[PATCH v3 2/3] winex11.drv: Activate window when restoring from iconic state.

[Zhiyi Zhang]

Some games, e.g., Project CARS waits for WM_ACTIVATE
when restoring from iconic state.

Signed-off-by: Zhiyi Zhang 
---
 dlls/winex11.drv/event.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/dlls/winex11.drv/event.c b/dlls/winex11.drv/event.c
index 6054f645c4..5c465aa033 100644
--- a/dlls/winex11.drv/event.c
+++ b/dlls/winex11.drv/event.c
@@ -1266,6 +1266,8 @@ static void handle_wm_state_notify( HWND hwnd, XPropertyEvent *event, BOOL updat
             {
                 TRACE( "restoring win %p/%lx\n", data->hwnd, data->whole_window );
                 release_win_data( data );
+                if ((style & (WS_MINIMIZE | WS_VISIBLE)) == (WS_MINIMIZE | WS_VISIBLE))
+                    SetActiveWindow( hwnd );
                 SendMessageW( hwnd, WM_SYSCOMMAND, SC_RESTORE, 0 );
                 return;
             }

[Dmitry Timoshkov]

Zhiyi Zhang  wrote:
> @@ -1266,6 +1266,8 @@ static void handle_wm_state_notify( HWND hwnd, XPropertyEvent *event, BOOL updat
>              {
>                  TRACE( "restoring win %p/%lx\n", data->hwnd, data->whole_window );
>                  release_win_data( data );
> -
>            if ((style & (WS_MINIMIZE | WS_VISIBLE)) == (WS_MINIMIZE | WS_VISIBLE))
> -
>                SetActiveWindow( hwnd );
>            SendMessageW( hwnd, WM_SYSCOMMAND, SC_RESTORE, 0 );
>            return;
>        }

Your test doesn't contain the code to check the (WS_MINIMIZE | WS_VISIBLE)
statement, I'd guess a not visible window won't receive an activation event
before window pos messages, but that should be verified with a test that calls
ShowWindow(SW_RESTORE) instead of Alt-tabbing. Also from your test it looks
like that it's the Explorer who actually activates the window before restoring
it, probably you should add similar code to programs/explorer/systray.c to
make your app work also in desktop mode.

[Zhiyi Zhang]

On 6/4/19 4:19 PM, Dmitry Timoshkov wrote:
> Zhiyi Zhang  wrote:
> > @@ -1266,6 +1266,8 @@ static void handle_wm_state_notify( HWND hwnd, XPropertyEvent *event, BOOL updat
> >              {
> >                  TRACE( "restoring win %p/%lx\n", data->hwnd, data->whole_window );
> >                  release_win_data( data );
> > -
> >            if ((style & (WS_MINIMIZE | WS_VISIBLE)) == (WS_MINIMIZE | WS_VISIBLE))
> > -
> >                SetActiveWindow( hwnd );
> >            SendMessageW( hwnd, WM_SYSCOMMAND, SC_RESTORE, 0 );
> >            return;
> >        }
>
> Your test doesn't contain the code to check the (WS_MINIMIZE | WS_VISIBLE)
> statement, I'd guess a not visible window won't receive an activation event
> before window pos messages, but that should be verified with a test that calls
> ShowWindow(SW_RESTORE) instead of Alt-tabbing. Also from your test it looks
> like that it's the Explorer who actually activates the window before restoring
> it, probably you should add similar code to programs/explorer/systray.c to
> make your app work also in desktop mode.

ShowWindow(SW_RESTORE) with invisible minimized window is already tested in user32/tests/msg.c#L5135
with WmShowRestoreMinOverlappedSeq message sequence.

Thanks. I will add it to explorer as well.

[Dmitry Timoshkov]

Zhiyi Zhang  wrote:
> > > @@ -1266,6 +1266,8 @@ static void handle_wm_state_notify( HWND hwnd, XPropertyEvent *event, BOOL updat
> > >              {
> > >                  TRACE( "restoring win %p/%lx\n", data->hwnd, data->whole_window );
> > >                  release_win_data( data );
> > > -
> > >            if ((style & (WS_MINIMIZE | WS_VISIBLE)) == (WS_MINIMIZE | WS_VISIBLE))
> > > -
> > >                SetActiveWindow( hwnd );
> > >            SendMessageW( hwnd, WM_SYSCOMMAND, SC_RESTORE, 0 );
> > >            return;
> > >        }
> >
> > Your test doesn't contain the code to check the (WS_MINIMIZE | WS_VISIBLE)
> > statement, I'd guess a not visible window won't receive an activation event
> > before window pos messages, but that should be verified with a test that calls
> > ShowWindow(SW_RESTORE) instead of Alt-tabbing. Also from your test it looks
> > like that it's the Explorer who actually activates the window before restoring
> > it, probably you should add similar code to programs/explorer/systray.c to
> > make your app work also in desktop mode.
>
> ShowWindow(SW_RESTORE) with invisible minimized window is already tested in user32/tests/msg.c#L5135
> with WmShowRestoreMinOverlappedSeq message sequence.

I see, thanks. However there is no an existing sequence for calling
DefWindowProc(WM_SYSCOMMAND, SC_RESTORE) for a minimized visible window
to compare it with, so that it could be more clear when and how the window
gets activated. Could you please add it as well?

[Zhiyi Zhang]

On 6/4/19 5:30 PM, Dmitry Timoshkov wrote:
> Zhiyi Zhang  wrote:
> > > > @@ -1266,6 +1266,8 @@ static void handle_wm_state_notify( HWND hwnd, XPropertyEvent *event, BOOL updat
> > > >              {
> > > >                  TRACE( "restoring win %p/%lx\n", data->hwnd, data->whole_window );
> > > >                  release_win_data( data );
> > > > -
> > > >            if ((style & (WS_MINIMIZE | WS_VISIBLE)) == (WS_MINIMIZE | WS_VISIBLE))
> > > > -
> > > >                SetActiveWindow( hwnd );
> > > >            SendMessageW( hwnd, WM_SYSCOMMAND, SC_RESTORE, 0 );
> > > >            return;
> > > >        }
> > >
> > > Your test doesn't contain the code to check the (WS_MINIMIZE | WS_VISIBLE)
> > > statement, I'd guess a not visible window won't receive an activation event
> > > before window pos messages, but that should be verified with a test that calls
> > > ShowWindow(SW_RESTORE) instead of Alt-tabbing. Also from your test it looks
> > > like that it's the Explorer who actually activates the window before restoring
> > > it, probably you should add similar code to programs/explorer/systray.c to
> > > make your app work also in desktop mode.
> >
> > ShowWindow(SW_RESTORE) with invisible minimized window is already tested in user32/tests/msg.c#L5135
> > with WmShowRestoreMinOverlappedSeq message sequence.
>
> I see, thanks. However there is no an existing sequence for calling
> DefWindowProc(WM_SYSCOMMAND, SC_RESTORE) for a minimized visible window
> to compare it with, so that it could be more clear when and how the window
> gets activated. Could you please add it as well?

Sure. My pleasure.