Re: DOCBOOK-APPS: Sorting and non-en_US indexes

[Togan Muftuoglu <toganm at users dot sourceforge dot net>]

* Jirka Kosek; <jirka@kosek.cz> on 24 Sep, 2002 wrote:
> default (usually English) I must specify language. E.g. if I want to
> sort something using Czech collating, I must specify it:
>
> <xsl:sort select="..." lang="cs"/>
Ok so far means I got it correct, however the output is not sorted
according to the Language attribute. I must be doing something wrong ?

Here are the things I have done :

1) Create Compare_tr class  
package com.icl.saxon.sort;

import java.text.Collator;
import java.util.Locale;

// Copyright © International Computers Limited 1998
// See conditions of use

/**
* A Comparer used with lang="tr".
*
* author Jirka Kosek (jirka@kosek.cz)
* <p>
* First seen in docbook-apps mailinglist
*
* 
* Modified for use with Turkish used with lang="tr"
*
* @author Togan Muftuoglu (toganm@users.sourceforge.net>
*
*/
public class Compare_tr extends TextComparer {

   int caseOrder = UPPERCASE_FIRST;

   /**
   * Compare two string objects, in three phases:
   * (a) ignoring accents and case
   * (b) if still equal, ignoring case
   * (c) if still equal, taking case into account
   * @return <0 if a<b, 0 if a=b, >0 if a>b
   */

   public int compare(Object a, Object b) {

       Collator trCollator = Collator.getInstance(new Locale("tr",
"TR"));

       return trCollator.compare(a, b);

   }

   public Comparer setCaseOrder(int caseOrder) {
       this.caseOrder = caseOrder;
       return this;
   }

}

2) Modified autoidx.xsl as follows when there was <xsl:sort.. > I added
the lang="tr" 

<?xml version="1.0"?>
<!DOCTYPE xsl:stylesheet [

<!ENTITY lowercase "'abc&#231;defg&#287;h&#305;ijklmno&#246;prs&#351;tu&#252;vyz'">
<!ENTITY uppercase "'ABC&#199;DEFG&#286;HI&#304;JKLMNO&#214;PRS&#350;TU&#220;VYZ'">

<!ENTITY primary   'normalize-space(concat(primary/@sortas, primary[not(@sortas)]))'>
<!ENTITY secondary 'normalize-space(concat(secondary/@sortas, secondary[not(@sortas)]))'>
<!ENTITY tertiary  'normalize-space(concat(tertiary/@sortas, tertiary[not(@sortas)]))'>

<!ENTITY sep '" "'>
]>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform";
               xmlns:fo="http://www.w3.org/1999/XSL/Format";
		version="1.0">


<!-- ==================================================================== -->
<!-- Derived from Jeni Tennison's work in the HTML case -->
<xsl:key name="letter"
        match="indexterm"
        use="translate(substring(&primary;, 1, 1),&lowercase;,&uppercase;)"/>

<xsl:key name="primary"
        match="indexterm"
        use="&primary;"/>

<xsl:key name="secondary"
        match="indexterm"
        use="concat(&primary;, &sep;, &secondary;)"/>

<xsl:key name="tertiary"
        match="indexterm"
        use="concat(&primary;, &sep;, &secondary;, &sep;, &tertiary;)"/>

<xsl:key name="see-also"
        match="indexterm[seealso]"
        use="concat(&primary;, &sep;, &secondary;, &sep;, &tertiary;, &sep;, seealso)"/>

<xsl:key name="see"
        match="indexterm[see]"
        use="concat(&primary;, &sep;, &secondary;, &sep;, &tertiary;, &sep;, see)"/>

<xsl:template name="generate-index">
 <xsl:variable name="terms" select="//indexterm[count(.|key('letter',
                                    translate(substring(&primary;, 1, 1),&lowercase;,&uppercase;))[1]) = 1]"/>
 <xsl:variable name="alphabetical"
               select="$terms[contains(concat(&lowercase;, &uppercase;),
                                       substring(&primary;, 1, 1))]"/>
 <xsl:variable name="others" select="$terms[not(contains(concat(&lowercase;,
                                                &uppercase;),
                                            substring(&primary;, 1, 1)))]"/>
 <fo:block>
   <xsl:if test="$others">
     <fo:block font-size="16pt"
               font-weight="bold"
               keep-with-next.within-column="always"
               space-before="1em">
       <xsl:call-template name="gentext">
         <xsl:with-param name="key" select="'index symbols'"/>
       </xsl:call-template>
     </fo:block>
     <fo:block>
       <xsl:apply-templates select="$others[count(.|key('primary',
                                    &primary;)[1]) = 1]"
                            mode="index-symbol-div">
         <xsl:sort select="&primary;" lang="tr" />
       </xsl:apply-templates>
     </fo:block>
..........
..........
</xsl:stylesheet>


3) I have imported the autoidx_tr.xsl into my customization


<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform";
               xmlns:fo="http://www.w3.org/1999/XSL/Format";
		version="1.0">

<xsl:import href="http://docbook.sourceforge.net/release/xsl/snapshot/fo/docbook.xsl"/>
<xsl:import href="/home/toganm/sgml/xsl/autoidx_tr.xsl"/>
.....
</xsl:stylesheet>

4) LANG=tr ~/sbin/parser.sh indexturkce3.fo indexterm.002.xml \
~/sgml/xsl/fop-common.xsl >error.txt 2>&1

where parser.sh includes all the JAR paths including Compare_tr location
along with Catalog reseolvers  ( I use Saxon 6.5.2 with SunJava version "1.3.1_02"
Under Linux I normally have tr_TR-UTF-8 locale I have tried width the
LANG=tr and LANG=tr_TR-UTF-8 
5) When I do grep Compare error.txt tehreis no sign of Compare_tr being
loaded

[Loaded java.io.ObjectStreamClass$CompareClassByName from
/usr/lib/SunJava2-1.3.1/jre/lib/rt.jar]
[Loaded java.io.ObjectStreamClass$CompareMemberByName from
/usr/lib/SunJava2-1.3.1/jre/lib/rt.jar]
[Loaded com.icl.saxon.sort.NodeOrderComparer]
[Loaded com.icl.saxon.sort.LocalOrderComparer]
[Loaded com.icl.saxon.sort.Comparer]
[Loaded com.icl.saxon.sort.TextComparer]
[Loaded com.icl.saxon.sort.StringComparer]
[Loaded com.icl.saxon.sort.Compare_en]
[Loaded com.icl.saxon.sort.DoubleComparer]
[Loaded com.icl.saxon.sort.DescendingComparer] 
6) Result is wrong order of the alphabet ? So what is that I am missing  
--

Togan Muftuoglu