Re: why/does bigloo/skribe needs hinting into testing ?
On Tue, Dec 20, 2005 at 09:41:44PM -0800, Steve Langasek wrote:
> On Tue, Dec 20, 2005 at 03:10:24PM +0100, Yann Dirson wrote:
>
> > For some reason bigloo and skribe have not yet migrated into testing.
> > The only reason seems to be their quasi-circular build-dependency
> > relationship, so maybe they need hinting. OTOH, shouldn't
> > katie/whoever see them as a single migratable set anyway ?
>
> trying: bigloo
> skipped: bigloo (14 <- 1257)
> got: 4+0: a-4
> * alpha: skribe
>
> http://ftp-master.debian.org/testing/update_output.txt.gz
I saw that already.
> This has nothing to do with build-dependencies, it has to do with not being
> able to update either package alone without leaving skribe uninstallable in
> testing. For computational reasons, this is not done automatically by
> britney.
Hm, I'm not sure I understand what's going on here. The version of
skribe in testing depends on the version of bigloo in testing, and the
skribe in sid depends on bigloo in sid. It's sure that migrating only
one of them will break "skribe in testing", but I see nothing
preventing them to migrate simultaneously - it's a classical soname
migration, the only difference I see is that bigloo has a
build-depends-indep on skribe, so I thought that might be the reason.
If not, I'd be happy to understand what the problem is :)
There is a new skribe release already, and a new alpha release of
bigloo, and I'd *really* like to have the current versions in etch
before I start messing with the sid packages.
Best regards,
--
Yann Dirson <ydirson@altern.org> |
Debian-related: <dirson@debian.org> | Support Debian GNU/Linux:
| Freedom, Power, Stability, Gratis
http://ydirson.free.fr/ | Check <http://www.debian.org/>
Reply to: