29526 2016-03-09 23:27:22 +0000 [XQ31] Incompatible change to ExtensionExpr 2016-03-18 15:28:33 +0000 1 1 1 Unclassified XPath / XQuery / XSLT XQuery 3.1 Candidate Recommendation PC All RESOLVED FIXED P2 normal --- 1 mike jonathan.robie andrew_coleman jmdyck josh.spiegel public-qt-comments oldest_to_newest 125422 0 mike 2016-03-09 23:27:22 +0000 In XQuery 1.0 and 3.0, It was defined that the construct (# pragma #) {} raised a static error if the pragma was not recognized by the query processor. In XQuery 3.1 this now returns an empty sequence when the pragma is not recognized. Although we don't normally treat the abolition of a static error as an incompatibility, I'm worried that we have removed a useful capability here. This construct could be used to make it clear that a query relied on a vendor extension and that no fallback was available. It's a construct that was valid and had a well-defined meaning, and we have changed that meaning. Can we please think again? 125566 1 andrew_coleman 2016-03-18 15:07:47 +0000 At the meeting on 2016-03-22, the WG agreed to revert this to its previous behaviour. The grammar will also be reverted so that ExtensionExpr does use EnclosedExpr. 125567 2 jmdyck 2016-03-18 15:23:52 +0000 (... so that it does *not* use EnclosedExpr, actually.) 125568 3 josh.spiegel 2016-03-18 15:28:33 +0000 Apologies, looks like I got it wrong in the minutes: ACTION A-636-09: Michael Dyck to revert the grammar so that ExtensionExpr does use EnclosedExpr (see bug 29526). Should say *does not*