Support `is_letter()` on `char`.

[daschl]

Unicode defines the category L for letter similar to others. We already have different accessors on char based on unicode and it would make sense to also check if the char is on a letter boundary (so for example you can split on non-letter boundaries). Go has a similar function available in its unicode package: https://golang.org/pkg/unicode/#IsLetter

Right now I'm using the unicode_categories crate but I think it has some bugs and I wonder if such basic unicode properties on char make sense to be part of stdlib directly.

[estebank]

Doesn't char::is_alphabetic fulfill this need? After cursory look at the code, it seems to be doing exactly what you'd want.

[frewsxcv]

From what I understand, "alphabetic" and "letter" are different:

Alphabetic: Lu | Ll | Lt | Lm | Lo | Nl | Other_Alphabetic

-- http://unicode.org/reports/tr44/#Alphabetic

Letter: Lu | Ll | Lt | Lm | Lo

-- http://unicode.org/reports/tr44/#General_Category_Values

[daschl]

I think @frewsxcv is right and I also think it makes sense to expose those "basic" unicode categories on char since its incredibly useful for tokenizers and the like.

[estebank]

#38137 tracks the stabilization is_letter, which is being introduced in PR #38125.

[SimonSapin]

Right now I'm using the unicode_categories crate but I think it has some bugs

Have you opened issues for these bugs, or checked that they’re already in the issue tracker?

[daschl]

@SimonSapin yes and did address them subsequently swgillespie/unicode-categories#5 . That said, I still think is_letter has its place int he standard library

[alexcrichton]

Proposed in #38125 the libs team decided to close, so I'm going to close this as well.