GNU bug report logs - #10639
solaris: test failures with coreutils-8.15

Previous Next

Full log

Message #32 received at 10639 <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Bruno Haible <bruno <at> clisp.org>
Cc: "JONES, BILL" <rj7252 <at> att.com>,
	"10639 <at> debbugs.gnu.org" <10639 <at> debbugs.gnu.org>, bug-gnulib <at> gnu.org
Subject: Re: acl code rewrite
Date: Wed, 22 Feb 2012 11:39:18 -0800

[Message part 1 (text/plain, inline)]

On 02/19/2012 07:18 PM, Bruno Haible wrote:
> You would think that commit is fine, right? But it introduces a test failure
> on Solaris 10 over NFS:
> 
> files tmpfile0 and tmpfile2: different ACE-ACL entry #0: different access masks 1400207 and 5400607
> FAIL: test-copy-acl-2.sh
> 
> What's happening, is that the test-copy-acl copies wrong ACL entries.

Thanks for looking into this.

I am not observing this problem on my Solaris 10 + NFS host.

$ uname -a
SunOS kiwi.cs.ucla.edu 5.10 Generic_141444-09 sun4u sparc SUNW,Sun-Fire-280R Solaris
$ mount | grep eggert
/home/eggert on elephant:/export/elephant/f/eggert remote/read/write/setuid/devices/xattr/dev=55522d5 on Wed Feb 22 10:24:33 2012

For file-has-acl.c I'd like it if the code didn't have the arbitrary limit
around half the maximum allocated size, and the code could be simplified
and regularized a bit so that the two Solaris cases are as similar as
possible, so I propose the attached patch.  The same idea can be applied to
copy-acl.c but of course it's more important to fix the bug.
(Maybe applying the idea will fix the bug? ...)

[file-has-acl-patch.diff (text/x-patch, attachment)]

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.