[PATCH weston v2 00/24] Initial Xwayland window positioning, with XWM reordering

Daniel Stone daniel at fooishbar.org
Tue Jan 17 15:44:32 UTC 2017

Hi Pekka,

On 17 January 2017 at 15:22, Pekka Paalanen <ppaalanen at gmail.com> wrote:
> You can see the state of the current series at:
> https://patchwork.freedesktop.org/series/17096/
> And I have a rebased branch at:
> https://git.collabora.com/cgit/user/pq/weston.git/log/?h=xwm-reorder-WIP
> In the branch I have squashed "desktop-shell: debug
> set_position_from_xwayland" into "shell: implement
> set_xwayland_position".
> Daniel, I think you can skip looking at all patches already R-b
> Quentin, except those where the commit message highlights a regression.
> I'd like some opinions if the regressions are bad and whether we should
> block things for the last patch.
> The last patch "xwm: use _XWAYLAND_ALLOW_COMMITS" cannot be merged
> until the Xwayland patch adding the feature has been accepted.

The first four ('add set_toplevel_with_position to internal API' ..
'implement set_xwayland_position') seem safe to push and are:
Reviewed-by: Daniel Stone <daniels at collabora.com>

'react to geometry changes' makes me a little nervous; is there a way
you can see that this can be resequenced or moved such that we don't
have the intermediate regression?

'split weston_wm_window_draw_decoration()', 'move fini near init in
set_pending_state' and 'split out
weston_wm_window_set_pending_state_OR()' are also R-b me, and seem
like they should be safe to push even without reacting to geometry
changes ... ?

'schedule repaint from MapRequest' looks OK for the second hunk, but
the assert added in the first hunk makes the conditional immediately
above it dead code. Oops.

'debug when weston_surface gets created' and 'explain the
read_properties() in xserver_map_shell_surface()' are R-b me and safe
to push.

With these, we just have the difficult ones left: react to geometry
changes, schedule repaint from MapRequest / do not draw decor twice on
map, and _XWAYLAND_ALLOW_COMMITS. I don't have the mental bandwidth to
deal with those today, but if you're able to gang the easier ones
together and push them, that at least makes this series very tractable

Thanks for this!



More information about the wayland-devel mailing list