The relase notes of XXE 10.4 [1] metntions this bug fix:
«Made the language fixup of the XInclude 1.1 implementation more
conforming to the specification.»
The new behavior is related to "language fixup" as defined by XInclude
1.0 and 1.1 [2]. And the new behavior is that when the language fixup
algoritm concludes that the language of the embedded element is unknown,
the empty attribute xml:lang="" is applied to the included element.
But I wonder if the implementation is too liberal in applying
xml:lang="", possibly due to unclear language of the XInclude spec.
It seems to me as if XXE 10.4 treats <html> and <html xml:lang=""> as
equal. Let us say tha we want to embed a <div> that lives inside a
<html> element. I will show two variants of this. In the first case, the
<html> element has no xml:lang. In the second varaint, there is an
xml:lang="".
Variant 1 : no xml:lang attribute.
<html http://www.w3.org/1999/xhtml"; >
<div id="embed-me">Hello world!</div>
</html>
Currently, XXE 10.4 will include the div element like so, "because" it
considers the language unknown:
<div id="embed-me" xml:lang="" >Hello world!</div>
Variant 2 : with xml:lang="<empty>"
<html http://www.w3.org/1999/xhtml"; xml:lang="" >
<div id="embed-me">Hello world!</div>
</html>
This is simple. XXE 10.4 will simply copy - or inherit - the xml:lang
attribute from the <html> element:
<div id="embed-me" xml:lang="" >Hello world!</div>
Just to be explicit: The behavior in Variant 2, is correct. But I am
honestly uncertain about Variant 1.
XML 1.0 says this - https://www.w3.org/TR/xml/#sec-lang-tag :
«The language specified by xml:lang applies to the element where it is
specified (including the values of its attributes), and to all elements
in its content unless overridden with another instance of xml:lang. In
particular, the empty value of xml:lang is used on an element B to
override a specification of xml:lang on an enclosing element A, without
specifying another language. Within B, it is considered that there is no
language information available, just as if xml:lang had not been
specified on B or any of its ancestors.»
So based on XML 1.0, it does seems as if <html> and <html xml:lang="">
are considered equal.
There is still a practical difference, though, which is that <html> may
inherit a language from a process higher up - such as from HTTP. Whereas
<html xml:lang=""> will not inherit anything from above.
A practical benefit of the new behavior, is that it becomes possible to
discover included/embedded elements for which there is no language
specified. This can proably come in as handy.
But there are situations where one wants the included element to inherit
the language from the parent element in the composed document (just as
one may want HTML to inherit the language from HTTP).
XML 1.0 and 1.1 says that «An XInclude processor may, at user option,
suppress xml:base and/or xml:lang fixup.» [3]
And so, because of the practical difference, it might make sense to be
able to disable langage fixup - at leaset with regard to the application
of xml:lang="" in the Variant 1 situation described above.
[1] https://xmlmind.com/xmleditor/changes.html#v10.4.0
[2] https://www.w3.org/TR/xinclude-11/#language
[3] https://www.w3.org/TR/xinclude-11/#creating-result
Leif Halvard Silli
--
XMLmind XML Editor Support List
xmleditor-support@xmlmind.com
http://www.xmlmind.com/mailman/listinfo/xmleditor-support
