RangedFileReader.__init__ is slow on large file

[philippe-docourt]

Version

0.1.2

File / Feature

class RangedFileReader(object):
    def __init(...)

Expected behaviour

Obtaining the size of the file should be much quicker, and less CPU intensive. We should probably use something like this:

os.fstat(self.f)
self.size = statobj.st_size

Actual behaviour

The current approach for obtaining the size of the file is terribly slow on large files : it uses CPU and disk I/O
self.size = len(self.f.read())