On Tue, Oct 4, 2016 at 10:43 AM, Henri Verbeet hverbeet@gmail.com wrote:
On 3 October 2016 at 17:27, Józef Kucia joseph.kucia@gmail.com wrote:
I think this fix is correct. The device keeps a reference to the implicit swapchain. The implicit swapchain is destroyed just before leaving from wined3d_device_uninit_3d() so it should be safe to call context_acquire().
This problem occurs only for additional swapchains and that's why it happens only in test_swapchain_flip().
You're probably right. (Although that's not to say the code wouldn't benefit from being a little more robust. I think it's hard to reason about.) Is state_unbind_shaders() required? I assume it's because unbinding the shaders needs to happen before destroying the shader backend, but would delaying the context_acquire() call until after the wine_rb_clear() call also work?
Yes, delaying context_acquire() would also work.