multiple wl surfaces commit in one client

[Simon Ser]

On Friday, November 6, 2020 10:38 AM, Jonas Ådahl <jadahl at gmail.com> wrote:

> On Fri, Nov 06, 2020 at 07:51:26AM +0000, Simon Ser wrote:
>
> > Hi,
> > On Friday, November 6, 2020 3:14 AM, zou lan nancy.lan.zou at gmail.com wrote:
> >
> > > Hi Simon & pekka
> > > Thank you for your reply!
> > >
> > > > > The OS could pre-empt the client after
> > > > > the first wl_surface_commit is flushed on the wire and before the
> > > > > second one is.
> > >
> > > I want to ask if after first wl_surface_commit, but I don't call
> > > wl_display_flush/wl_display_dispatch, is the first wl_surface_commit
> > > flushed to server because  out buffer of the connection is full?
> >
> > That is correct.
> >
> > > https://github.com/wayland-project/wayland/blob/master/src/connection.c,
> > > can we increase the buffer size in this structure wl_buffer to reduce
> > > the possibility of this?
> >
> > I'd rather not. This would not fix the issue, and it's unclear if it'd
> > really help.
> >
> > > By the way, for the WIP protocol,
> > > https://gitlab.freedesktop.org/jadahl/wayland-protocols/-/blob/wip/transactions/unstable/transactions/transactions-unstable-v1.xml,
> > > do we have any patchset in weston to implement it? I don't find it in
> > > weston master branch.
> >
> > I'm not aware of any Weston implementation as of yet. As you can see,
> > the protocol hasn't yet been merged in wayland-protocols. Writing a new
> > implementation would certainly help getting the protocol merged.
>
> weston, mutter and gtk implementations are linked to from this comment:
>
> https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/26#note_554405

Damn, I completely missed that. Thanks for the heads-up!