I am trying to run a Sikuli script in headless mode on a Ubuntu docker on which I installed Sikuli. The Sikuli script for now on, is rather basic. It starts emacs and ends it by selecting the close icon of the emacs window.
The Sikuli script works nicely if I run it on my local session.
If I try to run this Sikuli script on the Ubuntu Docker on which I installed Sikuli, it simply fails to identify the close icon of the emacs window.
Here is how I run the Sikuli on the docker: xvfb-run --listen-tcp --server-num 44 --auth-file /tmp/xvfb.auth -s "-ac -screen 0 1920x1200x24" runsikulix -r bla.skl &
the execution of the bla.skl script stops when it tries to find the close image. [error] FindFailed ( 1482778281253.png: (34x29) in S(0)[0,0 1920x1200] E:Y, T:3.0 )
I checked that indeed the emacs window starts correctly. To do this, I recorded my Xvfb session with the ffmpeg utility and I can see the Emacs windows.
Does someone have some experience with running Sikuli in headless mode on a docker. I will be grateful for any suggestion on how I should proceed here to have a more robust approach for running Sikuli scripts in headless mode and ensure that image based identification is functional. thanks