Owen Parry

+1 We could also add [black](https://github.com/psf/black) for the Python files and [cmake-format](https://cmake-format.readthedocs.io/en/latest/) at the same time? If you're a VSCode user, there are extensions for [clang-format](https://marketplace.visualstudio.com/items?itemName=xaver.clang-format), [cmake-format](https://marketplace.visualstudio.com/items?itemName=cheshirekow.cmake-format) and [black](https://marketplace.visualstudio.com/items?itemName=ms-python.black-formatter) that...

Very weird test failure. On the CI, the integration tests are ostensibly configured with -DCMAKE_BUILD_TYPE=Release (i.e. -O3), but the diagnostics computed in the new test all match the reference data...

@mikekryjak I've rejigged this test so that it applies strict (same double) tolerances when using -O0 or -01, but allows larger differences with -O2, -O3 or -Ofast. I don't fully...

Following advice from the dev meeting yesterday, I'll try and make this test more robust by comparing at steady state instead.

> Can we add an overwrite flag and include it in this simulation for testing purposes? We can then use it to test #399. This seems to increase the test...

Currently I'm running from flat initial profiles with ``` nout = 3 timestep = 5000 ``` Could certainly try restart files instead... I'll take a look at the CMake changes...