[ 
https://issues.apache.org/struts/browse/WW-1936?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Don Brown updated WW-1936:
--------------------------


As I said, these should be broken up into individual tickets.  Don't worry 
about the implementation at first, just write the ticket from the standpoint of 
the user.  We can then discuss how they should be implemented.  My quick 
comments below:

1) Lyfecycle

- @PreAction and @PosAction annotations
- @PreResult and @PosResult annotation

:: These look useful, but would be best written as an interceptor, and 
therefore a plugin.  We are going to consider some new annotations for 2.1, so 
this might fit nicely.

2) Scope

- Action scope
- Conversation scope
@Persistent (to inject (load and save) scoped variable into actions)

:: This will likely be a larger effort, combined with something like JSR 303 (I 
think), the web beans jsr.  Again, worth a ticket.

3) Configuration

- @Action annotation
- @Result and @Results annotations
- @Interceptor and @Interceptors annotations

:: Take a look at the tickets already against 2.1 and add new ones where 
appropriate

4) Interceptor

- TypeGeneric converter interceptor

:: Sounds good, worth a ticket

5) Injection

- @Inject (to generic injection)
- EJB 3 capabilities to run the code inside and outside an EJB container

:: This I'm not too sure about.  We already have an @Inject annotation.  What 
version of Struts 2 are you using?  Since 2.0.2, we've had our own internal IoC 
engine, based on an early version of Google Guice.  Still, I think user-level 
IoC should be done by something like Spring and not by a Struts-specific 
feature.

6) Ajax

- An embedded and speedy ajax solution and capabilities

:: Struts 2.0.x has the ajax theme and Struts 2.1.0 has an ajax plugin.  How 
would this be different than what we have?

7) Validation

- A java based (without xml) client and server side validation approach.

:: Could this be built on our existing validation system?  I'd probably not be 
in favor of ditching it completely, or I should say, the replacement better be 
damn convincing :)

8) View

- A java based view designer
- A propotype capabilities

We think a JSP taglibs code based is not a good approach to Struts. It cannot 
compete against JSF. We have developed a pure java approach using similars 
Struts taglibs classes, but without JSP.

:: Well, first of all, we aren't "competing" with JSF.  In fact, we have a JSF 
plugin that lets you use JSF components in a Struts 2 application.  How would 
this be different than Struts' current Freemarker, Velocity, and JSP support?  
Definitely worth a ticket as I'd like to hear more.

So again, please separate these features into different tickets so we can hold 
more focused discussions.

> New Java 1.5 features
> ---------------------
>
>                 Key: WW-1936
>                 URL: https://issues.apache.org/struts/browse/WW-1936
>             Project: Struts 2
>          Issue Type: New Feature
>          Components: API
>            Reporter: Jonatas Rodrigues
>            Priority: Critical
>
> I am the leader on a brazilian government project called Atena: an Struts 2, 
> Velocity and EJB 3 extension. We have developed many features that we think 
> could be incorporated into the Struts source, like:
> 1) Lyfecycle
> - @PreAction and @PosAction annotations
> - @PreResult and @PosResult annotation
> 2) Scope
> - Action scope
> - Conversation scope
> @Persistent (to inject (load and save) scoped variable into actions)
> 3) Configuration
> - @Action annotation
> - @Result and @Results annotations
> - @Interceptor and @Interceptors annotations
> 4) Interceptor
> - TypeGeneric converter interceptor
> 5) Injection
> - @Inject (to generic injection)
> - EJB 3 capabilities to run the code inside and outside an EJB container
> 6) Ajax
> - An embedded and speedy ajax solution and capabilities
> 7) Validation
> - A java based (without xml) client and server side validation approach.
> 8) View
> - A java based view designer
> - A propotype capabilities
> We think a JSP taglibs code based is not a good approach to Struts. It cannot 
> compete against JSF. We have developed a pure java approach using similars 
> Struts taglibs classes, but without JSP.
> If it could be intersting, please, write me.
> Jonatas Rodrigues
> [EMAIL PROTECTED]

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to