Custom 404

Was returning 200 OK

[grandpa]

I recently added some sites to my Webmaster Tools dashboard, and was setting the preferred domain to mydomain.com vs www.mydomain.com. I had a custom 404 page, and to my surprise the tool told me that my custom 404 page was returning a status 200 OK. What!? How could that be? I had this bit of code at the top of my custom 400 page:

<?php
header("HTTP/1.0 404 Not Found");
header("Status: 404 Not Found");
?>

Someone is messing with me, I think. So I decided to check the server headers for http://www.mydomain.com/404.php

What I saw was a 301 Permanent redirect, from www.mydomain.com to mydomain.com, followed by a 200 OK at mydomain.com. The custom page was completely overlooked.

Now is was time to take a closer look at my .htaccess file. Here's the bit where I told Apache to rewrite everything away from the subdomain www

RewriteCond %{HTTP_HOST} ^www.mydomain.com [NC]
RewriteRule ^(.*)$ [mydomain.com...] [R=301]

Racking my brain (an easy task these days) I looked at this code.. the problem had to be there. Perhaps I needed to tell Apache that this instruction was to be last, so I added that flag.

RewriteCond %{HTTP_HOST} ^www.mydomain.com [NC]
RewriteRule ^(.*)$ [mydomain.com...] [R=301,L]

Nothing doing. Time to dig deeper.

Aha! I wasn't rewriting the variable, in this case the error page. This is a simple matter of adding the variable (.*) to the end of the string, as $1

RewriteCond %{HTTP_HOST} ^www.mydomain.com [NC]
RewriteRule ^(.*)$ [mydomain.com...] [R=301,L]

Now my custom 404 page displays the information I want, and actually returns a 404 error code. Life is good.

[jdMorgan]

Minor tweaks: Always escape literal periods in patterns by preceding them with a "\", and no need to anchor a ".*" pattern standing alone:

RewriteCond %{HTTP_HOST} ^www\.example\.com [NC]
RewriteRule (.*) http://example.com/$1 [R=301,L]

An un-escaped period in a regular-expressions pattern is interpreted as a single-character wildcard, meaning "match any single character."

Now you can look into canonicalizing index.php, and all that other fun stuff... :)

Jim

[grandpa]

Thanks Jim. The fine-tweaks are duly noted.

Note to self: Next time I leave a job with 4 years of fine tuning the .htaccess file, send self a copy of said file.

>> canonicalizing index.php, and all that other fun stuff

There's more!?! :)

[jdMorgan]

Lots more. Of course, some sites don't have much of a problem with legacy links or malicious linkers, and so don't need all of it.

Some of the things we've discussed recently (and repeatedly) are redirecting

from example.com/<any_directory_or_subdirectory>/index.html to example.com/<any_directory_or_subdirectory>/

from //images/logo.gif to /images/logo.gif (note double slashes)

and from /images//logo.gif to /images/logo.gif

to remove spurious query strings when the page doesn't even use them (e.g. a static HTML page)

to remove trailing slashes from extensionless filenames

to add missing trailing slashes to directory names

to remove spurious characters from the end of URLs (often added by forum/blog auto-linking routines or because of typos, for example, "GET /foo.html." with the trailing period actually included in the URL

to remove spurious characters from the end of URLs caused by improperly-closed link tags in HTML code, for example "GET /bar.php>a%20really%20useful%20page%here</a>" where the closing quote was left off the URL in the <a href> tag.

But wait, there's even more! Basically, there are so many I can't even think of them all at once.

Jim

[g1smd]

We listed them in a thread just a few weeks back to save further brain ache jd!

[webmasterworld.com...]