01:31 swung0x48: I'm continue working on this MR which intends to hook up Kopper to Android stuff: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30766 . During the process a question arises: seems like the dri2/platform_android stuff is designed against ANativeWindowBuffer*, which isn't a public NDK API?
01:33 swung0x48: There are functions like droid_window_dequeue_buffer() which calls into ANativeWindow_dequeueBuffer() which isn't available publicly
01:39 swung0x48: I don't know the intended designed choices for Kopper here, but I suppose it should be like this on Android, aside from Mesa's internal structure: obtain ANativeWindow* from EGL API -> create VkSurface from that ANW -> create vk swapchain on top of that
01:40 swung0x48: Do I get that correct?
05:53 tjaalton: mesa ftbfs with glibc 2.42 https://gitlab.freedesktop.org/mesa/mesa/-/issues/13797
10:24 zmike: tarceri: are you going to MR that uniforms fix?
10:26 zmike: oh you already did and I missed it
10:26 zmike: strong work
17:03 bl4ckb0ne: is there anything in the vulkan spec regarding VkQueue functions (like vkQueueWaitIdle) once the VkDevice has been destroyed
17:03 bl4ckb0ne: i have a layer tracking queue access and I just hit that case
17:08 Sachiel: last paragraph of https://registry.khronos.org/vulkan/specs/latest/html/vkspec.html#fundamentals-objectmodel-lifetime says that's a no-no
17:09 bl4ckb0ne: ahh lifetime section, i was on the vkGetDeviceQueue part
17:09 bl4ckb0ne: thanks
17:10 bl4ckb0ne: so that should pop in the validation layer