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

Re: priorities and package relations



On Sat, 28 Nov 1998, Wichert Akkerman wrote:

> A while ago someone (Santiago iirc) filed a bugreport about packages
> depending on other packages with a lower priority. This made me
> wonder about allowed relations between packages. Reading the policy
> document does not give any explicit demands. The only thing that
> I am sure of is that a package may not depend on a packages with a
> lower priority. If this also holds for recommends is not specified,
> although it might be logical considering that a Recommend is handled
> the same as a Depend with respect to DFSG compliance.
> 
> I have updated my dependency-checker to also check Depends for
> priorities, and the report is available via www at
> http://master.debian.org/~wakkerma/unmet.html . 

Good work!

Yes, a Recommends should be treated the same as a Depend with respect to
DFSG compliance. [ You will find this in policy 2.5.0.0, section 2.1.2,
"The main section" ].


Also, packages in the base system (i.e. those in base2_1.tgz)
should not require a package outside of the base system to work, because
otherwise the base system would be "broken". Whether this "requires" 
is just Depends or both Depends and Recommends I don't know (if you are
going to write a dependency checker for that, you may try both and see
what happens, there are not many packages in the base system).

If this is not policy yet, maybe we should make it policy.


Also, since "optional" means "all the software that you might
reasonably want to install if you didn't know what it was or don't have
specialised requirements", and since the user will not reasonably want to
install packages that conflict at each other, I think we should consider a
big mistake that we ship package A and package B if both A and B are
optional but one of the two conflicts with the other.

Again, if this is not written in the policy yet, maybe it should, IMHO.

Thanks.

-- 
 "17c2d6170cc60beb55fb5f360c5e2d84" (a truly random sig)


Reply to: