Re: chmod/chown -R - maybe an ITP

[Craig Sanders <cas@taz.net.au>]

On Mon, Dec 13, 1999 at 11:37:09PM +0000, Jules Bean wrote:
> On Tue, 14 Dec 1999, Craig Sanders wrote:
> > you could write a safe-chown program, but even that would be vulnerable
> > to races, since there is no atomic test-and-chown operation...it would
> > just reduce the window of opportunity.
> 
> You're wrong, of course, as was pointed out earlier in the thread.

yeah, i noticed that...read the rest of the thread after sending my
messages.  should have read first, written later.

> Once you open a file, you can check it's the right one with fstat(), and
> chown it with fchown(), and it can't be replaced in between (since you
> have an fd open to it, and it will stay open to the original file).

yep.  so a safe-chown program is possible.  

probably the best way to implement would be to add an '-a' or '--atomic'
(alternatively, '-s' or '--safe') option to the chown in GNU fileutils,
and contribute it back upstream.

craig

--
craig sanders