5703 2008-05-20 21:27:19 +0000 extra space in the result of functx-fn-namespace-uri-for-prefix-all 2008-06-24 15:11:03 +0000 1 1 1 Unclassified XML Query Test Suite XML Query Test Suite unspecified PC Windows XP CLOSED FIXED P2 normal --- 1 andrew.eisenberg frans.englich public-qt-comments oldest_to_newest 20185 0 andrew.eisenberg 2008-05-20 21:27:19 +0000 The expected result of this test case is " http://datypic.com/pre http://datypic.com/unpre http://datypic.com/pre". The leading space in this result would be justified if namespace-uri-for-prefix('', $in-xml) returned a zero-length string ($in-xml has no default namespace). Section 11.2.5, fn:namespace-uri-for-prefix, in F&O, says: "If $prefix is the zero-length string or the empty sequence, it returns the namespace URI of the default (unnamed) namespace. Otherwise, it returns the empty sequence." I think that one could interpret this as requiring that an empty-sequence be returned, rather than a zero-length string. This interpretation would remove the leading space from the expected result. Am I missing something? Should this be brought to the WGs? 20187 1 mike 2008-05-20 21:45:38 +0000 I agree that the spec could be clarified here. However, I think that if there is no default namespace, namespace-uri-for-prefix("", $e) should return "" as the expected results suggest. This is consistent with the results of other functions that return the pseudo-URI of the non-namespace, for example namespace-uri($e) and namespace-uri-from-QName("", "fred") both return "". 20197 2 frans.englich 2008-05-21 08:15:55 +0000 Yes, I think #1 sounds sensible. I think that's verified by other tests currently, as well. Maybe close this report, and open an editorial one on the spec? 20805 3 frans.englich 2008-06-24 15:10:52 +0000 Ok, closing the report on the test suite. Feel free to open up on the spec.