rec:id index is now of type "urx", which normalises case. (It

probably shouldn't, but it does.  And it probably shouldn't break on
spaces, but it does that, too.)  The proper answer is to make my own
index-type that does case normalisation and nothing else, but this
interrim solution should help Per.

diff --git a/zebra/zeerex2index.xsl b/zebra/zeerex2index.xsl
index 925a971..7b04304 100644 (file)
--- a/zebra/zeerex2index.xsl
+++ b/zebra/zeerex2index.xsl
@@ -1,5 +1,5 @@
 <?xml version="1.0" encoding="UTF-8"?>
-<!-- $Id: zeerex2index.xsl,v 1.10 2007-03-02 11:04:10 mike Exp $ -->
+<!-- $Id: zeerex2index.xsl,v 1.11 2007-03-02 14:27:53 mike Exp $ -->
 <!-- See the ZeeRex profile at http://srw.cheshire3.org/profiles/ZeeRex/ -->
 <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
                 xmlns:z="http://indexdata.dk/zebra/xslt/1"
@@ -37,7 +37,7 @@
     <xsl:value-of select="@authoritative"/>
    </z:index>
 
-   <z:index name="rec:id" type="0">
+   <z:index name="rec:id" type="u">
     <xsl:value-of select="$id"/>
    </z:index>