What does extended F30 cycle mean for F29?

[Kalev Lember]

If F31 is delayed by 6 months and F30 is supported for 6 months longer,
does it mean F29 *also* automatically gets a longer cycle since it by
policy becomes EOL when F31 is out + 1 month?

Can we EOL F29 6 months before F31 is out to not have *two* long term
branches to maintain?

-- 
Kalev

[Peter Robinson]

On Wed, Nov 28, 2018 at 7:14 AM Kalev Lember <kalevlember(a)gmail.com&gt; wrote:
>  If F31 is delayed by 6 months and F30 is supported for 6 months longer,
>  does it mean F29 *also* automatically gets a longer cycle since it by
>  policy becomes EOL when F31 is out + 1 month?
>
>  Can we EOL F29 6 months before F31 is out to not have *two* long term
>  branches to maintain? 
So going back to memory of what we did in the F-21 cycle, we EOLed
F-19 at the 13 months line and then F-20 continued until a month after
F-22 was out.

[Tom Hughes]

On 28/11/2018 08:45, Peter Robinson wrote:
>  On Wed, Nov 28, 2018 at 7:14 AM Kalev Lember
> <kalevlember(a)gmail.com&gt; wrote:
> >
> > If F31 is delayed by 6 months and F30 is supported for 6 months longer,
> > does it mean F29 *also* automatically gets a longer cycle since it by
> > policy becomes EOL when F31 is out + 1 month?
> >
> > Can we EOL F29 6 months before F31 is out to not have *two* long term
> > branches to maintain?
>  
>  So going back to memory of what we did in the F-21 cycle, we EOLed
>  F-19 at the 13 months line and then F-20 continued until a month after
>  F-22 was out. 
End of life is not technically defined as a given number of months but
rather it is defined as one month after the release of the next but one
release so unless that definition was changed that is what would happen:

https://fedoraproject.org/wiki/Fedora_Release_Life_Cycle#End_of_Life_.28E...

Tom

-- 
Tom Hughes (tom(a)compton.nu)
http://compton.nu/

[Matthew Miller]

On Wed, Nov 28, 2018 at 08:45:55AM +0000, Peter Robinson wrote:
>  > If F31 is delayed by 6 months and F30 is supported for 6 months
> longer,
>  > does it mean F29 *also* automatically gets a longer cycle since it by
>  > policy becomes EOL when F31 is out + 1 month?
>  > Can we EOL F29 6 months before F31 is out to not have *two* long term
>  > branches to maintain?
>  So going back to memory of what we did in the F-21 cycle, we EOLed
>  F-19 at the 13 months line and then F-20 continued until a month after
>  F-22 was out. 
I'd support doing this again.

-- 
Matthew Miller
<mattdm(a)fedoraproject.org&gt;
Fedora Project Leader

[Kevin Kofler]

Peter Robinson wrote:
>  So going back to memory of what we did in the F-21 cycle, we EOLed
>  F-19 at the 13 months line and then F-20 continued until a month after
>  F-22 was out. 
No, we did not! F19 was supported until a month after F21 was released:
https://fedoraproject.org/wiki/Releases/21/Schedule ← F21 release 2014-12-09
https://fedoramagazine.org/fedora-19-eol-01-06-2015/ ← F19 EOL 2015-01-06
and that it what needs to happen to meet the promises we made to users. The 
whole point of the support cycle is to allow upgrading to Fn+2 directly. 
EOLing the release early because Fn+2 was delayed totally defeats the point.

If we are not prepared to do that, we simply must not prolong the F30 cycle 
arbitrarily. I don't see why we need to do that to begin with. 
Infrastructure work can be done in parallel with distribution development. 
Extending the release cycle is only useful if it actually leads to extended 
support times for the releases!

        Kevin Kofler

[Artur Iwicki]

No, we did NOT do that.

A quick search on the devel list archives shows that F19 was released in July 2013:
https://lists.fedoraproject.org/archives/list/[e-mail].o...
...and EOL'd on in January 2015:
https://lists.fedoraproject.org/archives/list/[e-mail].o...
...which gives F19 a 19-month support window.

I wasn't a packager back then, so it may be easy for me to say this, but either way -
I agree with Kevin Koffer. The current Release Life Cycle states that release N is
supported for one month after N+2 is released. This is a promise we make to our users, and
breaking this promise would be a very nasty thing to do. If we arrive at the conclusion
that we don't have the manpower to extend F29's support window past the usual 13
months, then I argue that we should NOT prolong the F30 development cycle.