Hi
2017-01-19 16:08 GMT-05:00 arjan tijms <arjan.tijms_at_gmail.com>:
> Hi,
>
> On Thu, Jan 19, 2017 at 7:28 PM, Leonardo Uribe <leonardo.uribe_at_irian.at>
> wrote:
>
>> - https://java.net/jira/browse/JAVASERVERFACES_SPEC_PUBLIC-1260
>> Extensionless paths: It was discussed a year ago, but no idea how it
>> looks like.
>> Some things added to the spec. Was META-INF/views (specific directory
>> for views) approved?
>>
>
> Unfortunately a fully featured out of the box extensionless support with
> nice convention folders such as WEB-INF or META-INF/views was not realised.
> For that you'd still need to use PrettyFaces or OmniFaces.
>
>
I see. I would like to have META-INF/views because this is one missing
point to modularize JSF applications.
JSF has resource library contracts to handler resources, templates or
composite componentes, faces flows to handle modules with well defined
entry and exit points, but we do not have something to handle "sets of
views".
In this moment, with the current spec, you can't bundle views, manage beans
and configuration files in the same jar, something that looks too useful to
ignore. But the issue is more complex than just bundle a bunch or files. It
is how to group modules in a JSF app in a way that makes sense. For
example, If there is a menu with each module, how to register each module
into the menu in a declarative way, without write all modules in a list.
The point is how to make it easy, without write the same generic code in
JSF apps over and over.
> 1260 was broken up from the bigger issue, and it just gives JSF a basic
> awareness of "exact mapping" next to the prefix and suffix mapping that it
> knows of today.
>
> This means that if you map the FacesServlet via a servlet-mapping (in e.g
> web.xml) to a URL pattern like "/something", and then invoke say
> localhost:8080/theapp/something, then JSF can handle that. The spec PDF
> for ViewHandler and the JavaDoc for ResourceHandler detail what JSF should
> do.
>
>
Ok, good to know that. Thanks for mention it.
> To more directly answer Manfred's question; an extra unit test still has
> to be committed for this (I have a basic one locally, but not suited to be
> committed at the moment). I started with migrating the so-called Cactus
> tests for the ViewHandler and ResourceHandler but this appeared to be way
> more work than anticipated. Cactus is somewhat difficult to comprehend at
> time.
>
> I'll make sure to commit this last test before 27/01/2017.
>
> Kind regards,
> Arjan Tijms
>
regards,
Leonardo Uribe