J. Bruce Fields wrote:

>On Wed, Jun 07, 2006 at 11:50:31AM -0400, Peter Staubach wrote:
>  
>
>>The Red Hat BZ number is 193621.
>>    
>>
>
>"You are not authorized to access bug #193621", it tells me....
>
>  
>

Hmm.  That ones seems to be restricted for some reason.  I think that
this happens when we get escalated bugzillas from customers.


>>The description is that when zero length files are copied, even over
>>an existing zero length file, the mtime on the target file does not
>>change.
>>    
>>
>
>Is the server-side patch sufficient on its own?
>  
>

The server side patch isn't quite sufficient on its own.  A RHEL-4 patch
is also required for the client side.  I could construct the RHEL-4 patch
so that it alone would be sufficient to address the particular problem
that that customer is having, but that isn't the entire situation.
Non-Linux clients would still have a problem with the current upstream
Linux server.  For example, in my testing, a Solaris 10 client mounting
an FC-5 server fails.  When running the attached script, the mtime on
the file, bar, should change by about 1 minute, 3 times.

    Thanx...

       ps