January 05, 2008
One of the potential drawbacks to switching content management systems for my website was losing the permanent link to specific postings. I receive a fair amount of traffic due to some Google searches, and I have posted comments on a couple of sites that generate a few links per week. I didn’t want to lose this traffic simply because the link had changed names.
In my case there were three issues I needed to address:
Here’s how I managed to create redirects for all 1330 of my Moveable Type entries.
If you search the Wordpress Codex you’ll find an article about importing Moveable Type entries to Wordpress, which contains a section on Preserving Permanent links. Using the template suggested in the article, I created a new Moveable Type template that output a line for each Moveable Type entry, like this:
When I rebuilt the site, causing this template to be run, the resulting file had a line for each Moveable Type entry that looked something like this:
If I had been just switching to Wordpress on the same host, I could have inserted this file in to my Wordpress directory and everything would have worked. However, without the Moveable Type installation on the new server, I would have to do more.
What I needed was a pure .htaccess file to permanently redirect visitors to old URL to the new URL. In order to do that I needed to strip out the php references and manually sanitize the titles. You see the URL scheme I had chosen for Wordpress uses a sanitized version of the posting title in the URL. Quotes, dollar signs, exclamation points, parenthesises, et cetera, aren’t allowed in these sanitized titles.
Making a backup of the redirect.php file the template code above had produced, I set out to accomplish this by hand. (NB: I did try, through the help of a friend to use awk and regular expressions to automate the title sanitazation. While this was largely successful, I ultimately ended up using a file I had converted in TextMate using a series of find and replace commands.)
The resulting file had 1330 lines that looked like this (and, yes, had I been quicker thinking I could have come much closer to this format with the template described above):
Redirect 301 /blogs/######.html https://zanshin.net/yyyy/mm/dd/title/
Points of interest in this format:
Upon loading this file to the root of my site on the new server I promptly got an internal server error (500), meaning that the .htaccess file had a mistake (or two or three). I am not ashamed to admit that I used a brute-force method for finding the errors: I commented all 1330 lines out, and then in blocks of 100 lines each un-commented them and reloaded the site. When it again threw the internal server error between lines 600 and 700, I knew where the problem was located. Two lines had problems. One was missing a “-“ for a space, and the other was missing the closing “/” at the end of the URL.
Hopefully now people coming to my site via old links, be they from a search engine, an embedded link on another site, or a bookmark, they will be redirected to the new location of the page.