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 28335 - [XT30] test match-82a has wrong XSLT version and assumes different priority for parenthesized patterns
Summary: [XT30] test match-82a has wrong XSLT version and assumes different priority f...
Status: CLOSED FIXED
Alias: None
Product: XPath / XQuery / XSLT
Classification: Unclassified
Component: XSLT 3.0 Test Suite (show other bugs)
Version: Last Call drafts
Hardware: PC Windows NT
: P2 normal
Target Milestone: ---
Assignee: Abel Braaksma
QA Contact: Mailing list for public feedback on specs from XSL and XML Query WGs
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-03-25 18:05 UTC by Abel Braaksma
Modified: 2015-05-06 21:19 UTC (History)
1 user (show)

See Also:

Attachments

Description Abel Braaksma 2015-03-25 18:05:15 UTC 
According to the spec, parentheses are removed prior to assessing for default priority (internal current draft), this test assumes that the priority is different.

Also, the test is written for 3.0, but the stylesheet has version 2.0.
Comment 1 Michael Kay 2015-03-25 19:19:56 UTC 
I think the test match-082 was originally written for XSLT 2.0, and was then forked because a 3.0 processor gives a different result. It's perfectly legal to run a stylesheet with version="2.0" with a 3.0 processor, and the results are well-defined.

You're right that at some point the rules changed so that match="(doc|cod)" is now treated as a UnionPattern for priority purposes, which means the second template rule should be chosen in preference. I propose we simply change the expected results.
Comment 2 Michael Kay 2015-03-25 22:14:15 UTC 
Fixed the second problem.
Comment 3 Abel Braaksma 2015-03-26 12:16:57 UTC 
We were both fixing the same test, but differently, I have merged the changes.
Comment 4 Abel Braaksma 2015-03-26 12:44:18 UTC 
> It's perfectly legal to run a stylesheet with version="2.0" with a 3.0 
> processor, and the results are well-defined.

Yes, but if it tests backwards-compatibility, it should probably be placed in a different category. I split the test-case in three variants to cover the same-priority situation more clearly.
Comment 5 Abel Braaksma 2015-05-06 21:19:02 UTC 
Was resolved > 30 days ago, closing.