refactor(parser): improve safety of char to bytes conversions (#13193)

`to_bytes` converts a `char` to a series of `N` bytes. Add an assertion
to make sure that the character *is* actually `N` bytes long.

Also, const assert the value of `LOSSY_REPLACEMENT_CHAR_FIRST_BYTE`,
rather than just a comment stating what we expect it to be.

Author: overlookmotel

crates/oxc_parser/src/lexer/string.rs

@@ -11,14 +11,16 @@ use super::{
 
 /// Convert `char` to UTF-8 bytes array.
 const fn to_bytes<const N: usize>(ch: char) -> [u8; N] {
+    assert!(ch.len_utf8() == N);
     let mut bytes = [0u8; N];
     ch.encode_utf8(&mut bytes);
     bytes
 }
 
 /// Lossy replacement character (U+FFFD) as UTF-8 bytes.
 const LOSSY_REPLACEMENT_CHAR_BYTES: [u8; 3] = to_bytes('\u{FFFD}');
-const LOSSY_REPLACEMENT_CHAR_FIRST_BYTE: u8 = LOSSY_REPLACEMENT_CHAR_BYTES[0]; // 0xEF
+const LOSSY_REPLACEMENT_CHAR_FIRST_BYTE: u8 = LOSSY_REPLACEMENT_CHAR_BYTES[0];
+const _: () = assert!(LOSSY_REPLACEMENT_CHAR_FIRST_BYTE == 0xEF);
 
 const MIN_ESCAPED_STR_LEN: usize = 16;
 

crates/oxc_parser/src/lexer/template.rs

@@ -13,6 +13,7 @@ const MIN_ESCAPED_TEMPLATE_LIT_LEN: usize = 16;
 
 /// Convert `char` to UTF-8 bytes array.
 const fn to_bytes<const N: usize>(ch: char) -> [u8; N] {
+    assert!(ch.len_utf8() == N);
     let mut bytes = [0u8; N];
     ch.encode_utf8(&mut bytes);
     bytes