This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: 4.2 Backport fix for PR libfortran/31196


On Mon, Apr 16, 2007 at 02:58:07PM -0700, Mike Stump wrote:
> On Apr 16, 2007, at 1:22 PM, Steve Kargl wrote:
> >Your response above is just another example of why 4.2 has lingered  
> >in a catatonic state.

(A bunch of irrelevant dialogue removed)

> So, if you want to blame me, you can blame me for any area I maintain  
> and for any regression I put in.  If you want to blame me for  
> regressions I didn't put in, or for an area I am not a maintainer  
> for, well, let just say, I disagree with your assessment.

Mike, go back and read what I wrote!  I did not blame you for anything. 
I responded to your hypothetical scenario where a developer could not
commit bug fix to mainline, and move on to the next bug.

I'll summarize my position once again:

1) I think the GCC developement model is fundamentally flawed.  The
   4.2 branch was created 6 months ago.  There were 124 regressions on
   this branch when I initially replied in this thread.  It appears
   likely that a 4.2 release will not occur for several more weeks
   (months? ever?).  I believe further delays are detrimental to
   gfortran's public exposure/image.

2) To address this problem, I briefly described how I think the 
   developement could be improved.  1) Mainline is branched for
   release X.Y.0.  2) Mainline is closed to all commits while the
   all effort is directed at getting X.Y.0 released.  If everyone
   with commit access is on-board with this development model,
   then mainline would be closed for a week.  If a developer can't
   contribute a patch during that week, so be it.  She can review
   patches or bootstrap GCC and run regression tests or simply take
   a break from GCC.

3) Mainline does not need to be closed for a micro release, eg., X.Y.2.
   Supposedly, only regression and documentation fixes should be going
   into X.Y.2.  

In looking at the GCC timeline, this development model would have
mainline closed for 1 to 2 weeks every 12 to 14 months.  The impact
of this closure on GCC development would be in the noise.

-- 
Steve


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]