silently (great testcase by Sadrak). |
|
|
61 | |
|
|
62 | 1.27 Sat Mar 20 21:11:16 CET 2010 |
|
|
63 | - avoid exceptions in the transport code if the connection |
|
|
64 | fails during TLS negotiation. |
|
|
65 | - delay new address broadcasts in seednodes, to reduce the |
|
|
66 | likelyhood of nodes connecting at the same time, leading to |
|
|
67 | message loss. |
|
|
68 | - improve general message passing efficiency by 8%, and special |
|
|
69 | case json framing for a 50% speed increase. |
|
|
70 | - add hints how to interface to aemp nodes from other programming |
|
|
71 | languages using the aemp protocol in AnyEvent::MP::Transport |
|
|
72 | (it can really be quite simple). |
|
|
73 | - handle unsupported remote framings more gracefully. |
|
|
74 | |
|
|
75 | 1.26 Wed Dec 30 16:48:37 CET 2009 |
|
|
76 | - INCOMPATIBLE CHANGE: the high-level node protocol changed, |
|
|
77 | all nodes need to upgrade. |
|
|
78 | - remote kil's did not work (reported by Eugene Grigoriev). |
|
|
79 | - documented monitor_timeout, connect_interval, auth_offer, |
|
|
80 | auth_accept, framing_format, autocork and nodelay options |
|
|
81 | and make them configurable via bin/aemp. |
|
|
82 | - some doc updates. |
|
|
83 | - remote kil's were not synchronous to message sends which affected |
|
|
84 | nobody, as remote kils were broken. |
|
|
85 | - remote mon/unmonitor requests were not synchronous to message |
|
|
86 | sends either sometimes, although it likely didnt matter. |
|
|
87 | |
|
|
88 | 1.24 Wed Dec 9 14:59:52 CET 2009 |
|
|
89 | - codename "the foul egg". |
|
|
90 | - increased "unmonitored local port died" message level to 2, so |
|
|
91 | it is displayed by default now. |
|
|
92 | - generate random seed at configure time, not module loading time, |
|
|
93 | so fork&configure works as expected. |
|
|
94 | - decreased "seed"-related message priority in Global from 0 (d'oh) |
|
|
95 | to 9, as they are not signifcant. |
|
|
96 | - new experimental module AnyEvent::MP::DataConn to establish |
|
|
97 | additional data connections between nodes. |
|
|
98 | - bin/aemp now loads AnyEvent::Watchdog::Util before trying |
|
|
99 | to call restart. |
|
|
100 | - bin/aemp setcert didn't properly handle the filename argument. |
|
|
101 | - removed empty and unused timeout= parameter from protocol greeting. |
|
|
102 | |
|
|
103 | 1.23 Fri Nov 6 18:46:26 CET 2009 |
| 9 | - listener-less nodes were misinformed by their masters about |
104 | - listener-less nodes were misinformed by their masters about |
| 10 | the location of new nodes, and therefore didn't connect |
105 | the location of new nodes, and therefore didn't connect |
| 11 | properly in all cases. |
106 | properly in all cases. |
| 12 | - messages send during node-up processing could sometimes get lost, |
107 | - messages send during node-up processing could sometimes get lost, |
| 13 | which would hamper global's ability to mesh the network. |
108 | which would hamper global's ability to mesh the network. |
