We are using oracle flexcube and we want to automate some stuff.
The way flexcube works is by opening a single browser tab and INSIDE it open multiple virtual windows. Each window is its own HTML component. So you have:
<html>
....
<html>component 1</html>
.....
<html>component 2</html>
.....
</html>please note that the html tag is in under a “document”
the “windows” are dragable and can overlap (z-axis overlap i guess).
Flexcube automatically opens the “windows” with a prefixed size and position, so most of the time, the window headers titles and minimize/close buttons overlap perfectly.
At runtime, when Record Capture tries to click a button, it looks in the first inner <html> tag even if the focus (max z-axis available index) is on another window, which is fine since I am using HTML, but it is unintuitive for other users. It always ends up finding the wrong button.
I have tried removing DOMxPath, and path. I also tried adding OuterHTML and HTML InnerText, but no dice.
The click command does occur and it clicks in a consistent place, not in random places. I am not sure if it is clicking the top window only, or whether it is hitting all stacked windows in the same place but on different z-axis. I can’t predict that effect so i can test.
if i pause the bot and close other windows and leave the only window i want, the bot continues as intended.
I have no idea from even where to start. Any advice or help would be very much appreciated