[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [Adrian Bunk <bunk@fs.tum.de>] Do the HPPA "binary-only" NMUs violate the GPL?

Matt Taggart <taggart@carmen.fc.hp.com> writes:

> I suspect the majority of the cases where this is occurring is when
> config.{sub,guess} needs updating for hppa. [...] We figured that
> there was no reason to force the other archs to recompile and
> everyone to download the new package, etc.

I disagree; apart from anything else, hppa is not the only
architecture that needs newer config.{sub,guess}.  If you do things
this way, e.g. sh and s/390 both miss out.

> What do you think? Would a source-NMU be better?

Definitely.  The porting problem (at least for the released
architectures) is not keeping up with the amount of uploads, but
rather all the uploads that don't build cleanly.
> The only other source-changes binary-hppa NMU that I've done was for
> apt. We needed the latest CVS version and didn't want to subject all
> the other archs to the new bits.

Better to slap some sense into the CowMan I think; tho in this case, I
guess no one was going to chase us for not complying with the GPL ;-)

> We'll knock it off, no need to add code(although I wouldn't object
> to such code being added).

It was a bluff ;-) I can't really detect sourceful changes, at best
all I could do was put a procedure in place so that one couldn't do a
source-changing bin-only NMU without being fully aware that you
weren't meant to be doing it and essentially deliberately lieing in
the changelog.


Reply to: