404 PHP handler

The problem: What to do with "404 not found" errors.
These can be quite common because of Linux/Unix case-sensitivity.

Case-sensitivity on Linux/Unix server

It is common to have your website on a Linux/Unix server with a case-sensitive file system.
On such a server, a URL may directly reference a matching file on the disk.
But then typing that URL with the wrong case will fail and give "404 Not Found".

Here are different approaches to dealing with URL case-sensitivity:

User Relationship between URL and file system What happens if case wrong Example of correct case Example of incorrect case

Normal user's personal webspace on:
www.computing.dcu.ie
Direct 404 Not Found Correct Incorrect

Sysadmins running department webspace on:
www.dcu.ie
Indirect. Content management system. URL seems to be address of an object in its database. Content management system can implement its own case-insensitive addressing. Correct Incorrect

My own website:
humphryscomputing.com
Direct 404 triggered, but I re-direct the error to my own 404 PHP handler program to do case-insensitive and partial matching. Correct Incorrect

To summarise, I set up for my site a case-insensitive site using a 404 PHP handler.

You can try this out on your own account on:

 student.computing.dcu.ie

Recipe is as follows.

ErrorDocument in .htaccess

The Apache web server allows a "404 not found" error to be redirected to a program (rather than output a standard error page).

Put a .htaccess file in:

$HOME/public_html/.htaccess

This .htaccess file has a line to redirect 404 to a program:

ErrorDocument 404 /~myuserid/404.php

Help files:

Apache docs
- Apache Core Features
  - ErrorDocument
- Custom error responses - notes that REDIRECT_URL is probably the environment variable we are interested in
- In general, to re-write URLs on Apache: mod_rewrite

Some notes:

ErrorDocument can redirect to a full "http://" URL. But it seems that REDIRECT_URL is then empty. This means the error page does not know the string you typed, so cannot really help you. So do not use this.
Some non-existent URLs seem to give 403 instead of 404. So redirect both error codes.

Conclusion, use this:

ErrorDocument 404 /~myuserid/404.php
ErrorDocument 403 /~myuserid/404.php

404.php

For the recipe for 404.php see Lab - 404 PHP handler

Testing

See the good URL:

https://humphryscomputing.com/Notes/UNIX/intro.html

And try some bad URLs:

Apache module

Apache module mod_speling

User	Relationship between URL and file system	What happens if case wrong	Example of correct case	Example of incorrect case
Normal user's personal webspace on: www.computing.dcu.ie	Direct	404 Not Found	Correct	Incorrect
Sysadmins running department webspace on: www.dcu.ie	Indirect. Content management system. URL seems to be address of an object in its database.	Content management system can implement its own case-insensitive addressing.	Correct	Incorrect
My own website: humphryscomputing.com	Direct	404 triggered, but I re-direct the error to my own 404 PHP handler program to do case-insensitive and partial matching.	Correct	Incorrect