[Fedora-spins] Fedora 10 Beta Freeze (Spins)

John Poelstra poelstra at redhat.com
Tue Sep 23 15:34:27 UTC 2008

Bryan Kearney said the following on 09/23/2008 05:48 AM Pacific Time:
> John Poelstra wrote:
>> Jeroen van Meeuwen said the following on 09/22/2008 02:29 PM Pacific 
>> Time:
>>> John Poelstra wrote:
>>>> What process are you proposing so that I know:
>>>> 1) When a spin is officially being proposed as a feature?
>>>> 2) All of the required previous required steps or approvals were 
>>>> completed correctly?
>>> There's more to be tracked on a spin even before it even becomes a 
>>> feature;
>>> - technical review => kickstart pool inclusion
>>> - board trademark approval (which doesn't require a feature page as 
>>> of yet because it can be the last stage a spin wants to go through; 
>>> just being able to remove the rebranding stuff from the kickstart in 
>>> the pool) => fedora spin
>>> - Feature approval => officially composed, released and supported 
>>> fedora spin
>>> Take into account with any design of any process that the former two 
>>> stages need to be tracked, too.
>> This is exactly what I have been asking for.
>> John
> Fair comments...
> Can I assume that the technical review and the trademark approval only 
> occur a single time if the SPIN will not be distributed by fedora? If 
> so, then I would suggest that all spins start with a Feature Page _or_ a 
>  Spin Page (where the former can replace the latter). This would allow 
> for a signoff to be made and recorded by the wiki. The only other 
> alternative I see would be to move it through bugzilla.. and I would 
> tend to prefer the wiki route since all the content could be added to 
> the feature page.
> Thoughts?
> -- bk

"please NOT bugzilla" to the thought of tracking in bugzilla :)

I like the wikipage idea a lot and making it dual purpose as you 
suggest!  So to push the idea a little further... you could start with 
unique page Category names that fall in *before* the Categories we use 
for the feature process now:

Then a quick look at that the history and page change comments and user 
making said change could make it clear who has approved.

This would assume that the approval process is serial AND that each 
approver is relying on the previous approver to have "done the right 
thing".  IOW when a page lands in "FeatureReadyForWrangler" all I would 
have to do (representing the regular feature process) is make sure there 
was approval from the {person, entity, group, SIG} directly preceding 
me.  I would NOT go farther back, beyond that, as that defeats the whole 
chain of trust and efficiency we would seek to gain from it.


More information about the rel-eng mailing list