Import Succesful!
Switch completed, Radio UserLand Archives up on Movable Type! Still some cleanup that is to be completed:

1. rss feed templates and validator link
2. set up of some editing environment hooking up with the blog api? (mozblog for firebird? eclipse plugin?)
3. change the blog-links in my email-sig and on our company website
4. do an honest effort to skin this over to our company style
5. anything I forgot?

PS: people interested in the sax-processor for the generation of the import-format for MT can drop me a line. It is quite likely t need some specific tuning, but it could be of help. # Posted by mpo at 02:34 PM | Comments (0) | TrackBack (0)

MT and images?
One jar-scp-ssh-unjar further I should be able to look at the images from the past. Hehe, when this one passed in the jar-output I knew which one to take for testing :-) # Posted by mpo at 11:10 AM | Comments (0) | TrackBack (0)

Radio UserLand Template
Hm, looks like the custom template I used on the previous blog (pls vote if you liked it so much that you would want me to get it over here as well) is the biggest reason why this import/export needs to be tweaked to personal needs anyway.

This seems to be my pattern to dig for (hm, SAX processing exercise!)

<div class="date">
  <a href="{RADIO_ROOT_URL}/{year}/{month}/{day}.html">
    <img ... /></a>&nbsp;{date in ddd dd mmmm yyyy}
</div>

<div class="newsItems">
  <a name="{entry_id}"></a>

  <table width="100%" cellpadding="1">
    <tr>
      <td valign="top">
        <br><b>{Title}</b>
{Body}
        <font class="small" size="-1" color="gray">
          {time in hh:mm:ss am/pm}&nbsp;&nbsp;
          <a href="{RADIO_ROOT_URL}/{year}/{month}/{day}.html#{entry_id}">
            <img .../></a>&nbsp;&nbsp;
        </font>
      </td>
      <td valign="top" align="right" nowrap></td>
    </tr>

  </table>
</div>

Observations:

1. The fun doesn't start until there is a //div[@class=newsItems]
2. In this diff one entry is a group of ./a and ./table
3. I could use the ./a/@name for some silly check, and to clear some buffers for the new entry.
4. The title is over at ./table/tr/td/b[1]
5. After that there is the Body up until some font[@class=small and @size="-1" and color="gray"
6. Which is holding the parts of the date inside its nested content and a/@href
7. This looks like Date stuff could be done with just some simple string manipulation (rather then convert to actual Date object)
8. I will need to hold onto the accumulated body untill the date is available since movable wants them in the different order. (Beware of the passed-event-syndrome in SAX :-))
9. I have to dig up some sample with an image to check up on the relative URI's there

# Posted by mpo at 10:47 AM | Comments (0) | TrackBack (0)

Moveable Type Input Format
{Actual Blog Entry HTML snippet} # Posted by mpo at 10:15 AM | Comments (0) | TrackBack (0)

How how, here comes the Neko show!

I didn't, right? But I should have, already a long time ago.

So here is my shameless plug concernig the CyberNeko tools from Andy Clark. Utterly great stuf.

To get my RadioUserLand-2-MT conversion done I need to parse the radio html files I exported, so I went on using cyberneko-html. I'm quite a dedicated JAXP and SAX addept. So finding out that I can use the Cyberneko-HTML-Parser just like any other JAXP/SAX thingy makes me a happy man. Just slide in the jar into your classpath up-front of the xerces jar and the magic of the META-INF services does the rest. Only now you get a SAXParser that cleans up the parsed HTML (as if it were welformed XML) by balancing out tags and the like. Yummy!

            SAXParserFactory spf = SAXParserFactory.newInstance();
            SAXParser sp = spf.newSAXParser();
            this.htmlReader = sp.getXMLReader();
            this.htmlReader.setContentHandler(this.defaultHandler);
            this.htmlReader.setErrorHandler(this.defaultHandler);

# Posted by mpo at 09:54 AM | Comments (0) | TrackBack (0)

Export from RadioLand
Here is how to get your old posts out of Radio Userland.

1. Get into radio Prefs | advanced | check the box for creating backups locally (and edit the path if needed)
2. Open Radio client and go for Radio | Publish | Entire WebSite
When it's finished you will find the files in the specified dir. # Posted by mpo at 12:31 PM | Comments (0) | TrackBack (0)

Doesn't work (Export Test)
Adding a second entry to see how this infects the export format. The python scripts I've found for migrating from radio to MT are not really helping (well, they are not running OTB, and when faced with the issue how to tweak them for my needs, I at least need to understand the issue at hand, the outputformat is one of them.) grmbl, why is stuff not working? # Posted by mpo at 12:26 PM | Comments (0) | TrackBack (0)