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

Re: education-common: con't fulfill the Recommends on !i386



On Thu, Sep 30, 2004 at 12:31:11PM +0200, Martin Schulze wrote:
> Jeroen van Wolffelaar wrote:
> > On Mon, Sep 27, 2004 at 11:38:02PM +0200, Adrian Bunk wrote:
> > > Package: education-common
> > > Version: 0.801
> > > Severity: serious

> > > education-common recommends grub which is available only on i386.

> > education-common is an arch: all package, therefore, it is technically
> > impossible to have an arch-specific Recommends:

> > Because this bug isn't one used as a reason for britney to keep a package out
> > of testing, and the package is otherwise installable, this package is now in
> > Sarge.

> > Release Managers, what do you think of this bug?

> > 1) Do you agree it is RC?
> >    - I'm personally tending to say: yes. If a package is installable on an
> > 	 arch, it should also be able to fulfil recommends.  Post-sarge, all
> > 	 packages that are not installable on certain archs should not be in the
> > 	 Packages.gz files for their uninstallable archs, or something analogous
> > 	 to that

> > 2) How should it be resolved?
> >    - Personal favourite, make this package arch:any, and have per architecture
> > 	 the needed boot loader for that architecture. I think any
> > 	 moderately-complex meta-package that depends on very architecture
> > 	 dependent stuff (I.e., hardware specific drivers, kernels and
> > 	 bootloaders) should be architecture: any, to be able to have
> > 	 architecture-dependent package relations. Maybe this part of
> > 	 education-common should be split out into education-boot, so that only
> > 	 that package needs to be any, but that's something the debian-edu folks
> > 	 should decide.

> I'd rather investigate why education-common needs to recommend grub at all.

> The name makes me think that it's a "task" package, basically consisting
> on dependencies.  Does it need to have a dependency to grub at all?

Quite agreed.  If there were a clearer reason for having such a
recommends, I'm not sure it would be RC (but I'd have to see such a
case to know for sure); here, it seems the relationship should
definitely be dropped.

-- 
Steve Langasek
postmodern programmer

Attachment: signature.asc
Description: Digital signature


Reply to: