Besu's Retesteth implementation failed transition tests. Two reasons:
* It wasn't adding the DaoForkBlock to genesis files correctly
* It wasn't using the protocol schedules og the block being imported,
but instead to the current chain head.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* #1066 Switched to use unprefixed hex strings for memory and stack values
Signed-off-by: David Mechler <david.mechler@consensys.net>
* Disable flaky tests per Ben Burns(Yeti) request
Signed-off-by: David Mechler <david.mechler@consensys.net>
* Revert last commit and enable ignored tests.
Signed-off-by: David Mechler <david.mechler@consensys.net>
* #1157 - updated to create 2 agents so that proper bonding can occur
Signed-off-by: David Mechler <david.mechler@consensys.net>
* #1162 - Updated test to mock the local peer PING packet creation so that the hash can be managed.
Signed-off-by: David Mechler <david.mechler@consensys.net>
Occasionally, a callback given to a subscriber will complete a future
which then runs another task which adds another callback to the
subscriber list of the same type. By the time the first callback is
finished, the new callback (which presumably wants another "real"
instance of the event to happen) could be next up in the list of the
first's subscribers callback iteration. This takes a snapshot of the
subscribers when the desired event happens so there's no risk of
following down the rabbit hole.
I collected some rough data about the size of the subscriber lists to
see the penalty for copying them and I found that by the time a full
sync began or a fast sync had started downloading headers, the maximum
subscribers any one Subscribers class had was 7.
Signed-off-by: Ratan Rai Sur <ratan.r.sur@gmail.com>
Our current EVM loop splits out cost, execution and halt evaluation. Flattening these three into one method results in a significant speedup in reference tests, which are EVM heavy.
Mostly the cost, exceptionalHaltCondition, and execute are merged into one method, sharing calculations between the three.
In fixed cost operations the merger is very simple AddOperation is a representative fixed cost operation.
Check the gas, check other exceptional halts, do the work, return the result.
In variable cost operations there is some value reads to be done before the cost is calculated. This is where a lot of the de-duplication occurs. CodeCopyOperation is a representative variable cost operation.
JumpIOperation shows where this merger pays off. If the condition is zero then some exceptional halts don't need to be considered. But with the three way split each step couldn't consider such optimizations because the local data was lost between each call.
Some cleanup was enabled by this. The old exceptional halt predicates were deleted and moved into each operation. Gas costs must be checked by the operation instead of globally, or we would lose state if we had to split into two methods and do the gas check shared.
The OperandStack was flattened into a single class instead of an interface plus a single implementation. stack underflow and overflow are signaled via named exceptions and handled via catches instead of pre-checking the stack height. Since overflow/underflow is exceedingly rare in mainnet transactions java exceptions are the more performant means.
Some of the APIs had lingering impacts on how some tests were run and the EVMTool (we cache operation cost in the message frame now.)
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Update to the reference tests used on hivetests.etherdevops.io and
retesteth.etherdevops.io. Both are using a more current release of the
reference tests that includes performance and subroutines items.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Previously we were returning success with a zero length hash, meaning we
understood the response but there were problems with the RLP. Instead
now we return one of two errors "Could not decode RLP for Block" or
"Could not import Block".
For #1130
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Produce error response if any of Address, Bytes or Bytes32 values
inside GraphQl requests are not prefixed with 0x as defined by the spec.
Fix old tests that used values that are not prefixed with 0x and add
additional tests to confirm Address, Bytes and Bytes32 values will be
rejected if not prefixed with 0x.
Signed-off-by: Mak Muftic <mak@nodefactory.io>
- Add compatibility with ClusterIP services in the kubernetes nat manager
- Add new Xnat-method-fallback-enabled flag
Signed-off-by: Karim TAAM <karim.t2am@gmail.com>
This PR fixes an error when downloading chain (Async operation failed)
When there are several blocks which are very large (> 12M) and which are requested in the same segment (by default 200) we can have timeouts and never manage to synchronize. This modification will make it possible to gradually reduce the size of the segment with each attempt. Then the segment resumes its default size for the next blocks
If the reduction is not enough at the last attempt we try with a single block
Signed-off-by: Karim TAAM <karim.t2am@gmail.com>
* use P2P consistently in help
Signed-off-by: Sally MacFarlane <sally.macfarlane@consensys.net>
Co-authored-by: mark-terry <36909937+mark-terry@users.noreply.github.com>
* Add check for spoofed IP in ping message
* Add logging message when ping request is rejected.
