Re: [Gdbheads] Re: Feb's patch resolution rate

[Ian Lance Taylor]

Andrew Cagney <address@hidden> writes:

> Jim, lets stick to this question.  Would the pair:
> 
>       Elena
>       David
> 
> make for a better symbol table leads than:
> 
>       JimB
>       Elena

Well, what about
    JimB
    Elena
    David
?

> The symbol table is a critical piece of infrastructure and as such
> needs a steady hand, and a long term commitment by developers.  Both
> through overhaul, and through a disciplined and diligent review of
> patches. What the area does not need is random tugs in differing
> directions based on the wim of "maintainer of the day".

If you feel that Jim has been doing random tugs in differing
directions, then that seems like a good reason to ask him to step down
from being a maintainer.  But if his decisions have been reasonable
even if he hasn't made very many of them, then I don't see why he
should step down.

> Now, through discipline and review, we see a very different situation
> -
>   the number of [non-deprecated] interfaces is dropping (5.3 peaked at
> 139, vs 104 in current).  Since I'm responsible for this area I'm
> expected to not just review an occasional and selective patch, rather
> I need to ensure that all patches are viewed in a timely manner and in
> addition, where necessary, respond to problems addressing them as we
> as a group see fit (cf my most recent addition of
> gdbarch_register_pre_init).
> 
> Adding more reviewers to the mix, be it by spreading responsibility
> more thinly, or by giving core-developers carte blanche, is not going
> to service this need.

But not all patches are the same.  Don't confuse being a gdb
maintainer--somebody who can approve patches, including their
own--with being responsible for ongoing architectural overhaul of gdb.
Those responsibilities are not the same.  Different people can
effectively contribute on both.

> > So I didn't see how it was in GDB's
> > interests to remove myself from the list of people able to approve
> > symtab patches.
> 
> GDB needs people that _do_ review symtab patches, not people that are
> listed as being "able to approve" symtab patches.  There is a very
> real difference.
> 
> Denying this is not being honest with either themselves or the GDB
> community, they are treating their position as one of power and glory,
> and not responsibility.

That does not address Jim's issue.  Why is it in gdb's interest for
Jim to step down as a maintainer?  Sure, there is a difference between
people who review patches and people able to approve, but how does the
existence of the latter affect the former?

And what's wrong with a position of power and glory?  People need some
motivation to work on gdb.  Power and glory is not the worst.

What specific benefit will gdb gain if Jim steps down?  In what
specific way will gdb become better?

Here, I'll give you one for free: if Jim steps down, it would take the
personality conflict between the two of you out of the realm of gdb,
and could conceivably smooth relations between the remaining gdb
maintainers.  Is that an argument you would be prepared to make?

Ian