Regexp is slow -- could use codepoint ranges for consecutive emojis

[mjwillson]

Currently the regexp is quite slow, likely due to it containing a separate union term for each emoji. A lot of the emojis are consecutive codepoints, and some like the national flag emojis could be specified concisely as [\U0001F1E6-\U0001F1FF]{2}

[hugovk]

Should this be closed now get_emoji_regexp has been removed in version 2.0?

https://carpedm20.github.io/emoji/docs/#regular-expression