On Sat, 28 Mar 2009, Pascal Terjan wrote:
> On Sat, Mar 28, 2009 at 10:44 PM, Szabolcs Szakacsits <sz...@ntfs-3g.org>
> wrote:
> >
> > Thank you for the followup.
> >
> > As far as I see, this is a mount(8) bug. Canonicalize_mountpoint should be
> > called before mount(2), not after.
>
> Well the second mount process starts by doing canonicalize_mountpoint
> before updating the mtab (and never calls mount(2)).
If mount(8) calls the mount.foo helpers with resolved mountpoint then no
need to resolve it again when it's called with the -i option.
But this breaks things which call mount -i with unresolved mountpoint.
Of course there could be a new mount(8) option for the resolved mountpoint
case but the problem is, there are about half dozen different mount(8)
already. We had a lot of problems with '/bin/mount -i' in the past.
It seems it would be more reliable and maintainable to copy back the
relevant mount mtab update code to FUSE (and fuse-lite). How much code
is it?
Szaka
------------------------------------------------------------------------------
_______________________________________________
ntfs-3g-devel mailing list
ntfs-3g-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ntfs-3g-devel
