00:38 <DemiMarieObenour[m]> dottedmag: indeed that was what I was referring to. The compositor knows which application has focus at any given time, for example.

07:12 <zubzub> https://gitlab.freedesktop.org/wayland/weston/-/blob/main/libweston-desktop/xwayland.c#L154

07:13 <zubzub> shouldn't that be oldgeom.y ?

07:36 <ofourdan> looks like it should :)

07:53 <pq> DemiMarieObenour[m], soreau, X11 child windows are clipped to their parent IIRC, Wayland sub-surfaces are not clipped. So that's a clear mismatch.

07:56 <pq> DemiMarieObenour[m], wayland-scanner is not the only scanner in existence. I think most (all) other language bindings have their own scanners while using libwayland still (for library compat like being able to use EGL).

07:57 <vsyrjala> redirected child windows are not clipped iirc. or maybe i'm misremembering this

08:02 <pq> vsyrjala, I'm not sure how that ever works. What pulls their content into the parent's image? Can they actually be visible outside parent dimensions?

08:03 <pq> *even

08:06 <vsyrjala> hmm. it was probably that their content is defined outside the parent's clip, but when presenting them they still get clipped by the parent

08:13 <pq> vsyrjala, that sounds like it.

10:10 <wlb> wayland-protocols/main: Alex Richardson * tests: allow cross-compiling the tests https://gitlab.freedesktop.org/wayland/wayland-protocols/commit/78f654ed9578 tests/meson.build

10:10 <wlb> wayland-protocols Merge request !119 merged \o/ (tests: allow cross-compiling the tests https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/119)

10:21 <wlb> wayland-protocols/main: Alex Richardson * tests: check whether -Wl,--unresolved-symbols=ignore-all is supported https://gitlab.freedesktop.org/wayland/wayland-protocols/commit/82c89e8fe739 tests/meson.build

10:21 <wlb> wayland-protocols Merge request !118 merged \o/ (tests: check whether -Wl,--unresolved-symbols=ignore-all is supported https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/118)

10:29 <X52> hi, I currently have a problem with libinput. I'm trying out a new Compositor (I want to swich from X11 to Wayland) and the new Compositor has not implemented a config tool for libinput yet. Because there is no config for libinput by default, I tried using pylibinput to change the settings. When I activate tap-to-click for example on my touchpad, I get a Success message, but the behavior dosn't change and libinput list-devices still

10:29 <X52> shows the default settings. Am I doing something wrong or is what I'm trying not possible?

10:30 <pq> X52, it's not possible. When you configure "libinput", you are configuring the instance of libinput in the current process. Changes inside one process do not reflect to other processes like the compositor.

10:34 <pq> X52, FWIW, https://wayland.freedesktop.org/libinput/doc/latest/faqs.html#how-do-i-configure-my-device-on-wayland

10:36 <X52> I saw the documentation, but there is no information on how to change the settings, if the compositor dosn't provide any tools, and thats my current problem.

10:36 <X52> (atleast in wayland)

10:37 <pq> It's a libinput design decision to not offer a configuration store that would be competing and conflicting with compositor and DE settings.

10:37 <pq> so compositor or DE has to provide the way

10:41 <ifreund> X52: this project might be a fine workaround for you until your compositor implements proper input config if you don't mind LD_PRELOAD hacks: https://gitlab.com/kirbykevinson/libinput-config

11:36 <wlb> weston Issue #550 opened by () How to temporarily take control of DRM from Weston? https://gitlab.freedesktop.org/wayland/weston/-/issues/550

13:52 <any1> pq, emersion: How do we want to fix this issue that DemiMarieObenour[m] was talking about earlier and I've summed up in a comment to https://github.com/swaywm/wlr-protocols/issues/62?

13:53 <emersion> the drm lease protocol is affected too

13:54 <pq> any1, sorry, I can't concentrate on that now.

13:57 <any1> pq: No worries. We'll get to it later. I have a work-around for it.

13:59 <any1> I think that it would be easiest to modify the affected protocols to work around the issue, but the problem should be solved in libwayland eventually.

17:20 <DemiMarieObenour[m]> If I understand the problem correctly, even `wl_callback` is affected. I believe that a fix will require annotating all destructor events in the XML and recompiling clients 😞.