How does QTP and other test tools capture objects?
I have built an engine that houses all of the code that would normally go into a script so that there is not as much script maintenance. Now to create a script you just plug in the window name and object name as well as the action to perform into an Excel spreadsheet.
But my question is: I would like to eliminate the need for QTP completely. Of course this means being able to somehow capture a users actions when in record mode. How does QTP and other tools capture the objects and their related properties when recording? I can capture the users mouse movements using the win32 api which amounts to analog mode in QTP (a bunch of coordinates). But how do you actually get the object its acting on and the action its performing?
Are they doing that in C++ or using the win32 api?
Re: How does QTP and other test tools capture objects?
Probably/conservatively, several hundred man-years of development. This is the "family jewels" for all of the company's doing automated test tools. If it was easy we would all be rolling our own.
But many have set off down this path. Search the Automated Test forum for other brave souls.