RE: Canon to canon logs

httpLocationTable.xml must be updated by hand, but there are only a dozen
entries and as I got partway down an XSLT implementation to parse and
replace templates I realized I was crazy to do so much work to duplicate
programmatically what I had already evaluated by hand.  Please do check the
values in the table - they are essentially hand-crafted assertions on what
the final address should look like given the {http location} and the sample
data the testcase requires.

By the way I just updated the table for data of myfirst, mymiddle, and
mylast in MessageTest-2G, consistent with the Canon tests.

Also there was a bug in the mustUnderstand assertions - missing namespace
and didn't allow "1" as an alternative for "true".

For those not on the side thread, we removed the dates from the log files so
that CVS can manage our history instead of us.

I added a nice feature that you can look at the message test log, click on
the "log" link and go right to the analysis of assertions for that test, and
then click again on "view log" which will take you directly to the message
log (including fragment id) that the analysis ran against.  By activating
that last link as "open in new window" you can quickly get a side-by-side
view of the message, the assertions, and see why they failed.

However, this feature requires Ant 1.7.0 for those brave enough to build the
results - I added an faq on that:
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/test-suite/faq.html#dashboa
rd


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

> -----Original Message-----
> From: Philippe Le Hegaret [mailto:plh@w3.org]
> Sent: Tuesday, January 09, 2007 7:54 AM
> To: Jonathan Marsh
> Cc: 'Youenn Fablet'; 'www-ws-desc'
> Subject: RE: Canon to canon logs
> 
> On Mon, 2007-01-08 at 16:02 -0800, Jonathan Marsh wrote:
> > I added a little indirection on the {http location} values.  I created a
> > little substitution table named httpLocationTable.xml, which simply
> contains
> > a list of pre-substitution and expected post-substitution values, and
> does a
> > swap when generating the assertions.
> 
> Nice one! Is httpLocationTable.xml generated from elsewhere or do you
> have to update it by hand?
> 
> > I also suppressed assertions checking the HTTP method and URL on
> response
> > messages - if that's not simply an oversight I am misunderstanding
> something
> > fundamental...
> 
> Good catch.
> 
> Philippe

Received on Wednesday, 10 January 2007 05:57:15 UTC