James Mitchell

Thanks @orlitzky we will prepare a fix

Aha, so this issue is resolved already? If so please close!

Thanks for the contribution @Tianrun-Y, can you please give an example which indicates why we should prefer the new method over the older method?

Can you please add the example to the relevant test file? Then I'll be happy to merge, and thanks very much!

As we discussed, it's not completely clear if this is a bug in Semigroups or in GAP itself, I'll investigate and report again here.

One final comment: shouldn't there be some tests for this?

https://arxiv.org/html/2503.06329v1

Looking at the gss at: https://github.com/ciaranm/glasgow-subgraph-solver it requires C++20 + boost (at least maybe more), which I think rules it out for our use cases at the moment, which is...

Looks good to me @Joseph-Edwards

This will fail in jobs with `--with-external-planarity` until: https://github.com/conda-forge/planarity-feedstock/pull/21 or a successor is merged (i.e. until the latest version of libplanarity is available via conda).