[Rssview-developers] Re: AW: RSSView: Problem with XML encoding?

[Martin <cin...@gm...>]

Am Mon, den 15.03.2004 schrieb sfk...@hs... um 23:04:
> Regaring the problem reading http://www.heise.de/newsticker/heise.rdf:
>  
> 0. To your question: The encoding declaration ISO-8859-1 in the prolog refers to the content. The tags/attributes are following the rules of XML names (see chap. 2.3 Common Syntactic Constructs for XML names and chap. 4.3.3 Character Encoding in Entities for encoding declarations).
>  
> 1. What is strange, is, that when I wget this XML file and save it to my localhost directory (where my local apache is running), rssview and crimson are able to read it! 
>  
> Another example, where rssview can flawlessly read is http://xml.newsisfree.com/feeds/29/629.xml <http://xml.newsisfree.com/feeds/29/629.xml>  which also has  ISO-8859-1 encoding. So, it is not really an encoding problem. 
>  
> What makes me suspicious is rather the following:
>  
> 2. wget reports "Length: unspecified" which usually is "Length: 1,696": whatever length it really is, crimson perhaps does'nt like unspecified lengths?

Maybe it's a problem with the HTTP-headers that are sent by heise.
I have seen a similar problem with simple HTML-pages. Heise is
sending them in ISO-8859-1 (HTTP-header) and they often still
contain the Euro symbol. My browser has problems with it.

I have to find out more about the communication while
crimson is fetching the pages. You can look at the
headers that are sent by heise with "wget -S".
I don't see any encoding headers there.

Maybe crimson says, it prefers to have UTF-8 and heise
says "OK" and does not specify the encoding?

(My tests show nothing suspicious until now.
 Maybe it's really the length.)
 
> 3. The top level exception of the esxeption stack report is issueing:
> org.xml.sax.SAXParseException: Zeichenumwandlungsfehler: "Malformed UTF-8 char -
> > - is an XML encoding declaration missing?" (Zeilenzahl m÷glicherweise zu niedrig)
> The german message in parantheses does'nt really have something to do with encoding (free translation it reads "Number of lines perhaps too small")??

I really don't have an idea. If you find something out, please 
be so kind and tell me.

Martin