https://bugs.winehq.org/show_bug.cgi?id=17982
Alexander Almaleh sashoalm@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |sashoalm@gmail.com
--- Comment #7 from Alexander Almaleh sashoalm@gmail.com --- Created attachment 48984 --> https://bugs.winehq.org/attachment.cgi?id=48984 Patch that makes a "naive" solution to the problem by calling Polygon() inside PatBlt()
I have a patch that "naively" substitutes PatBlt() with an appropriate call to Polygon(), which does support world transform (I set a NULL_PEN temporarily because Polygon() uses the pen while PatBlt() does not).
This fixes the behavior of PatBlt(), however it might have performance consequences, depending on whether Polygon() is optimized when drawing normal rectangles.
If anyone is interested, I can improve it to call Polygon() only when needed - it will check if the rectangle is actually going to be inclined and call Polygon() only then.
I already tested that this approach works for PatBlt(), but only when using PAT_COPY. I have not yet tested with the other ROPs.