Signed-off-by: David Mechler <david.mechler@consensys.net>
The IBFT2 configuration section of a genesis file has been updated
to allow the specification of a block reward (defaults to 0) and also an explicit
recipient (defaults to the block proposer).
The block reward can be specified as a hex string (with 0x prefix) or a decimal
string (no prefix), and is defined in Wei.
These values are not modifiable for the duration of the network.
Signed-off-by: Trent Mohay <trent.mohay@consensys.net>
- Added `--rpc-tx-feecap` command line flag.
- Maximum transaction fees (in Wei) accepted for transaction submitted through RPC.
- Defaulted to 1 ether.
- Updated `TransactionPool.addLocalTransaction` method: performs an additional check to verify if transaction fees don't exceed user defined fee cap (if cap is set to 0 then it is ignored and means no cap).
Signed-off-by: Abdelhamid Bakhta <abdelhamid.bakhta@consensys.net>
* rename more whitelist occurrences; change allowlisted to allowed and reword where we ended up with allowlisting
Signed-off-by: Sally MacFarlane <sally.macfarlane@consensys.net>
* change Perm*Whitelist JSON-RPC methods to extend the Perm*Allowlist methods to remove dupe
* added unit test for Perm JSON RPC methods collection
Signed-off-by: Sally MacFarlane <sally.macfarlane@consensys.net>
* Remove ROADMAP.md and link Wiki Roadmap in README.md
Signed-off-by: Tim Beiko <t.beiko23@gmail.com>
* Update changelog to add EIP 2315
Signed-off-by: Tim Beiko <t.beiko23@gmail.com>
* Update Changelog for 1.5
Signed-off-by: Tim Beiko <t.beiko23@gmail.com>
* Remove developper advocate contact
Signed-off-by: Tim Beiko <t.beiko23@gmail.com>
* Remove ROADMAP.md and link Wiki Roadmap in README.md
Signed-off-by: Tim Beiko <t.beiko23@gmail.com>
* Update changelog to add EIP 2315
Signed-off-by: Tim Beiko <t.beiko23@gmail.com>
* Update Changelog for 1.5
Signed-off-by: Tim Beiko <t.beiko23@gmail.com>
Internally we used to use an enum set to track halt reasons and we would
track multiple halt reasons. However, what the halt reason is does not
matter to reference tests and tracing, only that a halt occurred.
Repalace the EnumSet with an Optional and trace only one revert reason.
This saves us time in enumset management and also allows us to return
quicker once any halt is detected.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Upgrade to ErrorProne 2.4.0
* public constructors on abstract classes are removed
* Javadoc must have meaningfull documentation
* lambdas should not be variables
* Added to the list of confusing inner class names (Entry and Type)
* no assert keyword in tests
* Obsolete JDK classes produce errors now
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Upgrade Library Versions
Upgrade library versions to most recent version
Exceptions:
* Errorprone - lots of new violations need fixing
There will be a follow on PR for this.
* picocli - non-trivial conformance changes
* EthSigner&Orion - similar non-trivial API changes, plus it's only for
integration testing
* Kubernetes - significant library revisions
* Web3j - Web3j changed some error semantics in 5.0 that broke
acceptance tests. Expect an update in a follow on PR.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
MadelineMurray
Antoine Toulme
Danno Ferrin
David Mechler
Ratan (Rai) Sur
Lucas Saldanha
matkt
mace
Sally MacFarlane
Usman Saleem
magooster
Abdelhamid Bakhta
Trent Mohay
Tim Beiko
Joshua Fernandes