Recently I am working with the book "The C++ Programming Language 3rd Edition By Bjarne Stroustrup". In this book, the exercise 4.11 question 5. It ask me to print some largest and smallest value. I use the bit method and others work well except float, double and long double.
My system is Ubuntu 12.04 64bit. g++ version: 4.6.3
I check the IEEE's definition of double. It is something like this:
sign bit biased exponent fraction from normalized mantissa
1bit 8 bits 23 bits
My largest float is: Max = 0x 7f7f ffff
Another float is: Test = 0x ffff ffff
The output is: sizeof(long) is 4, Max = 2.1391e+09, the Test = 4.29497e+09
It is confused me that: 1st, the Test should be a negative but output is positive. 2nd, the Test is even larger than the Max.
Is there anyone could help me out here? Thank you!