Hans Moritz Günther

There is a reason for many of those not to be covered by the fixtures (or maybe there is a reason for all of them, but I almost certainly was...

> So, I can create a simplified version of the failing test in sherpa/astro/ui/tests/test_astro_ui_plot.py which should have the advantage of not needing XSPEC to trigger: This is something I think...

It might be easier for me to actually continue from this PR then from mine, since you've already gone through some of the pain of rebasing. Of course, by now...

OK, I may have read that wrong and we might be able to do it. That raises the question: Do we want to? I originally thought "yes" but the more...

related: #895 I think there is a line between reasonable validation and stopping the user from intentionally breaking things. We should to the former (e.g., see #895) it is very...

Now I get a conda error that I can't reproduce locally. Based on some google which brought me to https://github.com/conda/conda/issues/9672 I think might might be due to the fact that...

I could split out the CI changes into a separate PR. This PR is useful in it's own right as it (a) improves some examples in the documentation and (b)...

Note to self: As discussed in the developer meeting: - remove CI aspects, so that PR is only about doc changes - add docs for developers on how to run...

@DougBurke Do you have a chance to review this? I originally had changes to CI in here to run the doctests automatically, but removed them after the discussion we had...

@anetasie Did you check out this branch first? Differences in white space and in the last few digits of a floating number should be ignored automatically through doctest option flags...