1

If you are familiar with sikuli scripting, you know there is a pair of functions built in called "observe()" and "onAppear()". The Observe function works by essentially starting a thread that continually scans for specific images in a given region of the screen. On the appearance of said image. a handler is called that handles the image (pop up) and ends that specific thread.

In my sikuli script I have about 30 of these observations running at one time (all on different region objects) This works fine most of the time, but I run into a problem when testing an application that uses java. I start to see weird bugs in the application that are only reproducible when I am running the observations. This leads me to believe that the 30 or so observations are slowing down the processor and thus causing issues in the application I am testing.

I am working to cut down the observations I need but even so I would still need at least 20 running during a given test. Is there any way I can allocate more resources for the observations, or is there any other solutions out there that I have not noticed?

switch201
  • 587
  • 1
  • 4
  • 16
  • Try setting the affinity of the java and python process on different CPU cores. This can be done by using task manger, right click process, affinity. – Ayy Aug 21 '15 at 19:55
  • You said that you have multiple regions that you are observing. Are you expecting the same trigger on each ones of these regions? In other words, are you waiting for the same image to appear in each one of them? Thanks – Eugene S Aug 22 '15 at 05:53
  • No each region has it's own unique image it is waiting for. So essentially there are no duplicate regions and no duplicate "onAppear" images – switch201 Aug 25 '15 at 14:32

0 Answers0