When validating a page served with an HTTP header reporting an encoding of 'ascii', and a meta tag reporting an error of 'utf-8', warning W19 is shown:

> The character encoding specified in the HTTP header (ascii) is different
> from the value in the <meta> element (utf-8). I will use the value from the 
> HTTP header (ascii) for this validation. 

This is using the parameter charset=(detect+automatically)

However, if an encoding is forced with the parameter charset=utf-8, the wording of W19 becomes confusing and incorrect:

> The character encoding specified in the HTTP header (ascii) is different from
> the value in the <meta> element (utf-8). I will use the value from the HTTP 
> header (utf-8) for this validation.

The warning manages to report the HTTP header encoding as being both 'ascii' and 'utf-8'.

The warning assumes that the 'used' encoding was determined according to the specification, and not forced by the validator.

I assume W18 and W20 have the same problem.