Hans Moritz Günther
Hans Moritz Günther
I just noted that I had thought about this before. It's listed as "step 5" in #1375, but this is a more detailed description, so I'll leave both of them...
I have a strong preference to leave the first commit untouched, because it's not hand-written. It's generated using advanced git-foo that I don't exactly understand: It's made by reverting the...
On the other hand, in CI, there is no change, e.g. compare: https://github.com/sherpa/sherpa/actions/runs/4286652711/jobs/7466499212#step:13:120 and https://github.com/sherpa/sherpa/actions/runs/4287930568/jobs/7469345751#step:13:120. This could be changed easily, just by adding doctestplus to the "install-required", but given the...
squashed commits (except the first one), so this should be ready for @DougBurke to have a look again.
About leaving files around locally after running doctests, see discussion in https://github.com/astropy/pytest-doctestplus/issues/68
Should be solved by updates that CM did in the last release.
This is not a PR, just an issue that note the problem, but no, it would not remove that functionality. In that case, you still need to load and assign...
@wmclaugh merged your typo fixes
I'd be in favor of using matplotlibs defaults for simplicity. But, if I remember correctly, some people felt strongly about having astropy-orange in some plot etc. That's (I think) why...
I would say that proposal to remove it all is "under discussion". We're trying to figure out how many people actually use this feature, so what is the impact of...