Redirecting from old Wordpress URLs — getting query strings

Question

I'm migrating the content of a Wordpress blog to EE. The old blog has this URL structure:

/news/2012/08/08/2013-schedule/

And I need to redirect to this:

/news/post/2013-schedule/

At the moment I am using this .htaccess redirect:

RedirectMatch 301 /news/([0-9]+)/([0-9]+)/([0-9]+)/(.*)$ /news/post/$4

But it's resulting in this URL:

/news/post/2013-schedule/?/news/2012/08/08/2013-schedule/

This is not the end of the world, because I get the right results, despite the extra URL segments, but I'd really rather they went away.

Any help?

FWIW, I'm using the EllisLab-approved redirect method for removing index.php.

Answer 1

Few things I'd try first:

1. Have you tried adding the ? to the approved method after index.php?
2. If that doesn't work, have you tried adding ? before /news/post/$4 to see if that changes anything?
3. Check your $config['uri_protocol'] or add it to your config and change from AUTO to one of the other settings (check http://devot-ee.com/ee-config-vars to see the different options)

Any of that help?

Edit: I think this is related to an issue I had which was outlined here, hence Option 3: http://ellislab.com/forums/viewthread/202909/#957589

Answer 2

I think your rule is being evaluated at least twice and returns a match twice, once before and once after you do your index.php hiding.

You must be more narrow about what you are matching on. I think this will do the trick:

RedirectMatch 301 ^index.php/news/([0-9]+)/([0-9]+)/([0-9]+)/(.*)$ /news/post/$4 [L]

Answer 3

This is not marked as answered, did you find a solution? I don't know off the top of my head why this would be happening, but one thing to try would be to put an absolute url instead of a relative one at the destination of your redirectMatch statement. The docs ask for the full URL.

Try replacing that last (.*) with (([\/\w \.-]*) if a query string is somehow being added internally before reaching that step, that should strip it out by matching only numbers letters underscores dots and hyphens.

You may also want to move that up earlier in your list of redirects, it seems likely that you are getting unexpected input into your redirect. You could turn on redirect logging in Apache to try to get a better sense of what is happening.