Re: embedded target breakage [Was: Re: Patch commit requirements]

[Toshi Morita <tm2 at best dot com>]

> Toshi writes:
> 
> > >From my viewpoint, the bottleneck is in the patch review and approval process.
> > I've sent a patch which fixes the h8300-hms bootstrap problem TWICE to 
> > gcc-patches, and it was never reviewed or accepted for inclusion.
> > 
> > In addition, there are about six other patches I've posted which fix various
> > other problems for which I'm waiting for approval before doing any other work.
> > 
> > I would do more work on gcc, but right now it feels like all my patches are
> > being posted into a huge black hole, so it's pretty discouraging.
> > I'm sure other people feel the same way, since I see them post messages
> > entitled "My Unreviewed Patches" enumerating their unreviewed patches.
> 
> I know that this is frustrating.  Patch reviewers get overwhelmed and
> things slip through the cracks that don't match the priorities of the
> developers, which, for 3.0, was to meet the release criteria as much as
> possible.  Because of this, it is necessary to be persistent.
> 
> One of our priorities for 3.0.1 is fixing breakage on embedded
> platforms.  So please re-submit your patches (which may need changes to
> work against 3.0), as I'm sure that they'll now get more attention.
> 
> 

Most of them were approved by geoffk last night, except for this one:

http://gcc.gnu.org/ml/gcc-patches/2001-04/msg01267.html

I originally spent 30+ hours over four days tracing through gcc
in my attempt to grok this problem. I believe I understand this 
problem fairly well. geoffk disagreed with my analysis of the 
problem, but didn't specify why it was wrong.

If I don't understand the problem, then I obviously can't generate
a correct patch. So, any explanations would be greatly appreciated
of precisely why my analysis of the problem is incorrect.

Toshi