Hi All!
I've been working on a process for some time now that utilizes multiple objects to interact with an application that runs through Chrome. It is an internal system that we access through a URL in Chrome. As you work the screens in the application, various pop up windows appear. I had the bot working relatively well and then we upgraded from BP 6.4.2 to BP 6.10.1 and everything went sideways. My objects can no longer see any elements on the pages of the "pop up" windows. They're not little pop up boxes - they're the guts of the application. Clicking a link on one screen opens a new window, then the bot needs to click something there to expand and then click again to open another window... you get the point. I had all this broken out into separate objects with proper attach pages and it was working. Now on the upgraded BP when I attach, the bot can only "see" the outer Chrome window and Chrome navigation and those only on Win32 or UIA mode.
I have tried adding --force-renderer-accessibility to my initial chrome window launch but that did not help. I cannot add this on the subsequent objects because they don't "launch" anything, just attach to existing windows.
I tried increasing the browser timeout in the Automate config file as recommended in the release notes, but that did not help.
I tried adding a child index of 0, 1, or 2 but saw no change.
I have tried gathering all the chrome process IDs and looping through each, then "waiting" for a known element on the screen to confirm I'm attached correctly, but the bot was never able to successfully find the element.
I tried manually launching and navigating to the first pop up window, then attaching via standard attach or PID as described above but received the same result as before.
No matter what I do, the bot cannot find any elements on the subsequent screens. I would like to close the initial window so I only have one open, but that logs me out of the application so I can't do that, either.
I have confirmed that the Chrome extension is installed, enabled, and functional (because other websites work and even the initial page for this site works, too).
One other piece of information - This site is a .jsf, but responded to standard browser automation on version 6.4.2. The vendor for this application insists that it is "bot friendly" and until now it seemed as though it was. Because all of the things I interact with are through chrome, I have all my modellers set to Browser/Chrome/Firefox etc. Switching to Java automation does not work at all.
I would appreciate any assistance, suggestions, or guidance you can offer.
------------------------------
Diane Sanzone
------------------------------
