[Samba] Re: [Possible BUG] Samba v3.0.20b and permissions
gasch at eva.mpg.de
Thu Oct 20 14:43:11 GMT 2005
hi&big thx jerry,
two questions remain:
1) i'm using "store dos attributes" so (a)/(b) do not apply for me?
2) 'map read only = yes' should give you 3.0.14a behavior.
-> according to the docs wouldn't "map read only = Permissions or No"
give the <3.0.20 behaviour?
i'll play around with 3.0.21pre
thx so far
> In the absence of 'store dos attributes = yes' and the ability
> to represent DOS attribute bits separately from permission bits,
> we have two choices.
> (a) May the read only DOS attrib bit to the inverse of the
> user write bit (this is what we did prior to the 3.0.20 series),
> (b) represent the read only attribute based on whether the
> user can actually write to the file.
> If the 'R' DOS attribute is set, Windows will not allow you
> to delete the file IIRC which is probably what you are hitting.
> In 3.0.21pre1 'map read only = yes' should give you 3.0.14a behavior.
> The problem in the current 3.0.20 behavior (which might be
> considered a bug) is that we do not take the parent directory
> permissions into account when determining whether the user
> has write access or not to a given file. And hence this is
> another argument to stay away from user space access checks.
> But we have no choice with the changes in WinXP sp2 and file
> Does this help explain the behavior?
> cheers, jerry
> Alleviating the pain of Windows(tm) ------- http://www.samba.org
> GnuPG Key ----- http://www.plainjoe.org/gpg_public.asc
> "There's an anonymous coward in all of us." --anonymous
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.0 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
> -----END PGP SIGNATURE-----
Max Planck Institute for Evolutionary Anthropology
Department of Human Evolution (IT)
Deutscher Platz 6
Phone: 49 (0)341 - 3550 137
More information about the samba