Package EVR problems in Fedora 2007-10-31

[Rahul Sundaram]

Ralf Corsepius wrote:
> On Fri, 2007-11-02 at 10:28 -0400, Jesse Keating wrote:
>> On Fri, 02 Nov 2007 15:12:34 +0100
>> Ralf Corsepius <rc040203 at freenet.de> wrote:
>>
>>> The actual question behind this is: Are the EVRs of packages in
>>> testing supposed to be updates < testing < rawhide?
>>>
>>> I.e. are upgrades from testing to rawhide relevant?
>> IMHO yes.  Any package that could potentially go out for a release (if
>> it's in -testing, it could potentially go out) should have an nevra
>> that sorts lower than release+1.
> IMHO no. 
> 
> It renders the primary purpose of "testing" absurd: "testing" packages
> for "updates" (testing == volatile, scratch, ... irrelevant)
> 
> What matters, is packages which are being pushed from "testing" to
> "updates" containing appropriate EVRS at the very moment they are being
> pushed.

If the package has a lower E-V-R when it is pushed into the testing 
repository, it won't show up as an update for those enabling 
updates-testing repository. We need to fix it while it is in the testing 
repository itself for testers to find out about *other issues* in the 
package.

Rahul