strip_identifier can cause a panic with cjk identifiers

[mznet]

Calling strip_identifier with identifiers that contain cjk characters causes a Rust panic as shown below

assert!(is_valid_javascript_identifier("한글"));

thread 'js_identifiers::tests::test_is_valid_javascript_identifier' (1076350) panicked at src/js_identifiers.rs:49:12:
byte index 4 is not a char boundary; it is inside '글' (bytes 3..6) of `한글`
stack backtrace:
   0: __rustc::rust_begin_unwind
             at /rustc/f8297e351a40c1439a467bbbb6879088047f50b3/library/std/src/panicking.rs:698:5
   1: core::panicking::panic_fmt
             at /rustc/f8297e351a40c1439a467bbbb6879088047f50b3/library/core/src/panicking.rs:75:14
   2: core::str::slice_error_fail_rt
   3: core::str::slice_error_fail
             at /rustc/f8297e351a40c1439a467bbbb6879088047f50b3/library/core/src/str/mod.rs:69:5
   4: core::str::traits::<impl core::slice::index::SliceIndex<str> for core::ops::range::Range<usize>>::index
             at /Users/mjet.plane/.rustup/toolchains/1.91.0-aarch64-apple-darwin/lib/rustlib/src/rust/library/core/src/str/traits.rs:248:21
   5: core::str::traits::<impl core::slice::index::SliceIndex<str> for core::ops::range::RangeInclusive<usize>>::index
             at /Users/mjet.plane/.rustup/toolchains/1.91.0-aarch64-apple-darwin/lib/rustlib/src/rust/library/core/src/str/traits.rs:664:33
   6: core::str::traits::<impl core::slice::index::SliceIndex<str> for core::ops::range::RangeToInclusive<usize>>::index
             at /Users/mjet.plane/.rustup/toolchains/1.91.0-aarch64-apple-darwin/lib/rustlib/src/rust/library/core/src/str/traits.rs:751:24
   7: core::str::traits::<impl core::ops::index::Index<I> for str>::index
             at /Users/mjet.plane/.rustup/toolchains/1.91.0-aarch64-apple-darwin/lib/rustlib/src/rust/library/core/src/str/traits.rs:63:15
   8: sourcemap::js_identifiers::strip_identifier
             at ./src/js_identifiers.rs:49:12
   9: sourcemap::js_identifiers::is_valid_javascript_identifier
             at ./src/js_identifiers.rs:54:5

Using cjk characters in identifiers is not common, but I found examples where cjk characters are used as identifiers and the panic happened.
Javascript identifiers are not limited to ascii characters and can include unicode characters.

The current implementation stored only the start byte index of each character in end_idx while iterating, and then sliced the string using an inclusive range &s[..=end_idx].

For example, the string "한글", '한' occupies bytes 0–2 and '글' occupies bytes 3–5, but when processing '글', i = 3 is stored as end_idx and slicing with &s[..=3] breaks the UTF-8 character boundary.
This results in a byte index is not a char boundary panic.

To fix, the code now tracks the end position of each character instead of the start position by calculating end_idx = i + c.len_utf8(), and uses an exclusive range &s[..end_idx] when slicing.

This change covers not only CJK characters but also other non-ASCII Unicode identifiers.

[loewenheim]

Hi, thank you very much for the report and fix. Nice catch!

ETA: Can you please run cargo fmt and update the PR?

[mznet]

@loewenheim I applied cargo fmt to the code to improve readability.