pr83 random failure

[Michael Ritzert]

On Sat, Sep 25, 2004 at 10:32:45AM +0100, Andrew Haley wrote:
> Maybe.  However, I didn't see it with yesterday's CVS head.  Exactly
> what platform/gcc source version were you using?

Well, that's what random failures are about.  I cannot reproduce it myself
reliably.  The regression tester keeps on complaining that it fails.  Then
usually it's gone with the next run, sometimes without commits in between.
I'm running i686-pc-linux-gnu (FC1 and debian) if it makes a difference.

See http://gcc.gnu.org/ml/gcc-regression/2004-09/msg00197.html for the
ChangeLog heads from the latest failure.

Michael