W3C home > Mailing lists > Public > public-qt-comments@w3.org > February 2016

[Bug 29479] [XSLT30] Streaming and non-well-formed documents

From: <bugzilla@jessica.w3.org>
Date: Fri, 19 Feb 2016 17:44:04 +0000
To: public-qt-comments@w3.org
Message-ID: <bug-29479-523-7RVX3UlQjL@http.www.w3.org/Bugs/Public/>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=29479

--- Comment #2 from Abel Braaksma <abel.braaksma@xs4all.nl> ---
> If the user wants a complete well-formedness or validity check then they can 
> always add one to the processing pipeline.

I'm not sure how this can be done, unless by forcing the processor to go over
every node, but that will hardly ever be needed and if no action is going on on
a node (say, it is simply dismissed) then there's still the issue of closing
the root element, following your logic it may never be needed to find that
close tag, making it hard to detect an error in a document like:

<foo>
  <bar />
  <bar />
<!-- missing end </foo> -->

> Why? To shallow-copy an element you only need to know the name of the element.
I agree in principle, but isn't it also true that XSLT instructions close over
(if that's the term) their beginning and end tags?

My worry is mainly with the large processor-dependent behavior in this area, as
Martin Honnen's example shows. If a processor looks ahead one item, it may
fail, if another is not looking ahead, it may succeed, leading to
processor-independent behavior.

I'm not sure if there's an easy way to fix this, but if there is, I think it is
worthwhile to have a go at it.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
Received on Friday, 19 February 2016 17:44:07 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:57:59 UTC