Remove remaining instances of proto.clone()

[garyschulteog]

TODO(you): choose "part of" or "resolves" and the associated github issue #.

Resolves #4757

---

Description

• use existing Copy* methods from state/getters.go
• add Copy* methods to state/getters.go to remove proto.Clone() from initial-sync
• resolve test breakage fallout
  • most breakage seems due to the different resulting object between proto.Clone and the Copy* methods; the latter yields deeper copies of sparse objects which serialize differently

Write why you are making the changes in this pull request

Write a summary of the changes you are making

Link anything that would be helpful or relevant to the reviewers

[claassistantio]

All committers have signed the CLA.

[terencechain]

Thanks for opening this up!

[codecov]

Codecov Report

Merging #4806 into master will not change coverage by %.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #4806   +/-   ##
=======================================
  Coverage   26.45%   26.45%           
=======================================
  Files         187      187           
  Lines       13348    13348           
=======================================
  Hits         3531     3531           
  Misses       9215     9215           
  Partials      602      602           

[prestonvanloon]

Thanks for doing this @garyschulte. Can I recommend that we merge this PR incrementally? If you have some changes that are ready to go, lets put this PR for review and continue progress in a follow up PR.

[garyschulteog]

Thanks for doing this @garyschulte. Can I recommend that we merge this PR incrementally? If you have some changes that are ready to go, lets put this PR for review and continue progress in a follow up PR.

I removed the use of CopySignedBeaconBlock in ReceiveBlockNoVerify for now. reflect.DeepEquals reports that objects cloned via CopySignedBeaconBlock(block) are identical to those cloned with proto.Clone(block).(*ethpb.SignedBeaconBlock). However, during testing, intial sync always panics at slot 860 during testing if using the copy method rather than proto.Clone there.

I am have been running this branch for a few hours on the test net on a synced node, and am have nearly completely synced from 0 on another machine with this branch. To the best of my knowledge it is safe, but I honestly don't understand the root cause for the panics when using that clone method during sync.

Edit: resolved! so pleased :)

[0xKiwi]

Fixed the suggestions for ya @garyschulte 😄