New schedule for FC5 Test2 ISO roll-out?

[D Canfield]

Michael A. Peters wrote:
> On Wed, 2006-01-11 at 09:14 -0500, D Canfield wrote:
>
>   
>> Maybe I'll be laughed off the list, but I do think it would be great if 
>> a *goal* (not a requirement) of at least test3 of a given release would 
>> be to upgrade to final.
>>     
>
> Impossible because some applications may need to be downgraded for final
> - which can not really be done easily with yum (not w/o using epoch's -
> but that would be a silly reason to use an epoch)
>   
Has it been consistently true that packages are commonly downgraded 
between test3 and final?  I thought t3 was intended to be more "final" 
than that.  Again, I said *goal*, not *requirement*.

>>   Might help get a bit more testing on more 
>> diverse hardware if there wasn't the expectation that we will likely 
>> have to re-install a few weeks later.
>>     
>
> No test release should be used without expectation of re-install.
>   
That's the claim now.  But my point is that at least in my personal 
experience, I'm more likely to test something (and find useful 
feedback/bugs) if it's something I can use on one of my primary 
machines.  If I'm sure I'm going to have to reinstall the things 4 times 
(t1, t2, t3, final), I'm not likely to bother testing it beyond maybe 
running through the installer in vmware and looking to see what's new.  
Maybe the preference is not to have "casual" testers, but I think that's 
the only way to get diverse testing on real hardware.
>> Similarly, it would be nice if there were some notes between test 
>> releases (and final) pointing out any changes that might not have been 
>> handled by rpm -Fvh.
>>     
>
> If updating is done between test releases - it should be done clean
> install because the installer needs testing, and furthermore, the
> software should be tested in a clean slate - as that may reveal issues
> that need to be fixed, and might avoid false issues caused by previous
> test release.
>
>   

Doesn't the upgrade functionality need tested too?  And don't we need to 
know if there are issues caused by upgrades?   It would be important to 
know that someone had upgraded when tracking bugs, but again this seems 
like a good way to eliminate bugs early that will catch especially the 
lower-end users.

Just a thought.  If it's totally stupid, it of course won't happen. :-)

DC