From aa4593cab3d60468e5ea4318012c5252ebbc7d13 Mon Sep 17 00:00:00 2001
From: Alex Beregszaszi <alex@rtfs.hu>
Date: Tue, 7 Jun 2016 19:23:19 +0100
Subject: Support Unicode escape characters in string literals ('\uUUUU')

Fixes #638
---
 libsolidity/parsing/Scanner.cpp | 43 +++++++++++++++++++++++++++++++++++++++++
 libsolidity/parsing/Scanner.h   |  2 ++
 2 files changed, 45 insertions(+)

(limited to 'libsolidity')

diff --git a/libsolidity/parsing/Scanner.cpp b/libsolidity/parsing/Scanner.cpp
index d630d0ab..616e6a96 100644
--- a/libsolidity/parsing/Scanner.cpp
+++ b/libsolidity/parsing/Scanner.cpp
@@ -177,6 +177,41 @@ bool Scanner::scanHexByte(char& o_scannedByte)
 	return true;
 }
 
+bool Scanner::scanUnicode(unsigned & o_codepoint)
+{
+	unsigned x = 0;
+	for (int i = 0; i < 4; i++)
+	{
+		int d = hexValue(m_char);
+		if (d < 0)
+		{
+			rollback(i);
+			return false;
+		}
+		x = x * 16 + d;
+		advance();
+	}
+	o_codepoint = x;
+	return true;
+}
+
+// This supports codepoints between 0000 and FFFF.
+void Scanner::addUnicodeChar(unsigned codepoint)
+{
+	if (codepoint <= 0x7f)
+		addLiteralChar(codepoint);
+	else if (codepoint <= 0x7ff)
+	{
+		addLiteralChar(0xc0 | (codepoint >> 6));
+		addLiteralChar(0x80 | (codepoint & 0x3f));
+	}
+	else
+	{
+		addLiteralChar(0xe0 | (codepoint >> 12));
+		addLiteralChar(0x80 | ((codepoint >> 6) & 0x3f));
+		addLiteralChar(0x80 | (codepoint & 0x3f));
+	}
+}
 
 // Ensure that tokens can be stored in a byte.
 BOOST_STATIC_ASSERT(Token::NUM_TOKENS <= 0x100);
@@ -607,6 +642,14 @@ bool Scanner::scanEscape()
 	case 'v':
 		c = '\v';
 		break;
+	case 'u':
+	{
+		unsigned codepoint;
+		if (!scanUnicode(codepoint))
+			return false;
+		addUnicodeChar(codepoint);
+		return true;
+	}
 	case 'x':
 		if (!scanHexByte(c))
 			return false;
diff --git a/libsolidity/parsing/Scanner.h b/libsolidity/parsing/Scanner.h
index cd60aff8..3dde42b3 100644
--- a/libsolidity/parsing/Scanner.h
+++ b/libsolidity/parsing/Scanner.h
@@ -175,6 +175,7 @@ private:
 	inline void addLiteralChar(char c) { m_nextToken.literal.push_back(c); }
 	inline void addCommentLiteralChar(char c) { m_nextSkippedComment.literal.push_back(c); }
 	inline void addLiteralCharAndAdvance() { addLiteralChar(m_char); advance(); }
+	void addUnicodeChar(unsigned codepoint);
 	///@}
 
 	bool advance() { m_char = m_source.advanceAndGet(); return !m_source.isPastEndOfInput(); }
@@ -185,6 +186,7 @@ private:
 	inline Token::Value selectToken(char _next, Token::Value _then, Token::Value _else);
 
 	bool scanHexByte(char& o_scannedByte);
+	bool scanUnicode(unsigned& o_codepoint);
 
 	/// Scans a single Solidity token.
 	void scanToken();
-- 
cgit 


From 6db12c4f882411293da71c5951ba7088a9712c13 Mon Sep 17 00:00:00 2001
From: Alex Beregszaszi <alex@rtfs.hu>
Date: Mon, 1 Aug 2016 14:10:46 +0100
Subject: Rename addUnicodeChar to addUnicodeAsUTF8

---
 libsolidity/parsing/Scanner.cpp | 4 ++--
 libsolidity/parsing/Scanner.h   | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)

(limited to 'libsolidity')

diff --git a/libsolidity/parsing/Scanner.cpp b/libsolidity/parsing/Scanner.cpp
index 616e6a96..d730210a 100644
--- a/libsolidity/parsing/Scanner.cpp
+++ b/libsolidity/parsing/Scanner.cpp
@@ -196,7 +196,7 @@ bool Scanner::scanUnicode(unsigned & o_codepoint)
 }
 
 // This supports codepoints between 0000 and FFFF.
-void Scanner::addUnicodeChar(unsigned codepoint)
+void Scanner::addUnicodeAsUTF8(unsigned codepoint)
 {
 	if (codepoint <= 0x7f)
 		addLiteralChar(codepoint);
@@ -647,7 +647,7 @@ bool Scanner::scanEscape()
 		unsigned codepoint;
 		if (!scanUnicode(codepoint))
 			return false;
-		addUnicodeChar(codepoint);
+		addUnicodeAsUTF8(codepoint);
 		return true;
 	}
 	case 'x':
diff --git a/libsolidity/parsing/Scanner.h b/libsolidity/parsing/Scanner.h
index 3dde42b3..708adf8f 100644
--- a/libsolidity/parsing/Scanner.h
+++ b/libsolidity/parsing/Scanner.h
@@ -175,7 +175,7 @@ private:
 	inline void addLiteralChar(char c) { m_nextToken.literal.push_back(c); }
 	inline void addCommentLiteralChar(char c) { m_nextSkippedComment.literal.push_back(c); }
 	inline void addLiteralCharAndAdvance() { addLiteralChar(m_char); advance(); }
-	void addUnicodeChar(unsigned codepoint);
+	void addUnicodeAsUTF8(unsigned codepoint);
 	///@}
 
 	bool advance() { m_char = m_source.advanceAndGet(); return !m_source.isPastEndOfInput(); }
-- 
cgit