Mozilla strange behaviour during autodetection of character encodings

I have just discovered something odd. I have a web page that is hosted locally
and duplicated externally. When I pointed the browser at the local page, it
contained funny symbols. However, the page hosted externally displayed
differently. After a bit of investigation, I discovered the following:

From the browser when viewing the local page:

Menu:View,Character Encoding (This is set to Western ISO-8559-1).

From the browser when viewing the remote page:

Menu:View,Character Encoding (This is set to Unicode UTF-8).

Why is this autodetecting? I have the autodetect switched off:

Menu:View,Character Encoding, Autodetect, Off

And why does the browser think that the page on the remote web server is in
UTF-8, but the exact same page on the internal web server is ISO-8559-1?

Where does the browser get the character encoding information from?

There is a difference in that internal web pages bypass the filtering proxy,
whereas the externally hosted pages have to pass through the filter.

Also, I use a different webserver internally. The character encoding
information does not separately from the server does it?


Mark Hobley
Linux User: #370818