fraud: add peer that sent an invalid fraud proof to black list

[vgonkivs]

Resolves #954

[Wondertan]

Nice!

One last thing we should to care about is testing/mocking implementations of Proofs so that we can test network behaviours, like in this tests, without relying on a specific Proof type.

[vgonkivs]

One last thing we should to care about is testing/mocking implementations of Proofs so that we can test network behaviours, like in this tests, without relying on a specific Proof type.

Yes, I agree. But let's make it in scope of separate PR. IMO, it's not related to this one. Will create and post here an issue.

[vgonkivs]

#971

[codecov-commenter]

Codecov Report

Merging #966 (0896179) into main (da5a5a4) will decrease coverage by 0.02%.
The diff coverage is 83.33%.

@@            Coverage Diff             @@
##             main     #966      +/-   ##
==========================================
- Coverage   58.52%   58.50%   -0.03%     
==========================================
  Files         131      132       +1     
  Lines        7766     7947     +181     
==========================================
+ Hits         4545     4649     +104     
- Misses       2747     2820      +73     
- Partials      474      478       +4     

Impacted Files	Coverage Δ
fraud/service.go	69.56% <76.92%> (+8.69%)	⬆️
fraud/helpers.go	76.92% <100.00%> (ø)
fraud/store.go	64.70% <100.00%> (+2.20%)	⬆️
cmd/flags_misc.go	42.59% <0.00%> (-11.80%)	⬇️
node/settings.go	31.37% <0.00%> (-4.19%)	⬇️
node/services/service.go	80.66% <0.00%> (-2.79%)	⬇️
cmd/env.go	76.47% <0.00%> (ø)
service/state/core_access.go	18.32% <0.00%> (ø)
header/metrics.go	0.00% <0.00%> (ø)
ipld/plugin/nmt.go	41.53% <0.00%> (+1.43%)	⬆️
... and 4 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[Wondertan]

Nice!

Further #971 should prettify things a bit.

[renaynay]

Everything looks generally okay but I don't understand the reasoning behind adding the from peer.ID param to the validator func instead of just getting the peer from the msg.ReceivedFrom field -- what does this change?

Also need to go through tests - they are quite intense so i'm gonna do another pass through just for the test

[renaynay]

So the tests are quite long and hard to follow what's going on and why - it would be nice to clean them up a bit, reduce LOC as much as possible (for example, no need to define bServ if it's only used in one place, just pass mdutils.Bserv() directly as a param). The tests inside sync_test.go are really nice to follow, for example.

[vgonkivs]

msg.ReceivedFrom field is the address of node that have created a Fraud Proof. But it could be a case when other node receives a Fraud Proof and regossips it further

@renaynay