4.2 Backport fix for PR libfortran/31196
Steve Kargl
sgk@troutmask.apl.washington.edu
Tue Apr 17 00:43:00 GMT 2007
On Mon, Apr 16, 2007 at 11:52:19PM +0000, Joseph S. Myers wrote:
> On Mon, 16 Apr 2007, Steve Kargl wrote:
>
> > This is the proposed development model
> >
> > |
> > +-- GCC 4.2 branch created -- (20 Oct 06) ----+
> > | |
> > Mainline Frozen (20 Oct 06) |
> > | v
> > | GCC 4.2.0 release (4 Nov 06)
>
> Without relaxing the quality standard for releases to essentially that
> applied for branching (100 P1/P2 regressions), this is hopelessly
> unrealistic.
If the rules for Stage 1, Stage 2, and Stage 3 are actually followed,
then it is realistic. If a project meant for 4.3 Stage 1 misses the
Stage 1 time window, then it simply is not allowed to be merged in
Stage 2 and definitely not in Stage 3.
> To release a week after branching, you must branch at a point that is
> release ready. Thus, we'd have ended up branching around now, with
> mainline in release branch mode for six more months.
If we strictly follow the Stage guidelines, then I think it's doable.
> > This model can work if all the cats can be herded for 1 or 2 weeks.
>
> Volunteer effort is not redirectable that way, although the fallacy that
> it is is not original to GCC. Freezing mainline will lose the effort that
> would have been directed there, or maybe direct some of it to the same
> work but on development branches; the proportion redirected to release
> branch work will be negligible.
>
I freely admit that it is a paradigm shift, but the current model isn't
working. Having Mark essentially begging people to look at 4.2 regressions
isn't a very good development model.
--
Steve
More information about the Gcc-patches
mailing list