Include separate instances in model is not working as expected

Salute,

during our migration from orbeon 3.9 to orbeon 4.4 we encountered a
problem including instances into a model using xf:include. We used to
have xf:includes containing one or more xf:instances together with
xf:actions that would fill those instances on ev:event
xforms-model-construct-done. Those fragments were reused in several
models of different xforms by way of including them in a xf:model
element in the 'head'-tag of the xform. This worked with orbeon 3.9 as
we surrounded the instances and actions with a root tag - an useless
dummy element necessary only to make the include well-formed xml.
With orbeon 4.4 this ceased working. The xforms coding inside the dummy
root-element is not being processed any more. The instances do not
appear in the xforms inspector and the actions are not executed.
This seems to imply that any xf:include in orbeon 4.4 can only contain
one single element, which would reduce the possibilities of
encapsulation a lot. Is this the case or are we missing out on
something? We would appreciate any help on how to get our includes to
work again, as they are pretty numerous.



Attached a working example for orbeon 3.9.
For 4.4 please adapt the page-flow.xml.


Kind regards,
Heinrich

--
You received this message because you are subscribed to the Google Groups "Orbeon Forms" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To post to this group, send email to [hidden email].

Hi Heinrich,

Have you tried using the xpointer attribute on the xi:include (see link below)? This looks like a cleaner way to do it, as it would allow you to include multiple elements, and get a proper resulting file where the xf:instance are directly inside the xf:model.

http://wiki.orbeon.com/forms/doc/developer-guide/processors-xinclude#TOC-The-xpointer-attribute

Alex

Hoi Alex,

thanks for your answer. No we didn't try the xpointer-attribute yet. But
now we do! And it's working perfect.

The doc says that the entire document will be loaded before performing
the xpath. Is there some kind of caching for this document, so that we
can imagine some sort of single library-document which we pull out the
element-fragments where ever we need them?

Cheers

Heinrich

On 01/08/14 03:23, Alessandro Vernet wrote:

--
Before printing this e-mail, think about our environmental responsibility.

--
You received this message because you are subscribed to the Google Groups "Orbeon Forms" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To post to this group, send email to [hidden email].

Heinrich,

Caching depends on which piece of code is using the result of the XInclude transformation. In the best case scenario, caching takes place based on the URLs and last modified dates of all the files used to produce the result.

But if XInclude processing needs to take place, then the document referred to by the xpointer attribute will be loaded every time. You can see this in the code:

http://goo.gl/kMlsHf

-Erik