You may try the patch (ATC.diff) I sent a few days ago. I would be interested if this would fix it. I wonder if you see any of the diagnostics "..... already known" in the output.
Olaf
2006/3/18, Chris Metzler <[EMAIL PROTECTED]>:
On Fri, 17 Mar 2006 19:19:19 +0100
Melchior FRANZ wrote:
>
> * Melchior FRANZ -- Thursday 09 March 2006 19:43:
> > (A) tower.cxx/AI
>
> Probably not yet fixed, but it has become *very* rare. I'll keep Olaf
> informed next time it happens.
I think it still happens to me on about one out of every 20
approaches or so. Happened to me night before last.
