The followng is a couple typical examples (see: Apache's documentation on RedirectMatch)
RedirectMatch permanent ^index.*.html /
RedirectMatch permanent ^about.html /about
RedirectMatch permanent path/to/some/article.html http://example.com/node/7865
Unfortunately this doesn't work very well when it's a Drupal site. I believe the problem lies with the core FAIL in that Drupal (by default) uses UGLY URL's and uses Apache's rewrite module as a stupid hack to get around their inability to support clean URL's out of the box. That step in the Drupal installation where it checks for a capability to support clean URL's? Why do they have to check for that? Why can't they just do clean URL's from the getgo without having to do weirdo unnatural things with mod_rewrite?
| Scenario | URL |
|---|---|
| Default Drupal URL | http://example.com/index.php?q=node/7865 |
| Desired URL | http://example.com/user-name/article-title |
| URL when using RedirectMatch and mod_rewrite based Drupal Clean URL's | http://example.com/user-name/article-title?q=path/to/original/url |
The last row of that table is the subject of this tutorial. I've got several forms of the behavior in my sites, and have a cure for one of them.
| Scenario | Result |
|---|---|
| Redirect old URL to new URL within the same site | Fix known |
| Redirect old URL to new URL on a different Drupal site | Fix not known, FAIL |
Consider
RedirectMatch permanent ^about.html /about
RedirectMatch permanent node/1234 /node/4321
The first is a case of converting a site built with static HTML pages into one built with Drupal. A similar scenario (the second) is if you've written a page on a Drupal site (http://example.com/node/1234) and no longer want that page to exist, have significant traffic going to the old page, and have a replacement page with better content.
The result will be
| User visits URL | Browser redirects to (with FAILURE) |
|---|---|
| http://example.com/about.html | http://example.com/about?q=about.html -- plus, the user is told it's an unknown page |
| http://example.com/node/1234 | http://example.com/node/4321?q=node/1234 -- plus, the user is told it's an unknown page |
I found a solution which works at: http://drupal.org/node/142947 (note it dates from 2007) where the suggestion is to use something like this before the mod_rewrite section of Drupal's .htaccess file:
RewriteRule ^keywords/(.*)\.html$ /tags/$1.html [R=301,L]
This works for the following cases within the same site:
RewriteRule ^index.*.html / [R=301,L]
RewriteRule ^about.html /about [R=301,L]
But not for the following case that redirects the browser to a different (Drupal) site:
RewriteRule path/to/some/article.html http://example.com/node/7865 [R=301,L]
And of course because you've modified one of Drupal's core files you have to remember to not overwrite .htaccess when you update Drupal core (which occurs about every 2-3 months or so depending on the rate of security patches).
I should also discuss the Path Redirect module. It performs a similar service to the above Apache directives, but as a Drupal module. It's useful and probably does the above redirects without the failure modes above. However what made me not use it is that one enters the URL's one at a time in the admin form, and there is no ability to supply a pattern match on the URL. It depends on the number of URL's you have on hand to redirect. If you have only a few URL's to redirect then path_redirect will do the job fine. In my case I had hundreds of URL's and the RedirectMatch redirective reduced the number of mappings to a few dozen patterns.