why:
Rlp errors throw exceptions which cause the dispatcher loop to
terminate the current session immediately.
details:
The DisconnectionReasonList message requires a single entry list.
Observed and now accepted deviations are:
Geth: single byte number
bor(a Geth fork): blobbed single entry list containing a number
- add eventLoop to control all incoming events
- change semantic of write to asynchronously block only when send buffer is full, and not when bytes do not fit into send window
- change handling of receive buffer, to start dropping packets if the reorder buffer and receive buffer are full. Old behaviour was to async block unless there is space which could lead to resource exhaustion attacks
The ENR code used to be solely exception based, and these
exceptions where a left-over of that. They are useless as later
calls use Result anyhow.
Additionally, they cause quite the performance loss because they
are used in the "common path" for the toTypedRecord call, e.g.
when reading the fields of ip6, tcp6 and udp6.
Currently only setting `--styleCheck:hint` as there are some
dependency fixes required and the compiler seems to trip over the
findnode MessageKind, findnode Message field and the findNode
proc. Also over protocol.Protocol usage.
This removes the outdated copy of the SSZ code. It became incorrect
over time (e.g., empty SSZ list elements), and is no longer in use by
GitHub projects: https://github.com/search?q=extension%3Anim+eth%2Fssz
The canonical SSZ implementation resides at `nim-ssz-serialization`.
Compared to `nim-eth`, these changes were made meanwhile:
- `bitseqs` was extended with JSON serialization support,
and with the new functions `isZero` and `countOnes`.
- `bytes_reader` was renamed to `codec`, extended with a few additional
SSZ type conversions as well as support for `SingleMemberUnion`.
- The simplified merkle tree implementation in `merkle_tree.nim`
was removed. It was not used by other projects.
- `merkleization` was extended with support for `HashArray`, `HashList`
and `SingleMemberUnion`. The `isValidProof` functionality has been
moved to `nimbus-eth2` and replaced with the EF defined function
`is_valid_merkle_branch`. The test was also moved to `nimbus-eth2`.
There are no other GitHub projects using `isValidProof`:
https://github.com/search?q=extension%3Anim+isValidProof
Furthermore, a definition for `GeneralizedIndex` was added.
- `ssz_serialization` was moved one directory up, and improved with
bug fixes and `HashArray`, `HashList` and `SingleMemberUnion` support.
- `types` was extended with JSON serialization and new type support for
`Uint128`, `Uint256`, `HashArray`, `HashList` and `SingleMemberUnion`.
There is also a new `getBit` function for `BitList`.
* Add few missing top level raises Defect in uTP
- Add top level {.push raises: [Defect].}
- remove some local raises, including some unneeded
CatchableErrors.
- Don't export messageHandler (avoiding annoying naming collisions)
- export utp_router as those connection callbacks are in the API
* Add some missing copyright clauses
* Some ident and max line length cleanup
* Rename utp_discv5_protocol.nim to be more consistent
* Add separate datastructure to keep track of window
* Asynchronously block write until until new space in snd buffer
* Introduce write loop
* Properly handle write cancellation
* Proper handling of sending fin packet
* Reset remote window after configured amount of time
toBytes for NodeId wasn't selected by compiler byt if it does
get selected, it will fail on the test cases due to the
countdown that is done in logDistance.
Set to toBytesBE properly now and do countup, that should make
it correct also for BE architecture.
Removed toBytes to avoid confusion and avoid this one being
selected ever. The only place toBytes for NodeId was used is in
sessions.nim makeKey func and there also the stint one
(thus native endianness) was selected in Nim 1.2.x.
Native endianness is fine there as it is only an internal
representation.
* Refactor tests and move socket to separate file
* Move sockets handling to separate class
* Abstract over underlying transport
* Fix bug with receiving duplicated SYN packet
* Fix race condition in connect
* Modify outbuffer
Each element of outbuffer keeps encoded packet ,number
of transmissions of givern packet and information if
given packet needs to be re-send.
* Add initial handling of timeouts
* Add tests for syn re-sends
* Don't fail ENR decoding when value is an RLP list
* Store RLP raw list in the ENR field pair instead
* Add ENR kList FieldKind so lists can be treated differently
Treated differently now when printing out the ENR, mentioning
that it is a raw RLP list
* Add initial handling of acks
Add implemetaion of circular buffer based on reference implementation
Add way to test number of packet in flight
Add acking of initial syn packet
* Allow for tcp/udp ports to always be configured
- Allow for an ENR to be build with tcp and udp ports also when
no IP address is provided
- In the address set-up always provide best efforttcp and udp ports
also when configuration of external ip (and/or ports) fails.
* Add initial impl of utp over udp
* Add more comments
* Add licenses and push declarations
* Add tests to nimble task
* Pr comments
Use better random generator
Raise assert error in case of buffer io exception
* Modify nodes verification
* Move nodes verification to separate module
By moving verification to separate module it can be re-used
in different contexts not only in discoveryv5.
Storing large blobs in a "WITHOUT ROWID" table turns out to be extremely
slow when the tree must be rebalanced.
* Split out keystore capability into separate interface, making each
keystore a separate instance
* Disable "WITHOUT ROWID" optimization by default
* Implement prefix lookup that allows iterating over all values with a
certain prefix in their key
It's been confirmed that the test fails if the Kademlia crash workaround isn't
in `findNode`, and passes with it there.
Signed-off-by: Jamie Lokier <jamie@shareable.org>
* Add build_dcli target and add it to CI
* Fix local imports for dcli
* And use local imports for all other files too
* Use local imports in tests and rlpx protocols
* Add search for best route and refactor setupNat to setupAddress
* Update setupAddress and make enr ports in discovery optional
* Add specific error log when no route is found
* Use bindIP if it is public
* Adjust some log levels
* Add ip limits to routing table and routing table buckets
* Fix order of ip limit check and duplicate check for replacement
* Fix ip limit for node with updated ip in ENR
* Fix bug where address wouldn't update on ENR update
and update some comments
* Reuse some add/remove code in routing table
* Fix seen bug on ENR update in routing table
* Rework addNode to make sure to do address check always
and adjust some logs.
* More documentation on the ip limits in routing table [skip ci]
* use bearssl rng throughout
* seeder can fail
* imports and exports
* modules, sigh
* one more try
* move var
* even fewer thread vars
* remove out-of-date genrated files
- routing table metrics + option in dcli
- only forward "seen" nodes on a findNode request
- setJustSeen & replace on ping AND findnode
- self lookup only at start
- revalidate 10x more
- use bitsPerHop (b) of 5
- small fix in resolve
- small fix in bucket split
* Discv5: More error handling improvements
- More results usage and raises pragma annotations
- Remove ENode related code and adjust Node object
- Misc.
* Add sendMessage and catch RlpError when decoding WhoAreYou
* Make the receive proc exception free
Except for `Exception` hah...
* Address review comments
* And another bunch of results and raises annotations
* Send Nodes Message also on 0 nodes and remove usage of broken require
* port kvstore from nim-beacon-chain
* remove old database backends
* use kvstore in trie database
* add sqlite dep
* avoid template param double evaluation
* clean up heterogenous lookup todo
Jordan Hrycaj
KonradStaniec
Kim De Mey
jangko
Jacek Sieka
Etan Kissling
Ștefan Talpalaru
Jamie Lokier
Mamy André-Ratsimbazafy
Zahary Karadjov