On Fri, May 17, 2002 at 09:44:40AM +0100, Philip Hazel wrote:
> I'm afraid I can't remember the exact detail of this. One would have to
> run comparisons on the code for the different versions. Sorry, but my
> brain is now running well ahead, on Exim 4, as you might guess...
I was pretty much sure that would be the case, now.
> > The bug does not occur 100% of the time. I tried about 10 times and got this
> > behaviour on perhaps 8 occasions.
>
> I would be suspicious of actual disk errors. Do you have any logs that
> might record actual hardware problems?
This was my reaction too, although there is no evidence, yet, to support this
hypothesis.
> The first step, I am sure, is to upgrade to the latest Exim 3 (3.36) and
> see if the error persists. If it does, see if you can reproduce it while
> connected to a testing daemon (on a non-standard port) that has -d9 set
> so as to capture the debugging output. [You don't actually have to
> upgrade; just compiling 3.36 and running it for testing would do.]
This is exactly how I gleaned the output from before.
The same behaviour is not apparent under 3.36 so I guess whatever was causing
this in 3.22 has been fixed since.
Thanks for taking the time to get back to me.
Yours,
Ollie
--
Oliver Cook Systems Administrator, ClaraNET
ollie@??? 020 7903 3065
