[Automated-testing] [Openembedded-architecture] Yocto Project LTS Proposal/Discussion

[Richard Purdie]

On Mon, 2019-10-28 at 15:12 +0300, Petr Nechaev wrote:
>  I am sorry if this is off-topic, I can only speak as a user of
> yocto.
> But what is the goal of doing half-year releases along with LTS ones
> if we have limited and possibly declining amount of resources ?

There are changes made to the upstream software we derive from all the
time. The goal with master is to try and stay close to that. This means
if we find bugs, its easy to report upstream and is considered best
practise for open source.

Having a release focuses the minds of developers and means we stabilise
the project periodically. We need periods of instability where new
features and changes come in and periods of stabilisation where things
are bugfixed and moved to release quality. The optimal period for doing
this we've found is around every six months. Even within the six months
we have around four milestones which are like mini releases. We've
tried fewer and it led to instability, more frequent and it overloads
people, so again its an empirical feel of what works. The milestones do
again focus minds and help ensure quality, even if they're not made an
official release.

The people keen to develop against the latest things are not the people
who'd do best at stable releases either, the skillsets and interests
are different.

So whilst I understand the question, I think longer release times would
hurt the project and we'd need to do what we do with master anyway,
whether we call it a release or not.

Cheers,

Richard