| 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. |
|
|
| 7 | |
6 | |
| 8 | TODO: explain service sin aemp |
7 | faq: can't se anything |
| 9 | TODO: document autocork and nodelay options, probably others. |
8 | faq: all is asynch |
|
|
9 | faq: how to interface to non-perl nodes? |
| 10 | |
10 | |
|
|
11 | TODO: maybe disbale warnings by default? |
|
|
12 | |
|
|
13 | - listener-less nodes had trouble sending keepalive |
|
|
14 | messages on write timeouts (found by Sadrak). |
|
|
15 | - the monitor guard could cause a memleak due |
|
|
16 | to autovivification if a mon was cleared after |
|
|
17 | the port was gone (analysed by Sadrak). |
|
|
18 | |
|
|
19 | 1.28 Thu Apr 1 21:23:54 CEST 2010 |
|
|
20 | - accepted conenctions didn't corretcly set up the SRCNODE |
|
|
21 | in some cases, leading to intra-node messages (such as monitoring) |
|
|
22 | to fail silently (great testcase by Sadrak). |
|
|
23 | |
|
|
24 | 1.27 Sat Mar 20 21:11:16 CET 2010 |
|
|
25 | - avoid exceptions in the transport code if the connection |
|
|
26 | fails during TLS negotiation. |
|
|
27 | - delay new address broadcasts in seednodes, to reduce the |
|
|
28 | likelyhood of nodes connecting at the same time, leading to |
|
|
29 | message loss. |
|
|
30 | - improve general message passing efficiency by 8%, and special |
|
|
31 | case json framing for a 50% speed increase. |
|
|
32 | - add hints how to interface to aemp nodes from other programming |
|
|
33 | languages using the aemp protocol in AnyEvent::MP::Transport |
|
|
34 | (it can really be quite simple). |
|
|
35 | - handle unsupported remote framings more gracefully. |
|
|
36 | |
|
|
37 | 1.26 Wed Dec 30 16:48:37 CET 2009 |
|
|
38 | - INCOMPATIBLE CHANGE: the high-level node protocol changed, |
|
|
39 | all nodes need to upgrade. |
| 11 | - remote kil's did not work (reported by Eugene Grigoriev). |
40 | - remote kil's did not work (reported by Eugene Grigoriev). |
|
|
41 | - documented monitor_timeout, connect_interval, auth_offer, |
|
|
42 | auth_accept, framing_format, autocork and nodelay options |
|
|
43 | and make them configurable via bin/aemp. |
|
|
44 | - some doc updates. |
|
|
45 | - remote kil's were not synchronous to message sends which affected |
|
|
46 | nobody, as remote kils were broken. |
|
|
47 | - remote mon/unmonitor requests were not synchronous to message |
|
|
48 | sends either sometimes, although it likely didnt matter. |
| 12 | |
49 | |
| 13 | 1.24 Wed Dec 9 14:59:52 CET 2009 |
50 | 1.24 Wed Dec 9 14:59:52 CET 2009 |
| 14 | - codename "the foul egg". |
51 | - codename "the foul egg". |
| 15 | - increased "unmonitored local port died" message level to 2, so |
52 | - increased "unmonitored local port died" message level to 2, so |
| 16 | it is displayed by default now. |
53 | it is displayed by default now. |
