Hi, thanks for all the trouble explaining. I'll look into this some more. What I already did that didn't really work: - make sure there is a <?xml ..> declaration at the start of the file - make sure there is a <meta content-type UTF-8> as the first item in <html><head>
In the latter case: IE simply adds it's own <meta content-type iso-8859-1> above mine. :-( Thanks and bye, Helma > -----Original Message----- > From: J.Pietschmann [mailto:[EMAIL PROTECTED] > Sent: Thursday, 08 January 2004 21:07 > To: [EMAIL PROTECTED] > Subject: Re: Sudden difference in interpretation of #160 - update > > > [EMAIL PROTECTED] wrote: > > Internet Explorer is on "encoding: auto-select" which > > defaults to ISO 8859-1 with or without <metatag > content-type UTF-8> in the > > page header. Same goes for Opera. > > There are multiple ways for a browser to get hints for the encoding: > 1. The HTTP content-type header, which optionally specifies a > character > set. Unfortunately, the default character set is > iso-8859-1. See, for > example > http://www.jorendorff.com/articles/unicode/unicode-http.html > 2. An XML declaration, in case the content type is > application/xhtml+xml > (or XML or another XML based content). Unfortunately, if there is an > XML declaration without an encoding, the default is either UTF-8 or > UTF-8/UTF-18 autodetection, at the discretion of the consumer > 3. A META-element in the (X)HTML head defining an encoding. > There is no > default in case there is no such element. > IEx adds its own guess, and all browsers add manual overide > by the user. > > If you send UTF-8 encoded content, and both IEx and Opera detect it as > ISO-8859-1, most likely there is no character set defined in the HTTP > content type header, no XML decl (applies only to XHTML) and no META > element declaring an encoding. The HTTP header should be set by the > serializer, but there may be a servlet container override in place. > > I'd sniff the HTTP headers in order to check whether the content type > header properly declares an UTF-8 encoding, absent this, I'd either > - track down why this isn't set and fix this or > - add a META element declaring an UTF-8 encoding, which probably works > only if the content-type doesn't define a character set. > > J.Pietschmann > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
