On 21.10.2012 14:17, Henri Verbeet wrote:
2012/10/21 Rico Schüller kgbricola@web.de:
We've generally been moving in the opposite direction for the D3D tests, having each set of tests create its own device, window, etc. instead. This makes the tests much more self-contained, which in turn makes it much easier to debug if you have failures on some setup. I.e., you don't have to worry about some obscure render state that was set 15 tests ago, but affects every test in between, or some memory corruption due to a driver bug that only becomes visible when you're adding a new test. The ddraw/d3d8/d3d9 visual tests are particular bad in that regard. See also e.g. ddraw:ddraw7, etc., or to a lesser extent some of the more recent d3d8/d3d9:device tests.
Ok, makes sense. I'll create the device separately for each test then.