mod_rewrite question

[andreas-kalt]

I've tried to get my head around this but it seems I'm stuck, some help is appreciated.

Here's my problem: I've redsigned my site and want to redirect requests to the old URLs to the new ones.

The old URLs were like this:

www.domain.com/section/seven_habits.htm
or
www.domain.com/section/category/seven_habits.htm

the new ones are

www.domain.com/section/seven-habits
or
www.domain.com/section/category/seven-habits

Is there a way I can rewrite old ones to new ones and send out permanent redirect headers with mod_rewrite? I want the new ones to show up in the address field of the browser.

Thanks for any hints.

[dklynn]

andreas,

Yes, quite possible!

First, I'll assume that you've changed the links within your site. That will give you the "extensionless" URLs for your htm(l) pages.

Now, to make things simpler (for me), I'll assume:

1. You just want this redirection to occur in the section subdirectory
2. You only need to redirect .htm pages
3. You'll put the following in the .htaccess file in your root directory

RewriteEngine On
RewriteRule ^section/([a-z0-9/]+)$ section/$1.htm

This code looks at the {REQUEST_URI} (in the root directory) and checks for the section subdirectory, then checks to see that one or more lowercase letters, digits and the slash exist at the end of the string (i.e., NO dothtm). If those conditions are met (as they would be with either of your OLD examples), then the redirect is created with the .htm extension.

Does that clarify for you?

Regards,

DK

[andreas-kalt]

I'm not sure if I understand correctly of if I explained properly. So, here again:

The old site was static HTML pages with extensions. The new site runs on a CMS (Textpattern) which creates extensionless URLs. Additionally, I used underscores in the old filename scheme and Textpattern uses hyphens.

I want visitors with bookmarks to old-style URLs to be redirected to the new URLs and I want the new URLs to show up in their address field.

If I understand your example correctly, it's just the other way round than what I want. There are in fact new URLs, the old ones no longer exist once I go live but people might have bookmarks and the search engines will have old URLs in their index.

So I want a permanent redirect but without having to specify every single URL. That's why I thought of mod_rewrite and regexs.

I hope I explained this properly, I still have problems with the concept behind mod_rewrite, especially with old/new and which side of the expression stands for what.

Thanks for your help.

[dklynn]

andreas-kalt,

Hmmm, that's the way that I created my response the first time then talked myself out of it .

Without the conversion from - to _, the code would be

RewriteEngine On
RewriteRule ^section/([a-z0-9/]+)\.htm$ section/$1 [L]

Adding hyphens and converting to underscores to the above code is a little trickier:

RewriteEngine On
RewriteRule ^section/([a-z0-9_/]*)-([a-z0-9_/]*)\.htm$ section/$1_$2.htm [N]
RewriteRule ^section/([a-z0-9_/]+)\.htm$ section/$1 [L]

The new rule will look for a hyphen amid lowercase letters, digits, underscores and slashES and replace them ONE AT A TIME with underscores then restart and look for another hyphen.

Note 1: If you were sure that only one hyphen was in your URLs, you could drop the Next flag ([N]) (and the _'s in the first rule) and Apache would only have to make a single pass through this code.

Note 2: Regular expressions (preg_replace) are far more powerful as they'll change every occurance of a string in another string as directed but I believe that more care is required with Apache's implimentation. My use of the Next flag may be overkill on my part but it will force Apache to restart the rewrite rules with the updated {REQUEST_URI} and that's what we need to be sure we've changed all -'s to _'s.

By the time Apache gets to the second (and Last [L]) rule, all hyphens would have been converted to underscores and that's been added to that rule so you should be set.

The important thing is that you can understand what's going on with the regex used so you can do it the next time.

Regards,

DK

[andreas-kalt]

Thanks for your help, DK. In the meantime I realised that my CMS uses some rewrite rules that don't go well together with additional ones. So I stuck to normal redirects.

But you gave me some more insight into mod_rewrite. By and by, I will certainly get my head around it. Thanks.

[dklynn]

andreas-kalt,

That's great! Glad to have been of help.

Regards,

DK