Emoji "eye in speech bubble" is not interpreted correctly

[yantene]

There is an emoji named "eye in speech bubble" in the Unicode Emoji List.

eye in speech bubble - Emoji List, v5.0

According to this list, the emoji consists of codepoint U+1F441 U+FE0F U+200D U+1F5E8 U+FE0F,
but the following code returns nil.

Twemoji.find_by(unicode: "\u{1f441}\u{fe0f}\u{200d}\u{1f5e8}\u{fe0f}")

This is because the emoji has a codepoint U+1F441 U+200D U+1F5E8 in this gem.

Codepoint U+FE0F is "emoji presentation selector", a variation selector for emojis.

Similar issues are also found in keycap emojis. (:hash:, :asterisk:, etc.)
On the other hand, :{man,woman}-{kiss,heart}-{man,woman}: emojis contains U+FE0F in this gem.

I think this is a problem.
I think this issue should be fixed, but I have no idea.
What do you think?

[renechz]

Having a related issue, I think:

Twemoji.find_by(unicode: "☝️") # => nil
Twemoji.find_by(unicode: "✌️") # => nil

This seems to consistently happen with emojis that include the U+FE0F as a trailing presentation codepoint. e.g. the :point_up: has a codepoint U+261D in this gem, but is parsed as "261d-fe0f".

Twemoji.unicode_to_str("☝️") # => "261d-fe0f"

As far as I can tell from emoji variation sequences, the presentation codepoint is to say "this is a graphic (e.g. non-text) emoji". Wouldn't it be safe to add the U+FE0F suffix and assume all emojis are "non-text"?

This gem is using emoji-unicode.yml to build Regexp to match the emojis unicode and text. Maybe the fe0f codepoint can be added as an optional group? e.g ":point-up:": 261d(-fe0f)?. (Not good with regex, so I don't really have an idea).

That would likely bring some other issues in itself as I'm not completely familiar with the gem, but it could be a starting point.

[yantene]

@renechz Sorry for the delay in reaction.
Thank you for your mention.

I think that your proposal is not bad and should be implemented.
@JuanitoFatas What do you think?