REGR: if someone wants to find where this was fixed would be gr8

[jreback]

Regression from 0.16.2 to 0.16.1, but fixed in master, doced here: 2748e82

xref #10979 #9345

Ideally want identify where this happened, and if we have enough test coverage.

[kawochen]

#10625 looks relevant

---

edit: oops I think I read it wrong.

[jreback]

@kawochen yeh that didn't affect this. just trying to see if we are not testing this adequately as the regression slipped thru.

[jreback]

the note is close enough