I deployed my 4th robot and I during deployment I always came to conclusion that some activities are not very reliable on remote. Maybe I am missing something so please, if you think I am wrong, show me where.
From my experience, some activites I commonly use in my script, that work perfectly during multiple tests come out as unusable when the robot start to run remotly:
- Load picture
- Save picture (screenshot)
- Image exist
- Text exist
These are I think the most common. What is specific for the behaviour is:
-They all work flawless during development and anytime tested on local machine or when loged into remote machine and run robot / script localy
- They fault when run remotly, either as a whole script or in separate conditions - ( to be sure there is truly what we expect to be going on on the screen)
- Boolean returning activities always return False, also if they are in a the loop to check multiple times.
- Image activities fault as “activity failed” or “reference not set to an instance of an object” - where I simultaniously check if object type is truly the one activity expect, which return OK.
- I did set up my resolution via orchestrator to fit the one during development
- I use delays and type of object to be sure there is no null value as an input
- Screenshot after error also approves that specific element is visible on the screen, however activity seem to be blind towards it.
This all come strange to me and I always invest a lot of time into workarounds which are not as specific as I would like to use for my conditioning - for example multiple pop up windows, which text contain is crutial to the decision logic, etc…
Any suggestions anyone?
Thnx in advance!