[
https://issues.apache.org/jira/browse/MYFACES-2629?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12851509#action_12851509
]
Lewis Gass edited comment on MYFACES-2629 at 3/30/10 7:47 PM:
--------------------------------------------------------------
After more investigation, it turns out this problem also comes up in the SUN
RI. I sincerely believe this should have been part of the public API in some
way, but its probably too late for that. In any case, this just means that each
integration library I write for each respective JSF implementation will be more
involved. I have already did preliminary integration with the SUN RI
FaceletContextImplBase class and its working now with the TemplateClient
internal API, which seems to be replicated between both the SUN RI and MyFaces
implementations. TemplateClient, or at least the basic concept of the insert()
and define() UI tags seem to belong really in the public API.
However, for the moment, unless you all just want to make the above proposed
changes for other reasons, I will have to integrate by all means with your
FaceletContext implementation. In fact if you put the TemplateClient methods
(pushClient(), popClient() and extendClient(), includeDefinition()) somewhere
else it will break the ability of Gracelets to get involved in that process,
since Gracelets views can be used as TemplateClient's.
One thing I would ask is if you made the DefaultFaceletContext public, and not
package private, so that I do not have to reinvent the AJAX based extensions
you have added to the AbstractFaceletContext.
was (Author: elponderador):
After more investigation, it turns out this problem also comes up in the
SUN RI. I sincerely believe this should have been part of the public API in
some way, but its probably too late for that. In any case, this just means that
each integration library I write for each respective JSF implementation will be
more involved. I have already did preliminary integration with the SUN RI
FaceletContextImplBase class and its working now with the TemplateClient
internal API, which seems to be replicated between both the SUN RI and MyFaces
implementations. TemplateClient, or at least the basic concept of the insert()
and define() UI tags seem to belong really in the public API.
However, for the moment, unless you all just want to make the above proposed
changes for other reasons, I will have to integrate by all means with your
FaceletContext implementation. In fact if you put the TemplateClient methods
(pushClient(), popClient() and extendClient(), includeDefinition()) somewhere
else it will break the ability of Gracelets to get involved in that process,
since Gracelets views can be used as TemplateClient's.
> Accept abstract FaceletContext, do not force AbstractFaceletContext
> -------------------------------------------------------------------
>
> Key: MYFACES-2629
> URL: https://issues.apache.org/jira/browse/MYFACES-2629
> Project: MyFaces Core
> Issue Type: Improvement
> Components: General, JSR-314
> Affects Versions: 2.0.0-beta-3
> Environment: Tomcat 6.0+, MyFaces 2.0.0-beta3 API/Impl.
> Reporter: Lewis Gass
> Assignee: Leonardo Uribe
>
> I am the main coder on the Gracelets project
> (http://gracelets.sourceforge.net/) and have recently began integration of
> Groovy with JSF 2.0. In order for Gracelets to harness the already existing
> Facelets libraries it needs access to the TagLibrary class and the actual
> libraries loaded by the JSF 2.0 implementation. Since that library is not
> part of the JSF 2.0 public API, I have to write an extension for each
> different JSF 2.0 implementation in order to load them. I have been able to
> successfully integrate with the SUN RI with minimal code. However, in MyFaces
> Core implementation this code appears on line 135 of the
> org.apache.myfaces.view.facelets.tag.jsf.ComponentTagHandlerDelegate:
> AbstractFaceletContext actx = (AbstractFaceletContext) ctx;
> Gracelets has its own FaceletContext (which is part of the public API) in
> order to mimimize integration between different JSF 2.0 implementations.
> Since in MyFaces this is forced to be a particular sub class here, it breaks
> portability. Is there anyway this can be avoided?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.