Sorry I didn't get to this sooner. I left some comments with the general idea of keeping the d3dx10/11 tests as close as possible, now and going forward. Ideally it would be possible to do `diff -u dlls/d3dx10_43/tests/d3dx10.c dlls/d3dx11_43/tests/d3dx11.c` and quickly discern trivial differences (mainly all the s/10/11/) from "intended" ones.