C++ compiler interpretation of long doubles

HCL Informix® Object Interface for C++ provides data type conversion functions in the value interface ITConversions to enable conversion of C++ type long double. The intent is to permit fetching floating point values into C++ long double variables. However, the HCL Informix Client Software Development Kit does not allow for conversion of long double values into HCL Informix decimal or float types. Thus, Object Interface for C++ applications should always ensure that any floating literal passed to ITConversions::ConvertFrom(long double val) is within the double range. Otherwise, ConvertFrom(long double val) will return FALSE for value objects that contain SQL MONEY, FLOAT, and SMALLFLOAT values.

Object Interface for C++ is written with the assumption that a floating literal without the ANSI C++ specified suffixes l or L (example: 12.988 instead of 12.988L) assigned to a long double variable will be treated by the C++ compiler as a long double. This assumption agrees with the ANSI C++ Draft Standard (Doc No: X3J16/94-0027, WG21/N0414, 25 January 1994), which states that the type of a floating literal is double unless explicitly specified by a suffix. The suffixes f and F specify float; the suffixes l and L specify long double. Thus, the suffix l or L must be applied to a floating literal in order for it to be interpreted by the C++ compiler as a long double value.

long double d = 12.988;
double dasd = (double) d;
if( dasd == d )
return 0;
else return 1; 

The following table compares support for the ANSI C++ draft standard referenced above among several versions of Sun C++ compilers. The table shows how the different compiler versions evaluated the expression (dasd == d). If the expression evaluates to FALSE, the values are not equal.