Pack the files into the smallest number of bins.  This is an
approximation - the problem is NP-complete and no known algorithm can
guarantee a solution better than ‘(11/9) * OPTIMAL + 4’.

To make things even worse, for large files, there is no portable way
to predict how many blocks the file system implementation will require
for indirect blocks, directories, etc.  So keep in mind that this is
only an approximation.

Bin packing can be useful when you want to put a bunch of files on
floppies or zip disks and you want to optimize things a little bit
(‘bin-packing’).

‘gitfm’ assumes that you want to pack all the files in the current
directory - if there is any selected file in that directory it will be
unselected first.  Then ‘gitfm’ will ask for a bin size, and select
the files that should go in the first bin.  You are supposed to place
those files in the first bin (e.g. a tar archive), remove them from the
current directory, then run ‘bin-packing’ again, to obtain the list
of the files that should go into the second bin, etc.