I'm building a PHP application using CodeIgniter. It is similar to Let Me Google That For You where you write a sentence into a text input box, click submit, and you are taken to a URL that displays the result. I wanted the URL to be human-editable, and relatively simple. I've gotten around the CodeIgniter URL routing, so right now my URLs can look something like this:
http://website.com/?q=this+is+a+normal+url
The problem right now is when the sentence contains a special character like a question mark, or a backslash. Both of these mess with my current .htaccess rewrite rules, and it happens even when the character is encoded.
http://website.com/?q=this+is+a+normal+url? OR
http://website.com/?q=this+is+a+normal+url%3F
What does work is double-encoding. For example, if I take the question mark, and encode it to %253F (where the ? is encoded to %3F and the % sign is encoded to %25). This url works properly.
http://website.com/?q=this+is+a+normal+url%253F
Does anyone have an idea of what I can do here? Is there a clever way I could double encode the input? Can I write a .htaccess rewrite rule to get around this? I'm at a loss here. Here are the rewrite rules I'm currently using for everyone.
RewriteEngine on
RewriteCond %{QUERY_STRING} ^q=(.*)$
RewriteRule ^(.*)$ /index.php/app/create/%{QUERY_STRING}? [L]
Note: The way CodeIgniter works is they have a index/application/function/parameter URL setup. I'm feeding the function the full query string right now.
If your’re using Apache 2.2 and later, you can use the B flag to force the backreference to be escaped:
RewriteCond %{QUERY_STRING} ^q=.*
RewriteRule ^ /index.php/app/create/%0? [L,B]
I usually do human readable urls like this
$humanReadableUrl= implode("_",preg_split('/\W+/', trim($input), -1, PREG_SPLIT_NO_EMPTY));
It will remove any non-word characters and will add underscores beetween words