- From: Jeremy Carroll <jjc@hpl.hp.com>
- Date: Mon, 30 Apr 2007 17:49:53 +0100
- To: ogbujic@ccf.org
- CC: GRDDL Working Group <public-grddl-wg@w3.org>
Oh yes, I agree that the earlier resolution isn't really valid. I believe that the current library code deals with this correctly: by ignoring the xml:base on the root element, and only doing something with xml:base's lower down the tree. Jena is passing this test. I propose (re-)approving it. === I am also happy if we delete it, if that is the easier option. Jeremy Chimezie Ogbuji wrote: > On Mon, 2007-04-30 at 14:50 +0100, Jeremy Carroll wrote: >> I am happy with these modifications, except: >> - the aboutTests change, >> - inclusion of base-detail, superceded by events >> - minor aspects of description of #xmlbase3 >> >> Also, no approval seems to have been added to the grddl result for >> embedded-rdf4. > > I wasn't sure where we stand with regards to approval on this test. > See: > http://lists.w3.org/Archives/Public/public-grddl-wg/2007Apr/0255.html > where I indicate how this test might be conflicting with the 'current > reading'. > >> aboutTests >> ========== >> >> I'll explain my concerns, and make a suggestion as to achieve both our >> goals. >> >> My understanding is that a URI such as >> >> http://www.w3.org/TR/grddl-tests/titleauthor.html >> >> are not part of the WG's intent. > > Yes. Note in the current 'draft' the href link is absolute, so this will > not be a problem if / when it is published into > http://www.w3.org/TR/grddl-tests/ space - at least with regards to the > links in the source document. > >> The GRDDL result of the test cases document, when published at >> >> http://www.w3.org/TR/grddl-tests/ > > Yes, when gleaned 'on the fly' from http://www.w3.org/TR/grddl-tests/. > >> will include such a URI; hence this is unsatisfactory. >> I also believe that the GRDDL results will be published at >> http://www.w3.org/TR/grddl-tests/grddl-tests.rdf > > One option is to *not* publish that 'pre-gleaned' RDF graph from /TR > space. I much prefer this. However, if we go with your first > suggestion from below then this becomes a non-issue. > >> and once again, this should not include relative URIs into >> http://www.w3.org/TR/grddl-tests/ >> >> If >> http://www.w3.org/2001/sw/grddl-wg/td/grddl-tests.rdf >> is copied as is into the TR directory, we will have such references to >> non-resources. >> >> It would be possible to do either of the following: >> >> a) use the aboutTests as modified, but with an >> xml:base="http://www.w3.org/2001/sw/grddl-wg/td/" >> and then someone wanting to do local testing, can simply delete that base. > > I also prefer this. It requires only a little more effort than prior to > get an environment for local testing (which I think is crucial - > especially considering my painful experience with being blocked from W3C > space for repeated requests) > >> b) include the aboutTests as modified as aboutTestsLocal.xsl, and revert >> aboutTests.xsl to only give the full URLs > > I'm against this option as I think we can achieve both goals of not > relying on the current .htaccess redirect of test material as well as a > relatively painless way to run local tests. > >> Personally, I think either of these is unnecessary, but I wouldn't mind. >> >> For me, the implementor who is testing against a local copy of the test >> material is using a cache, and that is their responsibility and not the >> WGs. >> (a) above, in particular, seems to be making enough allowance for >> such an implementor > > I agree. I do think *some* allowance is prudent. > >> >> Suggested change: >> >> <xsl:template match="/"> >> <r:RDF> >> <xsl:apply-templates /> >> </r:RDF> >> </xsl:template> >> >> to >> >> <xsl:template match="/"> >> <r:RDF xml:base="http://www.w3.org/2001/sw/grddl-wg/td/"> >> <xsl:comment>If you wish to run the tests using local copies of the >> files then it is possible to modify the above xml:base as >> appropriate.</xsl:comment> >> <xsl:apply-templates /> >> </r:RDF> >> </xsl:template> > > Done.. > >> base-detail >> =========== >> >> The text is: >> [[ >> # A transform that does not respect xml:base: >> input output >> >> The output of a transform is processed with the retrieval IRI, not that >> given by an xml:base. >> ]] >> This text is not consistent with the descriptions and decisions of the >> xmlbase and htmlbase tests. >> >> Since we have not approved this test, I suggest simply deleting it. >> We may want to review our coverage of the base issues later. > > I agree (re: deleting it), I just wanted to confirm with you (as it was > in the pending list). Done .. > -- Hewlett-Packard Limited registered Office: Cain Road, Bracknell, Berks RG12 1HN Registered No: 690597 England
Received on Monday, 30 April 2007 16:50:07 UTC