Topic - you know your stuff

new topic post reply	Development -> Solutions Developer -> you know your stuff
moderators: pjr tab

you know your stuff

Gynwer

Joined: 22-April-2007

Posts: 79

Location: Belgium

Posted: 5-September-2008 13:57

Menzo,
I said it before, you definitely know your XML (and related technologies). I got dojo working in exactly the way you described. But the XRL problem remains. Take these two definitions :

 <link>
  <name>framework-dojo</name>
  <ext>/framework-dojo</ext>
   <int>active:xrl-html+operator@ffcpl:/etc/XRLLinks.xml+template@ffcpl:/resources/framework_dojo.xml</int>
 </link>

 <link>
  <name>framework-dojo2</name>
  <ext>/framework-dojo2</ext>
  <int>ffcpl:/resources/framework_dojo.htm</int>
 </link>

Where both the htm-file and the xml-file contain the basic example from the Dojo manual :

<html xmlns:xrl="http://1060.org/xrl">
  <head>
    <title>Dojo: Hello World!</title>

    <!-- SECTION 1 -->
    <style type="text/css">
        @import "/dojo/dijit/themes/tundra/tundra.css";
        @import "/dojo/dojo/resources/dojo.css"
    </style>
    <script type="text/javascript" src="/dojo/dojo/dojo.js" 
      djConfig="parseOnLoad: true"></script>
    <!-- SECTION 2 -->
    <script type="text/javascript">
       // Load Dojo's code relating to the Button widget
       dojo.require("dijit.form.Button");
    </script>
  </head>

  <body class="tundra">
   <button dojoType="dijit.form.Button" id="helloButton">
    Hello World!
    <script type="dojo/method" event="onClick">
      alert('You pressed the button');
    </script>
   </button>
  </body>
</html>

Then the framework_dojo2 works, framework_dojo doesn't. On looking at the source you can see that it gets translated (for the failing example) to :

<html xmlns:xrl="http://1060.org/xrl"  >
  <head >
    <title >Dojo: Hello World!</title>

    <!-- SECTION 1 -->
    <style type="text/css"  >
        @import &#34;/dojo/dijit/themes/tundra/tundra.css&#34;;
        @import &#34;/dojo/dojo/resources/dojo.css&#34;
    </style>
    <script djConfig="parseOnLoad: true" src="/dojo/dojo/dojo.js" type="text/javascript"  />
    <!-- SECTION 2 -->
    <script type="text/javascript"  >
       // Load Dojo&#39;s code relating to the Button widget
       dojo.require(&#34;dijit.form.Button&#34;);
    </script>
  </head>

  <body class="tundra"  >
   <button dojoType="dijit.form.Button" id="helloButton"  >
    Hello World!
    <script event="onClick" type="dojo/method"  >
      alert(&#39;You pressed the button&#39;);
    </script>
   </button>
  </body>
</html>

Anyway ... I'm making progress.
Thanks again Menzo !
Regards,
Tom

oops

Gynwer

Joined: 22-April-2007

Posts: 79

Location: Belgium

Posted: 5-September-2008 13:58

This should of course have been a reply to
http://www.1060.org/forum/topic/439/1
not a separate topic.

This javascript is getting to me. Time for the weekend.
Sorry,
Tom

menzo

Joined: 14-March-2005

Posts: 60

Location: Amsterdam, The Netherlands

Posted: 5-September-2008 14:41

Is this escaping some feature of the xrl-html accessor, does it also happen when you use just the xrl accessor?

only when casting

Gynwer

Joined: 22-April-2007

Posts: 79

Location: Belgium

Posted: 8-September-2008 08:18

Hi Menzo,

No, it doesn't happen when I use the following ...

 <link>
  <name>framework-dojo</name>
  <ext>/framework-dojo</ext>
   <int>active:xrl+operator@ffcpl:/etc/XRLLinks.xml+template@ffcpl:/resources/framework_dojo.xml</int>
 </link>

This gives me a correct XML page. However, any way I cast that page it insists on the encoding ...

 <link>
  <name>framework-dojo</name>
  <ext>/framework-dojo</ext>
   <int>active:cast+operator@xrl:cast.xml+operand@active:xrl%2Boperator@ffcpl:/etc/XRLLinks.xml%2Btemplate@ffcpl:/resources/framework_dojo.xml</int>
 </link>

With the following in cast.xml

<cast>
  <mimetype>text/html</mimetype>
</cast>

And I'm once again stuck with the encoding problem ...
Strange ... I can't remember wrestling with encoding that much before ...

