Rewrite of the graphics and present queue index determination

[asuessenbach]

Replaces #93.

[asuessenbach]

Note, that this should not be merged yet!

In case the graphics and the present index are different, the following code would fail:

       vk::DeviceQueueCreateInfo deviceQueueCreateInfo{ .queueFamilyIndex = graphicsIndex, .queueCount = 1, .pQueuePriorities = &queuePriority };
        vk::DeviceCreateInfo      deviceCreateInfo{ .pNext = &featureChain.get<vk::PhysicalDeviceFeatures2>(),
                                                    .queueCreateInfoCount = 1,
                                                    .pQueueCreateInfos = &deviceQueueCreateInfo,
                                                    .enabledExtensionCount = static_cast<uint32_t>(requiredDeviceExtension.size()),
                                                    .ppEnabledExtensionNames = requiredDeviceExtension.data() };
        device = vk::raii::Device( physicalDevice, deviceCreateInfo );
        graphicsQueue = vk::raii::Queue( device, graphicsIndex, 0 );
        presentQueue = vk::raii::Queue( device, presentIndex, 0 );

It's just the gaphicsIndex used to create a queue, but the presentIndex is used to get a queue out of the device. That should fail!

The question now is: should the index determination be simplified, such that just one index is determined, and just one queue is requested? Or should the device creation part be extended to correctly handle different graphics and present queue?

[SaschaWillems]

Tbh. the no. of devices that don't support present on graphics is so miniscule and if compared to how much more code that adds, I'd just skip that and simply mentioned it in the docs. I don't think anyone doing the tutorial will ever run into such a configuration.

[asuessenbach]

@SaschaWillems Totally agree.

Just asking for a queue with both graphics and present support makes it much easier.

[gpx1000]

Yep, agreed that this is easier and there's little chance anyone will ever really run into a need to do it the current way.

[asuessenbach]

OK, carried the modifications over to all the other chapters.

Please pay special attention to chapters 33 and 34, which were slightly different than the others.

Note, that I could not run any of the programs, as I'm currently not able to build them. That is, would be great if someone else could verify, that they are still alive.

[gpx1000]

I'll review and run all of them. Working on it now.

[gpx1000]

Alright, it seems everything is working for me here in Linux.

[SaschaWillems]

Working fine for me on Windows

[SaschaWillems]

Since we don't have a strict "three approvals" rule over here: Should we just merge this?

[gpx1000]

Yes, we absolutely should. I'll merge it now. I was waiting for the docs call that was yesterday in case we were going to go over them.