Bug#3133: quota files not sparse after running quotacheck

[Mark Eichin <eichin@cygnus.com>]

Marek Michalkiewicz <marekm@i17linuxb.ists.pwr.wroc.pl> writes:
> Hmm.  My copy of the info documentation says that it does, but only
> if the source file contains holes.  After quotacheck, it doesn't...
> Is there any reason why cp doesn't always make holes?

Sigh. I tried, it and then checked the source... it turns out that cp
stats the file to see if it has *any* holes (which is all it can tell,
really, it can't find them) and then makes holes for any block of
zeros it finds. If there aren't any, it doesn't try.

The reason is probably that there was a bug report, a long while back
(189 or 1990 if I remember right), that the sunos boot loader couldn't
read files with holes, so if you useed gnu cp on a kernel, you'd be
unable to boot the system. I'd never checked how it was resolved, I
guess this is it.

I wonder if they'd take a command line option to do it... though I'd
suspect that dd is a better place for it. Regardless, it's quotacheck
that needs to be fixed...