TexturePacker: use std::hash to check for dupes

[lrusak]

This changes TexturePacker to use std::hash instead of the included md5 implementation. This seems to still provide the same Textures.xbt file.

I'm not sure if my implemented std::hash method is really as good as the md5 implementation.

How has this been tested?

produces the same sha256 for Textures.xbt

What is the effect on users?

none

[fritsch]

Some questions:
What's the reasoning behind it? Simplification? Runtime Speed? Code Volume?
You implemented a reversible hash now, when the magic pattern is known, why?
Also seeding with the size when most pictures are maybe 1920x1080 is questionable.
I take this is just for comparing and nothing cryptographically relevant.

Why not use a proper upstream hash implementation? Is it cause of speed constraints? A simple crc32 is fast as hell.

[jenkins4kodi]

I've made some formatting changes to meet the current code style. The diffs are available in the following links:

• PR23296.diff

For more information please see our current code style guidelines.

[neo1973]

Overall nice reduction in code size, just one suggestion to think about.

[neo1973]

Suggested change

	for (auto x : frame.rgbaImage.pixels)
	{
	x = ((x >> 16) ^ x) * 0x45d9f3b;
	x = ((x >> 16) ^ x) * 0x45d9f3b;
	x = (x >> 16) ^ x;
	seed ^= x + 0x9e3779b9 + (seed << 6) + (seed >> 2);
	}
	auto frameHash = std::hash<std::string_view>()(
	std::string_view(reinterpret_cast<const char*>(frame.rgbaImage.pixels.data()),
	frame.rgbaImage.pixels.size()));
	seed ^= frameHash + 0x9e3779b9 + (seed << 6) + (seed >> 2);

Not an expert on hash functions, but it looks to me as this one is intended for integers? My suggestion is to use the std::string_view hasher, I assume it will give good results.

[jenkins4kodi]

@lrusak this needs a rebase