0.5.15lorg3.1 loses epochs on rh7.3/rh8.0/rh9 (promoteepoch?)

[Panu Matilainen]

On Wed, 9 Aug 2006, Axel Thimm wrote:
> On Wed, Aug 09, 2006 at 06:14:53PM +0300, Panu Matilainen wrote:
>> Actually I'm curious how the heck yum & smart manage to work on those
>> older releases?
>
> If you have an example you want to see whether yum/smart can handle I
> have RH9, RH8.0 and RH7.3 with apt/smart/yum installed and also
> zero-epoched repomd (e.g. the default createrepo mode). Upgrade
> operations work, but these are minimal systems, so there aren't many
> packages involved (~80 packages).
>
> yum/smart install <fill out this form>

I'd suppose fiddling with the packages involved in these failures would 
trigger problems if somebody wants to test a bit:
http://lists.laiskiainen.org/pipermail/apt-rpm-laiskiainen.org/2006-April/000024.html
http://lists.laiskiainen.org/pipermail/apt-rpm-laiskiainen.org/2006-April/000070.html

What just occurred to me ... something you said in an earlier thread: zero 
vs no epoch is different from epoch promotion. Maybe, just maybe, if rpm 
>= 4.2.1 is used (like atrpms has) this could be dealt with. If you want 
to test it, here's a patch to make the epoch hiding configurable for newer 
rpms. Set RPM::HideZeroEpoch to false and see what happens - IF this 
happens to work with both repomd and old style metadata on a system 
requiring promoteepoch but with rpm >= 4.2.1 I'll just apply it, 
if not I'm going to rest this case for good :)

 	- Panu -

--- apt-pkg/rpm/rpmsystem.cc    (revision 257)
+++ apt-pkg/rpm/rpmsystem.cc    (working copy)
@@ -215,7 +215,7 @@
        Opt++;
     }
     _rpmds_nopromote = NoPromote;
-   HideZeroEpoch = (NoPromote == 1);
+   HideZeroEpoch = _config->FindB("RPM::HideZeroEpoch", (NoPromote == 
1));
  #else
     HideZeroEpoch = false;
  #endif