Things to do:

	--	finer grained percent updates (per loop)

	--	make file renaming a seperate option from random passes

	--	add more support for systems w/o a random dev

	--	use stdin as an alternate entropy source

	--	stir the hash input buffer in hashin_update()

	--	support for device wiping --  FINISHED BUT ** UNTESTED **

	--	random length incremental truncation
		to try and screw with indirect blocks

	--	pass shuffling

	--	make wipe more modular - next major release

	--	add getdents() support
	--	add sendfile() support

	--	open the file (maybe earlier than we do now), stat() it
		and compare the new stat struct to the old one (to avoid
		races)

	--	inode/directory scrambling
		randomly create new files in directory containing target(s)
		create random (dir local) hard links.

		basically, try to wipe everything in struct stat

	--	switch to a total percentage for all the files selected,
		kind of like how apt-get does it.

		something like these:

		[blah.txt: 60%] 24%

		blah.txt: 60% [24%]

		[24%] blah.txt: 60%

		probably go with the last example (or maybe the first, or
		maybe even make it optional).

