Pierre said:
Since it works the same with 2 browsers and 2 web servers, I'm almost sure it's not dependant on the configuration - but if others can tests on different configurations I'd like to know the result

So I showed in my just previous messages (after the one you are responding to) my output from Live HTTP Headers, where it seems that Firefox is using UTF-8 transmission, both for header values (filename) and data values (euro character).  Without specifying Content-Type (for the data) or doing RFC 2047 encoding as would be expected from reading the various standard documents (RFC 2045, W3 HTML 4.01, RFC 2388).  I wonder now if Live HTTP Headers is reporting the logical data, prior to encoding for transmission.  But I was getting UTF-8 data inside my CGI script... 

So now I tweaked the server to save the bytes it transfers its rfile to the cgi process (had already tweaked that to be binary instead of having encodings), and it is clearly UTF-8 at that point also.  Looks just like the Live HTTP headers.  Now that I have data-capture on the server side, I can run the same tests with other browsers... so I ran it with Opera 11, IE 8, Chrome 8, and the only differences were the specific value of the boundaries... all the data was in UTF-8, both filename, and form data value.

I can't now find a setting for Firefox to allow the user to control the encoding it sends to the server, but I can't rule out that I once might have, and set it to UTF-8.  But I'm quite certain I don't know enough about the other browsers to adjust their settings.  I don't have Apache installed on this box, so I cannot test to see if it changes something.

Is there a newer standard these browsers are following, that permits UTF-8?  Or even requires it?

Why is Pierre seeing cp-1252, and I'm seeing UTF-8?  I'm running Windows 6.1 (Build 7600), 64-bit, the so-called Windows 7 Professional edition.