Testnet

[meyer9]

To get the testnet running we have to do the following things:

• [x] Reconfigure validator assignments including block proposer role.
• [x] Setup validators to attest to data when it's their turn
• [x] Share data to committee leader to aggregate
• [x] Share aggregation to block proposer to include in next block

EDIT: a couple more goals before this is finished

• [x] Mempool to keep track of attestations
• [x] Initial sync process
  • [x] Direct peer communication
  • [ ] Ban system
• [x] Process slot transitions separately from block transitions
• [x] Validators should wait for time to create block instead of waiting for slot (allows skipping of blocks)
• [x] Validators should get the proposer skips from the beacon state instead of keeping track of it in the keystore.
• [x] Beacon chain should update slot and epoch transitions based on time, not blocks.

[codecov[bot]]

Codecov Report

Merging #29 into master will decrease coverage by 10.94%. The diff coverage is 40.86%.

@@             Coverage Diff             @@
##           master      #29       +/-   ##
===========================================
- Coverage   62.13%   51.18%   -10.95%     
===========================================
  Files          16       19        +3     
  Lines        2052     2868      +816     
===========================================
+ Hits         1275     1468      +193     
- Misses        627     1210      +583     
- Partials      150      190       +40

Impacted Files	Coverage Δ
chainhash/hash.go	100% <ø> (ø)	:arrow_up:
csmt/node.go	100% <ø> (ø)	:arrow_up:
beacon/mempool.go	0% <0%> (ø)
beacon/db/badger.go	0% <0%> (ø)
primitives/state.go	31.56% <1.26%> (-1.72%)	:arrow_down:
csmt/csmt.go	81% <100%> (+1.87%)	:arrow_up:
beacon/db/memory.go	28.12% <21.81%> (-34.38%)	:arrow_down:
beacon/blockchain.go	47.14% <37%> (-13.28%)	:arrow_down:
beacon/statemanager.go	63.32% <63.32%> (ø)
bls/interface.go	80% <77.77%> (+2.97%)	:arrow_up:
... and 7 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 31c1625...c25b3d9. Read the comment docs.

[meyer9]

OK, so this is close. I'm not quite sure how to guarantee that the proposer got all of the attestations and also the criteria for the proposer to include the attestations in the block.

[meyer9]

OK, so this is basically ready except the sync process and block broadcasting.

[meyer9]

Blocked by #31

[meyer9]

Again, this is pretty close. I want to get the sync process working better and more like Bitcoin before I merge this into master. Also, more tests would probably be good, but also would probably require a couple large refactors.

[codecov-io]

Codecov Report

Merging #29 into master will decrease coverage by 24.39%. The diff coverage is 21.23%.

@@            Coverage Diff             @@
##           master      #29      +/-   ##
==========================================
- Coverage   62.13%   37.73%   -24.4%     
==========================================
  Files          16       20       +4     
  Lines        2052     3230    +1178     
==========================================
- Hits         1275     1219      -56     
- Misses        627     1873    +1246     
+ Partials      150      138      -12

Impacted Files	Coverage Δ
chainhash/hash.go	100% <ø> (ø)	:arrow_up:
csmt/node.go	100% <ø> (ø)	:arrow_up:
beacon/mempool.go	0% <0%> (ø)
beacon/syncmanager.go	0% <0%> (ø)
beacon/db/badger.go	0% <0%> (ø)
primitives/state.go	17.01% <0.65%> (-16.26%)	:arrow_down:
csmt/csmt.go	81% <100%> (+1.87%)	:arrow_up:
beacon/db/memory.go	28.12% <21.81%> (-34.38%)	:arrow_down:
beacon/blockchain.go	46% <43.7%> (-14.42%)	:arrow_down:
beacon/chain.go	52.29% <52.29%> (ø)
... and 9 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 31c1625...e579561. Read the comment docs.

[meyer9]

Sync process is working much better, but there's a crash when initial syncing from a client, quitting, and restarting. I think it has something to do with loading state and processing epoch transitions. I'm fairly sure the epoch transitions are processed twice (once at the end of the imported block, and once at the start. I think we should rewrite the import process.