This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.

Bug 27711 - modules-none
Summary: modules-none
Status: CLOSED FIXED
Alias: None
Product: XPath / XQuery / XSLT
Classification: Unclassified
Component: XQuery 3 & XPath 3 Test Suite (show other bugs)
Version: Working drafts
Hardware: PC Linux
: P2 normal
Target Milestone: ---
Assignee: O'Neil Delpratt
QA Contact: Mailing list for public feedback on specs from XSL and XML Query WGs
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-12-28 10:17 UTC by Benito van der Zander
Modified: 2015-02-10 11:48 UTC (History)
1 user (show)

See Also:


Attachments

Description Benito van der Zander 2014-12-28 10:17:26 UTC
The test-case modules-none has a <module uri="http://www.w3.org/TestModules/none" file="ModuleImport/empty-lib.xq"/> dependency, to a module that does not exist.

Now the test-case should check the behaviour when a not existing module is encountered, but I do not think the not existing file of the not existing module should be mentioned in a <module/> element, since it is the test driver loading the modules of <module/> elements, not the test-case. And there is no point in testing if the test driver can handle not existing files.
Comment 1 Michael Kay 2014-12-28 17:28:51 UTC
I would have some sympathy with the argument were it not that the test was written 9 years ago, and therefore presumably many people have written test drivers that can handle it, and these test drivers might well have to change if it were done differently. Yes, the task of resolving module URIs is to some extent a matter for test drivers, and therefore handling of an absent module URI is, to some extent, testing the test driver rather than the product; but I think a little bit of fuzziness at the edges is something we have to accept as a fact of life. This test has been around for a long time and I don't think there's sufficient justification here for changing it.
Comment 2 Benito van der Zander 2014-12-28 20:26:31 UTC
> were it not that the test was written 9 years ago,


The test is so old, but the meta data about the module not.

That was changed for all tests with the new XQTS.

Back then the file was just mentioned in a global map that was not really linked to the test except for the description. The new metadata seems to say that a module with that namespace from that file is used in the test.
Comment 3 Michael Kay 2015-01-03 20:25:35 UTC
I vote for rejecting this bug. It's perfectly reasonable to have a test that attempts to import a non-existant module, and I can't think of any other way of writing the test. If it's the test driver that handles the error rather than the product-under-test, so be it.
Comment 4 Benito van der Zander 2015-01-03 20:45:33 UTC
>and I can't think of any other way of writing the test.

Remove <module uri="http://www.w3.org/TestModules/none" file="ModuleImport/empty-lib.xq"/> and keep the test itself unchanged?
Comment 5 O'Neil Delpratt 2015-02-10 11:47:45 UTC
The WG has decided to remove the test in question.