QA (public) » openQA Project (public)

So you want to skip the assert_screen/check_screen timeout if you know it is going to fail anyways. That makes sense.

I would add a button "Skip current assert_screen timeout" which would be shown if an assert_screen is ongoing. It only affects the current assert_screen (and not subsequent ones). After clicking the button, the same happens as on a regular timeout. The button itself would disappear and 'Open needle editor' would appear as soon as the latest screenshot has been uploaded.

It's not only to skip the waiting time on an assert_screen which you just expect to fail. I am not arguing with you about the code quality of the old interactive mode but what it promised was what many people including myself here still would like to see. In this case: I just want to pause the test execution as soon as possible after I requested it. One use case is to create a needle from the current screen, e.g. to prevent further waiting or to catch exactly the screen at the time before it changes again, e.g. because of screenlock or a dialog that times out. Another use case is that I want to interact with the system outside the context of openQA, e.g. VNC remote control, at this time, not later, e.g. while the installation is currently running.

I just want to pause the test execution as soon as possible after I requested it.

Pausing as soon as possible in general is another feature but shouldn't be hard to implement. Emphasis on "as soon as possible" which would be in general on the next command.

mkittler wrote:

Emphasis on "as soon as possible" which would be in general on the next command.

That's too late, it should be "abort the currently running timeout".

@okurz I was talking about the general case, e. g. we might not be able to interrupt an assert_script_run. Of course the timeout for assert/check_screen can be reduced to zero and that is also part of this specific issue.