Remove special case of limit for double printing

The case of numeric limit for floating point does not imply that the
decimal version is precise enough.
An example of that is 0x1p-1022 which would be printed as 2.22507e-308
in decimal but this is not precise enough.

Author: romainbrenguier

jbmc/src/java_bytecode/expr2java.h

@@ -79,8 +79,6 @@ std::string floating_point_to_java_string(float_type value)
     return raw_decimal;
   }();
   const bool is_lossless = [&] {
-    if(value == std::numeric_limits<float_type>::min())
-      return true;
     try
     {
       return std::stod(decimal) == value;