2015-02-20 13:29 GMT+01:00 Stefan Dösinger stefandoesinger@gmail.com:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi,
I am working on a more comprehensive test for V8U8, V16U16, L6V5U5, X8L8V8U8 and Q8W8V8U8. I have attached the current status of this test. I am still struggling with finding the right precision that makes all Windows machines happy. There are also some problems with our conversion code.
Yes, I've noticed that too. I haven't really looked into it but the float-to-signed conversion has always been a bit tricky. In OpenGL AFAIK the conversion formula changed at some point (e.g. see table 2.9 in the 2.1 spec vs paragraph 2.3.5.1 in the 4.5 spec), it might be we have to use one instead of the other in our fallback conversion functions too. No idea if something like that applies to Windows drivers too.
Your test tests the case of a non-even width, whereas my test doesn't do that. I can build that into my test as well.
We can still commit your test if you want, and I'll remove it again in my patch (together with x8l8v8u8_test).
I just wrote (more like copypasted and edited) this test to make sure I didn't screw up patch 1/5 and since I wrote it I though "why not send it too". It's fine for me either way.