Re: /var/lib, /var/mail

[Santiago Vila <sanvila@unex.es>]

On Fri, 30 Jul 1999, Joseph Carter wrote:

> [...]
> What I am objecting to is wording in policy to the effect that we should
> do something, followed by "don't do this for potato".  Especially if it
> doesn't really matter if it's done for potato or not.

So maybe we both agree that a policy which is not to be followed is not
good, after all :-)

> > > [...]
> > > Given that your entire objection to this proposal is that you don't want
> > > to see anybody implement it YET for some reason either I'm incapible of
> > > understanding or you're not explaining, I don't see the problem. [...]
> > 
> > I *did* explain: Lots of *Pre-Dependencies* on base-files. I think this is
> > bad, becuse upgrading without apt will make the upgrade to be *painful*.
> > 
> > "Pre-Dependency problem, foo not unpacked because it pre-depends on
> > base-files >= bar, base-files >= bar is not installed yet".
> 
> Um, so ....  upload a base-files that does this now and let it get
> installed into a few mirrors...  If the dselect methods cannot handle
> this, they are broken.  apt can handle it and is AFAIK the default with
> potato now that apt has a working cdrom method...

I didn't say dselect can't handle this. It can, but if you use a dselect
method which does not do package ordering, you will get lots of "errors"
and will need several [I]nstall passes.

BTW: The same will happen if lots of essential packages in potato
depend on libc 2.1. Do we plan to avoid it this time?

Thanks.

-- 
 "78b68ccdb75773a96ac160402d5dd9b4" (a truly random sig)