Bug#469035: libc6-i386: include mutiarch ld.conf from i386
El dg 11 de 05 de 2008 a les 20:50 +0200, en/na Aurelien Jarno va
escriure:
> The version in unstable doesn't use the multiarch paths.
The version in unstable allows multiarch compilation.
> No I suggest to do a "echo path > /etc/ld.so.conf.d/path" manually.
...
> There is an implicit support for those paths. There is no support for
> libc6-i386 being the multiarch i386 libc6 on an amd64 system.
And you'd rather obstruct someone that could use it. That's a wontfix
even after adding multiarch support.
Once again, that file makes no harm. Library packages may do harm
(that'd have to be proved).
> As long as they are not official, I don't care about the problem they
> can cause. Officialising them means we have to care about them.
So you could start closing many wishlist bugs to enhance packages and
support different environments, couldn't you?
> Instead it
> should allow the installation of a package from a different architecture
> than the one of the system.
And that means packages from one architecture can be installed on
different architectures.
> If you don't know how to handle the conflict, the best is to not create
> conflicts.
I don't know how to handle a conflict that doesn't exist.
> And again, if we officially support the ld.so configuration
> we then implicitely support such package.
What package would you implicitly support? I thought you didn't care
about unofficial packages.
> If you got a working multiarch system, then please share your patches.
I'm already sharing the system. I'm contributing patches. Are you
subscribed to debian-amd64? Haven't ia32-libs maintainers ever mentioned
this?
I repeat, I've got a working multiarch system. And it'll be that way
until you show me what's wrong with it.
If I can't change your mind and you don't want to prove me wrong, don't
bother to reply. It's your theory versus my working packages.
Reply to: