Hello all.
Suppose I'm reading some doubles from a binary file.
I can do this: myBinaryFile.read(reinterpret_cast<char *> (&myDouble), sizeof (double));
Yuck. That is nasty. The code is also nonportable to machines of different endianness, different sizes of types, or maybe it's the file itself that's not portable lol? Thanks C++.
So I'd like to know if there are better methods I could employ for reading things from binary files. Maybe there's something I forgot about?
Read into a char[sizeof(double)] and move bytes around as needed. Although for maximum portability, I would just not mix binary files and floating points at all, or I would find some way to decompose the number into mantissa and exponent and then store these myself.