[apparmor] [patch] use seperate fs with user_xattr for xattr regression tests

[John Johansen]

On 09/30/2010 10:28 AM, Steve Beattie wrote:
> On Thu, Sep 30, 2010 at 01:47:02AM -0700, John Johansen wrote:
>> On 09/29/2010 11:02 PM, Steve Beattie wrote:
>>> This patch modifies the xattr regression test to use a separate loopback
>>> mounted filesystem to operate on, to guarantee that the mount option
>>> user_xattr is enabled (it's disabled by default on Ubuntu).
>>>
>>> With this change, a number of the user xattr testcases that were
>>> expected to pass but weren't started working; however, some of the ones
>>> that were failing as expected are now passing. I've touched up the
>>> expectations as well.
>>>
>>> After this patch is applied, the number of testcases in it that are not
>>> matching their expected result is still annoyingly high:
>>>
>> right, this is do to a change in what can be mediated with respect to
>> xattrs since the rewrite.  We need to get some new hooks upstream before
>> we will be able to address these failures.
> 
> Sure, mostly what I was trying to address with this patch was that on a
> filesystem without user_xattr, the script aborts because setfattr is
> unable to set a user xattr:
> 
>   setfattr: /tmp/sdtest.20644-28154-2Vo8j7/testfile: Operation not supported
> 
> (or maybe it doesn't abort, but at least generates an error message;
> I doubt -e has been set on the scripts.)

Right sorry I was just commenting on why they are broken, so anyone causually
reading the list would know, and I just wasn't up to reviewing the patch.

Which looks good to me the only thing I would like to see added is an early
bail out if the mount fails, but that can be done separately.

ACK.