wpbonelli

Thanks @gha3mi can you update the CSVs and README? Also need to make sure this doesn't push the CI test workflow over the 256 job limit. I think for Intel...

> For nvidia, we could test only the latest version and the latest version from each year. For lfortran, we could support only the latest version until it reaches the...

@gha3mi can you update CSVs/README when you have a chance? then this could merge

Hey @tdeyster sorry, should have followed up here before now — see https://github.com/modflowpy/flopy/pull/2357, we decided to deprecate checks until 4.x due to problems like the ones you describe and more

Echoing @jlarsen-usgs's call for comments here, so we can pick a convention and stick to it for the next major version release. One thought: maybe we don't even need a...

Thanks @mjreno really helpful comments, it is not obvious to me which approach is best either — agreed on experimenting in new development before making changes I think besides discretization...

reopen if we determine worthwhile later on

@langevin-usgs this is generally according to our outline ~~with the main difference being `MODEL_XY` and conditioning the release point transform on grid offset/rotation, probably good to chat again~~ switched to...

This should be ready for a look when @aprovost-usgs is back

After chatting with @mjr-deltares we thought this should wait, at least until a non-bugfix release, since it will break API usages. Maybe also consider alternatives. @langevin-usgs