Re: Storing MaverickContext in request causing issues with portal processes

whoops... accidentally sent this in the middle of typing...

should have said:

it might make sense to have this set as an attribute on the Command nodes rather than as an InitParam, that way you could mix and match this behaviour if/when appropriate.

--jim

  ----- Original Message ----- 
  From: jim moore 
  To: [email protected] 
  Sent: Wednesday, May 07, 2003 3:11 PM
  Subject: Re: [Mav-user] Storing MaverickContext in request causing issues with portal processes


  it might make sense to have this set as an attribute on the Command nodes  rather than, that way you could mix and match this behaviour if/when appropriate.

  --jim
    ----- Original Message ----- 
    From: Schnitzer, Jeff 
    To: [email protected] 
    Sent: Wednesday, May 07, 2003 2:32 PM
    Subject: RE: [Mav-user] Storing MaverickContext in request causing issues with portal processes


    Take it easy folks, this is just the law of unintended consequences working against us :-)



    Clearly this behavior is going to have to be configurable.  I propose an init-param for the Dispatcher servlet called "reuseMaverickContext".  Given the confusion that might result, I believe the default should be "false".



    Thoughts?



    J2FF



    -----Original Message-----
    From: Eric Kreiser [mailto:[email protected]] 
    Sent: Wednesday, May 07, 2003 11:10 AM
    To: [email protected]
    Subject: RE: [Mav-user] Storing MaverickContext in request causing issues with portal processes



    so what you are saying is that you made a design decision to preclude the use of your product in a portal environment

      -----Original Message-----
      From: Doug Kirk [mailto:[email protected]]
      Sent: Wednesday, May 07, 2003 1:42 PM
      To: [email protected]
      Subject: Re: [Mav-user] Storing MaverickContext in request causing issues with portal processes

      On Wednesday, May 7, 2003, at 10:24 AM, Eric Kreiser wrote: 



      The issue is not efficiency in chaining or not chaining - the issue is is it correct to assume that if two controllers are processed using the same under lying httpservletrequest that they are being chained - and the answer is no. 



      While you may consider it incorrect and an assumption, it is not. It is a design decision. Personally, I like this decision better than having to manually chain.