[XSL-LIST Mailing List Archive Home] [By Thread] [By Date]

[xsl] Re: Unwanted namespace prefix _0


Subject: [xsl] Re: Unwanted namespace prefix _0
From: Martin Holmes <mholmes@xxxxxxx>
Date: Mon, 23 Dec 2013 12:11:41 -0800

On 13-12-23 12:00 PM, Graydon wrote:
On Mon, Dec 23, 2013 at 11:01:55AM -0800, Martin Holmes scripsit:
Here are some test files:

<http://web.uvic.ca/~lancenrd/test/test.xml>
<http://web.uvic.ca/~lancenrd/test/test.xsl>

I was doing the transforms in Oxygen, using both a Transformation
Scenario and using the XSLT debugger. But when I run this on the
command line, I don't have the problem.

<egXML xmlns="http://www.tei-c.org/ns/Examples"


in the data -- egXML the element is in this namespace, but the namespace
doesn't have an associated prefix.

xmlns:eg="http://www.tei-c.org/ns/Examples"

on the stylesheet; this namespace has a prefix, locally "eg".

My original XSLT did not have this. I added it while testing to see if it would help. I get the same results with or without it.



The output doesn't have a prefix defined for that namespace -- you've excluded all the result prefixes! -- but it knows the namespace has a prefix, so you get _0.

But even if the namespace doesn't have a prefix, and even if it's not excluded, I get the same results.


If you put the prefix on the namespace in the data, the _0 goes away
because you're not feeding the prefix/no prefix dichotomy to the
processor and confusing it.

I don't get to control the data, unfortunately; it's the TEI source code. But the input data is valid, and it's perfectly reasonable to declare a namespace on an element without a prefix, isn't it? There's something odd happening here, and it's especially odd because it only happens in the context of Oxygen.


Cheers,
Martin


-- Graydon


Current Thread
Keywords