Re: [Xsltforms-support] Experimental Component control (Was: A couple of questions about subforms)

[Alain C. <ala...@ag...>]

Eric,
>> Currently, the default instance of a subform is merged with the one 
>> in the main form while, from a component, it has its own id. That's 
>> why those functions don't work with subforms even if their name would 
>> say so ... ;-) 
> That's what I had guessed but can you elaborate on this instance merge?
For subforms, models already with the same ids in the main form are 
merged: instances are merged and bindings are added in the main form. 
Merging instances is similar: if the corresponding id already exists in 
the main form, it is ignored; if not, it is added.

The default model without an id and the default instance without an id 
are merged as if they had the same id.
>
> I had the feeling that the context node in the master form was becoming
> the instance of the subform, is that what you mean?
No, in a subform, instances are identical but scoping is effective and 
"." references the context node.
>   
> Couldn't subform-instance() and subform-context() be pointers to the
> subform default instance and context node without changing the behaviour
> of this "merge"?
Yes, that's possible but I would say that subform-instance() should be 
useless unless, depending on the different situations, a component would 
also be used as a subform.
>> XSLTForms is a versatile product which can very easily be extended at
>> any level (I know I'm not the only one doing this...)!
> Sounds impressive...
>
>
Well, XSLT is powerful and Javascript can now do a lot of things!

-Alain