----- Original Message -----
From: "Bernhard Huber" <[EMAIL PROTECTED]>
> Let me summarize the configuration:
> A new JavadocSource and -SourceFactory using uri-prefix javadoc:
> In the configuration of the JavadocSource you have to
> define the physical location of the java source, as noted
> by the source-root/@uri attribute.
> I don't see the reason for source-roots/@group, yet.
> What is the purpose for defining it?
It is not used yet, perheps it should be removed for now. The original idea
was to mimic javadoc 'package groups'. Since you bring this up and I rethink
it, it would probably be better to remove it.
> As JavadocSource sticks to the Source contract it may be used
> in any src component definition.
> One obvious usage is using it as source for the xml filegenerator.
> eg. <map:generator src="javadoc:{javdoc-path}"/>,
> you may use it in cinclude, too.
Yes, indeed.
> The JavadocSource implements the javadoc: protocol, assuming
> that javadoc-path specifies a FQN java class, eg. java.util.HashMap.
>
> Resolving the FQN to a java source is done by JavadocSource using
> its source-root, and source-roots definitions.
Indeed again.
>
> Another question:
> Must javadoc-path specify a FQN only, is a package name allowed, yet?
No, not yet. Only (inner)class and (inner)interface documents are created
for now.
> A specification question: Shall be a package name allowed?
Yes! This is a planned feature.
> Next I'd like to ask about the XML namespace, and schema emitted by the
> JavadocSource.
>
> I'd like to suggest that we should use namespace:
> http://org.apache/cocoon/javadoc/1.0, default prefix jd
Sure, whatever ;-)
> The schema shall be more or less a marshalling of the Qdox objects.
> This marshalling part is something which i have elaborated, in my
> opionion it covers the qdox objects quite complete. That's what i might
> contribute to the overall JavaDoc contribution.
The object marshalling was not as straightforward as I would like. For
instance, QDox has both Type classes and JavaClass, which are the same
subject to some extend but completely different objects. This caused some
'ugly' code unfortunately...
> any comments welcome,
>
> regards bernhard
I'll be posting shortly (if I get it working in the latest CVS version!)
Ciao,
Bart
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]
