Andreas Süßenbach

As I don't have an Android environment to test, I can't tell if this change is valid. But it looks a bit strange... If it actually is an issue, could...

One could introduce a new function `vkb::PhysicalDevice::get_extension_features` that could be used like this: ``` if (gpu.get_extension_features().hostQueryReset) { gpu.request_extension_features().hostQueryReset= VK_TRUE; } ``` ... and would need to adjust all usages of...

Well, I can't interpret what happens here or what should... The only thing that might be an issue here: you should catch by const-reference, not by value.

Have you actually created the DebugUtilsMessenger? ``` vk::DebugUtilsMessengerEXT debugUtilsMessenger = instance.createDebugUtilsMessengerEXT( vk::DebugUtilsMessengerCreateInfoEXT( {}, severityFlags, messageTypeFlags, &debugMessageFunc ) ); ```

I assume, you don't provoke that error from within a destructor? A destructor should not throw. Besides that, I can't repro your issue. Could you provide more sources for inspection?

Hm... I don't know if that matters, but the callback should be `VKAPI_ATTR VkBool32 VKAPI_CALL`.

It's been a while now... Did you have any luck resolving your issue?

Please sorry for the delay. I think, you're right, using `import std` seems to be the right move. Would be great if you could start a PR on it.

@eladmaimoni > opting out of certain library features - for example the UniqueHandle api is probably unnecessary if one uses the raii API. This is already possible by defining VULKAN_HPP_NO_SMART_HANDLE....

> I think that documenting the hpp generator code ... could be helpful. Totally agree. > allowing generation of thinner headers You mean for those "power users" that generate their...