loop regression in last 30 hrs.

[Robert Lipe]

> > +FAIL: gcc.c-torture/execute/loop-2b.c execution,  -O3 -g 
> 
> loop-2b.c fails or doesn't fail according to random influences.  The
> underlying bug got never fixed, so it depends on all kinds of
> whacky factors if it gets triggered or not.

I trust your analysis.  Since this is the first failure I've seen since
4/30, I thought I'd mention it.

> >  FAIL: gcc.dg/980312-1.c (test for excess errors)
> >  FAIL: gcc.dg/980313-1.c (test for excess errors)
> >  FAIL: gcc.dg/990119-1.c execution test
> 
> What does gcc.log say about the errors there?

Those are actually context from the context diff on the logfilesa

The first two are lameness in the native assemblers.  These are the
first testsuite cases to emit P5 (P6?) opcodes and they make my
assembler whine.  We're working on a plan to address this.

The 990119 case has failed since it was added.   No obvious clues in
the logfiles.

Thanx,

RJL