- From: Abel Braaksma <abel.braaksma@xs4all.nl>
- Date: Thu, 29 Sep 2016 15:20:15 +0200
- To: "Public XSLWG" <public-xsl-wg@w3.org>
- Message-ID: <280901d21a54$389ffdb0$a9dff910$@xs4all.nl>
Look good. I would like to suggest an additional column, though. If the initial-mode is not set, it defaults to the default-mode attribute on the xsl:package or xsl:stylesheet element. While a processor could override this by defining that the "default default" is the unnamed mode, if such a default is absent, or, better yet, if we define it simply as "when no initial-mode is set, the default-mode attribute of the outermost element of the top-level package", we have this case covered. You write: " or (b) mode M is implicitly declared by virtue of the presence of an xsl:template with mode="M"." But isn't the default of undeclared constructs that the visibility is private? Or is this not true for modes, in that they are implicitly public (or: if not public, at least invokable as initial-mode)? I would understand the case for the latter, for backwards-compatibility, but I am wondering how this relates. In the attachment is a copy of an old attempt of all possible scenarios for invocation with an initial-mode. Primarily to illustrate the numerous combinations that are possible. I will have to check if it is still in line with what we currently know about XTDE0045. Cheers, Abel From: Michael Kay [mailto:mike@saxonica.com] Sent: Wednesday, September 28, 2016 8:47 PM To: Public XSLWG Subject: ACTION 2016-09-22-003 - bug 29827 - initial mode and error XTDE0045 ACTION 2016-09-22-003 MK: try to express the rules in the form of a table. Abel phrases it in terms that if declared-modes="yes" then the mode "doesn't exist" unless it is declared and it then becomes an error to refer to it. (This relates to bug 29827) I have attempted a table giving the proposed rules at attachment 1656 Michael Kay Saxonica
Attachments
- text/plain attachment: Notes_on_initiate_stylesheet.txt
Received on Thursday, 29 September 2016 13:20:54 UTC