Re: Alexandre Julliard : Revert "user32: SetWindowPos() should check the final window position in screen coords before adding SWP_NOMOVE. "

[Dmitry Timoshkov]

Alexandre Julliard  wrote:
> Author: Alexandre Julliard 
> Date:   Sun May  6 11:09:33 2018 +0200
>
> Revert "user32: SetWindowPos() should check the final window position in screen coords before adding SWP_NOMOVE."
>
> This reverts commit 38a76ecb0386d37d30b06b7947ed52be77ec9f1e.
> It has no effect, the coordinates are identical either way.

Just for the record: the coordinates are not indentical after the SetParent()
call, and that patch was needed for SetParent() implementation at that time
because SetParent() called SetWindowPos() directly. Since then the SetParent()
implementation has changed and this no longer has an effect.

Alexandre Julliard : Revert "user32: SetWindowPos() should check the final window position in screen coords before adding SWP_NOMOVE. "

[Alexandre Julliard]

Dmitry Timoshkov  writes:
> Alexandre Julliard  wrote:
> > Author: Alexandre Julliard 
> > Date:   Sun May  6 11:09:33 2018 +0200
> >
> > Revert "user32: SetWindowPos() should check the final window position in screen coords before adding SWP_NOMOVE."
> >
> > This reverts commit 38a76ecb0386d37d30b06b7947ed52be77ec9f1e.
> > It has no effect, the coordinates are identical either way.
>
> Just for the record: the coordinates are not indentical after the SetParent()
> call, and that patch was needed for SetParent() implementation at that time
> because SetParent() called SetWindowPos() directly. Since then the SetParent()
> implementation has changed and this no longer has an effect.

Actually it never had any effect, even in the original state, because at
that point the parent has already been changed. What you meant to do was
to map to screen coordinates relative to the previous parent, but it's
lost by the time we get to SetWindowPos().