Planning for another Wayland release

[Derek Foreman]

On 2018-01-16 04:10 PM, Daniel Stone wrote:
> Hey Bryce,
> 
> On 12 January 2018 at 22:51, Derek Foreman <derekf at osg.samsung.com> wrote:
>> On 2018-01-12 04:00 AM, Daniel Stone wrote:
>>> I did a clearout of the review queue last month, but that's not yet
>>> complete, and I'd like someone to have a more structured look at the
>>> queue (or just to look at all ...) to find stuff that should make it
>>> into the next release. Currently, we have at least Derek's zombie
>>> series, my atomic series and Pekka's clone-mode series, which are all
>>> unlikely to land before Tuesday, but are seemingly very close to
>>> having at least their foundations land.
>>>
>>> I think it would be best to announce a date before FOSDEM, but make
>>> sure we have a good handle on the state of the tree before we do and
>>> make sure we don't unnecessarily drop things. That would include
>>> keeping a close eye on how it develops, rather than just running
>>> release.sh on a strict timer. I was pretty unhappy last time when we
>>> put out a release in the middle of discussing a possibly-critical
>>> patchset; reading the list would've pointed to a few hours' or one
>>> day's delay as being helpful, but in the end the release just abruptly
>>> went out and made the whole thing moot.
>>>
>>> Given how much we could potentially be landing, I think we would need
>>> someone more actively involved (patch-herding, review, testing,
>>> list/IRC discussion) to handle the release process; I'm happy to do
>>> that for this release.
>>
>> I'm able to dedicate time to these tasks for the foreseeable future.
>>
>> If there's interest, I can take on a few release cycles.
> 
> That would be great, thanks! I'm going to land the most invasive
> chunks of the atomic work (up to patch 8, using a drm_plane for the
> scanout/primary plane) tomorrow, and if we can go through rapidly
> enough to land the most dangerous part ('move repaint state
> application to flush'), then landing actual core atomic support should
> be a slam-dunk. It might make sense to try to land some of Pekka's
> output rework as well, depending on the magnitude of API change.

Now the wlroots is a thing, I'm wondering how many users libweston has 
and if those users are lurking here.

Would be nice to know if any of them have opinions on how much work 
dealing with API change is.

You've heard it before, but I'll try to review the remaining atomic bits 
soon.

> I think everything Emil nominated seems good, and I've been wondering
> about a few others to have a closer look at:
> https://patchwork.freedesktop.org/patch/163676/
> https://patchwork.freedesktop.org/series/16338/ (decent feature, been
> sitting for ages, might as well do it with all the other time reworks
> we've been doing)
> https://patchwork.freedesktop.org/series/22883/ (already landed most
> of the rest of it; if it's straightforward enough, might as well)
> https://patchwork.freedesktop.org/series/34624/ (needs rework I think,
> oh wait that happened in
> https://patchwork.freedesktop.org/series/35175/)
> https://patchwork.freedesktop.org/patch/191790/
> https://patchwork.freedesktop.org/patch/193526/ (and the one it was in
> follow up to)
> https://patchwork.freedesktop.org/patch/196619/
> https://patchwork.freedesktop.org/series/35944/ (might as well land
> ivi-shell patches, since we're only partly the effective upstream)

I've reviewed a bit of this now.  Most of it looks like it shouldn't be 
too scary to land, though I'm a bit nervous about the generated symbol 
prefixing stuff, since it's quite core and a long term annoyance if we 
get it wrong (oh, and I'm not sure a consensus was ever reached on how 
it should be accomplished?)

I'm not really capable of reviewing lfrb's xwm series, but it seems 
reasonable in scope to me.

The ivi-shell stuff is also not something I can provide adequate review 
for - you ack/rb it and I'll commit it.  If we're not really the 
upstream, does it make sense to attempt to split ivi-shell off into 
another project?

Michael Teyfel also seems to have a handful of ivi-shell patches patchwork.

> Not all of these might make sense, and there might be some others
> which would be good to land as well. Do you have any opinion on any of
> those, or feel like trying to push any over the line, or ... ?

There seems to be some trivial stuff caught up in patchwork, but maybe 
some of it is just detritus that's already landed or been respun, I'll 
try to grind through that in the coming days.

I'd like to see nerdopolis' seat related stuff get some consideration 
because he's also trying to push for that in enlightenment, and it would 
  be good for both compositors to handle this consistently.

Peter seems to have a patch to bump wl_pointer to v7 since August that 
needs consideration.  At a glance it looks straightforward.

Patchwork is a catastrophe right now.  This will take a bit of time.

> Cheers,
> Daniel
>