ntfs-3g and ntfsprogs

Gene Czarcinski gene at czarc.net
Tue May 27 15:51:57 UTC 2008


On Monday 26 May 2008 17:49:29 Tom "spot" Callaway wrote:
> On Mon, 2008-05-26 at 17:13 -0400, Gene Czarcinski wrote:
> > My suggestion:  Provide an additional binary package for the ntfsprogs
> > mount command (e.g., ntfsprogs-mount) which would have the mount
> > command and man-page.  Installation of this "new" package should be
> > made to conflict with ntfs-3g so that both could not be installed at
> > the same time.  For F10 (an probably F11) continue with the current
> > default installs ... that is, both ntfs-3g and ntfsprogs but not
> > ntfsprogs-mount.
>
> I disagree. Having two methods for ntfs mount seems like a recipe for
> failure. I looked at both of them, and determined that the ntfs-3g mount
> mechanism was far more robust and better maintained. I don't really
> think we benefit at all from enabling the ntfsprogs mount functionality.

OK, you sold me.  I learned many years ago that it really does not matter 
how "fast" something runs if its operation is not reliably correct.
>
> Or, to put it more succinctly, when we've had any problems with ntfs-3g
> mount, Szaka has been extremely helpful in working with us to resolve
> the issues. We've also had issues with the ntfsprogs suite, and received
> zero help or feedback on our patches. "Supporting" the ntfsprogs mount
> will simply lead to more bugs, and I've got enough of those as is. :)

Even more sold.

I have already had a minor problem with ntfsclone where is would not restore 
to a (vmware virtual disk) partition which was exactly the same size/same 
geometry as the original.  The work around was to make the partition slightly 
larger, do the ntfsclone restore (which now worked), and then fix things up 
with ntfsresize.

>
> I'm hopeful that in time, the conflict/competition between ntfs-3g and
> ntfsprogs will all balance itself out. If we need to enable the
> ntfsprogs mount, it is only a minor amount of work, and could be done
> quickly.

I sure hope the issues are resolved and the two projects merge.

>
> ~spot
>
> p.s. I'm hedging my bets on ntfs-3g. They show community and growth,
> where ntfsprogs doesn't.

As I said, it was somewhat surprising to have a project be more or less 
stagnate and then, suddenly, issue a huge change.  This alone makes ntfsprogs 
a bit questionable.  The ntfsprogs mount command is a minor issue.  The real 
question to me is how good the code is in libntfs as compared to libntfs-3g.

Gene




More information about the fedora-devel-list mailing list