On 3/12/20 8:05 AM, Matteo Bruni wrote:
Signed-off-by: Matteo Bruni mbruni@codeweavers.com
I stared at invert_swizzle() an inordinate amount of time and I eventually convinced myself it's correct, or at least sensible. I'm going to require more swizzle tests when the compiler will actually start to work. I assume you already have a bunch for this but, obviously, tests with multiple LHS swizzles would be particularly interesting.
I'll admit I'm not happy about the code at all, but it's a complex operation and I haven't proven skilled enough to write anything better.
I'll at least add some more tests.