Peter Sobolewski

Here's my regular pyqt env: ``` ╰─ python test_pyside2.py (napari-dev) ─╯ qt new True pyqt6 False pyside6 False egl False attr True ``` and the env with `"napari=*=*pyside2*"` ``` ╰─...

But apparently just on arm64? Because @andy-sweet has no issues with pyside2 5.15.6 on x86?

If you have some ideas, can jump on a call/island whatever and screenshare.

I'm fine with blocking pyside2 on M1. I got the napari+pyside2 conda-forge combo when using (minus the dry run) `╰─ mamba create -n napari-clean -c conda-forge python=3.9 stardist-napari --dry-run` ```...

This excellent PR from @Czaki will help a lot in terms of allowing sane defaults (numbers and letters?) and allowing the user to make their own. https://github.com/napari/napari/pull/5018

Finally had a chance to play with this—really dig this PR. Once #5018 goes in it will be pure 🔥. While we're making icon order more logical, perhaps Labels should...

Just want to note that PR should close https://github.com/napari/napari/issues/3133

Can you clarify which tests are failing for you locally?

> @psobolewskiPhD It appears it was due to some cluttering of the python environment. Ok. The CI macOS fail seems like the Pint 0.20 fail: https://github.com/napari/napari/issues/5265#issuecomment-1291673288 A fix has been...

@Carreau On main I'm getting a warning when making a Labels layer: ``` /Users/piotrsobolewski/Dev/miniforge3/envs/napari-dev/lib/python3.10/site-packages/skimage/_shared/utils.py:394: UserWarning: Color data out of range: Z < 0 in 20 pixels! ``` Bisect pointed to...