This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.

Bug 28418 - [XT30] error-0047a XTDE0057 has been removed from spec (possibly erroneously)
Summary: [XT30] error-0047a XTDE0057 has been removed from spec (possibly erroneously)
Status: CLOSED FIXED
Alias: None
Product: XPath / XQuery / XSLT
Classification: Unclassified
Component: XSLT 3.0 (show other bugs)
Version: Last Call drafts
Hardware: PC Windows NT
: P2 normal
Target Milestone: ---
Assignee: Michael Kay
QA Contact: Mailing list for public feedback on specs from XSL and XML Query WGs
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-04-07 04:15 UTC by Abel Braaksma
Modified: 2015-10-29 09:50 UTC (History)
0 users

See Also:


Attachments

Description Abel Braaksma 2015-04-07 04:15:41 UTC
This error was still there in the 2012 WD, but disappeared in the 2013 WD. In the 2012 WD there were two identical errors with different descriptions:

   [ERR XTDE0047] It is a non-recoverable dynamic error if the invocation of 
   the stylesheet specifies both an initial mode and an initial template.

   [ERR XTDE0047] It is a non-recoverable dynamic error if the invocation of 
   the stylesheet specifies both an initial mode and an initial template.

I could not find anything in the mail archive or in Bugzilla, so I am not sure whether this was done by accident or purposely. The only mail with a reasonable comment about these dynamic errors is in https://lists.w3.org/Archives/Member/w3c-xsl-wg/2012Oct/0038.html

> Most of the ones that aren't catchable are "stylesheet invocation 
> errors" like not supplying a required parameter; in principle we could 
> invent another category for these, but I don't think it's worth the 
> hassle; just treat them as dynamic errors which can't be caught.

It could be argued that the error is redundant, because it is up to the API to decide what happens here. However, the above comment implies that there was no intend to remove the error.

From the testing standpoint, I think that, as long as this error does not exist, the test should not exist. I will tentatively mark it XSLT20 only, but if the error comes back, we should change the test back as well.
Comment 1 Abel Braaksma 2015-04-07 04:16:17 UTC
(this bug was originally reported by Vitaliy Yudenkov)
Comment 2 Abel Braaksma 2015-04-07 07:35:08 UTC
> I will tentatively mark it XSLT20 only, but if the error comes back, we 
> should change the test back as well.

Currently fixed as described. Leaving the bug open though, because I think it warrants some discussion.
Comment 3 Abel Braaksma 2015-04-15 18:12:10 UTC
Moving the bug from XT30 to XSLT 30, so it can be discussed by the WG.
Comment 4 Michael Kay 2015-04-23 08:09:02 UTC
I think the error condition disappeared as part of the reorganization of the "stylesheet invocation" section.

I think it was right to remove it. It's an API error. The API might be designed so that there is no way of triggering this error, e.g

public void setInitialModeOrTemplate(QName name, boolean isMode)

or in the case of invocation from an XML-based test suite, the schema for the test suite metadata might prevent both an initial mode and an initial template being present (in which case the error you get is not XTDE0047, but a validation error on validating the test suite).

I think the test case should be dropped. Chances are, any product that passes this test probably catches this error in the test driver, not in the product under test.
Comment 5 Michael Kay 2015-04-23 17:03:14 UTC
Agreed to retain the status quo in the spec and drop the test case.