StrPrelude::char_at() confuses people by accept byte index instead of char index

[liigo]

If we do not want to change its semantic, I would like propose to rename it to char_at_byte_position, or just remove the method. It is not very useful to get a char from byte index. Similar to char_at_reverse.

[steveklabnik]

/cc @aturon

[Gankra]

There is a standard here: all index-based operations work on bytes. This is the only reasonable choice for a "random" access operation on a UTF8 string. We could definitely consider renaming to code_point_at, or just remove the method, though.

I do not believe we have done a fine-grain stablization pass for str, only String. See #17438 for details.

[alexcrichton]

Yes, as @gankro mentions this is consistent with our current conventions with indexing and strings. I'm working on a patch for the fine-grained stabilization of str but I'm likely going to leave these methods as unstable for now as it's unclear how certain we are of them in a 1.0 world.

For now though I'm going to close this issue as the usage of indexing and lack of mentioning "byte" is consistent with our current conventions with string apis.