On 15 September 2015 at 01:37, Józef Kucia jkucia@codeweavers.com wrote:
There's an ugly check if *immediate_context is NULL. It will go away when we have a stub interface for the immediate context.
That usually means the ordering of the patches isn't quite right. :) Although I suppose it's not too bad in this case, but you should at least add some kind of FIXME comment.
- if (swapchain_out)
*swapchain_out = swapchain;- else
IDXGISwapChain_Release(swapchain);
Looking at the associated test, you may have misunderstood what I meant yesterday. If "swapchain_out" is NULL, you're creating a swapchain only to release it right away again. It may be the case that that's correct because of the side effects like the validation of swapchain_desc and potential window messages that end up being sent, but I suspect we should simply skip creating the swapchain.