Forum Discussion
j_smith
14 years agoOccasional Contributor
The only additional information I've been able to collect since my last question is that the PopupLookUpEditForm object is just underneath the system process in the hierarchy. It seems that with each individual LookUpEdit control from DevExpress, the actual popup part of the control is one object shared with all the LookUpEdit controls throughout the system. The options are just repopulated based off of which control has focus. The alias given to this PopupLookUpEditForm always seems to refer to the instance of this corresponding to the first LookUpEdit control that was focused or interacted with.
Now here is a workaround that I came up with that I would NOT recommend. After recording the test, using multiple LookUpEdit controls, if you follow the statement that calls PopupLookUpEditForm's Click() method with a statement that calls PopupLookUpEditForm's Dispose() method, the PopupLookUpEditForm seems to "recreate" itself each time it is called upon from another LookUpEdit control. This allows my tests to pass, but I do not like this workaround because it already causes my application to run differently under TestComplete than it does by itself at run-time. I would assume it is bad practice to have your test validate an aspect of your application by using the application in a manner that is different from an actual user making the same actions as the test.
Has anyone been able to come up with another workaround for this that still follows best practice?
Now here is a workaround that I came up with that I would NOT recommend. After recording the test, using multiple LookUpEdit controls, if you follow the statement that calls PopupLookUpEditForm's Click() method with a statement that calls PopupLookUpEditForm's Dispose() method, the PopupLookUpEditForm seems to "recreate" itself each time it is called upon from another LookUpEdit control. This allows my tests to pass, but I do not like this workaround because it already causes my application to run differently under TestComplete than it does by itself at run-time. I would assume it is bad practice to have your test validate an aspect of your application by using the application in a manner that is different from an actual user making the same actions as the test.
Has anyone been able to come up with another workaround for this that still follows best practice?