Ok, so I've got some HTML content with my own subroutine:

my $html = get_content($url);
print $html;

Result is I get the contents of $html printed, fine.

Then I want to decode it:

print HTTP::Response->decoded_content($html);

Result:

Odd number of elements in hash assignment at C:/Perl/lib/HTTP/Message.pm line 28
9.
Use of uninitialized value in print at myscript.pl line 269.

What's going on? I read the doc on decoded_content and messages and it says something about options I don't understand.

$mess->decoded_content( %options )

Scott

How do you expect to contruct an HTTP response from just the body of the response? You need the headers too. By creating an H::R response object, you don't magically get the headers your function previously discarded.

I mentioned that an HTTP response is a possible source of the encoding. I used an example where the HTTP response is provided via an HTTP::Reponse object. The H::R object reflects the HTTP response; it's not a decoding tool.

You're not starting from an HTTP response. You're starting from an HTML file. I told how to handle that situation too.

Either fix your function to return an HTTP response and not just the body of an HTTP response, or treat the value returned as a file.

You can probably tell I don't understand HTTP responses very well; I was considering this comment of yours:

That said, you don't need to do ANY of this. You just use HTTP::Response's ->decoded_content method and it will decode the content for you.

I was trying to figure out how to do that, but wasn't able to figure it out from reading the documentation. In any event, back to Plan A:

 print join " ",
     LWP::UserAgent->new->get("$url")->content_type, $/;
[download]

Oddly enough that works, but only returns this:

content type: text/html
[download]

The charset is not included, although when I use the same call on the MS or Google sites as in your example, it does.

Confused? yes. :-(

Thanks for your patience, and apologies for not getting it.