Carrot2

Well, I don't know if Opera supports quite fancy JavaScript that the 
default output component emits... I would be grateful if you could try 
it out -- go to:

carrot2-xslt-renderer/stylesheets/dtree.xsl

and alter the following code:

>         <script type="text/javascript" language="JavaScript">
>             <xsl:text disable-output-escaping="yes"><![CDATA[
>             // Load emulation layer for gecko.
>             if (navigator.family == "gecko")
>             {
>                 document.write('<link rel="stylesheet" type="text/css" href="]]></xsl:text><xsl:value-of select="concat($dtree.css.url,'-gecko.css')" />
>                 <xsl:text disable-output-escaping="yes"><![CDATA["></link>');
>                 document.write('<script type="text/javascript" language="JavaScript" src="]]></xsl:text><xsl:value-of select="concat($dtree.url,'/js/gecko-specific.js')" />
> 
>             <xsl:text disable-output-escaping="yes"><![CDATA["></sc' + 'ript>');
>             }
>             else
>             if (navigator.family == "ie4" && navigator.version >=5 && navigator.OS != "mac")
>             {
>                 document.write('<script type="text/javascript" language="JavaScript" src="]]></xsl:text><xsl:value-of select="concat($dtree.url,'/js/explorer-specific.js')" />
> 
>             <xsl:text disable-output-escaping="yes"><![CDATA["></sc' + 'ript>');
>             }
>             else
>             {
>             	alert("Browser type not supported: " + navigator.family + ", " + navigator.version);
>                 // Unsupported browser. Display the information.
>                 location.href = "]]></xsl:text><xsl:value-of select="concat($dtree.url, '/unsupported-browser.html')" /><xsl:text disable-output-escaping="yes"><![CDATA[";
>             }
>             ]]>
>             </xsl:text>
>         </script>

Try with either Gecko or IE - specific JavaScript code and tell me if it 
works with any of them. If it does, I will change browser-detection code 
and include Opera to the list of supported browsers.

I've also heard the 'detection' script does not properly detect Internet 
Explorer 5.5 -- this also needs to be fixed.

Thanks for any feedback on this.
Dawid

Asim Iqbal wrote:

> Hi,
> 
> I noticed Opera, 7.23 is not supported in carrot2. Is there any solution 
> for this ?.
> 
> Thanks.
> 
> _________________________________________________________________
> Lose those love handles! MSN Fitness shows you two moves to slim your 
> waist. 
> http://fitness.msn.com/articles/feeds/article.aspx?dept=exercise&article=et_pv_030104_lovehandles 
> 
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek
> For a limited time only, get FREE Ground shipping on all orders of $35
> or more. Hurry up and shop folks, this offer expires April 30th!
> http://www.thinkgeek.com/freeshipping/?cpg=12297
> _______________________________________________
> Carrot2-developers mailing list
> Carrot2-developers@...
> https://lists.sourceforge.net/lists/listinfo/carrot2-developers

Given enough time I was able to figure it out ;)

Inheriting from InputRequestProcess gives access to the methods in the 
AbstractRequestProcess superclass. This has a getServletConfig() method 
which returns a ServletConfig. A ServletConfig has in turn a 
getServletContext() method which does what I wanted.

So the solution, for anybody besides me who would ever need to do this 
kind of thing:
ServletContext context = getServletConfig().getServletContext();
within a method in InputRequestProcess subclass.

I'll stop rambling now :)

Sidenote: The Ant buildfile for the Carrot project has no Javadoc 
target. If anybody needs this documentation (as I did to figure out the 
above problem), add:
    <target name="javadoc" description="Make javadoc help files">
      <javadoc destdir="${javadoc.dir}" source="${basedir}"
             doctitle="${ant.project.name}" use="true"
             windowtitle="${ant.project.name}" author="true"
             version="true">
        <fileset dir="${basedir}" includes="**/*.java" />
      </javadoc>
     </target>
somewhere in the bottom of /build.xml before </project>.  Enjoy

/Laust

On 7/4-2004, at 12:02, Laust Rud Jensen wrote:

> Hi,
>
> Still trying to find my way around in the Carrot system.  I am 
> constructing a small search engine for Javadoc using LSI and I'm going 
> to use Carrot for result clustering.  Currently I'm trying to bridge 
> the gap between the two systems.
>
> I have a running search-service using a small number of JSP-pages 
> running within a Servlet (in Tomcat). The access data is through the 
> context for the servlet. This works pretty well as data is loaded once 
> when the webserver starts. All processing is going to take place on 
> the same Tomcat server for now.
>
> On the Carrot end of things I have a Servlet which is built similarly 
> to other Carrot input components. It is called javadoc-search and the 
> class JavadocSearch extends InputRequestProcessor.  This way I have 
> easy access to the search string and the wanted number of results, 
> plus everything is setup for easy output. Built using the Carrot build 
> process it maps to /carrot2-javadoc-search/.
>
> Problem:
> How do I, within the JavadocSearch.processQuery(query, count, output, 
> HttpServletRequest) method get the context for the current servlet so 
> I can access the shared data? Or is there some other way I can create 
> and access data easily between calls? The datafile is rather large (a 
> serialized class of  ~50MiB for the prototype), and should hopefully 
> only be loaded once when the webserver starts. The data will not 
> change so it can safely be accessed in parallel.
>
> Thanks for any help, ideas or pointers
>
> /Laust Rud