Re: [xmlblaster] bug or feature ?

[Marcel Ruff <mr at marcelruff.info>]

Jürgen Prinz wrote:
> The result of MessageUnit.toXml may not be a well-formed Xml-Document, 
> if the content includes characters as '<' or \000.

The C++ MessageUnit.toXml() and Java MsgUnit.toXml() methods
are mainly for logging. They are not used operational.

The Java variant uses a CDATA section:

  <content><![CDATA[BlaBlaBla]]></content>

which helps a bit but still may fail with binary contents
containing ']]>' or '<![CDATA[' tokens.

An approach would be to check the content and do a Base64
encoding when offending tokens are found:

 <content encoding='base64'>QmxhQmxhQmxh</content>

We than would need to update C++ and C to behave same as java.

Is this a reasonable approach?

>
> Are there "official"
> routines for encryption / decryption ?
Are you talking about our authentication plugin?

> Out of curiosity
>
> J. Prinz

best regards

Marcel

--
http://www.xmlBlaster.org