Thank you Dr. Kay for the explanation.
This provided answers to most questions I have, however there is still
this question:
What default namespace binding will be used in the second case (when
the namespace-context attribute is absent), if there is no
[xsl:]xpath-default-namespace attribute specified anywhere in the
stylesheet? The specification doesn't define this at all.
So, there is no [xsl:]xpath-default-namespace attribute anywhere in
the stylesheet.
There are two sub-cases here:
1. The <xsl:evaluate> element has no default-namespace binding.
2. The <xsl:evaluate> element has a default-namespace binding
I believe that in case 2, the default-namespace binding of the
<xsl:evaluate> instruction should become the the default-namespace
binding in the evaluation context (the alternative is to ignore this
and nave no default-namespace binding for the evaluation case -- as
obviously should be done in case 1.)
Is this conclusion right, or not? This isn't postulated specifically
in the specification.
Cheers,
Dimitre Novatchev
On Fri, Feb 13, 2015 at 3:44 AM, Michael Kay <mike(_at_)saxonica(_dot_)com>
wrote:
My question is why in the second case described in this section (when
the namespace-context attribute is absent), the binding for the
default-namespace is derived in a different way, than in the first
case?
To make it consistent with "static" XPath expressions embedded in the
stylesheet.
For example
<xsl:evaluate xpath="'doc/section/para[' || $predicate || ']'"/>
handles the default namespace in the same way as
<xsl:sequence select="doc/section/para"/>
What will happen if the namespace-context attribute is *present* and
it selects the <xsl:evaluate> instruction? Will then the rule for
deriving the default-namespace binding of the first sub-bullet be
used, or will the rule of the 2nd sub-bullet be used? As per the
current text it seems that the rule in the first sub-bullet must be
used. But this is exactly the same node as in the 2nd sub-bullet and
despite this identity, different rules for the default namespace
binding must be used. Was this really intended?
Actually, a stylesheet cannot select its own instructions. It can only select
a copy, by having a stylesheet that reads itself as a source document. When
stylesheets are read as source documents, they are never treated specially,
they always follow the same rules as other source documents - for example
they follow the rules for whitespace stripping in source documents, not the
rules for whitespace stripping in stylesheets. The direct answer to your
question is that stylesheets reading their own source code is not something
we considered as an important use case.
The typical use case for taking the namespace context from a node is where
the XPath expression itself is embedded in a source document (in a
spreadsheet-formula-like way), for example
<interest-rates>
<base-rate>0.02</base-rate>
<savings-rate>../base-rate + 0.03</savings-rate>
</interest-rates>
This might be executed by
<xsl:variable name="savings-rate">
<xsl:variable name="source" select="doc('interest-rates')//savings-rate">
<xsl:evaluate xpath="." namespace-context="."/>
</xsl:variable>
It seems appropriate here to take the default namespace for elements and
types from the XML default namespace.
Michael Kay
Saxonica
--~----------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
EasyUnsubscribe: http://lists.mulberrytech.com/unsub/xsl-list/1167547
or by email: xsl-list-unsub(_at_)lists(_dot_)mulberrytech(_dot_)com
--~--