QA (public) » openQA Project (public) » openQA Tests (public)

I didn't find anything related in changelog difference with

arch=s390x old=1616 new=1621 base=Beta2; vimdiff http://xcdchk.suse.de/browser/data/SLE-12-SP2-Server/Build${old}/${arch}/ChangeLog-${base}-Build${old}.txt http://xcdchk.suse.de/browser/data/SLE-12-SP2-Server/Build${new}/${arch}/ChangeLog-${base}-Build${new}.txt

The return value of the system call on xdotools might need to be checked but the resizing seems to work, just resizes afterwards.

I will try to reproduce the first bad, then the last good.

http://lord.arch/tests/1390 through http://lord.arch/tests/1394 reproduce this issue on Build1621 (5/5), http://lord.arch/tests/1395 through http://lord.arch/tests/1399 show that Build1616 is still working fine (5/5) so definitely product change that introduced this.

I have tried an SSH installation with build 1621 in a x86_64 VM. I noticed that during initialization the YaST window was closed and then reopened again.
However, the reopened window had the same size as the original window.

I have checked the logs and it turned out that YaST was restarted:

clients/installation.rb:92 Installation ret: restart_yast`

That's because of the self-update feature https://fate.novell.com/319716.

Unfortunately I have no idea why the reopened window has a different size :-(

The Welcome step calls "ensure_fullscreen" (https://openqa.suse.de/tests/446226/modules/welcome/steps/1/src), the same should be called also later (in https://openqa.suse.de/tests/446226/modules/skip_registration/steps/1/src ?)

Please fix this on the openQA side as the YaST solution would be too complicated and require too much effort with small gain in the end... :-(
Thank you!

PR https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/1478 as workaround to set window to fullscreen again in the step when application restarts after self-update. merged.

had to recreate the windowborder needle because of differing font. Initial needle was an icewm no my computer, the worker is running yet another font. https://openqa.suse.de/tests/447597 might prove it's working after the job finishes.