This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Even allowing for the tests flagged as parse-error, which I think are intentionally omitted from the XqueryX test suite, there are 234 queries listed in the catalog that only appear as XQuery. This includes all the queries specified using <input-query> so none of the external variable tests may be attempted by an XQueryX implementation. In addition the 13 modules in TestSources/*-lib.xq are only distributed in XQuery form, so an XQueryX implementation can not attempt any of the module tests. David
We are aware of these issues, but have not been able to scare up the necessary resources to correct them. We will be publishing XQTS 1.0 shortly. These fixes will appear in a point release in a couple of weeks.
> We are aware of these issues, yes I knew you were aware there were some missing, I wouldn't have reported it if it were just some missing test files but I thought perhaps it worth highlighting that all the input-query and modules are missing in case they'd fallen through the cracks. > but have not been able to scare up the necessary resources to correct them. You know you could always have used the xq2xml ones, every release of XQTS since 0.6 has had a full set of schema valid xqx files in the xq2xml distribution. There have been (and may still be) bugs in them of course but they have existed. I don't mind of they are used or not, I don't have any personal axe to grind here but it's rather frustrating to be told that the WG hasn't the resources to make some files that in some cases I've been freely distributing with a W3C licence for almost a year. > These fixes will appear in a point release in a couple of weeks. Hmm we'll see. David
The missing XQuery files are being worked on.. there are several bugs that need to be fixed so that they can be included. As far as input queries, and modules, I wasn't aware of the need to generate XQueryX for them... will need to clarify this w/ Andrew/XQTTF.
> As far as input queries, and modules, I wasn't > aware of the need to generate XQueryX for them... That's why I raised the report! If XQueryX is to get out of CR you presumably need XQueryX implementations to (collectively) pass the test suite, but without these files, xqueryx implementations can not even attempt the relevant tests still less pass them.
Also note that a number of the currently excluded XQueryX are correctly generated. We excluded them because they failed our automatic testing (which includes validation, comparing outputs of XQueries before/after conversion, etc). I just reviewed the outstanding issues manually, so I know which ones are correct, andwe can include these correct XQueryX queries shortly. The reason the automatic testing failed are bugs in the validation tool, use of non-deterministic functions in the queries (such at time), etc..
(In reply to comment #4) > > As far as input queries, and modules, I wasn't > > aware of the need to generate XQueryX for them... > > That's why I raised the report! > > If XQueryX is to get out of CR you presumably need XQueryX implementations to > (collectively) pass the test suite, but without these files, xqueryx > implementations can not even attempt the relevant tests still less pass them. > I wasn't aware that was the goal. Is that so, Andrew?
(In reply to comment #6) > I wasn't aware that was the goal. Is that so, Andrew? Yes. Our goal has always been to have an XQueryX equivalent for every test case that is not a parse error. The test cases for modules and external variables each require more than one XQuery expression, and therefore require more than one XQueryX document.
Apologies for letting this age for so long. Several months ago I fixed some bugs in our XQueryX generation and added the XQueryX files to our CVS repository. I've just added XQueryX versions of the modules in our TestSources directory. I believe that we now generate XQueryX for all but 39 XQuery tests. Most of these are negative test cases where the XQuery contains a duplicate prolog declaration and our XQueryX translator refuses to generate an XQueryX file. Please close this bug if you agree with this resolution.
sorry, slip of the mouse! reclosing.....