attn: DCB - reproducing Solaris ACL bug...

[John Trostel]

The XFS user function,'chacl', which calls acl_set_file, which calls acl_set
(which calls the system call)... works fine.

(spaces and <cr>'s added for clarity below)

[jt at jtsdell xfs_part]$ mkdir test

***This sets 'b'oth the acl and the dacl

[jt at jtsdell xfs_part]$ chacl -b
                         u::rwxdpo,g::r-x,g:user1:r-x,m::r-x,o::r-x
                         u::rwx,g::r--,g:user1:r-x,m::r--,o::r-- test

***This 'l'ists the acl and the dacl

[jt at jtsdell xfs_part]$ chacl -l test
                test
                [u::rwx,g::r-x,g:user1:r-x,m::rwx,o::r-x
                /u::rwx,g::r--,g:user1:r-x,m::r--,o::r--]

Is this not the behavior you see in your implementations?

Should I code up a test case at a lower level, or is the 'chacl' test
sufficient?

On 24-May-2001 Michael Davidson wrote:
> No - at least not the one I put together.
> 
> It was just a low level 20 line test program that calls the Solaris/UnixWare
> acl() system call with a particular test ACL.
> 
> md
> 
> John Trostel wrote:
>> 
>> What's the "test case" you are writting up? Is it running the acls through
>> the
>> Samba code?
>> 
>>

-- 
John M. Trostel
Linux OS Engineer
Connex
jtrostel at connex.com