Allow underscores in numbers.

5 files changed, 198 insertions, 2 deletions

diff --git a/.gitignore b/.gitignore
index 87a3e593..13d03600 100644
--- a/.gitignore
+++ b/.gitignore
@@ -46,3 +46,5 @@ browse.VC.db
 CMakeLists.txt.user
 /CMakeSettings.json
 /.vs
+/.cproject
+/.project
diff --git a/Changelog.md b/Changelog.md
index 0790c0ca..0fd967a8 100644
--- a/Changelog.md
+++ b/Changelog.md
@@ -147,6 +147,7 @@ Features:
  * General: Introduce new constructor syntax using the ``constructor`` keyword as experimental 0.5.0 feature.
  * General: Limit the number of errors output in a single run to 256.
  * General: Support accessing dynamic return data in post-byzantium EVMs.
+ * General: Allow underscores in numeric and hex literals to separate thousands and quads.
  * Inheritance: Error when using empty parentheses for base class constructors that require arguments as experimental 0.5.0 feature.
  * Inheritance: Error when using no parentheses in modifier-style constructor calls as experimental 0.5.0 feature.
  * Interfaces: Allow overriding external functions in interfaces with public in an implementing contract.
diff --git a/docs/types.rst b/docs/types.rst
index fbc83972..84e58fde 100644
--- a/docs/types.rst
+++ b/docs/types.rst
@@ -284,6 +284,9 @@ one side.  Examples include ``1.``, ``.1`` and ``1.3``.
 Scientific notation is also supported, where the base can have fractions, while the exponent cannot.
 Examples include ``2e10``, ``-2e10``, ``2e-10``, ``2.5e1``.
 
+Underscores can be used to separate digits of a numeric literal to aid readability.
+For example, ``123_000``, ``0x2eff_abde``, ``1_233e34_89`` are all valid. Underscores are only allowed between two digits.
+
 Number literal expressions retain arbitrary precision until they are converted to a non-literal type (i.e. by
 using them together with a non-literal expression).
 This means that computations do not overflow and divisions do not truncate
diff --git a/libsolidity/parsing/Scanner.cpp b/libsolidity/parsing/Scanner.cpp
index 801d2cc4..65189b19 100644
--- a/libsolidity/parsing/Scanner.cpp
+++ b/libsolidity/parsing/Scanner.cpp
@@ -726,8 +726,21 @@ Token::Value Scanner::scanHexString()
 
 void Scanner::scanDecimalDigits()
 {
-	while (isDecimalDigit(m_char))
+	if (!isDecimalDigit(m_char)) // avoid underscore at beginning
+		return;
+	while (isDecimalDigit(m_char) || m_char == '_') 
+	{
+		if (m_char == '_') 
+		{
+			advance();
+			if (!isDecimalDigit(m_char)) // avoid trailing underscore
+			{
+				rollback(1);
+				break;
+			}
+		}
 		addLiteralCharAndAdvance();
+	}
 }
 
 Token::Value Scanner::scanNumber(char _charSeen)
@@ -755,8 +768,19 @@ Token::Value Scanner::scanNumber(char _charSeen)
 				addLiteralCharAndAdvance();
 				if (!isHexDigit(m_char))
 					return Token::Illegal; // we must have at least one hex digit after 'x'/'X'
-				while (isHexDigit(m_char))
+				while (isHexDigit(m_char) || m_char == '_') // same logic as scanDecimalDigits
+				{
+					if (m_char == '_') 
+					{
+						advance();
+						if (!isHexDigit(m_char)) // avoid trailing underscore
+						{
+							rollback(1);
+							break;
+						}
+					}
 					addLiteralCharAndAdvance();
+				}
 			}
 			else if (isDecimalDigit(m_char))
 				// We do not allow octal numbers
diff --git a/test/libsolidity/SolidityScanner.cpp b/test/libsolidity/SolidityScanner.cpp
index 42e1b18e..b650d918 100644
--- a/test/libsolidity/SolidityScanner.cpp
+++ b/test/libsolidity/SolidityScanner.cpp
@@ -155,6 +155,172 @@ BOOST_AUTO_TEST_CASE(trailing_dot)
 	BOOST_CHECK_EQUAL(scanner.next(), Token::EOS);
 }
 
