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

Re: -= PROPOSAL =- Release sarge with amd64

* Thomas Bushnell, BSG (tb@becket.net) wrote:
> I'd be happy to think through it, but only if you give me details.


I'm not 100% sure that's the latest, hopefully others will correct me if
it isn't.  It's got the gist of it though.

> Details would be: which parts of LSB is the port not compliant with?

It doesn't have the i386 loader in the right place, it doesn't have
32bit libraries in /lib.  Actually, the i386 loader bit might not be
right anymore w/ ia32-lib...  Even so though, there's only a few 32bit
libraries installed.  The other thing is that /lib64 is a symlink and
not it's own directory into which 64bit libraries are placed..  It's not
immediately clear to me if this is really a violation of the LSB or not

> Why do the packages require changes to become compliant?  Why is the

They would have to be modified to install packages into /lib64 for amd64
instead of into /lib like every other arch.

> result in question considered inelegant?  I'm a smart guy: you can
> post technical details.

It doesn't solve all of the problem cases, and clutters up / to boot.
Lots of technical details *have* been posted, it's not like we developed
this in seclusion, there was a talk at DebConf about it, the proposals
have been posted here and other places previously, and it's received
comments from the dpkg, RM and other teams.

> A multi-arch system may or may not be a good idea, but regardless,
> it's irrelevant to the question at hand, which is about the inclusion
> of amd64 in stable now.  

I agree, let's not focus on multi-arch right now, but on the issues
regarding amd64 inclusion in sarge.

> My recollection is that they complied with the standard.

It's expected that the LSB is going to be changed.  Additionally,
compling with the LSB isn't really an option- it'd probably take a year
or more to modify all of the packages, not to mention the Debian
infrastructure packages.  Then we'd turn around and do it again to
support multiarch, not exactly a useful way to spend our time.

*Certainly* if this is the issue then please tell us and then listen to
our arguments as to why we don't see IA32/LSB compliance as a problem.


Attachment: signature.asc
Description: Digital signature

Reply to: