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

Re: Bug#535645: Bug #535645: Wrongfull removal of ia32-libs-tools



Steve Langasek <vorlon@debian.org> writes:

> On Sun, Aug 09, 2009 at 06:32:38PM -0600, Bdale Garbee wrote:
>> On Sun, 2009-08-09 at 16:37 -0700, Steve Langasek wrote:
>
>> > This assumes that users will continue to have ia32-libs-tools installed
>> > during the transition to multiarch.  What guarantees that this will be the
>> > case?  Does removing ia32-libs-tools also remove all the packages that were
>> > installed using it?
>
>> Interesting thought.  If all packages generated by ia32-libs-tools were
>> configured to depend on ia32-libs-tools, that might allow a much cleaner
>> conflict specification?  Just conflict with ia32-libs-tools?
>
> The conflict specification would be cleaner; the upgrade path would not, as
> conflicts at the base of a dependency tree tend to confuse apt into giving
> suboptimal upgrade solutions.

The prerm script could fail if ia32-* packages are left installed or I
could add a depends.

But if you want depends so you can conflict with a single package then
there is no need to depend on ia32-libs-tools. I already wrote,
multiple times, that the recent ia32-libs-tools generated packages all
"Provides: ia32-abi". So there is a single "package" to conflict
against. That was added for the benefit of ia32-libs and ia32-libs-gtk
to conflict with but works for anything.

I have absolutely no problem with multiarch conflicting ia32-abi if
Debian chooses that as upgrade path. Ia32-libs-tools should still be
in debian though, at least for a while longer, to give people the
chance to upgrade to where everything "Provides: ia32-abi".

MfG
        Goswin


Reply to: