| 03-02-2004, 06:46 PM | #1 |
Question says all. I know reals store more decimal places then three thanks to the return bug, but I don't know if they comply to the float specifcation. From my basic knowledge a float is made up of such 1 bit for sign, 9 bits for exponent, 22 bits for mantissa. The java specification (just read the Float wrapper classes comments about float to int bits) tells you how to compile such a number to its float form from bit but I don't know how to do such actions as bitwise and and the like in jass. Code:
(bits is the hex make up of the float)
int s = ((bits >> 31) == 0) ? 1 : -1;
int e = ((bits >> 23) & 0xff);
int m = (e == 0) ?
(bits & 0x7fffff) << 1 :
(bits & 0x7fffff) | 0x800000; |
| 03-02-2004, 07:34 PM | #2 |
That sounds about right for java. For c++ and other langs floats are differnt then doubles. Doubles are closer to reals in my gess since the dec does not float. In java or C++ i always use doubles never floats. |
| 03-03-2004, 03:09 PM | #3 |
Dont know if this might help you: Integer -2147483648 to 2147483647 (integer) (eg 2; 3;..) Real 5.0 x 10324 .. 1.7 x 10308 (float) (eg 1,342...) |
| 03-04-2004, 01:27 PM | #4 |
Erhm, they are most certainly floats.. or doubles.. but double is just float with double accuracy.. (since warcraft 3 is programmerd in c++ ) |