ethereum/pm

Ethereum Core Devs Meeting 78 Agenda

timbeiko opened this issue · 8 comments

Ethereum Core Devs Meeting 78 Agenda

Agenda

  1. Muir Glacier Updates
  2. Testing updates
  3. Eligibility for Inclusion (EFI) EIP Review
    1. EIP-2456
    2. EIP-1962
    3. EIP-2348
  4. EIPIP (EIP Improvement Proposal) Meeting
  5. Review previous decisions made and action items
  6. Next call

Can we add EIP-2456 to the discussion? At least as a springboard for a "are block based upgrade triggers the correct thing to do" discussion.

EIP 1962 status:

  • Rust and C++ implementations are complete feature wise
  • Both arithmetic functions and gas estimation functions are fuzzy tested
  • Moved to 10 separate “operations” that can be mapped to precompile addresses
  • C-like API is exposed on demand with universal signatures for Rust and C++ implementations
  • Go wrapper is on the way

What’s left:

  • decide whether use single (Rust) or independent (e.g. Rust for Parity, wrapped C++ for Geth, etc) implementations
  • make pipelines/builders for static libraries/go modules (would appreciate help with it) due to Rust and C++17 underneath
  • what fuzzer to try next? Honggfuzz is fast and convenient (especially multicore!), but in a meantime it may be interesting to try e.g. AFL

Short note: there was a critical bug in our MuirGlacier EthereumJS VM release v4.1.2 preventing the VM to working properly when instantiated with the MuirGlacier HF. We just released a bugfix v4.1.3 release. An update is mandatory if you want a working installation.

Also, EIP-2348, validated EVM contracts, as part of EFI review, but at the end as I feel the other two are more important. I added discussion points for two of the objections (transaction evm execution and why bytecode header) in the ethereum-magicians thread and would like to solicit responses.

@shamatar see https://github.com/ethereum/benchmarking/ -- that repo contians some old benchmarks for the crypto-precompiles

It also contains some analysis like https://github.com/ethereum/benchmarking/blob/master/analysis.md

Thanks! I think that I’ll need to modify those scripts and run on my reference PC cause I use 4 threads (for 6 physical cores) during generation of the test vectors, so CPU frequency is not on the maximum (Intel CPU frequency management)

Integration guide for EIP1962 for Geth and other who would want to use C++ implementation https://github.com/matter-labs/eip1962/blob/master/documentation/Integration.md

Closing in favor of #148.