Re: [Gdbheads] proposed change to GDB maintainership rules

[Ian Lance Taylor]

Andrew Cagney <address@hidden> writes:

> "If there are several maintainers for a given domain then responsibility
> falls to the first maintainer.   The first maintainer is free to
> devlolve that responsibility among the other maintainers."

You discussed this paragraph well, and everything you say makes sense
to me, but I have to say that I still don't understand how patch
review works.

I would say that patch review is the most important aspect of being a
GNU maintainer.  This is not to say that the other types of
responsibility are not important--they are very important.  But patch
review is the most important.  Without effective patch review, people
do not contribute.  When people do not contribute the pace of software
development is reduced to what a few people can accomplish, rather
than the pace which many people can accomplish.

Maintainers must take responsibility for looking over patches quickly,
and approving them, rewriting them to be acceptable, rejecting them
with an explanation, or suggesting changes.  Maintainers who don't
accomplish that are not effective maintainers.  That is not to say
that they can not be effective contributors, or that they can not be
very good at maintaining code and making technical decisions.

Of course, not everybody can be good at everything.  There is nothing
wrong with permitting certain people to contribute freely without
expecting them to also review patches promptly.  In such a case
conflicts are possible, and must be resolved by consensus.  The long
term maintainability of the code is very important.  Encouraging
contributors is also very important.

I was the sole GNU binutils maintainer for several years, around 1995
to 1999.  During most of those years people at Cygnus were able to
contribute freely, but I was the only person who handled patches
submitted from outside Cygnus.  I think I did an effective job of it,
and I think I greatly boosted the number of contributors and the pace
of binutils development.  To do that, I make reviewing and accepting
patches my first priority.  I don't think I lowered my standards;
instead, I regularly rewrote patches to make them acceptable.


Anyhow, my point is, who does handle patch review?  gdb is a big
project.  It receives many patches.  The gdb MAINTAINERS file doesn't
clearly explain who is entitled to approve patches.  (For that matter,
the bintuils MAINTAINERS file doesn't explain it either).  gcc's
policies are very clear:
    http://gcc.gnu.org/cvswrite.html#policies
These policies have proven very effective in practice.  They are
essentially the policies which are currently followed by the
binutils.

Ian