In our ActiveX application tests, sometimes the default state runs and decides that the main form doesn't exist - yet it plainly does. Then it invokes another one (successfully) and hell breaks loose. (OK, it's not that bad but it means all our subsequent tests don't run)
Anyone experienced this and found some kind of work-around?
I feel like I'm the only person running silktests against ActiveX - every time I post on the Borland boards I'm told 'very interesting, open a support ticket.' but we are on limited support, and they can't even resolve our biggest issue, with SetActive not working on a subset of our systems. Aaaaahhhh!
I don't feel like this is a real timing problem, though - I am running the test and I see the test machine in a completely clean and ready to go state.
I grant that Windows could be taking a mini-break, though, and not responding. However, when physically using the machine I don't notice it locking up for 3 minutes at a time.
That begs the question, though - what is going on if it's not timing? It feels almost as if the application stops responding to some operating system calls. I've seen a lot of strange behavior with the OS interacting with our application. (SetActive just causes the app's bar on the taskbar to flash, We can not create new registry entries from silk, but we can do everything else, that kind of thing)
From your description, it sure seems like environmental issues; maybe the test system image has problems? Does this happen on every test machine that you have configured? Is your test machine controlled by a domain (I'm thinking rights issues)?