Bug 2149338

Merge request proposed: https://gitlab.com/redhat/centos-stream/rpms/wayland-protocols/-/merge_requests/6

I realize that I just bumped wayland-protocols only a bit ago in bug 2139449, but I didn't know that the fractional scaling protocol was going to get merged so soon. That protocol immediately got an implementation that landed in kwin, which caused an unusual bump in the protocols version required again. I usually try to avoid doing this if I don't need to.

Also, at least wayland-protocols is a purely additive, backwards compatible upgrade. It's just a bunch of definitions that implementors scan to make Wayland protocol code.

Finally, there's even an implementation of this protocol coming for Mutter, which would mean this protocol would be of benefit to GNOME users too: https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2394

Given how many computers *require* fractional scaling for HiDPI (outside of MacBooks, most HiDPI screens are in configurations that need 125% to 175% scaling factors), I did not bother to push back on having the protocols upgrade so soon again.

(In reply to Neal Gompa from comment #3)
> I realize that I just bumped wayland-protocols only a bit ago in bug
> 2139449, but I didn't know that the fractional scaling protocol was going to
> get merged so soon. That protocol immediately got an implementation that
> landed in kwin, which caused an unusual bump in the protocols version
> required again. I usually try to avoid doing this if I don't need to.

Thanks for making an extra effort to provide context Neal :-)

> Also, at least wayland-protocols is a purely additive, backwards compatible
> upgrade. It's just a bunch of definitions that implementors scan to make
> Wayland protocol code.

True. Do mind that QE is always required to do some smoke testing, and that there is some process overhead as well for us, but you're right that in this case it should be fine.

> Finally, there's even an implementation of this protocol coming for Mutter,
> which would mean this protocol would be of benefit to GNOME users too:
> https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2394

Note that we usually don't rebase GNOME packages in a major RHEL version.

> Given how many computers *require* fractional scaling for HiDPI (outside of
> MacBooks, most HiDPI screens are in configurations that need 125% to 175%
> scaling factors), I did not bother to push back on having the protocols
> upgrade so soon again.

Note that fractional scaling is already possible from the compositor's side without the protocol. The protocol just allows clients to do the fractional scaling themselves

(In reply to Niels De Graef from comment #4)
> (In reply to Neal Gompa from comment #3)
> > I realize that I just bumped wayland-protocols only a bit ago in bug
> > 2139449, but I didn't know that the fractional scaling protocol was going to
> > get merged so soon. That protocol immediately got an implementation that
> > landed in kwin, which caused an unusual bump in the protocols version
> > required again. I usually try to avoid doing this if I don't need to.
> 
> Thanks for making an extra effort to provide context Neal :-)
> 
> > Also, at least wayland-protocols is a purely additive, backwards compatible
> > upgrade. It's just a bunch of definitions that implementors scan to make
> > Wayland protocol code.
> 
> True. Do mind that QE is always required to do some smoke testing, and that
> there is some process overhead as well for us, but you're right that in this
> case it should be fine.
> 
> > Finally, there's even an implementation of this protocol coming for Mutter,
> > which would mean this protocol would be of benefit to GNOME users too:
> > https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2394
> 
> Note that we usually don't rebase GNOME packages in a major RHEL version.
> 

I didn't know that, actually. RHEL 7 received several GNOME rebases over its life.

I didn't pay much attention to RHEL 8's GNOME stack, but I did see some GNOME components getting upgraded over its life. So from my point of view, I didn't have a reason to believe that GNOME *wouldn't* be upgraded at some point.

Even without rebases, that doesn't rule out the potential for backporting changes to enhance the experience.

> > Given how many computers *require* fractional scaling for HiDPI (outside of
> > MacBooks, most HiDPI screens are in configurations that need 125% to 175%
> > scaling factors), I did not bother to push back on having the protocols
> > upgrade so soon again.
> 
> Note that fractional scaling is already possible from the compositor's side
> without the protocol. The protocol just allows clients to do the fractional
> scaling themselves

While that is true, it essentially makes it possible to avoid having to scale to the next double and downscale (which leads to blurry/fuzzy applications). It's a huge complaint about Wayland environments and the biggest reason why Wayland fractional scaling isn't exposed by default in GNOME.

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (wayland-protocols bug fix and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2023:2300