Time for a new snapshot, some stuff accumulated.
The timer corruption bug is still not fixed...
No news on the bug, so to relax:
A DBus interface, some small fixes, refactoring
some small but important fixes (missing string termination), but...
I'm still chasing an ugly race condition. timeout tree corruption is the most visible result. Some connection gets removed, but the connection timeout not (properly) removed, and after the connection struct is cleared -> tree broken.
In valgrind it looks like use after free, which may point at hzp, but valgrind itself is more like "giving up". First you see nothing interresting, suddenly -> boom, valgrind blows up or the code, then valgrind. Sometimes it even looks like a double free, but that looks more like the result of valgrind totaly going belly up.
My debugging is failing me, looks like i should have not angered the gods:
"Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." Brian W. Kernighan
important fixes over the last snapshot, fixes for x86_64 and so on...
New features, new bugs, yiepiee!
Changes in UDP buffer handling and packet reception.
Expermimental support for the new Linux 2.6.33 recvmmsg().
Changes to the bind address handling.
New config option for the default port.
Some little changes, finally with a "working" install.
Also binary packages for several distros are now available
(note: They are prop. broken :-P ).
The last source snapshot contains some important bugfixes.
Also, thanks to autoconf, compilation is largely simplified.