Changed error message for for Unicode character in non-unicode string literal

Co-authored-by: Kamil Śliwak <[email protected]>

Author: 2 people

liblangutil/Scanner.cpp

@@ -76,6 +76,7 @@ string to_string(ScannerError _errorCode)
 		case ScannerError::IllegalHexDigit: return "Hexadecimal digit missing or invalid.";
 		case ScannerError::IllegalCommentTerminator: return "Expected multi-line comment-terminator.";
 		case ScannerError::IllegalEscapeSequence: return "Invalid escape sequence.";
+		case ScannerError::UnicodeCharacterInNonUnicodeString: return "Invalid character in string. If you are trying to use Unicode characters, use a unicode\"...\" string literal.";
 		case ScannerError::IllegalCharacterInString: return "Invalid character in string.";
 		case ScannerError::IllegalStringEndQuote: return "Expected string end-quote.";
 		case ScannerError::IllegalNumberSeparator: return "Invalid use of number separator '_'.";
@@ -844,7 +845,11 @@ Token Scanner::scanString(bool const _isUnicode)
 			// We are using a manual range and not isprint() to avoid
 			// any potential complications with locale.
 			if (!_isUnicode && (static_cast<unsigned>(c) <= 0x1f || static_cast<unsigned>(c) >= 0x7f))
-				return setError(ScannerError::IllegalCharacterInString);
+			{
+				if (m_kind == ScannerKind::Yul)
+					return setError(ScannerError::IllegalCharacterInString);
+				return setError(ScannerError::UnicodeCharacterInNonUnicodeString);
+			}
 			addLiteralChar(c);
 		}
 	}

liblangutil/Scanner.h

@@ -81,6 +81,7 @@ enum class ScannerError
 	IllegalHexDigit,
 	IllegalCommentTerminator,
 	IllegalEscapeSequence,
+	UnicodeCharacterInNonUnicodeString,
 	IllegalCharacterInString,
 	IllegalStringEndQuote,
 	IllegalNumberSeparator,
@@ -160,6 +161,8 @@ class Scanner
 	/// Called by the parser during FunctionDefinition parsing to clear the current comment
 	void clearCurrentCommentLiteral() { m_skippedComments[Current].literal.clear(); }
 
+	ScannerKind scannerKind() const { return m_kind; }
+
 	///@}
 
 	///@{

test/liblangutil/Scanner.cpp

@@ -123,7 +123,7 @@ BOOST_AUTO_TEST_CASE(string_nonprintable)
 		if (v == '\n' || v == '\v' || v == '\f' || v == '\r')
 			BOOST_CHECK_EQUAL(scanner.currentError(), ScannerError::IllegalStringEndQuote);
 		else
-			BOOST_CHECK_EQUAL(scanner.currentError(), ScannerError::IllegalCharacterInString);
+			BOOST_CHECK_EQUAL(scanner.currentError(),ScannerError::UnicodeCharacterInNonUnicodeString);
 		BOOST_CHECK_EQUAL(scanner.currentLiteral(), "");
 	}
 }

test/libsolidity/syntaxTests/string/control_character_returned_in_string.sol

@@ -0,0 +1,7 @@
+contract test {
+    function f() public pure returns (string memory) {
+        return "
;
+    }
+}
+// ----
+// ParserError 8936: (86-88): Invalid character in string. If you are trying to use Unicode characters, use a unicode"..." string literal.

test/libsolidity/syntaxTests/string/string_unicode_without_prefix.sol

@@ -4,4 +4,4 @@ contract test {
     }
 }
 // ----
-// ParserError 8936: (86-88): Invalid character in string.
+// ParserError 8936: (86-88): Invalid character in string. If you are trying to use Unicode characters, use a unicode"..." string literal.