Proposal: no more alphas/betas for Wayland releases

[Jonas Ådahl]

On Wed, May 21, 2025 at 05:43:09AM -0400, Neal Gompa wrote:
> On Tue, May 20, 2025 at 4:03 PM Simon Ser <contact at emersion.fr> wrote:
> >
> > Hi all,
> >
> > With years passing by, development in the main Wayland repository has
> > slowed down quite a bit, activity has moved over to wayland-protocols
> > and compositors. However, cutting a new Wayland release is still a
> > heavyweight process: it takes at least one and a half months with at
> > least 3 pre-releases. I'm also not sure about the value of all of these
> > pre-releases: historically they've been used to push the last features
> > over the fence before the RCs, but it's easy enough to talk and
> > coordinate over the bits that we want to wait on for the release.
> >
> > I would suggest to drop the alphas/betas from the release process, ie.
> > go straight to RC1. The process would then continue as usual, with
> > weekly RCs. As a release manager this would help reduce the load. This
> > is also what I've been doing for Sway and wlroots for a very long time.
> >
> > Would this make sense? Are there other reasons why alphas/betas were
> > valuable?
> >
> 
> Funnily enough, I think wlroots should probably have them since
> wlroots releases are so highly disruptive for basically every
> consumer...
> 
> That said, going straight to RCs for libwayland itself seems
> reasonable. But, do we even need those? Could we go straight to
> release since the churn is so low?

I think it's better to start with skipping alpha/beta for now, so that
there is at least *some* testing before things end up in a release. This
is different from wayland-protocols, since they don't end up causing any
actual changes on people's system until servers and clients make use of
and ship them.


Jonas

> 
> 
> 
> -- 
> 真実はいつも一つ！/ Always, there's only one truth!