RE: Redundant code for non-us?
The package in question was curl.
Root of this discussion:
http://www.debian.org/Lists-Archives/debian-mentors-9908/msg00059.html
the messages supporting my opinion:
http://www.debian.org/Lists-Archives/debian-mentors-9908/msg00074.html
http://www.debian.org/Lists-Archives/debian-mentors-9908/msg00076.html
which I have also expressed here:
http://www.debian.org/Lists-Archives/debian-mentors-9908/msg00095.html
but there was no more discussion after that, so perhaps it's
a bit preposterous to call it a result or a consensus.
I'd rather not to dig the decision(?) referred in my post,
but the problem seems to be a general one. I think this should
be clarified somewhere. For that reason I also Cc: -policy.
If this needs to be in BTS please forward it there.
t.aa
Peter S Galbraith <GalbraithP@dfo-mpo.gc.ca> 15. Nov 1999 18:07
>
> arto.astala@nokia.com wrote:
>
> > iirc this exact situation was discussed rather recently with
> > some other package, and the result was that there is one /main
> > source archive. The division to '/us' and /non-us is not
> > important, especially now when apt can hide it from the user.
> >
> > This would imply one source in /non-us and two binaries.
> >
> > Can't remember exact time or the package in question, however.
> > Still it might be useful to browse through archive to see
> > if anything turns up.
>
> I'm trying to find this discussion and not succeeding. Was it on
> -mentors?
>
> I grepped for non-us in recent archives of -devel, -legal and
> -mentors and didn't see anything like that.
>
> Thanks for any additional pointers (anything you can remember).
> Peter
>
> > Peter S Galbraith <GalbraithP@dfo-mpo.gc.ca> Fri Nov 12,
> 1999 8:03 PM
> > > The latest upstream version of powstatd comes in two favours,
> > > regular and one that use 128-bit TEA encryption for communication
> > > between master and slave.
> > >
> > > Currently, the upstream author is producing two source archives
> > > (but we are discussing this). The only difference is that one
> > > doesn't include xtea.c and the other one does. I planned to have
> > > my regular `powstatd' package supplemented by a new
> > > `powstatd-crypt' package.
> > >
> > > Do I have to make two source packages?
> > >
> > > Or can I upload everything to non-us and produce two binary
> > > packages (one for regular main and the other for non-us/main)?
> > > I have the feeling this is not allowed because it makes the
> > > source hard to find. Right?
> > >
> > > The irony is that xtea.c is freely available outside the US, does
> > > not originate from the US, and has only _32_ lines of code!
>
Reply to: