| 1 | Revision history for AnyEvent::MP |
1 | Revision history for AnyEvent::MP |
| 2 | |
2 | |
| 3 | TODO: global protocol fix? |
|
|
| 4 | TODO: use seednodes only to broadcast? |
|
|
| 5 | TODO: testsuite |
3 | TODO: testsuite |
| 6 | TODO: intro: maybe simple job pool example? |
4 | TODO: intro: maybe simple job pool example? |
| 7 | TODO: intro: mention watchdog... |
5 | TODO: intro: mention watchdog... |
| 8 | TODO: group monitoring |
6 | TODO: use gvpe method of making contact, if possible. |
| 9 | |
7 | |
| 10 | TODO: cal, cal_func |
8 | TODO: explain service sin aemp |
| 11 | TODO: mon_nodes should init-call for all up nodes |
9 | - fixed AnyEvent::MP::Global::grp_reg to return a guard also |
| 12 | TODO: known_nodes no longer relevant => connected_nodes? |
10 | in scalar context. |
| 13 | TODO: spawn does not create remote port dummy entry |
11 | - listener-less nodes were misinformed by their masters about |
| 14 | TODO: better autoseeding, timout*(nconnect+1) |
12 | the location of new nodes, and therefore didn't connect |
|
|
13 | properly in all cases. |
|
|
14 | - try to improve error reporting when automatically loading |
|
|
15 | a function - being unable to load a module will now |
|
|
16 | stop the process and report the error. |
|
|
17 | |
|
|
18 | 1.22 Sat Oct 17 03:41:47 CEST 2009 |
|
|
19 | - fix two rcv-bugs: after adding a tagged rcv, the default |
|
|
20 | callback would not be used or changed anymore (reported |
|
|
21 | by Felix Antonius Wilhelm Ostmann, with apologies for |
|
|
22 | taking so long to releae the fix). |
|
|
23 | - added Coro::MP introduction to AnyEvent::MP::Intro. |
|
|
24 | - add peval function. |
|
|
25 | |
|
|
26 | 1.21 Mon Oct 5 21:58:36 CEST 2009 |
|
|
27 | - always delay kil messages. |
|
|
28 | - remote monitor requests were never cleaned up (found by |
|
|
29 | Sadrak and analysed by elmex). |
|
|
30 | - move $VERSION to AnyEvent::MP and AnyEvent::MP only. |
|
|
31 | |
|
|
32 | 1.2 Thu Oct 1 22:59:35 CEST 2009 |
|
|
33 | - fix a rather big bug where message processing could be |
|
|
34 | inhibited completely. |
|
|
35 | - grp_mon could not return a guard due to a typo. |
|
|
36 | - implement primitive aemp shell. |
|
|
37 | - do not use bound address as listen length (reported by |
|
|
38 | Felix Antonius Wilhelm Ostman). |
|
|
39 | - no longer export an undefined "reg" function from AE::MP. |
|
|
40 | - the mon_guard return value no longer keeps an additional |
|
|
41 | reference to the passed refs. |
|
|
42 | |
|
|
43 | 1.1 Fri Sep 11 04:34:03 CEST 2009 |
|
|
44 | - bumped the transport protocol to version 1 - sorry, but there |
|
|
45 | were too many bugfixes/changes. |
|
|
46 | - new function AnyEvent::MP::cal for simple rpc. |
| 15 | - renamed AnyEvent::MP::Global functions to grp_reg|get|mon. |
47 | - renamed AnyEvent::MP::Global functions to grp_reg|get|mon. |
|
|
48 | - implemented monitoring for Global groups. |
| 16 | - removed all userspace time-outs from the transport, instead |
49 | - removed all userspace time-outs from the transport, instead |
| 17 | rely on tcp retransmit timeouts and tcp keepalive. |
50 | rely on tcp retransmit timeouts and tcp keepalive. |
| 18 | - spawn now delays spawns on the local node artificially, so |
51 | - spawn now delays spawns on the local node artificially, so |
| 19 | it can return before invoking the init function. |
52 | it can return before invoking the init function. |
| 20 | - break endless recursion between ping-pong ports on the local |
53 | - break endless recursion between ping-pong ports on the local |
| … | |
… | |
| 23 | - added aemp restart. |
56 | - added aemp restart. |
| 24 | - support for starting init functions with parameters added. |
57 | - support for starting init functions with parameters added. |
| 25 | - data_format, auth_offer and auth_accept are now configurable, albeit |
58 | - data_format, auth_offer and auth_accept are now configurable, albeit |
| 26 | not documented. |
59 | not documented. |
| 27 | - new service: AnyEvent::MP::LogCatcher. |
60 | - new service: AnyEvent::MP::LogCatcher. |
|
|
61 | - rely on perl srand'ing better than we could. |
|
|
62 | - try to backport to 5.8.2, unfortunately, recursive dependencies |
|
|
63 | rely on Module::Build, which fails at 5.8.7, so YMMV. |
|
|
64 | - use a vastly different reseeding strategy. |
|
|
65 | - use seed nodes to notify nodes about new nodes. |
|
|
66 | - in the case where a ndoe learns new addresses after it tried |
|
|
67 | connecting, prefer the new addrseses over the old ones. |
| 28 | |
68 | |
| 29 | 1.0 Wed Sep 2 20:40:43 CEST 2009 |
69 | 1.0 Wed Sep 2 20:40:43 CEST 2009 |
| 30 | - randomise more connection times, but reduce delay to 0.2s (seeds) |
70 | - randomise more connection times, but reduce delay to 0.2s (seeds) |
| 31 | and 0.05s (normal connects). |
71 | and 0.05s (normal connects). |
| 32 | - upgrade to and require AnyEvent::Watchdog 1.0. |
72 | - upgrade to and require AnyEvent::Watchdog 1.0. |
