Surprising rewriting of URL by htaccess rule - .htaccess

Question

I've zeroed my problem and I've specific question.
With only the following code in the .httaccess why index2.php gets called if I type in my URL as www.mysite.com/url2 ?
RewriteEngine On
RewriteCond %{REQUEST_URI} (.html|.htm|.feed|.pdf|.raw)$ [NC]
RewriteRule (.*) index2.php [L]
I've also tested it at http://www.regextester.com and should not replace it with index2.php:
In the end I want this rule to skip any URL starting with /url2 or /url2/*.
EDIT: I've made screen recording of this problem: http://screenr.com/BBBN

Answer 1

You have this in your .htaccess:
RewriteEngine On
RewriteCond %{REQUEST_URI} (.html|.htm|.feed|.pdf|.raw)$ [NC]
RewriteRule (.*) index2.php [L]
What it does? it rewrites anything that ends with html, htm, feed , pdf , raw to index2.php. So, if you are getting results as your URL is ends with those extensions, then there are two possible answers:
There is another rewrite rule in an .htaccess in upper directories (or in server config files) that causes the URL to be rewritten.
Your URL actually ends with those extensions. have in mind, what you enter in your address bar, will be edited and rewritten. For example, if you enter www.mysite.com/url2 in your address bar and that file doesn't exist on server, your server will try to load the proper error document. So, if your error document is /404.html, it will be rewritten to index2.php at the end.
Update:
I think it's the case. create a file named 404.php in your document root. Inside your main .htaccess (in your document root), put this:
ErrorDocument 404 /404.php
delete all other ErrorDocument directives.
inside 404.php , put this:
<?php
echo 'From 404.php file';
?>
Logic behind it:
When you have a weird behavior in mod_rewrite, the best solution in my experience is using rewrite log. to enable rewrite log put this in your virtualhost or other server config directives you may choose:
RewriteLogLevel 9
RewriteLog "logs/RewriteLog.log"
be careful: the code above will enable rewrite log and start logging at highest level possible (logging everything). It will decrease your server speed and the log file will become huge very quickly. Do this only on your dev server.
Explanation: When you try to access www.mysite.com/url2, Apache gives your URL to rewrite module. Rewrite module checks if any of RewriteRules applies to your URL. Because you have one rule and it doesn't apply to your URL, it tries to load the normal file. But this file does not exit. So, Apache will do the next step which is showing the proper error message. When you set a custom error file, Apache will run the test against the new address. For example if error document is /404.html, Apache checks whether your rule applies to /404.html or not. Since it does, it will rewrite it.
The point to remember is apache will do this every time there is change in URL, whether the change is made by rewrite module or not!

Answer 2

The rule you list should work as you expect if this is the only rule. Fact is that theory is fun, but apparently it doesn't work as expected. Please note that . will match ANY CHARACTER. If you want to match the full stop/period character, you'll need to escape it. That's why I use \.(html|htm|feed|pdf|raw)$ instead of (.html|.htm|.feed|.pdf|.raw)$ below.
You can add another RewriteCond that simply doesn't match if the url starts with /url2, like below. This might not be a viable solution if there are lots of urls that shouldn't be matched.
RewriteCond %{REQUEST_URI} !^/url2
RewriteCond %{REQUEST_URI} \.(html|htm|feed|pdf|raw)$ [NC]
RewriteRule (.*) index2.php [L]
To get a better understanding of what is happening you can alter the rule to something like this. Now simply enter the urls you dont want to be matched in the url bar and inspect the url bar after the redirect happens. In the url-parameter you now see what url actually triggered this rule to match. This screencast shows you a similar version working with a sneaky rewriterule that is working away on the url.
#A way of finding out what is -actually- matched
RewriteCond %{REQUEST_URI} \.(html|htm|feed|pdf|raw)$ [NC]
RewriteCond %{REQUEST_URI} !/foo
RewriteRule (.*) /foo?url=$1 [R,L]
You can decide to match the %{THE_REQUEST} variable instead. This will always contain the request itself. If something else is rewriting the url, this variable doesn't change, meaning you can use this to overwrite any changes. Make sure the url won't be matching itself. You would get something like below. An example screencast can be found here.
#If it doesn't end on .html/htm/feed etc, this one won't match
RewriteCond %{THE_REQUEST} ^(GET|POST)\ /.*\.(html|htm|feed|pdf|raw)\ HTTP [NC]
RewriteCond %{REQUEST_URI} !^/index2\.php$
RewriteRule (.*) /index2.php [L]