Re: [UDD] dehs does not necessarily contain latest version
On 05/08/10 at 14:01 +0200, Andreas Tille wrote:
> Hi,
>
> I wonder what principle might be used to obtain the unstable_version in dehs table.
> Example:
>
> udd=# SELECT source, unstable_version, unstable_upstream, unstable_parsed_version, unstable_status FROM dehs d WHERE source = 'r-cran-maptools' ;
> source | unstable_version | unstable_upstream | unstable_parsed_version | unstable_status
> -----------------+------------------+-------------------+-------------------------+-----------------
> r-cran-maptools | 0.7.26-1 | 0.7-34 | 0.7-26 | outdated
>
> udd=# SELECT r as release, version, archs, component
> udd-# FROM versions_archs_component('r-cran-maptools') AS (r text, version text, archs text, component text)
> udd-# JOIN releases ON releases.release = r
> udd-# ORDER BY releases.sort ASC, version DESC;
> release | version | archs | component
> ---------+----------+-------+-----------
> squeeze | 0.7.26-1 | i386 | non-free
> sid | 0.7.26-1 | i386 | non-free
> sid | 0.7-34-2 | amd64 | non-free
>
> which means: If there are different versions in unstable (for different
> architectures, which might be the case for non-free or also some not
> supported archs), what version is regarded as "latest". IMHO the right
> version would be MAX(version) and thus the dehs table is wrong (and
> probably the imported dehs data set (but I never looked into this
> directly).
>
> This fact leads to some false alarms for the Blends tools and I'd like
> to fix this if possible. If you agree that MAX(version) is the
> information we want to keep, I'd file a bug report. If you rather
> support the philosophy that MIN(version) should be stored (rationale
> would be that there is at least one arch which needs an update) I think
> we need to store two additional fields: max_version and max_outdated
> (or something like this).
>
> What do you think?
I'm not sure of the problem here. It can either be:
- In the dehs table, UDD shows the unstable_version as provided by DEHS,
which, due to synchronization delays, might not be the latest one in unstable.
=> I don't think that this should be fixed
- DEHS is not aware that there can be multiple versions in unstable.
=> This totally should be fixed
Raphael, comments?
--
| Lucas Nussbaum
| lucas@lucas-nussbaum.net http://www.lucas-nussbaum.net/ |
| jabber: lucas@nussbaum.fr GPG: 1024D/023B3F4F |
Reply to: