On Tue, 11 Jun 2024 at 09:36, Akihiko Odaki <akihiko.od...@daynix.com> wrote: > > […] I also > > definitely think host cursor integration is useful and valuable, at > > least in absolute pointing mode - for example, when the host system is > > itself being remote controlled, and also to avoid the cursor being > > cropped near the edges of the guest viewport. > > Can you elaborate more about the remote control scenario? I don't think > having extra code just to fix cropped cursor is not worthwhile (I even > feel a bit awkward to make the cursor overflow.)
If you're remote-controlling the host Mac, many VNC/RDP/whatever clients will use a local cursor and simply request cursor image updates from the server and apply them to the local native cursor. That way, there's no lag when moving the cursor even on slower connections, which you'd otherwise get if you had to wait for the regular screen image update, and which can make precise positioning difficult. At any rate, most other Qemu UI frontends likewise implement guest cursors by setting the guest-supplied cursor image on the host's native windowing system's cursor, e.g. gdk_window_set_cursor(). I don't really see a good reason why macOS should be different? Qemu would also hardly be the first VMM on the Mac to pass guest pointers through as NSPointers - Parallels Desktop appears to do the same, for example. > You can add Based-on: to the cover letter to clarify the dependency, and > add "RFC" to the subject if the code is not ready to pull. Please look > at docs/devel/submitting-a-patch.rst for more info. I've done that now, thanks. https://patchew.org/QEMU/20240625134931.92279-1-p...@philjordan.eu/ Phil
