W3C home > Mailing lists > Public > public-qa-dev@w3.org > August 2006

Re: [ucn] adding an observer (and somehow failing so far)

From: olivier Thereaux <ot@w3.org>
Date: Tue, 22 Aug 2006 13:41:10 +0900
Message-Id: <73E1D3AC-B303-49BD-9CB3-4DA89026D3D3@w3.org>
Cc: Damien LEROY <xelloss.leroy@gmail.com>, QA Dev <public-qa-dev@w3.org>
To: Jean-Guilhem Rouel <jean-gui@w3.org>

Hello Jean-Gui, Damien, all,

On Aug 22, 2006, at 05:58 , Jean-Guilhem Rouel wrote:
> I am not sure to understand what you want to add in the ant file.
> Building Unicorn is completely independent from the tasks and
> configuration files.

Yes, and no, if I only base myself on the documentation :)
Unless I am mistaken (or the documentation is wrong), the UI is being  
generated at the time the war file is being deployed. And since, for  
now at least, the output of the auto-deployment is not very useful  
nor easy to get to, the idea was that the ant build could be a good  
time to run a few checks, before passing potentially broken files to  

> Errors are reported in the servlet because xml-wellformedness of  
> config
> files and everything else is checked at run time, during the
> initialization of the servlet (to initialize the servlet, there are at
> least two ways: restarting tomcat or using the tomcat manager web
> interface).

are the HTML and JS file also regenerated then?

> So, you can add a target in ant, to check that, but I don't think it
> would be very helpful since one of the aims of Unicorn is to avoid
> having to re-build it after each change in the tasks, observers or
> templates. Improving messages in log files and splitting them (it  
> might
> already be the case, I don't remember) would be far more useful. You
> could have for example one file for errors in the tasklist, one for
> errors in the observers list, one for errors in contracts and one for
> errors in templates. Another interesting one would log all the  
> requests
> coming from the end-user and their associated errors. And something  
> very
> important would be to describe all these log files in a README or
> something like that :).

Absolutely, I agree with all this. I think this is something Damien  
has been working on. To some extent, the logs are here and already  
quite well separated, but not yet very usable, and not yet documented.

Received on Tuesday, 22 August 2006 04:40:27 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:36:26 UTC