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 3001 - orderDecl-2 is non-deterministic
Summary: orderDecl-2 is non-deterministic
Status: CLOSED FIXED
Alias: None
Product: XML Query Test Suite
Classification: Unclassified
Component: XML Query Test Suite (show other bugs)
Version: 0.8.6
Hardware: PC Linux
: P2 normal
Target Milestone: ---
Assignee: Carmelo Montanez
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-03-15 10:52 UTC by John Snelson
Modified: 2006-04-06 15:00 UTC (History)
0 users

See Also:


Attachments

Description John Snelson 2006-03-15 10:52:02 UTC
Since XQuery specifies that "the relative order of nodes in distinct trees is stable but implementation-dependent" (section 2.4.1), the test "orderDecl-2" has non-deterministic results, and certainly cannot test the ordering declaration in the manner in which it is supposed to.

Ordering of the constructed elements used in this test should be implementation defined, but the test seems to expect some defined ordering.

There may be other tests that are incorrect in this manner, I have not done a full check of the test suite.
Comment 1 Carmelo Montanez 2006-03-16 15:33:59 UTC
John:

Thanks for the comment.  This issues was raised by another commentor and a second expected outcome was added to the catalog file.  Please close the
bug if in agreement.

Thanks,
Carmelo
Comment 2 John Snelson 2006-03-16 18:11:31 UTC
I am assuming that the purpose of this test is to exercise the order declaration, in which case this is an unsatisfactory solution. With a second expected outcome the order declaration is not tested.
Comment 3 John Snelson 2006-03-16 18:14:48 UTC
What about a test like this:

declare ordering ordered;
let $a := <a><b>1</b><c>2</c></a>
return $a/b union $a/c

which does test the order declaration.
Comment 4 Carmelo Montanez 2006-04-06 14:53:53 UTC
Hey John:

Sorry for the delay.  I think your suggestion is just fine.  I implemented it.
Please close the bug if in agreement and when able to verify.

Thanks,
Carmelo
Comment 5 John Snelson 2006-04-06 15:00:34 UTC
Thanks.