- From: <bugzilla@wiggum.w3.org>
- Date: Sun, 06 Aug 2006 08:59:40 +0000
- To: public-qt-comments@w3.org
- CC:
http://www.w3.org/Bugs/Public/show_bug.cgi?id=2794 martin@x-hive.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |martin@x-hive.com ------- Comment #1 from martin@x-hive.com 2006-08-06 08:59 ------- Is there any way to weigh in on this issue? If I'm not misreading it, the current spec as of 2006-07-11 disallows expressions to return both values and pending update lists (by disallowing expressions that have updating and non-updating children). I'm not entirely sure about that, as the wording is pretty strange: 1. If any branch contains an updating expression, the conditional expression is an updating expression; otherwise it is a non-updating expression. ... 3.If the conditional expression is an updating expression, then both of its branches must contain either an updating expression, an empty expression ( ), or a call to the fn:error function; otherwise a static error is raised [err:XUST0001]. In this case, the test expression is evaluated and the then or else clause is selected and evaluated as specified in Section 3.10 of [XQuery 1.0]. The result of the conditional expression is the pending update list returned by the selected branch. This seems like a very complicated way of saying "If one of the branches of a conditional expression is an updating expression, the other has to be one too". I think Mary Holsteges comment in bug #2796 is very valid, disallowing this will forbid so many use cases that it renders the whole thing entirely useless.
Received on Sunday, 6 August 2006 08:59:43 UTC