clang -Wimplicit-fallthrough and missing breaks

[Stephan Bergmann]

On 05/25/2016 09:01 AM, Tor Lillqvist wrote:
>
>         --- a/vcl/opengl/win/gdiimpl.cxx
>         +++ b/vcl/opengl/win/gdiimpl.cxx
>         @@ -166,6 +166,7 @@ static LRESULT CALLBACK WndProc(HWND hwnd,
>         UINT message, WPARAM wParam, LPARAM l
>                  case VK_SPACE:
>                      break;
>                  }
>         +        SAL_FALLTHROUGH; //TODO ???
>              default:
>                  return DefWindowProc(hwnd, message, wParam, lParam);
>              }
>
>
> I did not write this code, but looking at it, it seems fairly obvious
> that the fall-through is intentional here. But it is beyond my
> understanding why such a temporary window, that is used only for some
> milliseconds, and never even displayed, would need to handle the Escape
> and Space keys specially.

But it does not handle VK_SPACE specially:

> static LRESULT CALLBACK WndProc(HWND hwnd, UINT message, WPARAM wParam, LPARAM lParam)
> {
>     switch (message)
>     {
>     case WM_CREATE:
>         return 0;
>     case WM_CLOSE:
>         PostQuitMessage(0);
>         return 0;
>     case WM_DESTROY:
>         return 0;
>     case WM_KEYDOWN:
>         switch(wParam)
>         {
>         case VK_ESCAPE:
>             PostQuitMessage(0);
>             return 0;
>
>         case VK_SPACE:
>             break;
>         }
>         SAL_FALLTHROUGH; //TODO ???
>     default:
>         return DefWindowProc(hwnd, message, wParam, lParam);
>     }
> }

In the inner switch (in case WM_KEYDOWN), VK_SPACE and the (implicit) 
default both fall through to the outer switch's default case, which 
looks rather dubious to me (why mention VK_SPACE explicitly, then?). 
The code is like that ever since 
<https://cgit.freedesktop.org/libreoffice/core/commit/?id=0f6fca34909535b48bad41e73e5d3d0e86c744b9> 
"add anti-aliasing init"; whose author is in CC now.