PsychoPhil

As you may have noticed by looking at the category lists over on the right (yeah I know, like anybody looks over there), I’ve finished importing all the posts from the geeklog version of this site over to WordPress. Given that GeekLog isn’t exactly a mainstream blog and WordPress doesn’t have an importer for GeekLog it was pretty much a major pain in the ass. If you’ve come across this entry looking for instructions on how to do this migration, I’m sorry to tell you there is no easy way to do it.

As I stated previously, I was able to find and install an rss2 class for GeekLog. I’m not sure why its not included in the main GeekLog install but being able to generate a ‘modern’ rss2 feed (so it would include basic information like say, time a date of the original post) was a great help. I did make a change to the rss2 feed class. The geeklog feed was putting the story contents in a < description > tag. I found that the WordPress importer worked better if it found story conent in a < content :encoded > tag. Easy enough to switch.

I also managed to find and fix two bugs in the rss-import.php file in WordPress. They are now listed on the WordPress Mantis bug tracking site as issues 0000878 and 0000879. Bug one was that the WordPress import did not strip the CDATA tag from $title on import. The second is that the import switched the value for post_date and post_date_gmt when inserting the imported story into the database. This means that stories had the GMT as their post date instead of EST. The only other change I made to the WordPress importer was to change the category number so all posts were imported into the ‘geeklog’ catagory instead of the default ‘weblog’.

After running a few import tests with approx 10 stories, everything looked to be working properly so I fired off an import of the entire feed. Thats when I remembered that GeekLog allowed you to choose to enter a post as either plain text or HTML. If the WordPress importer found any HTML in an entry (like a URL) it would strip ALL formatting from an entry. I had already planned to have to go back through and place images back into a few stories by hand. Because of the HTML stripping I also had to reformat a large number of stories by hand as well.

There where a limited number of comment threads I wanted to carry over as well (mostly on some of the political posts). Those where simply cut and paste into the stories. As a result, all the older comments are listed as being entered by ‘administrator’ but the original authors ID (if there was one) is included in the text of the comment.

7 Responses to “Geeklog Import”