W3C home > Mailing lists > Public > www-ws-desc@w3.org > January 2007

RE: Changes to test suite

From: Jonathan Marsh <jonathan@wso2.com>
Date: Thu, 4 Jan 2007 20:40:55 -0800
To: "'Youenn Fablet'" <youenn.fablet@crf.canon.fr>
Cc: <www-ws-desc@w3.org>
Message-ID: <010101c73083$b0d449a0$3301a8c0@DELLICIOUS>

Hmm.  I think it's broken in a different way.

According to [1], which I don't think has been incorporated into the spec,
the {http location} isn't "concatenated" but resolved.

The relative URI "EchoName" is resolved against a base URI
"http://example.org/MessageTest-3G/endpoint-1" resulting in
"http://example.org/MessageTest-3G/EchoName" which still isn't quite what I
intended. ;-)

I think the best solution is to append a "/" to the endpoint addresses,
which I will do shortly.

[1] http://lists.w3.org/Archives/Public/www-ws-desc/2006Sep/0034

Jonathan Marsh - http://www.wso2.com - http://auburnmarshes.spaces.live.com
 

> -----Original Message-----
> From: Youenn Fablet [mailto:youenn.fablet@crf.canon.fr]
> Sent: Thursday, January 04, 2007 6:28 AM
> To: Jonathan Marsh
> Cc: www-ws-desc@w3.org
> Subject: Re: Changes to test suite
> 
> Jonathan Marsh wrote:
> >
> > 1) LocationTemplate-1G. Removed the "#" characters from the
> > whttp:location on the "Escape" binding. These aren't legal at this
> > spot in a URI. And there isn't a requirement they be escaped during
> > the templating algorithm, so they aren't testing anything and
> > preventing the wsdlcm from having a valid anyURI for the {http
> > location} property.
> >
> > 2) MessageTest-2G. Augmented each whttp:location attribute to start
> > with the name of the operation. Without some kind of unique
> > whttp:location, operation dispatch in Axis doesn't work. I'll send a
> > new issue describing the limitations of this shortly.
> >
> > 3) MessageTest-3G. Added whttp:location attributes on each Binding
> > Operation, with a value equal to the operation name. Because unique
> > whttp:location attributes are needed by Axis2, generic http bindings
> > are not possible, so I added minimally-defined operations to the
> > SafeEchoHTTPBinding.
> >
> By concatenating these new locations with the @address value we end up
> with URLs like:
> - http://example.org/MessageTest-3G/endpoint-1GuaranteedFault, or
> - http://example.org/MessageTest-3G/endpoint-1EchoName
> It may be better, in the primer and these examples to use/promote URLS
> like (please note the added / or ?):
> - http://example.org/MessageTest-3G/endpoint-1*/*GuaranteedFault, or
> - http://example.org/MessageTest-3G/endpoint-1*?*GuaranteedFault,
> - http://example.org/MessageTest-3G/endpoint-1/EchoName, or
> - http://example.org/MessageTest-3G/endpoint-1*?*EchoName
> 
> Regards,
> Youenn
> 
> > We also have a known problem with evaluating gzipped messages, for
> > which I have an action item. I think it would be easier first to get
> > some initial results to pinpoint precisely which testcases can't be
> > evaluated, and for which of those a control group would be useful. So
> > I haven't touched this part yet.
> >
> > **Jonathan Marsh** - http://www.wso2.com -
> > http://auburnmarshes.spaces.live.com
> >
Received on Friday, 5 January 2007 04:40:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:45 GMT