This is the celebrated journal of Mr. Simon Collison A.K.A Colly

Rewrite Redirect, Apache paths, RSS

27th October 2004

How many of your visitors make the journey via RSS feeds? A quarter - a third - more than half - almost all? Until recently, I’d not placed a great deal of emphasis on maintaining CollyLogic‘s feeds, assuming they were important, but not vital to maintaining a high daily hits average.

Well, more fool me. A recent server swap left my URL structure slightly altered, rendering all those bookmarked feeds utterly useless. With no time to solve the problem through Apache or htaccess before a trip abroad, I was happy to ignore the issue until my return. Then I realised that my hits had decreased by approximately 85 percent!

So what do you do? Hundreds or thousands of your visitors have bookmarked your feeds in their favourite RSS readers, but those feeds - possibly the most important pages on your site - suddenly exist in a different location. How on earth do you get word to those readers that you are not dead, you are still blogging, but your URL structure has changed? Basically, unless you have them all in a mailing list, or run an alternative resource with a shared readership, you can’t. It’s gotta be a mod-rewrite manoeuvre - go with your new URL structure, but ensure you redirect to the new feeds.

For example, I needed to redirect from http://www.colly.com/index.php/weblog/rss_2.0/ to http://www.colly.com/index.php?/weblog/rss_2.0/

Here’s how:

Create a blank file in the root directory called htaccess (if you haven’t already).
Once you’ve uploaded the file, add a ‘.’ to it’s name, creating .htaccess
Paste in the following code and save it (Note: I have inserted a line break before the http:// and URL. Best to remove that):

RewriteEngine on
RewriteRule ^index.php/weblog/rss_1.0/$ 
http://www.colly.com/index.php?/weblog/rss_1.0/ [R=301,L]

Note the old URL has no http://www (relative), but the second (new) URL is full (absolute).

Then for each new redirect, copy the RewriteRule line, and adjust the URLs.

Now, this isn’t a new method, it isn’t entirely concocted by me, and it won’t get me a book deal. I’m flagging up this code because I feel there are far too many confusing mod rewrite methods out there, and for the most part, web builders simply need a straightforward method of safely redirecting to another page. In my opinion, this method is wonderfully easy, and will undoubtedly get you out of a sticky hole at some point. Maybe another site attempted to link to your killer CSS tip, but buggered up the URL. No problem, just redirect their visitors from the duff URL to the correct one.

So, why did the URLs have to change?

Incidentally, The reason this happened after switching to Media Temple is almost certainly due to a change in Apache versions.  Apache 1.x was more ‘lax’ on the CGI PATH_INFO/PATH_TRANSLATED specification, while Apache 2.x (on our MT server) is now conforming to the specification correctly.

If I decide to revert back, what I need to do is to make Apache 2.x use the old style PATH_INFO specification that Apache 1.x did. Huge thanks to David at MT for the info, and a full set of instructions detailing how to do it. If anyone wants the info, just email me. Gotta say, so far MT have been fantastic, and the customer support has been excellent. Good work, nerds…

Responses