From 9104b5e4086d02cb552c6edbffb4f816268a75dd Mon Sep 17 00:00:00 2001
From: yor1xd <gabrielhenriquenf1@gmail.com>
Date: Thu, 21 May 2026 21:43:32 -0300
Subject: [PATCH] lexer: implement \x and \u numeric escape sequences

---
 lexer/src/lib.rs   |  51 +++++++++++++++++++-
 lexer/src/tests.rs | 117 +++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 166 insertions(+), 2 deletions(-)

diff --git a/lexer/src/lib.rs b/lexer/src/lib.rs
index abe7084..f982b0f 100644
--- a/lexer/src/lib.rs
+++ b/lexer/src/lib.rs
@@ -387,6 +387,11 @@ fn parse_escape<const REGEX: bool>(
 ) -> Result<usize> {
     let mut count = 2;
     let is_oct = |x: char| ('0'..'8').contains(&x);
+    let is_hex = |i: usize| {
+        slice
+            .get(i)
+            .is_some_and(|&x| (x as char).is_ascii_hexdigit())
+    };
     let is_slice_oct = |i| slice.get(i).map(|&x| x as char).is_some_and(is_oct);
     let Some(to_escape) = slice.get(1).map(|x| *x as char) else {
         return Err(LexingError::UnexpectedEof);
@@ -417,8 +422,38 @@ fn parse_escape<const REGEX: bool>(
                 .take(count - 1)
                 .fold(0, |acc, digit| acc * 8 + digit - b'0') as char
         }
-        'x' if !posix_strict => todo!(),
-        'u' => todo!(),
+        'x' if !posix_strict => {
+            let num_digits = is_hex(2) as usize + (is_hex(2) && is_hex(3)) as usize;
+
+            if num_digits == 0 {
+                'x'
+            } else {
+                count += num_digits;
+
+                let value = parse_hex_digits(&slice[2..2 + num_digits]) as u8;
+                value as char
+            }
+        }
+        'u' if !posix_strict => {
+            let num_digits = (2..=9).take_while(|&i| is_hex(i)).count();
+
+            if num_digits == 0 {
+                'u'
+            } else {
+                count += num_digits;
+
+                let codepoint = parse_hex_digits(&slice[2..2 + num_digits]);
+
+                // FIXME: assumes UTF-8 locale; replacement character and encoding may differ
+                // for non-UTF-8 locales.
+                let c = char::from_u32(codepoint).unwrap_or('\u{FFFD}');
+                let mut buf = [0u8; 4];
+                let encoded = c.encode_utf8(&mut buf);
+                out.extend_from_slice_copy(encoded.as_bytes());
+
+                return Ok(count);
+            }
+        }
         // Unspecified by POSIX; we ditto GNU.
         c => c, // TODO: Output warning
     };
@@ -426,6 +461,18 @@ fn parse_escape<const REGEX: bool>(
     Ok(count)
 }
 
+fn parse_hex_digits(slice: &[u8]) -> u32 {
+    slice.iter().fold(0u32, |acc, &digit| {
+        acc * 16
+            + match digit {
+                b'0'..=b'9' => (digit - b'0') as u32,
+                b'a'..=b'f' => (digit - b'a' + 10) as u32,
+                b'A'..=b'F' => (digit - b'A' + 10) as u32,
+                _ => unreachable!(),
+            }
+    })
+}
+
 fn parse_ident<'a>(lex: &mut Lexer<'a>, index: impl SliceIndex<[u8], Output = [u8]>) -> &'a str {
     accept_operator(lex);
     // SAFETY: The regex matching ensures it is ASCII.
diff --git a/lexer/src/tests.rs b/lexer/src/tests.rs
index ebeb435..d56e8c1 100644
--- a/lexer/src/tests.rs
+++ b/lexer/src/tests.rs
@@ -259,3 +259,120 @@ fn lexer_test_regex_ambiguity() {
         ]
     );
 }
+
+#[test]
+fn lexer_test_hex_escape() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\x41\"", &arena, false, false),
+        &[Token::String(b"A".into())]
+    );
+}
+
+#[test]
+fn lexer_test_hex_escape_uppercase() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\x4F\"", &arena, false, false),
+        &[Token::String(b"O".into())]
+    );
+}
+
+#[test]
+fn lexer_test_hex_escape_single_digit() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\x9\"", &arena, false, false),
+        &[Token::String(b"\x09".into())]
+    );
+}
+
+#[test]
+fn lexer_test_hex_escape_posix_strict() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\x41\"", &arena, true, false),
+        &[Token::String(b"x41".into())]
+    );
+}
+
+#[test]
+fn lexer_test_unicode_escape_posix_strict() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\u0041\"", &arena, true, false),
+        &[Token::String(b"u0041".into())]
+    );
+}
+
+#[test]
+fn lexer_test_unicode_escape_ascii() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\u0041\"", &arena, false, false),
+        &[Token::String(b"A".into())]
+    );
+}
+
+#[test]
+fn lexer_test_unicode_escape_two_byte() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\u00e9\"", &arena, false, false),
+        &[Token::String("\u{00e9}".as_bytes().into())]
+    );
+}
+
+#[test]
+fn lexer_test_unicode_escape_three_byte() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\u4e2d\"", &arena, false, false),
+        &[Token::String("\u{4e2d}".as_bytes().into())]
+    );
+}
+
+#[test]
+fn lexer_test_unicode_escape_uppercase() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\u004F\"", &arena, false, false),
+        &[Token::String(b"O".into())]
+    );
+}
+
+#[test]
+fn lexer_test_unicode_escape_single_digit() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\u9\"", &arena, false, false),
+        &[Token::String("\u{9}".as_bytes().into())]
+    );
+}
+
+#[test]
+fn lexer_test_hex_escape_no_digits() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\x\"", &arena, false, false),
+        &[Token::String(b"x".into())]
+    );
+}
+
+#[test]
+fn lexer_test_unicode_escape_no_digits() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\u\"", &arena, false, false),
+        &[Token::String(b"u".into())]
+    );
+}
+
+#[test]
+fn lexer_test_unicode_escape_eight_digits() {
+    let arena = Bump::new();
+    assert_eq!(
+        &lex(b"\"\\u00000032\"", &arena, false, false),
+        &[Token::String(b"2".into())]
+    );
+}