[rfa/3.0.x] NetBSD/PPC support

[David Edelsohn]

>>>>> Andrew Cagney writes:

Andrew> If the patch was changed (brutally) to not affect other targets would it 
Andrew> be accepted?

>> GCC does not create sub-branches for each sub-release.

Andrew> You mean 3.0.x is a single branch so there is no 3.0.3 branch that
Andrew> I could commit to now.

	Yes, there only is a gcc-3.0 branch, not gcc-3.0.1, gcc-3.0.2,
gcc-3.0.3, etc.  I interpreted your original message as submitting your
code for gcc-3.0.3, so why does this matter?  Why are you now proposing to
create a kludge for gcc-3.0.2 when you previously implied gcc-3.0.3?

	If you want the change in gcc-3.0.2, I am not going to object
strenuously, but you should have submitted this change during the previous
two and a half months.  I don't have a lot of sympathy for an extremely
localized, hacked variant just to get the change in gcc-3.0.2.  There is
no technical reason for a special variant just for gcc-3.0.2.

	You will need to ask Mark Mitchell, the GCC Release Manager, for
permission to make this borderline intrusive change in the branch so late
in the gcc-3.0.2 release cycle.  I prefer that you either convince Mark to
accept the current version of the change into gcc-3.0.2 or wait until
gcc-3.0.3 instead of installing a kludge.  This isn't a case where we
could not accept the complete change into the release branch, this simply
is a case of timing and deadlines.

	I don't mind your committing the change to the branch at the
beginning of the gcc-3.0.3 cycle when there is plenty of time to test.

David