parent starting down solve too early bug

[dehann]

When solving Caesar1D or Hex with tree init, clique 1 starts downsolve before clique 2 completes it's upsolve

[dehann]

Hi @Affie , I was wondering if you could have a go at this issue perhaps please? Should be a basic redirect in old fetch CSM when doing hex init. You can just work from the current testBasicTreeInit.jl test file?

Aside: I think most of #954 will only happen in v0.17 timeframe. That is a larger job than I previously thought -- but at least a bit more standard CSM structure as part of the #754/#910 effort.

[Affie]

Should be a basic redirect in old fetch CSM when doing hex init

I didn't quite understand this comment but went with something similar to the take! CSM.
see #958