Re: [PATCH]: memory region attribute nits

[jtc at redback dot com (J.T. Conklin)]

>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
Andrew> J.T., A ramble,

Andrew> As an unofficial guideline, I'd be most worried about the
Andrew> targets I've listed as buildable in the MAINTAINERS file.  If
Andrew> there is a target that is known to not build then I personally
Andrew> think it is questionable to expect you to get your changes
Andrew> right for that target.

That's a good idea.  I've been using arm-elf, i386-aout, i386-netbsd,
i386-netbsdelf, m68k-coff, mips-elf, powerpc-eabi, powerpc-netbsd, and
just recently added sh-elf trying to get good coverage.  It's a pretty
good list, but the list of buildable targets in MAINTAINERS is better.
I'll add the missing ones going forward.  However, it's a bit
difficult for me to test many of the native configs because I don't
have access to them.  I'm sure that many other GDB hackers are in the
same.  Hopefully those interested in those configs will take the time
to ensure that they build.

Andrew> In the future, this problem - being able to reliably make global changes
Andrew> - is likely to become more of an issue.  We don't want to let GDB's
Andrew> development get bogged down because people fear the consequences of
Andrew> making even simple mechanical changes.  Things like -Werror and listing
Andrew> buildable targets in MAINTAINERS are all part of my secret agenda to
Andrew> make the process more tractable.

I agree with your now-not so secret agenda.  To me, the most
interesting changes that we'll be making to GDB are those that effect
all targets.  It is the nature of these that sometimes something will
get broken.  I believe it is the responsibility of those making those
changes to do their best to make them right and to help fix things up
if there is a problem.  But there is also a responsibility of those
who get bitten by such a change to understand that such things will 
happen.

        --jtc

-- 
J.T. Conklin
RedBack Networks