Include radix base prefix in large number representation

astral-sh · Sep 28, 2023 · fc3f47b · fc3f47b
1 parent 316f759
commit fc3f47b
Show file tree

Hide file tree

Showing 3 changed files with 21 additions and 6 deletions.
diff --git a/crates/ruff_python_ast/src/int.rs b/crates/ruff_python_ast/src/int.rs
@@ -41,15 +41,19 @@ impl Int {
     }
 
     /// Parse an [`Int`] from a string with a given radix.
-    pub fn from_str_radix(s: &str, radix: u32) -> Result<Self, std::num::ParseIntError> {
-        match i64::from_str_radix(s, radix) {
+    pub fn from_str_radix(
+        src: &str,
+        radix: u32,
+        token: &str,
+    ) -> Result<Self, std::num::ParseIntError> {
+        match i64::from_str_radix(src, radix) {
             Ok(value) => Ok(Int::small(value)),
             Err(err) => {
                 if matches!(
                     err.kind(),
                     std::num::IntErrorKind::PosOverflow | std::num::IntErrorKind::NegOverflow
                 ) {
-                    Ok(Int::big(s))
+                    Ok(Int::big(token))
                 } else {
                     Err(err)
                 }

diff --git a/crates/ruff_python_parser/src/lexer.rs b/crates/ruff_python_parser/src/lexer.rs
@@ -261,10 +261,14 @@ impl<'source> Lexer<'source> {
             'x' | 'o' | 'b'
         ));
 
+        // Lex the portion of the token after the base prefix (e.g., `9D5` in `0x9D5`).
         let mut number = LexedText::new(self.offset(), self.source);
         self.radix_run(&mut number, radix);
 
-        let value = match Int::from_str_radix(number.as_str(), radix.as_u32()) {
+        // Extract the entire number, including the base prefix (e.g., `0x9D5`).
+        let token = &self.source[self.token_range()];
+
+        let value = match Int::from_str_radix(number.as_str(), radix.as_u32(), token) {
             Ok(int) => int,
             Err(err) => {
                 return Err(LexicalError {
@@ -1462,7 +1466,8 @@ def f(arg=%timeit a = b):
 
     #[test]
     fn test_numbers() {
-        let source = "0x2f 0o12 0b1101 0 123 123_45_67_890 0.2 1e+2 2.1e3 2j 2.2j 000";
+        let source =
+            "0x2f 0o12 0b1101 0 123 123_45_67_890 0.2 1e+2 2.1e3 2j 2.2j 000 0x995DC9BBDF1939FA";
         assert_debug_snapshot!(lex_source(source));
     }
 

diff --git a/crates/ruff_python_parser/src/snapshots/ruff_python_parser__lexer__tests__numbers.snap b/crates/ruff_python_parser/src/snapshots/ruff_python_parser__lexer__tests__numbers.snap
@@ -77,8 +77,14 @@ expression: lex_source(source)
         },
         60..63,
     ),
+    (
+        Int {
+            value: 0x995DC9BBDF1939FA,
+        },
+        64..82,
+    ),
     (
         Newline,
-        63..63,
+        82..82,
     ),
 ]