This repository has been archived by the owner on Aug 8, 2018. It is now read-only.
forked from heikoheiko/pyethapp
-
Notifications
You must be signed in to change notification settings - Fork 604
Updating for mainnet
Bob Summerwill edited this page Mar 10, 2017
·
13 revisions
PyEth call - Fri March 10, 2017
The purpose of this call is to determine the technical and soft issues about getting the PyEth app running again with mainnet.
Participants:
- @vbuterin - Vitalik Buterin (Ethereum Foundation)
- @heikoheiko - Heiko Hees (Brainbot)
- @konradkonrad - Konrad Feldmeier (Brainbot)
- @janx - Jan Xie (Cryptape)
- @pipermerriam - Piper Merriam (ConsenSys)
- @bobsummerwill - Bob Summerwill (ConsenSys and EEA)
- @shahankhatch - Shahan Khatchadourian (ConsenSys and EEA)
Status of the project:
- Homestead support is there
- Moving towards Metropolis support, so there is something to test against
- Casper, and consensus abstraction, are being worked on
- pairing pre-compile is there (state revamp branch?)
- missing PV 62 (wip; not stable) https://github.com/ethereum/pyethapp/pull/176
- client is slow (in general); hard to get through DoS
- diverged branches: state_revamp vs develop; state_revamp doesn't work with pyethapp yet
- see A pyethereum revamp-in-progress: "Purification", Consensus Abstraction, State Snapshots (ie. rapid sync time and an end to hard-fork technical debt) and more for rationale for the state_revamp work
- it makes sense for an EIP to be created to standard the JSON format for state snapshots which has come out of this work.
- @cdetrio - Casey Detrio has been talking to the ethereumjs guys about using that JSON format for their needs. Parity also has some state snapshot format.
- Jan has a node that passed DDos block but before state clean block
- spurious dragon hf is on state_revamp
- missing PV 63/fast sync: https://github.com/ethereum/pyethapp/issues/89 -- necessary for Python client since it is slow in general, not as much an issue for other clients since they are faster
- Raiden relies on pieces of pyethereum
- used in general for reading keyfiles and generating pre-signed transactions
- Konrad has a release of pyeth based on 'develop' branch that allows its use as a pypi installable library, though could use some refactoring to integrate new changes (from 'state_revamp' branch)
Planned upcoming features:
- windows support; problems secp256k1 + leveldb (secp256k1 issue was felt in C++ client too)
Necessary changes to the architecture and project structure:
- extract solc stuff into extra library
- strip out "extra stuff" (tbd)
- make it easier to integrate CI and add code linting to CI
- there is progress on web3py, so perhaps remove overlap from pyethapp internals
- try to reduce dependencies on external utility apps
- make the in-memory EVM lighter so that it is easier to test with
- better documentation in order to be able to talk about breaking changes and how to modify/update the codebase
- Swarm, can be separated
- Whisper, can be used to support "configurable privacy" as a way to transfer a message off-chain
- not necessary to merge pv_62 with state_merge branch, since casper and consensus abstraction is not yet needed for public sync
Next steps:
- Goal: move towards getting pyethapp to sync with the public chain; become "viable" client again
- Piper: make a blog post asking for comments/feedback on pyeth client since there are plans for it undergo dramatic changes
- Jan: merge state_revamp (pv 62) + pyethapp (develop branch), timeline unknown due to resource constraints
- Shahan: to relay fulltime+dedicate resourcing need to EEA/ConsenSys "python, network, Eth stack nice to have"
- Use pyethereum gitter as dedicated comms channel: https://gitter.im/ethereum/pyethereum
- Talk via email/comms to schedule next call in two weeks