* improve logging
* [consensus] do not try to use negative index
* Revert "do not try to use negative index"
This reverts commit b434fd3f4af39f32650e909292cc0123bedba86e. We have
fixed the cause of the issue, which was time drift on a new cloud
provider's nodes. See `systemd-timesyncd.service`
Even if this fix had been merged, it would likely not have solved the
problem given those nodes with the correct time would pick a different
leader from those with time drift. Or, in other words, the view change
would not have gone through.
* improve logging
* chore(build): upgrade golang to 1.19
* chore(build): run `go mod tidy`
* chore(build): run `goimports -w -e ${file}`
* chore(build): revert github ci changes
* chore(build): pin golang version to 1.19.5
* chore(build): fix protoc version on gen files
* fix: elastic rpc new filters fix
* fix: add a small delay to give readers time to catch up
* fix: run the subscribe in a go routine
* fix: generate the filter id before
* fix: rpl not able to marshal negative big int
* fix: use json to get the message
* Resolve harmony-one/bounties#77: Staking precompiles
Create write capable precompiles that can perform staking transactions
Add hard fork logic (EpochTBD) for these precompiles
Tests for new code with at least 80% unit test coverage
Staking library + tests in MaxMustermann2/harmony-staking-precompiles
* Fix small typo in comment
* Run goimports on files to fix Travis
* Do not activate staking precompile on shard 0
* Cascade readOnly to WriteCapableContract
* No overlap in readOnly + writeCapable precompiles
* Use function selector instead of directive
From Solidity, use abi.encodeWithSelector and match it against the
exact ABI of the functions. This allows us to remove the need for
a directive (32) being encoded, and thus saves 28 bytes of data.
* Do not allow contracts to become validators
As discussed with Jacky on #3906
* Merge harmony-one/harmony/main properly this time
* Run goimports
* Update gas calculation for staking precompile
Please see comment in core/vm/contracts_write.go RequiredGas
* Do not allow contract to become validator (2/2)
* Cache StakeMsgs from precompiled transactions
Add the StakeMsgs to ProcessorResult and cascade them in insertChain
* Remove ContractCode fields from validators
Since smart contracts can no longer beecome validators,
this field is superfluous. Remove it from the Wrapper
structure, and do not assign it a value when creating
a validator. Build and goimports checked
* Update comments in response to feedback
(1) Comments to start with function names
(2) Comments for public variables
(3) Comment to match function name RunPrecompiledContract
(4) Clarify that CreateValidatorFunc + EditValidatorFunc are still used
* Fix Travis build by reverting rosetta change
* Add revert capability to 3 staking tx types
- Delegate
- Undelegate
- CollectRewards
* Fix build: Update evm_test for ValidatorWrapper
* Merge main into harmony-staking-precompiles
* Add gas for precompile calls and allow EOA usage
- Each time the precompile is called, charge the base gas fee plus data
cost (if data can be parsed successfully). A gas fee is added to
prevent benevolent contract deployers from subsidizing the staking
transactions for EOAs through repeated assembly `delegatecall`.
- Allow EOAs to use the staking precompile directly. Some changes to
the Solidity library are associated with this change.
- Remove bytes from parsing address, since the ABI unpacks it into an
address format correctly.
- Add or update tests. Test coverage report to be attached to the PR
shortly.
* Run goimports
* Check read only and write capable for overlap
* Handle precompile stakeMsgs for block proposer
The staking precompile generates staking messages which are cascaded to
the block via the EVM in `state_processor.go`. This change cascades them
in `worker.go` to allow block proposers and block verifiers to keep the
same state.
* Run goimports for cf2dfac4081444e36a120c9432f4e..
* Update staking precompile epoch to 2 for localnet
Bring it in line with staking epoch. Change effects all configurations
except mainnet and testnet. `goimports` included.
* Add read only precompile to fetch the epoch num
* Move epoch precompile to 250
* precompiles: left pad the returned epoch number
* chainConfig: check epochs for precompiles
panic if staking precompile epoch < pre staking epoch
* Add staking migration precompile
- Lives at address 251
- Migrates delegations + pending undelegations from address A to B
- Useful if address A is hacked
- Charges gas of 21k + cost of bytes for two addresses
- Does not remove existing delegations, just sets them to zero.
Replicates current undelegate setup
- Unit tests and `goimports` included. Integration test following
shortly in MaxMustermann2/harmony-staking-precompiles
* Migration precompile: merge into staking
Merge the two precompiles into one, add gas calculation for migration
precompile. Move epoch precompile to 251 as a result. When migrating,
add undelegations to `To`'s existing undelegations, if any match the
epoch.
* Add migration gas test, remove panic, add check
In response to review comments, add tests for migration gas wherein
there are 0/1/2 delegations to migrate. Add the index out of bound check
to migration gas calculator and remove panics. Lastly, re-sort
migrated undelegations if no existing undelegation in the same epoch was
found on `To`.
* Move undelegations sorting to end of loop
* Add revert mechanism for UpdateValidatorWrapper
Closesharmony-one/bounties#90
(1) Use LRU for ValidatorWrapper objects in stateDB to plug a potential
memory leak
(2) Merge ValidatorWrapper and ValidatorWrapperCopy to let callers ask
for either a copy, or a pointer to the cached object. Additionally,
give callers the option to not deep copy delegations (which is a
heavy process). Copies need to be explicitly committed (and thus
can be reverted), while the pointers are committed when Finalise
is called.
(3) Add a UpdateValidatorWrapperWithRevert function, which is used by
staking txs `Delegate`, `Undelegate`, and `CollectRewards`. Other
2 types of staking txs and `db.Finalize` continue to use
UpdateValidateWrapper without revert, again, to save memoery
(4) Add unit tests which check
a) Revert goes through
b) Wrapper is as expected after revert
c) State is as expected after revert
* Change back to dictionary for stateValidators
Since the memory / CPU usage saved is not significantly different when
using an LRU + map structure, go back to the original dictionary
structure to keep code easy to read and have limited modifications.
* Add tests for validator wrapper reverts
As requested by @rlan35, add tests beyond just adding and reverting a
delegation. The tests are successive in the sense that we do multiple
modifications to the wrapper, save a snapshot before each modification
and revert to each of them to confirm everything works well. This change
improves test coverage of statedb.go to 66.7% from 64.8% and that of
core/state to 71.9% from 70.8%, and covers all the code that has been
modified by this PR in statedb.go.
For clarity, the modifications to the wrapper include (1) creation of
wrapper in state, (2) adding a delegation to the wrapper, (3)
increasing the blocks signed, and (4) a change in the validator Name and
the BlockReward. Two additional tests have been added to cover the
`panic` and the `GetCode` cases.