Hi,
thank you all for the tests (also to those who haven't replied to
wine-devel). With the data I figured out how to solve the test results
in this test.
In general I'd like to be as exact as possible with the implementation
and in this case the values look always the same in all logs I got
Hi Rico,
I compiled the 32 bit wine tests on win7 64 bit using cygwin, and sure enough,
there are 40 test failures there. You should probably avoid doing direct
equality tests on floats; see e.g. compare() at line 47 of
d3dx9_36/tests/mesh.c.
You may have to tweak the slop value. (Google floating
