On 03/09/2016 06:37 PM, Henri Verbeet wrote:
On 9 March 2016 at 16:26, Paul Gofman gofmanp@gmail.com wrote:
Thank you for the review. I will fix that, and resubmit probably after I get the comments from Matteo for the other patches in the series. Regarding D3DXVECTOR4, it has floats. I wanted to compare exact values.
You store floats now too, although you initialise them with doubles. (Unless you're talking about the "result" values in test_effect_preshader_op_results[]?)
Yes, I am talking here solely about "result" values from test_effect_preshader_op_results. These are the only ones which are computed and then compared.
The other float constants which are introduced in natural float representation are either directly copied inside the effect framework and compared with the same table in the test, or converted to int in the effect and used as indices. Initializing them with doubles does not affect the result but it is my inaccuracy and I will fix this either.