implement unicode safe api for String

[Yu-zh]

Part of #1452

This PR adds the following APIs for String in builtin for general purpose usage:

charcode_at(String, Int) -> Int
charcode_length(String) -> Int
codepoint_at(String, Int) -> Char
codepoint_length(String) -> Int
unsafe_charcode_at(String, Int) -> Int // charcode_at without bound checking

And adds the following APIs for implementing pattern matching on string:

length_eq(String, Int) -> Bool
length_ge(String, Int) -> Bool
rev_get(String, Int) -> Char

The next step is to change the behavior of op_get and length to codepoint_at and codepoint_length, and fix the various misuses.

One subtlety is that the codepoint_at here has different behavior that code_point_at in JavaScript. Here the index is the count of code points, whereas for JavaScript, the index is the count of utf16 code units. @bobzhang

[peter-jerry-ye-code-review]

‼️ This code review is generated by a bot. Please verify the content before trusting it.

Based on the git diff provided, here are three potential issues I've identified:

1. Inconsistent String Length Behavior: There are two different length-related functions (charcode_length and codepoint_length) which might cause confusion. The commented-out code suggests there's a plan to change the default length() method to use codepoint_length, but this could break existing code that expects UTF-16 code unit counts. This inconsistency could lead to bugs in code that assumes a specific length behavior.

2. Potential Performance Issue: In the length_eq and length_ge functions, there's an O(n) complexity for what could potentially be an O(1) operation if using raw UTF-16 lengths. While this provides more accurate Unicode character counts, it might cause unexpected performance issues in code that frequently checks string lengths.

3. Error Handling Inconsistency: The code uses abort() for error handling in some places (like in codepoint_at and rev_get), but the error messages are inconsistent. For example:

   abort("index out of bounds")  // in some places
   abort("Index out of bounds: cannot access index \{index}")  // in others

   This inconsistency in error messages could make debugging more difficult.

These issues could impact code maintainability and reliability if not addressed.

[coveralls]

Pull Request Test Coverage Report for Build 4853

Details

• 71 of 101 (70.3%) changed or added relevant lines in 4 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.3%) to 82.944%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
string/view.mbt	20	23	86.96%
string/string.mbt	36	45	80.0%
builtin/string.mbt	12	30	40.0%

Totals
Change from base Build 4852:	-0.3%
Covered Lines:	5048
Relevant Lines:	6086

---

💛 - Coveralls