file into memory first, and then scan for newlines. The file might be very
big. This version instead scans through the file. This could probably be
further improved too, to not use this read-single-bytes approach.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@442 a1c6a512-1295-4272-9138-f99709370657
a lot faster. Also made the function accept a seed number as a function
argument, as we are likely to want to randomise using a known seed very
many times...
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@437 a1c6a512-1295-4272-9138-f99709370657
