- From: Dominique Hazaël-Massieux <dom@w3.org>
- Date: Wed, 02 Feb 2005 12:01:52 +0100
- To: connolly@w3.org
- Cc: public-rdf-in-xhtml-tf@w3.org
- Message-Id: <1107342112.10233.110.camel@stratustier>
Hi DanC, RDF/XHTML TF, I've started a (smallish for now) test suite for GRDDL implementations; right now it includes only test for the simplest case, ie XHTML document using the GRDDL profile: http://dev.w3.org/cvsweb/2005/grddl-ts/ The tests are described in RDF, using the RDF schema that was developed for the RDF Core Test Suite: http://dev.w3.org/cvsweb/2005/grddl-ts/test-lists.rdf Adding more tests to cover the other cases is simple enough, but before getting to that I'd like to complete the simple test harness I've started to develop: http://dev.w3.org/cvsweb/2005/grddl-ts/testHarness.py You invoke it à la: python testHarness.py --<action> "/home/dom/dev.w3.org/2005/grddl-ts/gleanWrapper" file:///home/dom/dev.w3.org/2005/grddl-ts/test-lists.rdf where <action> can be "run" or "update" (I've written a wrapper for glean.py and one for the XSLT-servlet based demo; they're straightforward shell scripts that takes a URI as sole argument; I'll probably add a wrapper for a RAP-based GRDDLer later on) "update" is implemented and makes it so that local files matching the output URIs be generated from the processor given in argument. "run", which is supposed to be the interesting part, only outputs the results of GRDDLing the various input files; it should instead compare the said output with the referenced output documents given in the tests list. DanC, you mentioned using cwm to do this: would this imply using delta.py as distributed in the cwm tarball? If so, I'm not sure whether this should be done by importing the said python code, or by using it as a command line tool; if the latter, it looks like delta.py doesn't return a different code depending on whether there was any diff or not, making it a bit more painful to use in this context. Any suggestions? (in the end, I'd like the test harness to output an EARL description of what tests failed/succeeded, but that's probably easy enough to code once the diff is implemented) Dom -- Dominique Hazaël-Massieux - http://www.w3.org/People/Dom/ W3C/ERCIM mailto:dom@w3.org
Received on Wednesday, 2 February 2005 12:31:45 UTC