feat(parser): detect binary files with TS1490 error

crates/oxc_parser/src/diagnostics.rs

@@ -48,6 +48,11 @@ pub fn overlong_source() -> OxcDiagnostic {
     OxcDiagnostic::error("Source length exceeds 4 GiB limit")
 }
 
+#[cold]
+pub fn file_appears_to_be_binary() -> OxcDiagnostic {
+    ts_error("1490", "File appears to be binary.")
+}
+
 #[cold]
 pub fn flow(span: Span) -> OxcDiagnostic {
     OxcDiagnostic::error("Flow is not supported").with_label(span)

crates/oxc_parser/src/lexer/unicode.rs

@@ -33,6 +33,10 @@ impl<'a> Lexer<'a> {
     pub(super) fn unicode_char_handler(&mut self) -> Kind {
         let c = self.peek_char().unwrap();
         match c {
+            // U+FFFD (replacement character) appears when a binary file is decoded as UTF-8.
+            // This is likely a binary file that cannot be parsed.
+            // <https://github.com/microsoft/TypeScript/blob/main/src/compiler/scanner.ts>
+            '\u{FFFD}' => self.handle_binary_file(),
             c if is_identifier_start_unicode(c) => {
                 let start_pos = self.source.position();
                 self.consume_char();
@@ -45,6 +49,13 @@ impl<'a> Lexer<'a> {
         }
     }
 
+    #[cold]
+    fn handle_binary_file(&mut self) -> Kind {
+        self.error(diagnostics::file_appears_to_be_binary());
+        self.source.advance_to_end();
+        Kind::Eof
+    }
+
     #[cold]
     fn handle_irregular_whitespace(&mut self, _c: char) -> Kind {
         self.consume_char();

crates/oxc_parser/src/lib.rs

@@ -808,6 +808,23 @@ mod test {
         }
     }
 
+    #[test]
+    fn binary_file() {
+        let allocator = Allocator::default();
+        let source_type = SourceType::default();
+
+        // U+FFFD as a standalone token — file appears to be binary
+        let ret = Parser::new(&allocator, "\u{FFFD}", source_type).parse();
+        assert!(ret.program.is_empty());
+        assert_eq!(ret.errors.len(), 1);
+        assert_eq!(ret.errors[0].to_string(), "File appears to be binary.");
+
+        // U+FFFD inside string literals — should parse fine
+        let ret = Parser::new(&allocator, "\"oops \u{FFFD} oops\";", source_type).parse();
+        assert!(!ret.program.is_empty());
+        assert!(ret.errors.is_empty());
+    }
+
     #[test]
     fn memory_leak() {
         let allocator = Allocator::default();

tasks/coverage/snapshots/parser_typescript.snap

@@ -3,7 +3,7 @@ commit: 95e3aaa9
 parser_typescript Summary:
 AST Parsed     : 9841/9841 (100.00%)
 Positive Passed: 9841/9841 (100.00%)
-Negative Passed: 1529/2557 (59.80%)
+Negative Passed: 1530/2558 (59.81%)
 Expect Syntax Error: tasks/coverage/typescript/tests/cases/compiler/FunctionDeclaration3.ts
 
 Expect Syntax Error: tasks/coverage/typescript/tests/cases/compiler/FunctionDeclaration4.ts
@@ -2312,6 +2312,13 @@ Expect Syntax Error: tasks/coverage/typescript/tests/cases/conformance/types/wit
  5 │ }
    ╰────
 
+  × Expected a semicolon or an implicit semicolon after a statement, but found none
+   ╭─[typescript/tests/cases/compiler/TransportStream.ts:1:2]
+ 1 │ G@�G@�G@�
+   ·  ▲
+   ╰────
+  help: Try inserting a semicolon here
+
   × Getters and setters must have an implementation.
     ╭─[typescript/tests/cases/compiler/abstractPropertyNegative.ts:16:9]
  15 │     abstract notAllowed: string;
@@ -4586,11 +4593,7 @@ Expect Syntax Error: tasks/coverage/typescript/tests/cases/conformance/types/wit
  293 │     class interface { }
      ╰────
 
-  × Invalid Character `�`
-   ╭─[typescript/tests/cases/compiler/corrupted.ts:1:1]
- 1 │ ���
-   · ─
-   ╰────
+  × TS(1490): File appears to be binary.
 
   × TS(2391): Function implementation is missing or not immediately following the declaration.
    ╭─[typescript/tests/cases/compiler/crashOnMethodSignatures.ts:2:5]

tasks/coverage/src/typescript/constants.rs

@@ -92,9 +92,6 @@ pub static NOT_SUPPORTED_TEST_PATHS: phf::Set<&'static str> = phf::phf_set![
     // TSC: Parse without error, they support BOM
     // OXC: We do not ignore or exclude BOM, will be invalid character error
     "bom-utf16be.ts",
-    // TSC: This is just a binary file, but their test project skips reading
-    // OXC: Try to parse, and fail
-    "TransportStream.ts",
     // TSC: Allows `catch({x}) { var x; }` (destructured catch param + var redeclaration)
     // OXC: Reports redeclaration error per Annex B §B.3.4 (only simple identifiers are exempt)
     "asyncWithVarShadowing_es6.ts",