Derive PartialEq for AlphaColor, OpaqueColor, PremulColor

[waywardmonkeys]

No description provided.

[waywardmonkeys]

A possible issue here is that DynamicColor has a manual impl for PartialEq where it directly does a comparison of the bits.

[raphlinus]

Riiight, that's the reason I didn't do this before.

I think we need a PartialEq, so I'm approving this. What's far less clear is what semantics it should have. In particular, I don't think we should guarantee that a color component of 0.0 is equal to a color component of -0.0, ie reserve the right to make that change. Of course we're making no stability guarantees at this point.

[tomcur]

I wonder whether it makes sense to follow #75 and impl PartialEq, Hash based on the bits. There is some divergence between the types, with DynamicColor not being bytemuck::Pod, so maybe that informs the decision. An alternative is to provide the required functionality on DynamicColor (maybe to_bits) so users can themselves implement Hash, Eq on some wrapping type.

[tomcur]

(Happy for this to land as-is.)

[waywardmonkeys]

Landed this as is ... I need to do some stuff for Default as well so that we can use this with Parley which requires that a Brush impl PartialEq and Default.