Making updates-testing more useful

Les Mikesell lesmikesell at gmail.com
Wed Dec 10 19:44:23 UTC 2008


Seth Vidal wrote:
> 
> 
>>> I think the idea is good. I don't think it helps us to store more
>>> information in PK, though.
>>
>> Agreed.
>>
>>> If we're going to be storing where a package is installed from. It 
>>> should
>>> probably be in rpm.
>>
>> We can't easily do that though, can we?  Given that the same binary RPMs
>> that are in updates-testing today will be in updates tomorrow without
>> changing...
> 
> which is, of course, the problem.

So what is the solution in terms of possibilities for yum?  Personally, 
I'd like to see a sane way to manage repeatable updates at any level to 
at least give an end user a chance to test on a spare machine or VM 
exactly what will be propagated in an update to a more critical box. But 
  even that might not be the right solution where an emergency fix needs 
to get pushed immediately.

There has to be some clever way to let the repos change at their own 
speed but give the end user control over what updates are installed on 
any particular machine by adjusting some sort of 'risk' factor.

-- 
   Les Mikesell
    lesmikesell at gmail.com






More information about the fedora-devel-list mailing list