Re: [ros-dev] [ros-diffs] [jimtabor] 73475: [Win32SS|RTL] - Fix clipping for Draw Text. Inspired by Katayama Hirofumi MZ patch and ComCtl32/RichEd20/UxThmems. See CORE-2201.

[Thomas Faber]

On 2016-12-22 11:39, [e-mail] wrote:
> +#ifndef _WIN32K_ ///// Fix CORE-2201.
> if (!(flags & DT_NOCLIP) )
> {
>   int hasClip;
>   hrgn = CreateRectRgn(0,0,0,0);
>   if (hrgn)
>   {
>      hasClip = GetClipRgn(hdc, hrgn);
>      // If the region to be retrieved is NULL, the return value is 0.
>      if (hasClip != 1)
>      {
>         DeleteObject(hrgn);
>         hrgn = NULL;
>      }
>      IntersectClipRect(hdc, rect->left, rect->top, rect->right, rect->bottom);
>   }
> }
>
> +#else
> if (!(flags & DT_NOCLIP) )
> {
>   int hasClip;
>   hrgn = NtGdiCreateRectRgn(0,0,0,0);
>   if (hrgn)
>   {
>      hasClip = NtGdiGetRandomRgn(hdc, hrgn, CLIPRGN);
>      if (hasClip != 1)
>      {
>         GreDeleteObject(hrgn);
>         hrgn = NULL;
>      }
>      NtGdiIntersectClipRect(hdc, rect->left, rect->top, rect->right, rect->bottom);
>   }
> }
>
> +#endif /////

Any particular reason not to simply do the NtGdi call in both user
and kernel? Looks like they should be doing the exact same thing.

[James Tabor]

Wow,
too much in my inbox!

Yeah, I guess we can just go with NtGDI only calls.

On Sun, Dec 25, 2016 at 10:35 AM, Thomas Faber [e-mail]
wrote:
> On 2016-12-22 11:39, [e-mail] wrote:
> > +#ifndef _WIN32K_ ///// Fix CORE-2201.
> > if (!(flags & DT_NOCLIP) )
> > {
> >   int hasClip;
> >   hrgn = CreateRectRgn(0,0,0,0);
> >   if (hrgn)
> >   {
> >      hasClip = GetClipRgn(hdc, hrgn);
> >      // If the region to be retrieved is NULL, the return value is
> >      if (hasClip != 1)
> >      {
> >         DeleteObject(hrgn);
> >         hrgn = NULL;
> >      }
> >      IntersectClipRect(hdc, rect->left, rect->top, rect->right,
> >
> > rect->bottom);
> >   }
> > }
> >
> > +#else
> > if (!(flags & DT_NOCLIP) )
> > {
> >   int hasClip;
> >   hrgn = NtGdiCreateRectRgn(0,0,0,0);
> >   if (hrgn)
> >   {
> >      hasClip = NtGdiGetRandomRgn(hdc, hrgn, CLIPRGN);
> >      if (hasClip != 1)
> >      {
> >         GreDeleteObject(hrgn);
> >         hrgn = NULL;
> >      }
> >      NtGdiIntersectClipRect(hdc, rect->left, rect->top,
> >
> > rect->right, rect->bottom);
> >   }
> > }
> >
> > +#endif /////
>
> Any particular reason not to simply do the NtGdi call in both user
> and kernel? Looks like they should be doing the exact same thing.
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev