--- Begin Message ---
|
Subject: |
Re: Gdb committee meeting |
|
Date: |
Mon, 8 Dec 2003 23:47:40 -0700 |
On 07 Dec 2003 12:19:08 -0500, Richard Stallman <address@hidden> said:
> [ Elena Zannoni <address@hidden> ]:
>> It has come to my attention that there is a schism happening within
>> the Gdb community.
> What do other people think? Who else feels there is a problem?
> If there is a split, what are the groups that are splitting?
There isn't a split, no. However, several of us have been discussing
ways in which we could improve GDB's maintenance rules, because a lot
of people feel that it is very hard to contribute to GDB.
A major obstacle is the fact that it takes too much time for patches
to get reviewed. There are many areas for which we currently only
have a single maintainer. Since these maintainers are volunteers who
don't have an endless supply of spare time to spend on GDB, these
maintainers have become bottlenecks. The fear of introducing even
more bottlenecks has made us hesitant to nominate new volunteers. As
a result, potential contributors seem to turn away.
Another major problem is the fact that we have no way to resolve
conflicts; some of us are under the impression that the party with the
longest breath wins.
With the above in mind we have reached consensus that the new rules:
* Should be simple.
* Should not affect the quality of the contributions that get
incorporated.
* Should increase the number of maintainers.
* Should decrease the turnaround time for patches.
* Should have a mechanism to resolve conflicts in a clear way.
We have come up with a proposal (see below) that is similar in some
respects to the rules under which the GCC project operates. (Their
rules seem to work fairly well for them.) By letting global
maintainers approve patches in all areas of GDB, we hope to decrease
the turnaround time for patches. This in turn should clear the road
for nominating new maintainers. By giving contributors and
maintainers the right to request a vote, we hope to provide a clear
way to resolve conflicts. Requiring a majority should guarantee that
only patches that are up to our standards will get in.
Specifically, we propose that the "Various Maintainers" section of
GDB's MAINTAINERS file be replaced by the following:
All non-obvious patches to GDB must be approved. Area maintainers
can approve patches to the relevant area of GDB; global maintainers
are treated as if they are area maintainers for all areas of GDB.
Maintainers are permitted to approve their own patches in areas
where they have authority to approve others' patches.
If maintainers disagree whether or not a patch should be approved,
and can't resolve that disagreement via discussion, it shall be
resolved by a vote. Any global maintainer or area maintainer for
the relevant area can vote; a majority of votes cast is required for
approval.
We welcome discussion of this proposal or of any other ways in which
GDB's maintenance process could be improved.
Signed,
Jim Blandy <address@hidden>
Joel Brobecker <address@hidden>
Kevin Buettner <address@hidden>
David Carlton <address@hidden>
Daniel Jacobowitz <address@hidden>
Stan Shebs <address@hidden>
Michael Snyder <address@hidden>
Corinna Vinschen <address@hidden>
Mark Kettenis <address@hidden> is out of e-mail contact for much
of this week, but he was extensively involved in preparing this
message, so we believe that he agrees with it as well.
--- End Message ---