Re: [Gdbheads] proposed change to GDB maintainership rules

[Michael Snyder]

Richard Stallman wrote:
> If people are proposing that we should change the way this
> part of the GNU system is maintained, the first question is,
> is there really a problem? 

I think there is.  At least seven of the twelve blanket write
maintainers are on record as thinking that there is.  3 or 4
non-blanket-write maintainers with whom I've conversed about
it think there is.


> What useful features or improvements
> does the GNU system lack, which it could have had if
> GDB maintenance were done differently?  How important are these
> features?

I should think that gdb development in general benefits
the GNU system.  Any specific bug fix or new feature may
or may not directly impact, say, debugging on the Hurd,
(or GNU/Linux), but some percentage of them will [1].

The proposed change relates to making gdb maintainance and
development easier and less painful, eliminating a bottleneck
that makes it slow, difficult and painful to get changes
approved.


[1] An illustrative example: a while ago I added a feature
allowing gdb to drop a corefile of the inferior on demand.
I didn't try to make it OS/ABI independant, but instead
added hooks to make it as easy as possible to port it to
other platforms.  Since then a number of other platforms
have implemented the hooks.  So far Hurd isn't one of them,
which might be interpreted as meaning that the GNU system
did not benefit from the work -- but someone with the right
knowledge and access could add the hooks to make the gcore
command work on the Hurd very easily, at any time, so in
potential the benefit is there, just waiting to be realized.