Revert "Refactoring; Reuse function in simplecpp"

This reverts commit dbe9eb2a27.

lib/mathlib.cpp

@@ -21,8 +21,6 @@
 #include "errortypes.h"
 #include "utils.h"
 
-#include <simplecpp.h>
-
 #include <cctype>
 #include <cmath>
 #include <cstdlib>
@@ -357,6 +355,35 @@ unsigned int MathLib::encodeMultiChar(const std::string& str)
     return retval;
 }
 
+static bool isoctal(int c)
+{
+    return c>='0' && c<='7';
+}
+
+MathLib::bigint MathLib::characterLiteralToLongNumber(const std::string& str)
+{
+    if (str.empty())
+        return 0; // <- only possible in unit testing
+
+    // '\xF6'
+    if (str.size() == 4 && str.compare(0,2,"\\x")==0 && std::isxdigit(str[2]) && std::isxdigit(str[3])) {
+        return std::strtoul(str.substr(2).c_str(), nullptr, 16);
+    }
+
+    // '\123'
+    if (str.size() == 4 && str[0] == '\\' && isoctal(str[1]) && isoctal(str[2]) && isoctal(str[3])) {
+        return (char)std::strtoul(str.substr(1).c_str(), nullptr, 8);
+    }
+
+    // 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')
+    const std::string& normStr = normalizeCharacterLiteral(str);
+    return encodeMultiChar(normStr);
+}
+
 std::string MathLib::normalizeCharacterLiteral(const std::string& iLiteral)
 {
     std::string normalizedLiteral;
@@ -506,7 +533,7 @@ MathLib::bigint MathLib::toLongNumber(const std::string & str)
     }
 
     if (isCharLiteral(str)) {
-        return simplecpp::characterLiteralToLL(str);
+        return characterLiteralToLongNumber(getCharLiteral(str));
     }
 
     try {
@@ -570,7 +597,7 @@ static double floatHexToDoubleNumber(const std::string& str)
 double MathLib::toDoubleNumber(const std::string &str)
 {
     if (isCharLiteral(str))
-        return simplecpp::characterLiteralToLL(str);
+        return characterLiteralToLongNumber(getCharLiteral(str));
     if (isIntHex(str))
         return static_cast<double>(toLongNumber(str));
     // nullcheck

lib/mathlib.h

@@ -127,6 +127,12 @@ public:
 
     static unsigned int encodeMultiChar(const std::string& str);
 
+    /**
+     * \param[in] str character literal
+     * @return Number of internal representation of the character literal
+     * */
+    static MathLib::bigint characterLiteralToLongNumber(const std::string& str);
+
     /**
      * \param[in] iCode Code being considered
      * \param[in] iPos A posision within iCode

test/testmathlib.cpp

@@ -304,6 +304,38 @@ private:
         ASSERT_EQUALS((int)('\100'), MathLib::toLongNumber("'\\100'"));
         ASSERT_EQUALS((int)('\200'), MathLib::toLongNumber("'\\200'"));
         ASSERT_EQUALS((int)(L'A'),   MathLib::toLongNumber("L'A'"));
+#ifdef __GNUC__
+        // BEGIN Implementation-specific results
+        ASSERT_EQUALS((int)('AB'),    MathLib::toLongNumber("'AB'"));
+        ASSERT_EQUALS((int)('ABC'),    MathLib::toLongNumber("'ABC'"));
+        ASSERT_EQUALS((int)('ABCD'),    MathLib::toLongNumber("'ABCD'"));
+        ASSERT_EQUALS((int)('ABCDE'),    MathLib::toLongNumber("'ABCDE'"));
+        // END Implementation-specific results
+#endif
+        ASSERT_EQUALS((int)('\0'),   MathLib::toLongNumber("'\\0'"));
+        ASSERT_EQUALS(0x1B,   MathLib::toLongNumber("'\\e'"));
+        ASSERT_EQUALS((int)('\r'),   MathLib::toLongNumber("'\\r'"));
+        ASSERT_EQUALS((int)('\x12'), MathLib::toLongNumber("'\\x12'"));
+        // may cause some compile problems: ASSERT_EQUALS((int)('\x123'), MathLib::toLongNumber("'\\x123'"));
+        // may cause some compile problems: ASSERT_EQUALS((int)('\x1234'), MathLib::toLongNumber("'\\x1234'"));
+        ASSERT_EQUALS((int)('\3'),  MathLib::toLongNumber("'\\3'"));
+        ASSERT_EQUALS((int)('\34'),  MathLib::toLongNumber("'\\34'"));
+        ASSERT_EQUALS((int)('\034'), MathLib::toLongNumber("'\\034'"));
+        ASSERT_EQUALS((int)('\x34'), MathLib::toLongNumber("'\\x34'"));
+        ASSERT_EQUALS((int)('\134'), MathLib::toLongNumber("'\\134'"));
+        ASSERT_EQUALS((int)('\134t'), MathLib::toLongNumber("'\\134t'")); // Ticket #7452
+        ASSERT_THROW(MathLib::toLongNumber("'\\9'"), InternalError);
+        ASSERT_THROW(MathLib::toLongNumber("'\\934'"), InternalError);
+        // that is not gcc/clang encoding
+        ASSERT_EQUALS(959657011, MathLib::toLongNumber("'\\u9343'"));
+        ASSERT_EQUALS(1714631779, MathLib::toLongNumber("'\\U0001f34c'"));
+        {
+            // some unit-testing for a utility function
+            ASSERT_EQUALS(0, MathLib::characterLiteralToLongNumber(std::string()));
+            ASSERT_EQUALS(32, MathLib::characterLiteralToLongNumber(std::string(" ")));
+            ASSERT_EQUALS(538976288, MathLib::characterLiteralToLongNumber(std::string("          ")));
+            ASSERT_THROW(MathLib::characterLiteralToLongNumber(std::string("\\u")), InternalError);
+        }
 
         ASSERT_EQUALS(-8552249625308161526, MathLib::toLongNumber("0x89504e470d0a1a0a"));
         ASSERT_EQUALS(-8481036456200365558, MathLib::toLongNumber("0x8a4d4e470d0a1a0a"));