fix(parser): fix panic on multi-byte characaters (#233)

* fix(oxc_parser): fix panic when EOF on a multi-byte character

relates #232

* fix(parser): fix panic on multi-byte char in private identifer

relates #232

crates/oxc_parser/src/cursor.rs

@@ -174,7 +174,7 @@ impl<'a> Parser<'a> {
     /// # Errors
     pub fn expect_without_advance(&mut self, kind: Kind) -> Result<()> {
         if !self.at(kind) {
-            let range = self.current_range();
+            let range = self.cur_token().span();
             return Err(
                 diagnostics::ExpectToken(kind.to_str(), self.cur_kind().to_str(), range).into()
             );
@@ -190,21 +190,6 @@ impl<'a> Parser<'a> {
         Ok(())
     }
 
-    #[must_use]
-    pub fn current_range(&self) -> Span {
-        let cur_token = self.cur_token();
-        match self.cur_kind() {
-            Kind::Eof => {
-                if self.prev_token_end < cur_token.end {
-                    Span::new(self.prev_token_end, self.prev_token_end)
-                } else {
-                    Span::new(self.prev_token_end - 1, self.prev_token_end)
-                }
-            }
-            _ => cur_token.span(),
-        }
-    }
-
     /// Expect the next next token to be a `JsxChild`, i.e. `<` or `{` or `JSXText`
     /// # Errors
     pub fn expect_jsx_child(&mut self, kind: Kind) -> Result<()> {

crates/oxc_parser/src/diagnostics.rs

@@ -15,13 +15,9 @@ pub struct Flow(#[label] pub Span);
 pub struct UnexpectedToken(#[label] pub Span);
 
 #[derive(Debug, Error, Diagnostic)]
-#[error("Expect token")]
+#[error("Expected `{0}` but found `{1}`")]
 #[diagnostic()]
-pub struct ExpectToken(
-    pub &'static str,
-    pub &'static str,
-    #[label("Expect `{0}` here, but found `{1}`")] pub Span,
-);
+pub struct ExpectToken(pub &'static str, pub &'static str, #[label("`{0}` expected")] pub Span);
 
 #[derive(Debug, Error, Diagnostic)]
 #[error("Invalid escape sequence")]

crates/oxc_parser/src/js/statement.rs

@@ -297,7 +297,7 @@ impl<'a> Parser<'a> {
             .then(|| self.parse_expression())
             .transpose()
             .map_err(|_| {
-                let range = self.current_range();
+                let range = self.cur_token().span();
                 diagnostics::ExpectToken(Kind::Semicolon.to_str(), self.cur_kind().to_str(), range)
             })?;
         self.expect(Kind::Semicolon)?;

crates/oxc_parser/src/lexer/mod.rs

@@ -726,11 +726,15 @@ impl<'a> Lexer<'a> {
                 self.identifier_unicode_escape_sequence(&mut builder, true);
             }
             Some(c) => {
-                self.error(diagnostics::InvalidCharacter(c, Span::new(start, self.offset() - 1)));
+                #[allow(clippy::cast_possible_truncation)]
+                self.error(diagnostics::InvalidCharacter(
+                    c,
+                    Span::new(start, start + c.len_utf8() as u32),
+                ));
                 return Kind::Undetermined;
             }
             None => {
-                self.error(diagnostics::UnexpectedEnd(Span::new(start, self.offset() - 1)));
+                self.error(diagnostics::UnexpectedEnd(Span::new(start, start)));
                 return Kind::Undetermined;
             }
         }

crates/oxc_parser/src/lib.rs

@@ -144,7 +144,7 @@ impl<'a> Parser<'a> {
         if self.cur_kind() == Kind::Undetermined {
             return Err(self.errors.borrow_mut().pop().unwrap());
         }
-        Err(diagnostics::UnexpectedToken(self.current_range()).into())
+        Err(diagnostics::UnexpectedToken(self.cur_token().span()).into())
     }
 
     /// Push a Syntax Error