From 78489f429355e2f852ab35b65099f7f53eea6a5c Mon Sep 17 00:00:00 2001
From: Alexander Mai <amai@users.sf.net>
Date: Sat, 28 Nov 2015 16:37:26 +0100
Subject: [PATCH] MathLib::characterLiteralToLongNumber can encode unicode char
 literals, though it's incompatible to gcc/clang

---
 lib/mathlib.cpp      | 25 +++++++++++++++----------
 test/testmathlib.cpp | 12 +++++++++---
 2 files changed, 24 insertions(+), 13 deletions(-)

diff --git a/lib/mathlib.cpp b/lib/mathlib.cpp
index 6fa4f9656..c75c14786 100644
--- a/lib/mathlib.cpp
+++ b/lib/mathlib.cpp
@@ -322,33 +322,38 @@ static bool isOctalDigitString(const std::string& str)
     return true;
 }
 
+static unsigned int encodeMultiChar(const std::string& str)
+{
+    unsigned int retval(str.front());
+    for (std::string::const_iterator it=str.begin()+1; it!=str.end(); ++it) {
+        retval = retval<<8 | *it;
+    }
+    return retval;
+}
+
 MathLib::bigint MathLib::characterLiteralToLongNumber(const std::string& str)
 {
     if (str.empty())
         return 0; // for unit-testing...
     if (str.size()==1)
         return str[0] & 0xff;
-    const std::string& str1 = str.substr(1);
     if (str[0] != '\\') {
         // C99 6.4.4.4
         // The value of an integer character constant containing more than one character (e.g., 'ab'),
         // or containing a character or escape sequence that does not map to a single-byte execution character,
         // is implementation-defined.
         // clang and gcc seem to use the following encoding: 'AB' as (('A' << 8) | 'B')
-        unsigned int retval(str.front());
-        for (std::string::const_iterator it=str1.begin(); it!=str1.end(); ++it) {
-            retval = retval<<8 | *it;
-        }
-        return (MathLib::bigint)retval;
+        return encodeMultiChar(str);
     }
+    const std::string& str1 = str.substr(1);
 
     switch (str1[0]) {
     case 'x':
         return toLongNumber("0x" + str.substr(2));
-    case 'u': // 16bit unicode character
-        throw InternalError(0, "Internal Error. MathLib::toLongNumber: Unhandled 16-bit unicode char constant \\" + str);
-    case 'U': // 16bit unicode character
-        throw InternalError(0, "Internal Error. MathLib::toLongNumber: Unhandled 32-bit unicode char constant \\" + str);
+    case 'u': // 16-bit unicode character
+        return encodeMultiChar(str1);
+    case 'U': // 32-bit unicode character
+        return encodeMultiChar(str1);
     default: {
         char c;
         switch (str.size()-1) {
diff --git a/test/testmathlib.cpp b/test/testmathlib.cpp
index aaa8c99b8..3dde41fd2 100644
--- a/test/testmathlib.cpp
+++ b/test/testmathlib.cpp
@@ -290,9 +290,15 @@ private:
         ASSERT_EQUALS((int)('\134'), MathLib::toLongNumber("'\\134'"));
         ASSERT_THROW(MathLib::toLongNumber("'\\9'"), InternalError);
         ASSERT_THROW(MathLib::toLongNumber("'\\934'"), InternalError);
-        ASSERT_THROW(MathLib::toLongNumber("'\\u9343'"), InternalError);
-        ASSERT_THROW(MathLib::toLongNumber("'\\U0001f34c'"), InternalError);
-
+        // that is not gcc/clang encoding
+        ASSERT_EQUALS(959657011, MathLib::toLongNumber("'\\u9343'"));
+        ASSERT_EQUALS(1714631779, MathLib::toLongNumber("'\\U0001f34c'"));
+#ifdef __GNUC__
+        // BEGIN Implementation-specific results
+        TODO_ASSERT_EQUALS((int)'\u9343', 959657011, MathLib::toLongNumber("'\\u9343'"));
+        TODO_ASSERT_EQUALS((int)'\U0001f34c', 1714631779, MathLib::toLongNumber("'\\U0001f34c'"));
+        // END Implementation-specific results
+#endif
         {
             // some unit-testing for a utility function
             ASSERT_EQUALS(0, MathLib::characterLiteralToLongNumber(std::string("")));