Proposed F19 Feature: New firstboot

[Adam Williamson]

On Mon, 2013-01-28 at 11:46 +0000, Jaroslav Reznik wrote:
> = Features/NewFirstboot =
> https://fedoraproject.org/wiki/Features/NewFirstboot
> 
> Feature owner(s): Martin Sivák <msivak at redhat.com>
> 
> This feature proposes new initial setup application with better integration to 
> the NewUI anaconda and to Gnome Initial Experience. 
> 
> == Detailed description ==
> Since the Anaconda installer moved to the NewUI Hub and Spoke concept, we can 
> reuse much of it's architecture and screens in the after reboot configuration 
> utility -- initial-setup. So the idea behind the firstboot replacement is that 
> we will have a new app that will use the same Hub and Spoke model and the same 
> API as Anaconda.
> 
> This will give us the possibility of letting the user configure his system 
> either during the package extraction or after reboot (important for OEMs). It 
> will also allow other teams (power management, security team, IPA) to prepare 
> their own screens for Anaconda and initial-setup and so further enhancing the 
> user experience.
> 
> Anaconda, initial-setup and Gnome Inital Experience will communicate to ensure 
> the screens are not shown multiple times. So for example the root password 
> setup or user creation process will be done only in one place, depending on 
> the installed system.
> 
> The old Firstboot will still stay as a fallback in case somebody still has his 
> old Firstboot plugins he needs to use. 

I am concerned about the timeframe on this. The completion percentage
seems rather optimistic:

"Percentage of completion: 70% (the engine is working, package
undergoing review, no configuration screens present)"

To me, 'initial framework coding is complete but we haven't written any
of the code that actually does stuff yet' is a long way short of 70%.
Especially since this is just re-using anaconda's hub/spoke setup, as I
read it, so presumably implementing the 'engine' was the *easy* bit of
the work.

Does FESCo agree that 70% is a realistic completion percentage for the
current state? Has a reliable evaluation of the likely amount of work
involved here, and the necessary time, been completed?

At the least I'd suggest we take care to make sure the old firstboot
works in the F19 context as part of testing, and Martin prioritizes
fixes for it if we find any, so the contingency plan is viable if work
on the new one is not sufficiently complete by Beta.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net