It is automatic, just not in the way that you're thinking =] PPCs (Mac) store data in bigendian style and x86s (Windows/Linux) store data in littleendian style.
So if the code works on x86 but not Mac, you need to specify littleendian on Mac. Really, if you're going to interpret data via files you need to specify what byte order you're reading in every time, for platform compatibility purposes.
It does make sense to 'automatically' read bytes in the order the processor is used to by default, but people need to be made more aware of it, like most of the omissions in the docs, I guess.
So if the code works on x86 but not Mac, you need to specify littleendian on Mac. Really, if you're going to interpret data via files you need to specify what byte order you're reading in every time, for platform compatibility purposes.
It does make sense to 'automatically' read bytes in the order the processor is used to by default, but people need to be made more aware of it, like most of the omissions in the docs, I guess.