Migrated randr protocol patch to weston git upstream

[Pekka Paalanen]

On Thu, 09 Apr 2015 17:07:42 +0800
Leslie Zhai <xiang.zhai at i-soft.com.cn> wrote:

> Hi Pekka,
> 
> Thanks for your reply!
> 
> Then how to set mirror (clone) mode for multi-monitors via 
> ~/.config/weston.ini [OUTPUT SECTION] by mode, transform or seat?
> It is often extend mode by default when plugin new monitor, so I 
> consider about XRandR naturally, the xrandr utility can set mirror mode 
> by --output XXX --same-as ...

Weston's compositing does not support mirroring. The damage tracking
does not account for it, AFAIK. More specifically, overlapping outputs
are not taken into account when clearing damage, I believe. It might
work by luck, but I don't think it's intentional or guaranteed. Proving
me wrong there would be nice (by code analysis; casual testing doesn't
count).

Without fixing that, the only way to mirror would be to put up the same
DRM FB up on multiple primary planes. I think there used to be patches
to that effect, where you could say in weston.ini that an output is a
clone of another. However, it doesn't seem those ever landed, perhaps
because of the quality of the implementation or no-one really saw the
benefit to bother reviewing it.

It is a huge leap from that to writing a Weston configuration protocol
and tools, and I don't think any of the security issues have been
solved either.

Unless you can come up with a viable use case that does not involve
desktop usage (because desktop is not a goal for Weston), I can't
really see this going through.

> If there is NO support for weston yet, why not randr protocol skeleton 
> implementation just for fun ;-)

Because it would be an extra maintenance burden. It is the same reason
as why making Weston (upstream) a nice user desktop is a specific
non-goal. If you want a user-friendly desktop, use a DE project that
actually aims for it. Weston only implements enough desktop features
to allow testing Wayland.

I suppose you could keep it as an external Weston module. If you need
to propose patches to Weston upstream to make it work, we might look at
those.


Thanks,
pq

> On 2015年04月09日 16:41, Pekka Paalanen wrote:
> > On Thu, 09 Apr 2015 14:42:33 +0800
> > Leslie Zhai <xiang.zhai at i-soft.com.cn> wrote:
> >
> >> Hi wayland developers,
> >>
> >> I found that Quanxian Wang had implemented randr protocol in 2014
> >> http://lists.freedesktop.org/archives/wayland-devel/2014-April/014091.html
> >>
> >> But I could not find randr relative branch in weston official git
> >> http://cgit.freedesktop.org/wayland/weston/?h=XXX
> >>
> >> I will forked weston official git to my github
> >> https://github.com/xiangzhai/weston to fix randr issue if there is ;-)
> >> And weston had already changed its architecture, so Quanxian`s patch
> >> failed to work for weston git upstream, then I simply migrated it shown
> >> as below:
> >>
> >> diff --git a/Makefile.am b/Makefile.am
> > Hi,
> >
> > sorry, I don't think it belongs in Weston as is anymore. See the
> > discussion about Weston's purpose:
> > http://lists.freedesktop.org/archives/wayland-devel/2014-December/018830.html
> >
> > What is the use case for weston-randr nowadays? Why would you want to
> > cause permanent video mode changes at runtime?
> >
> > If you have a viable non-desktop use case, maybe it could be considered
> > then.
> >
> >
> > Thanks,
> > pq
>