[PLUGIN] bzr uncommit updated

[Aaron Bentley]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

John Yates wrote:

> I do believe that if you reread the referenced
> article closely you will see that there is very
> little discussion of hard links and that all the
> nasty behavior centers around Windows' equivalent
> of symbolic links.

Yes, I'm aware that the nasty behaviour in that example was produced by
junctions.

As I said, my concern is that because these things are so rarely used,
other programs will make assumptions about them that are incorrect.

A trivial example would be disk useage programs.  Hard links have also
caused bugs in Windows auditing.  They may also cause confusion for users.

  Little surprise there given
> the chaos that occasioned the NTFS5 introduction
> of the junction or reparse point.

I haven't seen anything to indicate that Explorer's bad behaviour is
caused by the junction implementation itself.

> Hardlinks OTOH have been in NTFS since day one.

- From what I've read, they were introduced in NT 4.

> I encourage you to enable them on Windows as an
> experiment.

I'd be happy to allow users to override my conservative choice, but not
using it as the default behaviour.  But if someone else wants to submit
a patch and it's accepted, that's fine by me.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDOEk+0F+nu1YWqI0RAlXVAJoCVse79l/xOx6YquHr6clPWxl1xQCcCNFK
/+iTOKLmZl3xPuZXJiSJ8oU=
=R4Cf
-----END PGP SIGNATURE-----