Add benchmark tests for color parsing

[GPHemsley]

Spinning off from #314:

With the complexities involved with parsing colors, there are concerned about regressing the time it takes to parse them.

Per @emilio:

But it should be rather trivial to add a few benchmarks for color parsing in src/tests.rs like we have for data: uris etc?

We should look into that.

[sculpt0r]

Hi :)
Do you think this issue could be a good candidate to start contribute? If so, may I take this one?

[emilio]

Hey, sure thing, go ahead!

[sculpt0r]

It has been a long time...
If I'm not mistaken, there is no benchmarking at all in GH workflow, despite bench feature marks in the test code.

I'm fully rely on the docs: https://doc.rust-lang.org/unstable-book/library-features/test.html

So, part of this issue solve will be adding some custom workflow?

[emilio]

I don't think we necessarily need a custom workflow, just something that can be benchmarked easily locally.

[emilio]

Of course the more we can automate after that the better, but that's not a strict requirement imo.