Re: XML Namespace labels

[Martin Bjorklund <mbj@tail-f.com>]

Andy Bierman <ietf@andybierman.com> wrote:
> Speaking of which, how are you handling this special case problem:
> 
> The RPC processing mandates that all the attributes in <rpc> are
> returned in <rpc-reply>.  This includes all the xmlns declarations.
> 
> Inside data models, the label for a given namespace can change (over and 
> over)
> with yet more xmlns declarations.  
> 
> If an error occurs, and a QName needs to be returned in <error-info>,
> do you reuse the labels forced in the <rpc-reply> or create a new xmlns
> declaration with the label used in the QName? And if that label is
> already in use for a different namespace (in the <rpc> element)? Do you
> use the label from the xmlns in the original <rpc> element?

When you write 'label', do you mean the namespace prefix?  Actually I
don't see the problem, but in this case, our implementation uses new
prefixes and new xmlns declarations.

> Since the URIs are the same, all SW is expected to figure out that different
> labels map to the same NS, and applications should handle it no matter what.
> 
> XML is so fun...

Yes, but honestly I don't think namespace handling is the biggest
problem when implementing netconf... But then I am selling something :)

Anyway, you write:

> The RPC processing mandates that all the attributes in <rpc> are
> returned in <rpc-reply>.  This includes all the xmlns declarations.

Is this intentional?  Why is it important to return the xmlns
declarations?  Couldn't we just state that xmlns declarations are not
returned?


/martin


--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>