Regards,
Tom

menzo

Joined: 14-March-2005

Posts: 60

Location: Amsterdam, The Netherlands

Posted: 8-September-2008 13:26

Hi Tom,

If I understand correctly you want to do something like:

<script type="text/javascript" xrl:resolve="text">
/* <![CDATA[ */
... code goes here ...
<img src="xrl:test"/>
... code goes here ...
/* ]]> */
</script>

But the serializer won't keep the CDATA sections intact, and it doesn't have to. So I assume you end up with something like this as a result:

<script type="text/javascript">
/*  */
... code goes here ...
&lt;img src="http://www.example.com/test.gif"/&gt;
... code goes here ...
/*  */
</script>

I think you need to post process the XRL result and at that point add the JavaScript comments. So your template could be:

<script type="text/javascript" xrl:resolve="text">
... code goes here ...
<img src="xrl:test"/>
... code goes here ...
</script>

Which after XRL processing will be:

<script type="text/javascript">
... code goes here ...
&lt;img src="http://www.example.com/test.gif"/&gt;
... code goes here ...
</script>

Now you can run an XSLT stylesheet over this to hide the scripts from older browsers:

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="2.0">
   <xsl:template match="script">
      <xsl:copy>
         <xsl:comment>
            <xsl:value-of select="."/>
            <xsl:text>//</xsl:text>
         </xsl:comment>
      </xsl:copy>
   </xsl:template>
</xsl:stylesheet>

Which results in:

<script><!--
... code goes here ...
<img src="http://www.example.com/test.gif"/>
... code goes here ...
//--></script>

Maybe this helps ...

Menzo

pipe through xslt

Gynwer

Joined: 22-April-2007

Posts: 79

Location: Belgium

Posted: 9-September-2008 09:18

Hi Menzo,

Basically all special characters in my javascript were translated to their number values.
You did put me on the right track though. When I pipe the output of the xrl-html through an xsl transformation, everything reverts back.
So ... in my XRLlinks.xml that looks like ...

<int>active:xslt+operator@ffcpl:/resources/mime.xsl+operand@active:xrl-html@ffcpl:/etc/XRLLinks.xml%2Btemplate@ffcpl:/resources/mastertemplate_dojo.xml%2Bcontent@xrl:empty_content.xml</int>

And in mime.xsl I have :

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
  <xsl:output method="html" omit-xml-declaration="yes" indent="no" />
  <xsl:template match="@*|node()">
    <xsl:copy>
      <xsl:apply-templates select="@*|node()" />
    </xsl:copy>
  </xsl:template>
</xsl:stylesheet>

And presto, a correct page. Even better, I don't have to hide my javascript between any sort of tags anymore (although that's still better for old browsers of course).

Now, solution is there ... what I still do not understand however is the why. Why would an xrl-html start translating all my special characters ?

Regards,
Tom

menzo

Joined: 14-March-2005

Posts: 60

Location: Amsterdam, The Netherlands

Posted: 9-September-2008 09:32

That's indeed a riddle. I ran around in the XRLLinker.java a bit, and I can see it sets the correct MIME type when you use xrl-html or xrl-html-tolerant ... but I can't see it doing anything to escape 'special' characters. Probably it does it in the serialization phase somewhere ...

found it ...

menzo

Joined: 14-March-2005

Posts: 60

Location: Amsterdam, The Netherlands

Posted: 9-September-2008 09:48

It indeed happens during serialization. XRL returns a DOMXDA, when that gets serialized using DOMXDASerializer the following code is executed:

         if (isHTML)
         {   sw.write("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n");
            if (dt!=null)
            {   writeDocumentType(sw,dt);
            }
            SerializedDOMAspect.serialize(n,sw);
         }
         else
         {   if (dt!=null)
            {   writeDocumentType(sw,dt);
            }
            XMLUtils.getInstance().toXML(sw, n, false,false);
         }

So in case of a HTML mime type SerializedDOMAspect.serialize() is called. And this is the class that takes care of all HTML specifics. And for content it does the following:

         case Node.TEXT_NODE:
         case Node.CDATA_SECTION_NODE:
            writeEscaped(aWriter, aNode.getNodeValue());
            break;

And the writeEscaped method does the following:

   public static void writeEscaped(Writer aWriter, String aInput) throws IOException
   {   char[] input = aInput.toCharArray();
      int length=input.length;
      for (int i=0; i<length; i++)
      {   char c=input[i];
         switch(c)
         {   case '<':
               aWriter.write("&#60;");
               break;
            case '>':
               aWriter.write("&#62;");
               break;
            case '\'':
               aWriter.write("&#39;");
               break;
            case '"':
               aWriter.write("&#34;");
               break;
            case '&':
               aWriter.write("&#38;");
               break;
            default:
               aWriter.write(c);
               break;
         }
      }
      
   }

I think the escaping is a bit too eager there. Single and double quotes only need to be escaped inside attribute values, not inside regular content. Isn't it?

It's still a bit curious why browsers fail to execute the scripts with the encoded quotes, as I assume they would just resolve all entity references before interpreting the script ...

Menzo

new topic post reply

To find out about new replies to this post as they occur
please subscribe to one of these feeds:

moderate