3.3 regression, PR 11366 (was Re: GCC Release Status (2003-07-11))

[Mark Mitchell]

On Wed, 2003-07-16 at 19:58, Phil Edwards wrote:
> Mark,
> 
> What's your opinion about PR optimization/11366?  This was a killer for
> Debian users when 3.3 was released.  The only known workaround to date is
> to use -O1 instead of -O2.
> 
> It's got a wrong-code keyword and "[3.3 regression]" in the summary, but I
> don't know whether that shows up in the should-3.3.1-go-out-yet bug queries.

Yup, that's on the list. 

The problem with that PR is that there's no good test case in the PR. 
Until somebody comes up with one, it's going to be nearly impossible to
deal with the bug.

Thanks for pointing it out!

-- 
Mark Mitchell <mark@codesourcery.com>
CodeSourcery, LLC