[API design] Remove ability to operate on parts of a CharSequence via startIndex and endIndex?

[cketti]

I've been going back and forth on whether to include functions that support a start and end index to only operate on part of a CharSequence. I've landed on not including/removing such functionality.

Reasons for this are:

• It keeps the API surface small(er).
• CharSequence.codePointAt() can be used without special handling at the end of a partial sequence.
• For a caller it's easy enough to use CharSequence.subSequence().
• Most Java and Kotlin APIs don't support providing a start and end index either, e.g. CharSequence.forEach()

For feature parity with Java's String.codePointCount(int, int) our CharSequence.codePointCount() will be an exception and continue to support a start and end index.

Before changing CharSequence.codePointIterator(Int, Int) to remove the parameters, I'd like to hear arguments from people who wish for the library to keep supporting operations with a start and end index.