Re: Alarm Leaks (was: Re: [sa-list] Re: Spamd keeps getting hung up!)

[Justin Mason <jm...@jmason.org>]

Daryl C. W. O'Shea writes:
> Justin Mason wrote:
> > Dan Mahoney, System Admin writes:
> >> And 4) This whenever it gets hung up (those __alarm__ things couldn't be 
> >> more vague, could they?):
> >>
> >> Mar  2 12:53:07 quark spamd[94770]: __alarm__
> >> Mar  2 12:53:07 quark spamd[94770]: __alarm__
> > 
> > that's with the bug 4696 patches applied?  they should not be sneaking
> > through if so.
> > 
> > If *that* one issue still persists after the others are cleaned
> > up, it'd be worth commenting on bug 4696.
> 
> Oops, I forgot to mention this before. :(
> 
> Alarms have been leaking since M::SA::Timeout was added.
> 
> I haven't looked into exactly why it's leaking from the timeout code, 
> but I know that the last ditch "caller" check that I added to Logger.pm 
> isn't catching it since it's looking for caller[3] and not caller[4].

it might be worth opening this as a new bug.

--j.