> Mhr wrote on Thu, 13 Dec 2007 16:45:34 -0800:
>
> > If that's true, doesn't it break some kind of unwritten rule to supersede newer software from an OEM source with older software that has been rebuilt?
>
> I'm just "reading by" but if I understood you correctly you installed a
> 32bit rpm of the 2.3 version (which you missed to tell from the beginning)
> and now yum wants to "update" to a 2.0.4 64bit rpm version.
> I think you missed a vital point - there's no connection
> between the 32bit and 64bit packages, it will never attempt to "update" a
> 32bit package to 64bit. What actually happens is that it updates the
> installed older 64bit package to the newer one. It doesn't matter at all
> if you have a 32bit version installed or not. It's just a mystery why
> those packages are not shown in your rpm -qa run. But it's clear that
> yum thinks they are installed.
>
Interesting points, but not the one I was aiming at.
My point was that the 2.3.1 version from OOo is newer than the 2.0.4 version from CentOS (although there is an epoch difference which makes the CentOS version look newer), and it is modified from the original 2.0.4 OOo distribution in that the one that comes with CentOS is a 64-bit package.
IOW, the CentOS distribution is modified from the original that came from OOo, but it is still an older revision than the newest one from OOo.
Shouldn't it be the case that a newer revision is NOT updated with an older one, epochs notwithstanding (is the epoch an OOo thing or a CentOS thing?) when the old revision's newer release is still a rebuild of an older revision?
(Is this getting too convoluted? I wouldn't think so, but....)
What do you CentOS folks think?
mhr
_______________________________________________ CentOS mailing list CentOS@xxxxxxxxxx http://lists.centos.org/mailman/listinfo/centos
