This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Hi, It would be practical and helpful if XQTSCatalog.xsd was documented(with XML comments is sufficient for me), such that one didn't have to guess and reverse engineer what the vocabulary communicates. See also comment by Per Bothner, "Re: XML Query Test Suite version 0.8.6": http://lists.w3.org/Archives/Public/www-ql/2006JanMar/0033.html Regards, Frans
Created attachment 411 [details] commented XQTSCatalog.xsd I've made a first pass at commenting the XQTSCatalog.xsd file. I'm providing it here so that you don't have to wait until our next release.
That was a very kind thought of you. Thanks, Frans
Thanks. If would be better to use xs:document elements, rather than XML comments. Some tools, such as xs3p, can translate xs:document comments along with the rest of the xsd into a more readable html file. xs3p is the best of the best tools I've tried, but even it is limited. Perhaps W3C has better tools.
An important still-missing piece of documentation (and not one I can reliable figure out) is the "comparison-enum". What is the difference between XML, Fragment, Text, and Inspect? What kind of whitespace-normalization/-stripping is appropriate for each case? What serialization parameters should be used? The serialization spec leaves some aspects implementation-defined. For example: "Whether a serializer uses apostrophes or quotation marks to delimit attribute values in the XML output method is an example of such a detail." These have to specified, unless the assumption is that the implementation parses the expected result, and compares infosets. Another factor is floating-point. Formatting of float/double is still unclear, with inconsistencies between F&O and the Schema 1.1 draft. The testsuite specifies Text comparison for some results. Perhaps adding a new type "Doubles" for a space-separated list of xs:double values. Then an implementation could parse the expected result, and compare the results as numbers rather than Text.
(In reply to comment #4) > An important still-missing piece of documentation (and not one I can reliable > figure out) is the "comparison-enum". > > What is the difference between XML, Fragment, Text, and Inspect? > What kind of whitespace-normalization/-stripping is appropriate for each case? > ... I believe that these questions are answered in our Guidelines for Running the XML Query Test Suite [1], specifically the section on Comparing Results. > Another factor is floating-point. Formatting of float/double is still unclear, > with inconsistencies between F&O and the Schema 1.1 draft. The testsuite > specifies Text comparison for some results. Perhaps adding a new type "Doubles" > for a space-separated list of xs:double values. Then an implementation could > parse the expected result, and compare the results as numbers rather than Text. Our expected results are serialized as specified in XSLT 2.0 and XQuery 1.0 Serialization. We have an open issue concerning float and double, as several lexical representations map to the same value [2]. [1] http://www.w3.org/XML/Query/test-suite/Guidelines%20for%20Running%20the%20XML%20Query%20Test%20Suite.html [2] http://www.w3.org/Bugs/Public/show_bug.cgi?id=2895
Don't forget test-case/@version-drop! (the date of when the test case was created?) Regards, Frans
(In reply to comment #3) > Thanks. If would be better to use xs:document elements, rather than XML > comments. These elements have been added to the schema.
(In reply to comment #6) > Don't forget test-case/@version-drop! (the date of when the test case was > created?) We do not appear to be using this attribute. I'll add an annotation to this effect.