I confirm that the glibc bug is now fixed in svn and that the other "bug" depended from
my old version of SA. Thanks a lot for taking up this project.
Giorgio
On 10:38 Sat 06 Jan , Mike Frysinger wrote:
> On Saturday 06 January 2007 09:59, Giorgio Lando wrote:
> > Thank you! I will try it. Actually I need to update my SA.
> > As for the first bug, here I am able to use 4 facilities in one rule
> > without that error, just the fifth determines it.
>
> as i said, it's not really deterministic ... many factors decide when the
> thing will actually crash
>
> if you look at the code, you'd see that starting with the 2nd facitlity,
> metalog would start clobbering memory it didnt allocate
>
> > Probably, once fixed this problem, 0.8 could be released.
>
> for sure, i need to fix the hangs in the signal handlers ... that's a pretty
> nasty bug (if you use a command directive, there's a not-too-hard-to-trigger
> race condition which would hang metalog and thus hang anything that tried to
> do syslog())
> -mike
|
