Joe Germuska wrote the following on 2/18/2005 2:42 PM:
The ActionContext interface defines three maps, requestScope, sessionScope, and applicationScope. We decided that these were logically relevant to any Struts application, regardless of the request/response API. Thus, as long as you're using one ActionContext, you can use these logical maps without being bound to their implementations.
The fact is that the ServletActionContext class uses the HttpServletRequest, its Session, and the initiating Servlet's ApplicationContext to implement these scope maps, but most of your code should be insulated from it.
Does that help? Or did I dodge the real issue?
I believe that helps:) I know some of my questions are lame without having really looked at the source code at all, but I'm wondering if you could do something like this in a Command...
public boolean execute(Context context) throws Exception{ context.setAttribute("someList",someList);
...
}and then from a JSP be able to somehow have reference to this context...
c:forEach items="${handleToConntext.emloyees}"
Can you do the above?
I think using commons-chain is a very cool idea, I just think the reality is you'll end up still seeing much use of the ServletActionContext from within the Command objects which goes against your comment that you'll be able to write "http-free" components as the chain events. I don't really see how you are going to avoid this in a web application. Take for instance something as simple as wanting to create custom ActionMessages based on the result of something that happened in a Command or even worse - across multiple Commands?
-- Rick
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
