Using .htaccess rewrite rules to reflect a “fake” directory structure in the addres bar

Question

I'm working with an online encyclopedia and I am trying to achieve the following: Given the physical location of a file in http://example.com/articles/c/a/t/Cat.html , Get the location in the address bar to show http://example.com/encyclopedia/Cat.html

This also needs to work so that if a link is clicked or someone types in "example.com/encyclopedia/Cat.html" , the server will look for the file in "/articles/c/a/t/Cat.html" , yet still serve the shorter URI in the address bar.

I understand this may involve some heavy .htaccess voodoo to accomplish, or perhaps that it would be better to use a PHP script to serve this purpose.

So far I have the following in my .htaccess:

<IfModule mod_rewrite.c>
Options +FollowSymLinks
RewriteEngine On
RewriteRule ^encyclopedia/(.*)\.html$ articles/$1.html [NC]
RewriteCond %{THE_REQUEST} ^GET\ articles/(.*)
RewriteRule ^articles/(.*) /encyclopedia/$1 [L,R=301]
</IfModule>

However with this code, it only works by going to "example.com/encyclopedia/c/a/t/Cat.html" and showing the proper page, and when you go to "/articles/c/a/t/Cat.html it still doesn't rewrite it as "/encyclopedia/" , it just stays the same.

Edit - By removing the GET\\ part from the RewriteCond and removing the leading forward-slash from /encyclopedia/$1 in the following line, any requests to "/articles/c/a/t/Cat.html" are correctly redirected to "/encyclopedia/c/a/t/Cat.html" . I am still at a loss trying to remove the "/c/a/t" part though. **

I've tried using the following two rules to remove the "c/a/t/" part:

RewriteRule ^encyclopedia/((.)(.)(.).*)\.html$ articles/$2/$3/$4/$1.html [NC]
RewriteRule ^articles/(.)/(.)/(.)/(.*) /encyclopedia/$4 [L,R=301]

But with no success as I'm sure what's happening is I'm getting the capital "C" from "Cat.html" and putting that in as "/articles/C/a/t/Cat.html" which will obviously not work.

I've been looking around studying .htaccess RewriteRule and RewriteCond for days but I still haven't been able to figure this out and been BHOK enough to cause a few migraines.

Would this be better accomplished using a PHP script? Or can this voodoo be easily enough accomplished via only .htaccess rules?

Answer 1

First thing, forget about .htaccess files. .htaccess files is just an extension of Apache configuration files that you can put in some directories. They're really slowing down your apache server, he needs to check part of his configuration at runtime. It's done to allow some configuration on hosted environments.

Put everything you have in .htaccess files in <Directory> sections on your VirtualHost and use AllowOverride None to tell Apache to forget about trying to read .htaccess files.

So what you need is mod-rewrite voodoo, not .htaccess voodoo :-)

Now your rewrite problem is quite complex. If you need some mod-rewrite help do not forget to read this ServFault article : Everything You Ever Wanted to Know about Mod_Rewrite Rules but Were Afraid to Ask?

I assume that your Cat.html -> c/a/t/Cat.html is just an example and that you can have more than 3 letters : CatAndDogs.html -> c/a/t/a/n/d/d/o/g/s/CatAndDogs.html .

The part of mod-reqrite you need is (I think) RewriteMap . There you will find some helpers like lowercase: that coudl help you, but you will also find the prg: which means using an external program to perform the mapping. I would use perl examples of such rewriteMaps examples available via google and make some transformations. Should be quite easy and Fast in Perl to transform CatAndDogs.html in c/a/t/a/n/d/d/o/g/s/CatAndDogs.html .

Note that RewriteMap will never work inside a .htaccess . Forget .htaccess files. The prg: keyword will launch your perl program as a parallel daemon and will feed him with quite a lot of data, you shoudl really write something robust & fast. Do not forget to use the RewriteLock directive to avoid mixing results (some prg: mappers do not care about mixing results, think about load balancers for examples, but you do want to avoid mixing results for parallel queries)