on Thu, 17 Jul 2003 15:16:22 +0100 "John Wilson" <[EMAIL PROTECTED]> wrote:
> rufio wrote: > <cut> > > I don't see your problem. All you have to do is split what changes > > and what stay constant: generating string containing a XML messge is > > constant so framework does that; > > encoding may change, and besides it's rather low level issue, so > > transport layer should encode the message. > > No you don't see the problem I'm afraid. The actual Java strings > produced are different for different encodings. For UTF-8 and UTF-16 > everything is just fine and the standard Java Writer implementations > will do the encoding. For US-ASCII and ISO 8859-1 the XML writer has > to be aware of the encoding being uses and generate the appropriate > character references for Unicode characters which are not part of the > character set used for the encoding. You can't write a general XML > write without knowing what encoding is being used. I mentioned before that plugin should provide it's own Writer. I imagine this would work like this: - framework constructs invalid (prologless) message encoded in UTF8 (internal java encoding) - then writes the message to writer provided by plugin, which knows how to enode it. i.e. framework shouldn't care about encoding.
