CHANGELOG

changeset 13
3e2ea1e7b2e8
child 16
742a6620b378
equal deleted inserted replaced
12:a9b590350c03 13:3e2ea1e7b2e8
1 ======
2 0.1.2 - Revision 15 - 2007-08-18
3 + Setup system to use new coro management as described in COROUTINE_MANAGEMENT
4 The callbacks are called from the event_loop 'thread' rather than that which they are
5 created in. This will prevent the self-resume problem as well as dead-thread problems.
6 - Recognized issues to fix in next release:
7 * Socket/event closing needs to be cleaned
8 * luaevent.lua needs refactoring
9 * luaevent.[ch] need to be cleaned up
10 ======
11 0.1.1 - Revision 14 - 2007-06-13
12 + Fixed event-handling code to cancel events on nothing being returned
13 + Added socket/object cleanup.
14 + Filed bug to libevent about the strange valgrind-released errors
15 - Recognized following issues:
16 Timeouts needed
17 Need to handle events setup from inside a coroutine... need to get a global Lua state from a thread...
18 ======
19 0.1.0 - Revision 6 - 2007-06-10 22:00 EST
20 Completed mostly working version
21 * Moved to a mode where addevent calls a callback rather than it being instantiated within.
22 If the callback returns -1, then no event is ever setup,
23 Otherwise the integer value is used to setup the event.
24 This allows for using coroutine.wrap rather than a cooked-up wrapper
25 * Tests work, although there are a few remaining issues:
26 * Need to figure a good way of preserving the event object,
27 not sure if current method is good enough, since the socket
28 is the only anchor, and it is only held inside the coro..
29 circular reference, something that Lua 'handles' well.
30 * Doing more than the maximum sockets the process is allows
31 causes strangeness to occur in libevent.. somehow
32 it is getting around to epoll_add which is causing valgrind
33 to barf.
34 * Added cheap protection code for failures in callback handlers
35 ======
36 0.0.0 - Revision 2 - 2007-06-10 12:00 EST
37 Initial public version, was broken due to self-resume coroutines

mercurial