I recently had to migrate a bunch of pages from a custom CMS to a WordPress blog. Groovy to the rescue!

I used the wordpress java client library wordpress-java, but I had to checkout and build the latest version from SVN since the binary download didn’t include support for custom fields.  This method will connect to a WordPress blog, retrieve a template post, upload a couple of media files, then create a new post from the template that includes links to the uploaded files.  Remember to enable xmlrpc in your WordPress blog options before trying this!

def makePost(String title, String text, File image, File pdf){

    // your wordpress account details go here
    WordPress wp = new WordPress('username', 'password', 'http://example.com/blog/xmlrpc.php')

    // make a not of the post ID of the template post
    page templatePost = wp.getPost(3)

    // upload the media files - an image and a pdf
    // make sure to set the correct mime-types
    def uploadedImage = wp.newMediaObject('image/png', image, true)
    def uploadedPdf = wp.newMediaObject('application/pdf', pdf, true)

    // make html
    templatePost.setTitle(title)
    templatePost.setDescription(
"""I am using triple quotes so I can write a multi-line string.
link to the uploaded pdf
link to the uploaded image
""")

    String result = wp.newPost(templatePost, 'true')
    System.out.println("new post is at ${newPost.getPermaLink()}")
    }

Whenever I set up a new site, I want to make sure that example.com points to www.example.com. This is basic SEO – check out this link for a quick explanation. I know that this is trivial to do in Apache, but for the sake of simplicity I wanted to use Tomcat only as my web server.

After much searching around, it became clear that (1) Tomcat can’t do this and (2) I’m not the only person with this problem. The solution is a Java servlet filter called UrlRewriteFilter. It does the job perfectly, but it took me a while to get it set up right, so here’s a quick howto for anyone else in the same position.

The instructions on the project’s homepage will get you most of the way there. You unzip the contents of the binary download (a single .jar and a config file) into your context’s directory, then add a servlet mapping to make sure that UrlRewriteFilter gets to handle all requests. Make sure that the UrlRewriteFilter servlet mapping goes at the top of the web.xml file, to ensure that the filter gets to see HTTP requests before any other servlet.

The tricky part is getting the config file WEB-INF/urlrewrite.xml just right, so here’s the configuration that worked for me:

    
        seo redirect
        ^www.example.com
        ^/(.*)
        http://www.example.com/$1
    

The condition element tells us that we want to trigger this rule when the HTTP request host header is not equal to www.example.com. Because it’s a regular expression, we escape the dots with backslashes (and use a start-anchor for good measure). The from element uses parentheses to capture the whole of the path the the resource that’s being asked for, minus the starting forward slash. Of course, if the request is just for example.com (probably the most common scenario) then this will be empty. The captured value is then used in the to element to construct that path for the redirect. So example.com/tour.jsp should redirect to www.example.com/tour.jsp.

UPDATE
A quick update to mention a gotcha that I ran into; if you set up UrlRewriteFilter as described above then it will redirect you if you try to access your website on your development machine by going to localhost:8080. This can lead to very frustrating scenarios involving repeatedly hitting F5 while muttering “why aren’t you showing my changes, dammit?” and failing to notice that the address bar reads “www.mywebsite.com”. Because UrlRewriteFilter lets you have multiple conditions, it’s quite easy to get the correct behaviour:

    
        seo redirect
        ^www.example.com
        ^localhost
        ^/(.*)
        http://www.example.com/$1