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

Bug#341524: yaird: fails to upgrade from linux-2.4 to 2.6



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 1 Dec 2005 12:25:02 +0100
Sven Luther <sven.luther@wanadoo.fr> wrote:

> On Thu, Dec 01, 2005 at 12:17:25PM +0100, Jonas Smedegaard wrote:

> > On Thu, 1 Dec 2005 08:25:13 +0100
> > Sven Luther <sven.luther@wanadoo.fr> wrote:
> > 
> > > Maybe the best idea would be to have yaird as first default as
> > > now, but have the dependencies install both, which would
> > > guarantee it always works.
> > 
> > A better approach IMHO is to recommend (not depend on) alternative
> > ramdisk tool than the one picked as default.
> 
> Well, with tools like aptitude, there is not much of a difference.

Yes there is: Aptitude treats recommends as they are: Recommended but
acceptable to explicitly avoid. Or, as I do for some setups:
explicitly avoid recommended packages *always* by passing it
- --with-recommends.


> > initramfs-tools will pull in udev which is unwanted for some, and
> > yaird will similarly pull in perl dependencies that is unwanted for
> > some.
> 
> perl dependencies are hardly problematic, especially since we have
> perl-base, altough it would be good to make yaird depend on perl-base
> only, but we discussed this already. Udev might be more legitimate to
> not want though.

Stop spreading FUD, please. this has been discussed before and it does
not seem likely to have yaird use only perl-base without extensive
changes.


> > Sure, both can be seen as weird complaints (I have had my fights
> > with convincing the php4 maintainer to not depend on X11 by default
> > when
> 
> This is also a sane complaint, we divided ocaml packages between the
> x version and the nox stuff too.

Please note that I did not claim insanity. I wrote that it _can_ be
seen as weird.

Agreeing with me is no argument that others can see it as weird. If
interested then I'd appreciate help making libgd dependent packages
build-depend on the -noxpm variant development packages, but really
that is irrelevant for this bugreport.


> > sanely avoidable), but I don't think disliking udev should force
> > one to either recompile kernels or use equivs.
> 
> Yep, that makes sense. Too bad dependencies don't allow us to do
> something more advanced.

I just porposed a possible approach: When depending on yaird then
recommend initramfs-tools, and vice versa.

I am aware that apt-get ignores recommends, but that is a bug with apt,
and everytime people are hit by that bug we get a chance to advertise
aptitude (and if such advertisement is refused with "apt-get is good
enough for me", well - then it's their own fault they get hurt by
things like this).


 - Jonas

- -- 
* Jonas Smedegaard - idealist og Internet-arkitekt
* Tlf.: +45 40843136  Website: http://dr.jones.dk/

 - Enden er nær: http://www.shibumi.org/eoti.htm
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)

iD8DBQFDjwDMn7DbMsAkQLgRAnnZAJ9JPKevwJ6LklwoWJ4RPy5eUMSZnACdHMmh
hCrRrAG92MmkxDLEmG6ku9c=
=nzYT
-----END PGP SIGNATURE-----



Reply to: