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

Re: Question to the candidates: inclusion of the kFreeBSD-* ports



On Fri, Mar 02, 2007 at 09:50:24AM +0100, Petr Salinger wrote:
> >seems to list under three DDs actively involved in the port,
> OK, we miss one more current DD listed here.

Yup, though that shouldn't be much of a challenge. The other problem
is that the list doesn't seem active, so it's not incredibly clear that
people are actively maintaining the port.

> >I don't really see how it would benefit from being added to the archive,
> porters NMU, 

I'm not seeing why you need to be in the archive to do NMUs to improve
packages?

I mean, you need to make sure you don't break the package (on Linux,
Hurd, whatever) while you fix it for kfreebsd, and you don't get to
do 0-day NMUs, or ignore feedback from the maintainer, and such -- but
sending patches to the BTS, tracking them with a usertag, and NMUing the
ones that don't get added still seems entirely reasonable, and feasible.

That's still about improving Debian, in that case you're just making
the source we distribute able to be easily compiled for GNU/FreeBSD.

> posibility to start negotiation with RM to be included with 
> lenny (after etch is out, of course), ...

That doesn't look realistic to me, and seems a bit premature as a
justification for getting into the archive anyway.

> >Personally, due to things like [0],
> There are only different arguments passed to configure ...

It's extra complexity. Thinking about it in a different context, it's
not the right solution either -- that would be (afaics) to autodetect
the presence of SELinux in ./configure, so you only need an option if
you're being exceptional.

But what I mean is more that maintaining two simple Debian patches (one
for Linux, one for kFreeBSD) is probably simpler than maintaining one
complicated Debian patch (with some conditional make cruft). And probably
more importantly, it means that you can do immediately upload stuff for
kfreebsd without *any* risk of breaking the Linux version. And even if
you do autoimport patches from Linux to kFreeBSD, you have the option
of not uploading the source for BSD unless it does actually build for BSD.

Does that make it a bit clearer what I mean b?

> >I tend to think having different sources for different OSes
> >is likely to make sense; which isn't something we can manage
> >with the main archive as it stands.

?

The other side is just making the Debian mirror network available for
kfreebsd users. With only a couple of dozen apparent users, I'm not
really convinced that's particularly valuable though. YMMV.

Cheers,
aj

Attachment: signature.asc
Description: Digital signature


Reply to: