Re: winex11.drv: WM_SYSCOMMAND/SC_CLOSE should be posted not sent

[Alexandre Julliard]

Dmitry Timoshkov  writes:
> this patch fixes crash on exit reported in the bug #2918.
>
> In DefWindowProc we already use PostMessage for WM_SYSCOMMAND/SC_CLOSE
> in response to Alt+F4.
>
> Changelog:
>     winex11.drv: WM_SYSCOMMAND/SC_CLOSE should be posted not sent.

This is what we used to do, but it was changed to a SendMessage in
commit 43b5576998aea7accd9f4fb791043d89e9b83cad because some apps
depend on that.

winex11.drv: WM_SYSCOMMAND/SC_CLOSE should be posted not sent

[Dmitry Timoshkov]

"Alexandre Julliard"  wrote:
> > this patch fixes crash on exit reported in the bug #2918.
> >
> > In DefWindowProc we already use PostMessage for WM_SYSCOMMAND/SC_CLOSE
> > in response to Alt+F4.
> >
> > Changelog:
> >     winex11.drv: WM_SYSCOMMAND/SC_CLOSE should be posted not sent.
>
> This is what we used to do, but it was changed to a SendMessage in
> commit 43b5576998aea7accd9f4fb791043d89e9b83cad because some apps
> depend on that.

In the log message of that commit I don't see any app mentioned or bug number
referenced. Besides, not only DefWindowProc behaviour already contradicts the
above commit, but spy++ also clearly shows that WM_SYSCOMMAND/SC_CLOSE is
posted, not sent under Windows. What kind of other proof would you prefer to
see?