Optimise ColourInfo conversion and add conversion helper method

[smoogipoo]

The performance gains here are mostly theoretical and I can't benchmark them meaningfully in a real-world scenario. Consider this more of a refactoring that adds an optimal (and more user-friendly) method for checking whether a ColourInfo represents a single colour and then returning that colour.

This comes off the back of the osu!-side smoke implementation (where I initially noticed this), which potentially uses this method for tens of thousands of vertices (up to 30k is what I was able to achieve after the recent changes). Similar is also done with slider paths (e.g. at some point https://osu.ppy.sh/beatmapsets/756794#osu/1605148 draws some 20k vertices). So at most you'd see ~0.2ms frametime improvement.

Before

Method	Mean	Error	StdDev	Allocated
ConvertToSRGBColour	14.31 ns	0.139 ns	0.123 ns	-
ConvertToColor4	15.69 ns	0.122 ns	0.102 ns	-

After:

Method	Mean	Error	StdDev	Allocated
ConvertToSRGBColour	6.9628 ns	0.0684 ns	0.0534 ns	-
ConvertToColor4	11.1604 ns	0.1658 ns	0.1551 ns	-
ExtractAndConvertToColor4	0.4039 ns	0.0247 ns	0.0231 ns	-

It uses concepts I've learnt over the years:

• Methods with throw are not inlined - extract these methods to promote inlining.
• Having strings in methods, even if the code is not hit, leads to a significant increase in ASM putting pressure on the JIT.

For comparisons sake, this is the best "real-world" example I could make, by looping TestSceneSmoke over 30 seconds:

before	after

Again, I don't think these are indicative of anything so don't put much weight to them. There's too much noise.

[peppy]

Tests not happy 😢

[turbedi]

There already exists a throw helper for InvalidOperationException here.
Maybe we can use this one?

[frenzibyte]

While this is supposed to be a correct use case of that method, because it causes a string construction in the conversion method, which is what's being avoided in this PR.

[smoogipoo]

Technically it's not a string construction, but rather the code constructing the string results in a lot of dead (i.e. never hit) ASM code for the JIT to compile that could also potentially cause the function to not be inlined.

[smoogipoo]

Fixed tests, the assertion was added post-fact.