Unified named colors in RichTextLabel

[dalexeev]

Now the BB tag [color] uses the color names from the Color class.
Also, the use of the # symbol in an HTML color code has become optional.

---

Please note that there is backward incompatibility with several old color names.

[bojidar-bg]

Maybe, to lessen the compatibility breakage, grey and navy could be added to the list of Colors in Color?

[dalexeev]

Should the grey color be #bfbfbf or #808080? If gray and grey are different, it will be strange. If grey is #bfbfbf, then compatibility will still be broken.

[bojidar-bg]

Well, hm, didn't consider that. Feel free to leave it as is, in this case.

[dalexeev]

I watched https://en.wikipedia.org/wiki/X11_color_names and noticed the 'Alternatives' column in the colors table:

I'm not sure if we really need this, but in any case, this is a matter of a separate PR.

[dalexeev]

Rebased and ready to merge.

[aaronfranke]

I tested this, throwing everything I could think of at it (3, 4, 6, 8 digit hex colors, with and without #, and named), and it works perfectly.

Additionally, I love the implementation. I believe the idea was to replace ColorN and from HTML with a constructor that accepts a string, and Color::from_string works perfectly for this. Although that can be done in another PR, since it would also involve removing ColorN which would be a big change.

[dalexeev]

To be honest, it seems to me too that we have too many color constructors from a string: Color::html, Color::named, and now also Color::from_string. In GDScript, Color(string) uses Color::html.

It seems to me that it would be more appropriate if in GDScript Color(string) would use Color::from_string instead. In this case, it will be possible to use both Color("#ff0000") and Color("red"). Then ColorN("red") won't really be needed.

Also, I had a question if the methods find_named_color, get_named_color_count, get_named_color_name, get_named_color should be made private, but I decided not to touch that. First of all, I wanted to fix the inconsistency in RichTextLabel, and I made changes to the core only when necessary. For other changes, we probably need a separate PR.

[akien-mga]

It seems to me that it would be more appropriate if in GDScript Color(string) would use Color::from_string instead. In this case, it will be possible to use both Color("#ff0000") and Color("red"). Then ColorN("red") won't really be needed.

That's what I planned to implement after a recent discussion with @reduz. But if you want to do it in a follow up PR, that'd be very welcome :)

[akien-mga]

Thanks!

[akien-mga]

Welp, should have rebased before merging as there's some more uses of _get_color_from_string to convert.