29597 2016-04-26 22:00:15 +0000 [QT3] modules-pub-priv-24 2016-07-19 16:23:18 +0000 1 1 1 Unclassified XPath / XQuery / XSLT XQuery 3 & XPath 3 Test Suite Working drafts PC Linux RESOLVED FIXED P2 normal --- 1 benito oneil abel.braaksma public-qt-comments oldest_to_newest 126175 0 benito 2016-04-26 22:00:15 +0000 In #29589 I wrote local:test in modules-collide-var-001 could raise XPST0017, which is ofc nonsense. It was supposed to read: modules-pub-priv-24 (!) has <a>{local:test()}</a>, but there is no local:test() function. So it could throw XPST0017 instead the expected error 126200 1 benito 2016-04-27 16:08:22 +0000 Now I am starting to wonder: When an variable initialization expression has an (static or dynamic) error, but the variable is never used, is it still an error of the entire query? It might get evaluated lazily 126400 2 abel.braaksma 2016-05-10 14:09:15 +0000 (In reply to Benito van der Zander from comment #1) > but the variable is never used, is it still an > error of the entire query? It might get evaluated lazily Section "2.3.4 Errors and Optimization" of XP31 explains this. If the variable is never called, it need not be evaluated. In such cases it is implementation dependent whether the dynamic error is raised (i.e., with eager variable evaluation, the error may always be raised, with lazy evaluation only when actually evaluated). 126814 3 abel.braaksma 2016-06-21 16:08:01 +0000 > It might get evaluated lazily Note that my comment #2 applies to *dynamic errors*, static errors must always be reported. 126815 4 mike 2016-06-21 16:08:52 +0000 Decision after WG discussion: (a) there are two static errors in the test, we should eliminate the call on local:test() so there is only one error. (b) it then becomes mandatory to report the error. Static errors must be reported whether or not the code is evaluated. 126986 5 oneil 2016-07-19 16:23:18 +0000 Closing this bug as the fix has been applied in the test case.