Improve the AsciiStr trait

[nodakai]

AsciiStr::to_lower() and AsciiStr::to_upper() should be renamed to AsciiStr::to_lowercase() and AsciiStr::to_uppercase(), respectively, when Ascii::to_lower() and Ascii::to_upper() are deprecated in favor of Ascii::to_lowercase() and Ascii::to_uppercase().

Incidentally, we should perhaps import capitalize() and capwords() from Python.

[kmcallister]

cc #16801

[aturon]

cc me

[orium]

#17947 fixes the AsciiStr::{to_lower,to_upper}().

Incidentally, we should perhaps import capitalize() and capwords() from Python.

This should probably be a new RFC. If so this issue can be closed.

[nodakai]

Close this issue hoping capitalize() can be piggybacked to #19350

[SimonSapin]

I assume that capitalize means “make upper case the first letter of the string”, and capwords means “make upper case the first letter of each word of the string”.

The problem is that the concepts of “upper case”, “letter”, and “words” are not easy to define for international text. For example, the upper case of i in Turkish is İ, instead of I in other languages. Some languages don’t use spaces to separate words, or don’t have case. And a single string might contain text in multiple languages.

Variations of capitalize and capwords that only operate in the ASCII range are easy to define, but they’re not that useful. Many languages are mostly but not entirely ASCII, so ASCII-only is just wrong for them. E.g the upper case of é is É.

The Unicode Standard and the Unicode Common Locale Data Repository can help do this right, but there are still many possible variations. It’s not obvious which and how many of these should be included in Rust’s standard library.

Because this is a complex topic, I did not include it in rust-lang/rfcs#486. I suggest starting a separate discussion or RFC.

[nodakai]

@SimonSapin Right, I should have thought in the framework of Unicode from the beginning 😛