# Bug report ## Summary - [x] I have searched existing issues to ensure this hasn't been reported. **Expected Behaviour**: The unit test looking for axis inversion should find the expected value to start with **Current Behaviour**: The axis inversion appeared to have a different value to what was expected. This is happening in CI on the LabThings repo (see, for example, [CI log](https://github.com/labthings/labthings-fastapi/actions/runs/23746885098/job/69177865354?pr=278) - I apologise if that link dies, I've had some CI sorting to do and ended up doing a few force pushes). This also happened locally on my machine, but went away when I deleted the `openflexure` folder that had popped up in that repo. I have not yet had lots of time to debug, but I have a suspicion that maybe these tests are not using a temp folder for settings correctly, and are picking up an existing settings folder that's causing the problem. ## Configuration <!-- Adjust as appropriate. --> | Component | Value | | :-------- | :------- | | **Server Version** | v3 | | **Release/Branch** | v3-labthings-validate-properties-on-set branch branch | | **UI** | e.g. pytest | | **Operating System** | Ubuntu in WSL2 on Windows 11, `ubuntu-latest` CI runner on GitHub | | **Python Version** | CPython (3.11.15) | | **Camera Version** | simulation | | **Board Version** | simulation | ## Steps to reproduce I ran the workflow documented at [CI log](https://github.com/labthings/labthings-fastapi/actions/runs/23746885098/job/69177865354?pr=278): it happens when running `pytest` locally. **Reproducibility:** - [x] **Consistent:** Happens every time. - [ ] **Intermittent:** Happens roughly [X]% of the time or every [X] scans/clicks. - [ ] **One-off:** Only happened once so far. ## Relevant logs and/or screenshots [CI log](https://github.com/labthings/labthings-fastapi/actions/runs/23746885098/job/69177865354?pr=278) ## Additional details