W3C home > Mailing lists > Public > www-ws-desc@w3.org > June 2002

Fw: [soapbuilders] NMTokens, NCNames, PortTypes, and Bindings.

From: Sanjiva Weerawarana <sanjiva@watson.ibm.com>
Date: Tue, 11 Jun 2002 10:33:26 +0600
Message-ID: <00a401c21101$22eef3c0$524b6a20@lankabook2>
To: "WS-Desc WG \(Public\)" <www-ws-desc@w3.org>

Here's a bug in the 1.1 docs we need to fix ..

Gudge: I think it would be very good if you try to run a fine
toothcomb over the schemas before we publish our first doc!

Sanjiva.

----- Original Message ----- 
From: "Wes Moulder" <wes@themindelectric.com>
To: <soapbuilders@yahoogroups.com>
Sent: Tuesday, June 11, 2002 1:01 AM
Subject: [soapbuilders] NMTokens, NCNames, PortTypes, and Bindings.


> Oh my!
>  
> >From the soapbuilders conference, one bug came up with the WebMethods
> guys, and I wanted some clarification of understanding.
> The WSDL spec says that a PortType's name attribute is an nmtoken, but
> the Binding refers to this via a QName, which we have always understood
> to be a combination of the targetNamespace of the WSDL, and the name of
> the PortType.  If the PortType's name attribute is an nmtoken, it's
> allowed to have characters (such as ":") which are illegal for an
> NCName.
>  
> This same issue holds for binding and wsdl:port in the service
> description.
>  
> The issue is, all over the place in the WSDL spec, it uses nmtokens for
> name attributes, which can result in invalid qnames for types that refer
> to them.  Is this an inconsistency in the WSDL spec, an inconsistency in
> my understanding of the issues, or a "feature"?
>  
> Comments appreciated,
> --Wes
>  
>  
> 
Received on Tuesday, 11 June 2002 00:34:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:20 GMT