Just pass the tests

[malik-irain]

Finally solved!

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 64.28%. Comparing base (3a1a81b) to head (1499d2a).
⚠️ Report is 14 commits behind head on 5.0.x_dev.

Additional details and impacted files

@@              Coverage Diff              @@
##           5.0.x_dev     #113      +/-   ##
=============================================
+ Coverage      64.24%   64.28%   +0.04%     
=============================================
  Files             73       73              
  Lines           9137     9156      +19     
=============================================
+ Hits            5870     5886      +16     
- Misses          3267     3270       +3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[seb5g]

could you tell what you did a bit more explicitly?

[seb5g]

if it's not inheriting from QObject one cannot create and use signal. I guess because it's a mixin, it's no big deal if used within another class inheriting from QObject? Or maybe that was the issue?

[malik-irain]

As it's a Mixin class it's not supposed to be used by itself and it's actually used by another class inheriting QObject. So in the end, QObject was initialized twice. When done in normal execution it's fine (or at least it doesn't crash) but in pytest it seems to be the source of segfaults.

I tried several ways to solve the problem but this was the only functional one.

[seb5g]

ok understood! THere are still tests not working ;-)

[malik-irain]

One problem I have and can't seem to solve is with the test search_lineedit_test.py::TestSearchLineEdit::test_debounce_resets_on_new_input
https://github.com/PyMoDAQ/pymodaq_gui/blob/5.0.x_dev/tests/utils/widgets/search_lineedit_test.py#L47-L56

@Ashwolaa How is it different than the one directly before, test_rapid_changes_emit_once in what's tested. Both seem to test that only the last set text is taken into account as long as it done before debounce timeout, no?

[Ashwolaa]

One problem I have and can't seem to solve is with the test search_lineedit_test.py::TestSearchLineEdit::test_debounce_resets_on_new_input https://github.com/PyMoDAQ/pymodaq_gui/blob/5.0.x_dev/tests/utils/widgets/search_lineedit_test.py#L47-L56

Hum strange. With PyQt6, the tests are passing on my config (fresh install though).

@Ashwolaa How is it different than the one directly before, test_rapid_changes_emit_once in what's tested. Both seem to test that only the last set text is taken into account as long as it done before debounce timeout, no?

Yeah, that is correct. One is just checking by stacking all the signals while the other is only waiting for the first one to come but they are roughly equivalent. We can remove/merge it if you prefer.

[malik-irain]

Yes, the issue only appear with PySide6, but I'm not sure the test is wrong, because even when replacing it with pass, doesn't avoid triggered the Qt Error (RuntimeError: Internal C++ object (PySide6.QtWidgets.QWidgetAction) already deleted.) . So I'm thinking maybe some Qt Object isn't attached correctly in the implementation

[malik-irain]

Should be good, everything pass now!

[Ashwolaa]

You fixed pyside6!
Nice job !