Fix more incorrect dom pointers with sCU

[marvinhagemeister]

We missed a few cases where it would still keep old DOM around. One test is still failing.

Fixes #1864 🤞

[mzealey]

The issue I reported at bottom of #1871 still manifests with these changes.

[coveralls]

Coverage remained the same at 99.881% when pulling 30aaa1b on scu-fix-2 into 372c62f on master.

[mzealey]

I've found another similar issue with a node (material UI button) suddenly reappearing in -rc2 and -rc3 but it is a bit complex to trigger even within my application. I think there are still some edge cases missing... In this case I have a <div> with some buttons under which are conditionally rendered. Sometimes the non-conditional ones switch to being first and the conditional ones go to the end of the div. And then sometimes they duplicate after this.

[mzealey]

@marvinhagemeister was the above issue with the patch addressed prior to closing - i don't see any additional work on this?

[marvinhagemeister]

@mzealey yup, the same fix was included in #1888, which has landed in master already 🎉