Re: [PATCH v8 01/15] ui & main loop: Redesign of system-specific main thread event handling

[Phil Dennis-Jordan]

On Sun, 10 Nov 2024 at 08:10, Akihiko Odaki <akihiko.odaki@daynix.com> wrote:
>
> On 2024/11/10 16:08, Phil Dennis-Jordan wrote:
> >
> >
> > On Sun 10. Nov 2024 at 08:01, Akihiko Odaki <akihiko.odaki@daynix.com
> > <mailto:akihiko.odaki@daynix.com>> wrote:
> >
> >     On 2024/11/08 23:46, Phil Dennis-Jordan wrote:
> >      > macOS's Cocoa event handling must be done on the initial (main)
> >     thread
> >      > of the process. Furthermore, if library or application code uses
> >      > libdispatch, the main dispatch queue must be handling events on
> >     the main
> >      > thread as well.
> >      >
> >      > So far, this has affected Qemu in both the Cocoa and SDL UIs,
> >     although
> >      > in different ways: the Cocoa UI replaces the default qemu_main
> >     function
> >      > with one that spins Qemu's internal main event loop off onto a
> >      > background thread. SDL (which uses Cocoa internally) on the other
> >     hand
> >      > uses a polling approach within Qemu's main event loop. Events are
> >      > polled during the SDL UI's dpy_refresh callback, which happens to run
> >      > on the main thread by default.
> >
> >     GTK should also do the same as SDL and requires treatment; I forgot to
> >     note that in previous reviews.
> >
> >
> > Although it‘s possible to build Qemu with GTK support enabled on macOS,
> > that UI doesn’t actually work on macOS at all, and apparently hasn’t
> > been supported since 2018, see:
> > https://stackoverflow.com/a/51474795 <https://stackoverflow.com/a/51474795>
> >
> > I don’t think there’s any point making adjustments to the GTK code by
> > guessing what might be needed if someone did fix that to work with macOS
> > at some point.
>
> But there is a GitLab issue saying it "sometimes" crashes, implying it
> works otherwise:
> https://gitlab.com/qemu-project/qemu/-/issues/2539

I spent some time trying to get the GTK GUI working locally on macOS
with no success (with or without the pvg/vmapple patch set) until I
found various info online suggesting it was broken. But I guess I can
add a qemu_main = NULL; line to gtk_display_init(). This will retain
the existing behaviour both on other platforms (where qemu_main is
NULL anyway) and if anyone does manage to use it on macOS.