Hans Moritz Günther
Hans Moritz Günther
Isn't this exactly what `load_user_stat` does? Or do I not understand what the issue is that this issue is pointing out? ```Python >>> from sherpa.astro import ui >>> ui.list_stats() ['cash',...
I personally like using a meta class as a parent of the class we define, which will glitter the new class automatically. That that's a breaking change, more complicated, and...
This is almost but not exactly a duplicate of an issue I opened almost, but not exactly 100 issues ago: #1789
Test failure is unrelated; the failing test probably just needs a slightly larger numerical range to pass on all Python and numpy versions and all OS that we test. I'll...
Yes, this is the tolerance for this single, specific test.
A practical note: If we use hypothesis in CI, then it's possible that you see CI failures unrelated to a PR because hypothesis has discovered a new bug. That should...
I vote for post-processing (and using that in a few more of the Sherpa threads as example). My reasoning is mostly to prevent feature-creep and save us the maintenance and...
There already is a line `backend.set_jointplot(0, 0, self.rows, self.cols, create=create)` in there. I think that's where a backend (if possible) joins the axes. And, for the hypothetical second backend that...
That is correct. We always had a dummy backend for datastack (which had it's own IO backend) ,s this really is nothing new. I will end documentation in some form...
Related: #316