Eric Larson

From debugging with @olafhauk we determined that this is an issue with MESA 18.3.4 and there isn't much we can do there but wait for ancient machines to be upgraded....

> How long does the line profiling take? Just to make sure that we don't add much overhead... The overhead seems to be negligible in testing so far > @larsoner...

Yeah, not all tests pass when running in profile mode, which I'd like to fix

Yes this tutorial has always been the slowest at 2-3 minutes. I don't think this part is new. Any way to speed it up would be helpful! Keep in mind...

Failing test is unrelated and being tackled in #11034. Thanks for looking into this @Dod12 , I'll go ahead and merge! If you're feeling motivated, you could add similar tests...

> We will also need to find a way to allow you to specify the S-D distance manually (for beer-lambert law), as you don't have accurate spatial information in the...

Sure we could support `surf_maps=None` with these other parameters I guess. I think it's done this way because it can be a bit slow and you might want to reuse...

Okay looks like the failures thus far are just due to an `eeglabio` update. I'll restart this once we figure out how to deal with that

macOS failure is probably https://github.com/conda-forge/conda-forge-repodata-patches-feedstock/issues/310, a bug a caused by me messing up our recipe... hopefully I can fix it soon

@agramfort I added your shims for OpenMEEG-based BEM, I/O, and forward computations, with some tests that you can use when you start trying to port over code. PyPi and conda-forge...