You know the best way to nerd snipe a tester? Say "Test fails for an unknown reason."
Context: I& #39;ve been poking around the LibrePCB issues, looking for something I could maybe help with... and found this: https://github.com/LibrePCB/LibrePCB/issues/718">https://github.com/LibrePCB/...
Which then lead me to this:
#L13">https://github.com/LibrePCB/LibrePCB/blob/59be81367912a18e01af63197739415f5e8e2d65/tests/funq/librarymanager/test_create_local_library.py #L13">https://github.com/LibrePCB/...
#L13">https://github.com/LibrePCB/LibrePCB/blob/59be81367912a18e01af63197739415f5e8e2d65/tests/funq/librarymanager/test_create_local_library.py #L13">https://github.com/LibrePCB/...
So... of course.. I installed VirtualBox, installed macOS Catalina, set up a dev environment, checked out and compiled the code, ran the tests, and reproduced the same test failure... (Never mind the fact I& #39;ve never debugged a Qt app before...)
Looking at the source code... It& #39;s automating the "Create Library" widget. It& #39;s supposed to fill out the form, click the "Create library" button... and verify that count of projects has increased by one. But the count doesn& #39;t increase, so the test fails...
I noticed that right before the test reported the failure, the UI window was still on screen... So on a hunch, I wondered what would happen if I clicked the button myself while the test was running (after the data is entered, but before the test timeout...)
*Okay, I didn& #39;t actually "get" the little punk. I& #39;ve only found a way to get the punk off my lawn temporarily, metaphorically speaking. *BUT*, I now know the exact pivot point where things start going down hill.