RFC: Let cluster config/init bubble up exception groups

[elupus]

Let cluster config/init bubble up exception groups

• Only skip errors when re-joining

This shows that we have several faults we currently don't detect properly.

[elupus]

So i've gotten the tests in line i think. There where several cases where our tests where throwing unexpected exceptions due to unexpected responses on data. This exposes those errors and adds some solutions to them.

What do you guys think?

[elupus]

Ps. Some of the fixes should be broken out into separate changes perhaps.

[codecov]

Codecov Report

Attention: Patch coverage is 94.73684% with 1 line in your changes missing coverage. Please review.

Project coverage is 95.67%. Comparing base (dfe5ae7) to head (ebc708c).
Report is 1 commits behind head on dev.

Files	Patch %	Lines
zha/application/gateway.py	80.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##              dev     #120      +/-   ##
==========================================
+ Coverage   95.65%   95.67%   +0.01%     
==========================================
  Files          61       61              
  Lines        9255     9262       +7     
==========================================
+ Hits         8853     8861       +8     
+ Misses        402      401       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[puddly]

I think some of the tests are also failing because we are (slowly) migrating away from using lists (and lists-of-lists) to represent command responses to t.Struct subclasses that concurrently support a tuple interface (rsp[0]) in addition to normal attribute lookup (rsp.status). I think the unit tests right now have a lot of places that return lists-of-lists (return [[]]) that bubble up to code that uses attributes.