Thanks Dale, that was a very informative mail. I used chain result type just for a better separation of actions and till now it did not seem to have any undesirable effect. Anyway I 'll try a new approach to the problem and I 'll avoid using chain result type. Thanks all for the help.
On Wed, May 5, 2010 at 5:55 PM, Dale Newfield <d...@newfield.org> wrote: > I believe that your problem isn't with execAndWait -- it's with the chain > result type. The way that execAndWait works is that there's only one job > runnable at a time in any given session with any given action name. The > intermediate requests don't need the arguments as it's just using the action > name to look up the already-running action. The original thread is > continuing to run and should still have the parameter values from the > launching request. When it completes the next matching request to come in > will be handled with the original (launching) action instance. This means > that the intermediate "Is it done yet?" requests don't need any parameters. > I've basically always found that the chain result type causes many more > problems than it solves. Any reason you can't just make the SUCCESS result > for the execAndWait action the same as what you currently have as the > SUCCESS result for your chained action? > > -Dale >