+BOOST_AUTO_TEST_CASE(underscores_in_integer)
+{
+	Scanner scanner(CharStream("var x = 1_23_4;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Number);
+	BOOST_CHECK_EQUAL(scanner.currentLiteral(), "1234");
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Semicolon);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::EOS);
+}
+
+BOOST_AUTO_TEST_CASE(underscores_in_scientific_notation)
+{
+	Scanner scanner(CharStream("var x = 1_2e10;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Number);
+	BOOST_CHECK_EQUAL(scanner.currentLiteral(), "12e10");
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Semicolon);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::EOS);
+}
+
+BOOST_AUTO_TEST_CASE(underscores_in_scientific_notation_in_exp_part)
+{
+	Scanner scanner(CharStream("var x = 12e1_0;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Number);
+	BOOST_CHECK_EQUAL(scanner.currentLiteral(), "12e10");
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Semicolon);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::EOS);
+}
+
+
+BOOST_AUTO_TEST_CASE(underscores_in_hex)
+{
+	Scanner scanner(CharStream("var x = 0xab_19cf;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Number);
+	BOOST_CHECK_EQUAL(scanner.currentLiteral(), "0xab19cf");
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Semicolon);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::EOS);
+}
+
+BOOST_AUTO_TEST_CASE(leading_underscore_integer_is_identifier)
+{
+	Scanner scanner(CharStream("var x = _12;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+}
+
+BOOST_AUTO_TEST_CASE(leading_underscore_decimal_is_identifier)
+{
+	Scanner scanner(CharStream("var x = _1.2;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+}
+
+BOOST_AUTO_TEST_CASE(leading_underscore_decimal_after_dot_illegal)
+{
+	Scanner scanner(CharStream("var x = 1._2;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
+BOOST_AUTO_TEST_CASE(leading_underscore_exp_are_identifier)
+{
+	Scanner scanner(CharStream("var x = _1e2;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+}
+
+BOOST_AUTO_TEST_CASE(leading_underscore_exp_after_e_illegal)
+{
+	Scanner scanner(CharStream("var x = 1e_2;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
+BOOST_AUTO_TEST_CASE(leading_underscore_hex_illegal)
+{
+	Scanner scanner(CharStream("var x = 0x_abc;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
+BOOST_AUTO_TEST_CASE(trailing_underscore_integer_illegal)
+{
+	Scanner scanner(CharStream("var x = 12_;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
+BOOST_AUTO_TEST_CASE(leading_underscore_after_decimal_illegal)
+{
+	Scanner scanner(CharStream("var x = 1.2_;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
+BOOST_AUTO_TEST_CASE(leading_underscore_before_decimal_illegal)
+{
+	Scanner scanner(CharStream("var x = 1_.2;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
+BOOST_AUTO_TEST_CASE(trailing_underscore_exp_illegal)
+{
+	Scanner scanner(CharStream("var x = 1e2_;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
+BOOST_AUTO_TEST_CASE(trailing_underscore_exp_before_e_illegal)
+{
+	Scanner scanner(CharStream("var x = 1_e2;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
+BOOST_AUTO_TEST_CASE(trailing_underscore_hex_illegal)
+{
+	Scanner scanner(CharStream("var x = 0xabc_;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
+BOOST_AUTO_TEST_CASE(double_underscore_illegal)
+{
+	Scanner scanner(CharStream("var x = 1__2;"));
+	BOOST_CHECK_EQUAL(scanner.currentToken(), Token::Var);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Identifier);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Assign);
+	BOOST_CHECK_EQUAL(scanner.next(), Token::Illegal);
+}
+
 BOOST_AUTO_TEST_CASE(negative_numbers)
 {
 	Scanner scanner(CharStream("var x = -.2 + -0x78 + -7.3 + 8.9 + 2e-2;"));