Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com
Message-ID: <433BDB6D.1010806@byu.net>
Date: Thu, 29 Sep 2005 06:17:49 -0600
From: Eric Blake <ebb9 AT byu DOT net>
User-Agent: Mozilla Thunderbird 1.0.2 (Windows/20050317)
MIME-Version: 1.0
To: cygwin AT cygwin DOT com
Subject: Re: bug in rmdir(2)
References: <092820052226 DOT 6432 DOT 433B18B2000A598D0000192022007358340A050E040D0C079D0A AT comcast DOT net> <20050929075025 DOT GA19713 AT calimero DOT vinschen DOT de>
In-Reply-To: <20050929075025.GA19713@calimero.vinschen.de>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
X-IsSubscribed: yes

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

According to Corinna Vinschen on 9/29/2005 1:50 AM:
>>The fix to rmdir(2) is easy - check for a trailing / or /. or /..
>>before handing the name off to the complex path processing
>>code, and fail with EINVAL if so.  rmdir(2) isn't called often
>>enough for this to slow down everything else, and there are
>>no Windows API calls in this failure mode, and in return you
>>get POSIX compliance.
> 
> 
> Just a trailing slash is fine, btw.  Try this on any POSIXy system.

It is still inconsistent on POSIXy systems - compare Solaris 8 vs. Solaris 10:

8> mkdir a
8> ln -s a b
8> rmdir b/
rmdir: directory "b/": Path component not a directory

10> mkdir a
10> ln -s a b
10> rmdir b/
10> ls ?
b

However, I'm not sure in Solaris 8 whether the trailing slash was removed
by rmdir(1) before calling rmdir(2), or whether rmdir(2) errored out with
ENOTDIR.  But Solaris 10 certainly removed the referrant directory,
leaving the symlink dangling.

Further discussion on the coreutils list pointed out that POSIX has some
holes in this area (a strict reading of "rm -R symlink/", for example,
would remove the contents of the referrant directory, then fail with
EINVAL, but obviously no system does that - they just remove symlink and
leave the referrant directory intact).  In light of the misunderstandings
in interpreting POSIX, and the various behaviors on different POSIXy
platforms, I'm now in the process of asking the Austin group for
clarification (with a goal of mandating, or at least allowing, existing
behavior, where it is different from a strict reading of the
specification).  So for now I agree that you only need to check for
explicit trailing . or .., and don't worry about trailing / until the next
version of POSIX better defines what can/should be done.

- --
Life is short - so eat dessert first!

Eric Blake             ebb9 AT byu DOT net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDO9tt84KuGfSFAYARAmuTAKDQSxOctFpFJW5mukyYCdzIlNy1owCgvsLJ
f0wQVl/Go1ToSGYy+TOm4/0=
=WY/e
-----END PGP SIGNATURE-----

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/