Re: 3.0.1 regressions

[Mark Mitchell <mark at codesourcery dot com>]

--On Monday, August 13, 2001 02:04:18 PM -0400 David Ronis 
<ronis@ronispc.chem.mcgill.ca> wrote:

>
> Mark Mitchell writes:
>  >
>  >
>  >
>  > --On Monday, August 13, 2001 12:53:30 PM -0400 David Ronis
>  > <ronis@ronispc.chem.mcgill.ca> wrote:
>  >
>  > >
>  > >
>  > > Hi Mark,
>  > >
>  > > Have the problems with fomit-frame-pointer, optimization and ix86's
>  > > been fixed?  It doesn't seem so from gnats:
>  >
>  > These are not regressions from GCC 3.0, if I understand correctly.  So,
>  > while they are very important, and should definitely be fixed, they
> need  > not hold up GCC 3.0.1.
>
> The are not regressions from 3.0, in that 3.0 had these broken
> already.  On the other hand, I think I'm missing something.  3.0.1
> will only fix what? further regressions in the 3.0.1 snapshots?
> Wasn't it supposed to address the 2.95.3->3.0 regressions?  (that's
> what you get for not following that part of the mailing list).
>

There are two categories of things to fix on the 3.0 branch: regressions
from previous release series (e.g., 2.9x, 2.7.x) and regressions from
the preceding 3.0 releases.  I will try very hard to avoid shipping
a 3.0.x that regresses from another 3.0 release; that would mean that
even if the 3.0 series was working for you up until now you cannot
upgrade to the latest minor version.  This is very bad.

We also *want* to fix regressions from previous major releases, and we
should try hard to fix them -- but there's no reason not to ship 3.0.1
just because it doesn't fix these.  If 3.0 worked for you, so will 3.0.1.
If 3.0 didn't, then you're no worse off -- and we will try to fix them
in 3.0.2.


-- 
Mark Mitchell                mark@codesourcery.com
CodeSourcery, LLC            http://www.codesourcery.com