Segfaults on powerpc platform

[Panu Matilainen]

On Thu, 5 Jul 2007, Tim Verhoeven wrote:

> Hi,
>
> It's been a while that I had time to spend on this issue. But I've
> been testing again the last couple of days. I've pulled the latest
> devel version using git and I've been able to successfully compile AND
> run it on RHEL 5. So it looks like the multiarch changes fixed the
> segfault.

Ok, that's excellent news. I still don't know what the exact issue on ppc 
was, but if the latest version works it probably is related to the 
on-the-fly renaming where on ppc practically everything got renamed to 
.32bit although that's the preferred arch. If that's the case, the newer 
version essentially hides, not really fix, the bug.

> The problem is that all my production systems are RHEL 4. But trying
> to compile the devel version in RHEL 4 fails because it doesn't have
> sqlite and I wont want to have to provide it myself (security and
> bugfixes etc.). Is there any way to get the current devel version to
> compile without sqlite support. I've read the comments in the
> configure script that it should be possible. The question then
> ofcourse is, how easy or hard is it ?

Sqlite is at the moment tied to repomd support. If you don't need repomd 
(ie you're using genbasedir generated repositories) you can build it with 
--disable-repomd option. If you need repomd, you're currently out of luck.
It's not particularly hard to add the necessary checks and conditionals, 
mostly just boring work I haven't gotten around to. The sqlite build 
requirement is one of the remaining blockers for a stable release btw.

But as a first aid, you can get sqlite for RHEL 4 from for example Dag's 
repo (http://dag.wieers.com/rpm/) or ATrpms (http://atrpms.net/) without 
having to build and maintain them yourselves.