This content is no longer maintained on this Wiki. The Moin DocBook pages are canonical.
I got an idea of how to embed "inline" xml in to the wikisyntax (step 5 of the plan of attack).
In moinmoin there are custom parsers for highlighting program snippets and irclogs. The way to call these is to write
{ { {#java
public static void main(String args[] ) {
System.out.println("Hello world");
}
} } }
The "#java" will result in the wikiparser invoking the syntax highligher (aka parser) for java.
My idea would use a similar strategy, but the wikiparser would invoke a "parser" that would simply pass xml to the formatter. It would be invoked in a similar manner:
{ { {#xmlpassthrough
<db:bookinfo>
<db:revision>1.3</db:revision>
<db:date>2006-12-02</db:date>
</db:bookinfo>
} } }
This would get parsed by a real xml parser. The parser would then check if the formatter has support for a namespace, and if it does, it would start passing the info by, calling a special formatter method for each xml-element. A simplistic sax interface would be perfect, since each formatter could just implement the sax methods they are interested in.
My plan is to have every xml-element namespaced, and xmlpassthrough would have a hardcoded namespace list. This way there would be a single canonical list of the namespace keywords.
The xmlpassthrough would do the following before passing the result to the real xml parser:
- gather the whole xml string
- wrap it in <passthrough> </passthrough> so that there will be a single root element, as required by xml spec
- prepend the required declarations and namespaces