add StringView

[Yu-zh]

Basic usage:

let str = "Hello🤣🤣🤣"
guard let Some(start) = CharOffset::offset_by(str, 1)
guard let Some(end) = CharOffset::offset_by(str, 6)
let view = str[start:end]
inspect!(view, content="ello🤣")

To create a CharsView from a String, we need to use the CharOffset as indices. CharOffset is an opaque type that represents a position in the utf-16 String. It can be constructed via CharOffset::offset_by(str: String, offset: int) where the offset is the number of characters skipped between the start of the string and the returning position. It can also be constructed via CharOffset::unsafe_from_int(index: Int).

This is to avoid confusion with indexing String using indices, i.e. indices indicate utf-16 code unit offset when accessing String whereas when viewing String the indices represents offsets of characters.

Once we hold a CharsView, we can use integers directly to access chars or create subviews, and here the index always represents the codepoint offset.

---

Update:
renamed the types and methods:

CharsView -> StringView
CharOffset -> StringIndex
offset_by -> index_at

[peter-jerry-ye-code-review]

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

Here are three potential issues or observations from the provided git diff output:

1. Inconsistent Documentation for index_at Function:

   • The index_at function in view.mbt has a parameter start~ : StringIndex = 0, but the documentation example uses start=offset instead of start~=offset. This inconsistency in naming conventions between the function signature and the example could lead to confusion for users. The documentation should align with the actual parameter names used in the function.

2. Potential Performance Issue in op_as_view:

   • In the op_as_view function for StringView, there is a comment mentioning a TODO: "provide index_at_rev or index_at2 to avoid repeatedly iterate the string." This suggests that the current implementation might have performance issues due to repeated string iteration. This should be addressed to improve efficiency, especially for large strings or frequent operations.

3. Error Handling in op_get Function:

   • The op_get function in view.mbt checks for invalid surrogate pairs and aborts with an error message. However, the error message "invalid surrogate pair" is not very descriptive. It would be more helpful to include additional context, such as the position or the specific characters involved, to aid in debugging.

These issues should be reviewed and addressed to improve code clarity, performance, and maintainability.

[coveralls]

Pull Request Test Coverage Report for Build 4682

Details

• 38 of 44 (86.36%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.02%) to 83.217%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
string/view.mbt	38	44	86.36%

Totals
Change from base Build 4681:	0.02%
Covered Lines:	4765
Relevant Lines:	5726

---

💛 - Coveralls

[bobzhang]

try to move StringIndex/StringView into string package