-
-
Notifications
You must be signed in to change notification settings - Fork 16
Home
Johann Barbie edited this page Sep 3, 2018
·
8 revisions
π Weekly meeting (Mondays - 13:00 UTC) call doc. π
attendees: Jaynti, Sami, Esteban, Johba
- [Jaynti] updates on hydrated state op-code execution, see PR 9.
- [Kosta] Only storage tests left. Written the implementation (not committed) but hit out of gas again. see PR 5.
- [Esteban] move meeting to 1 hours earlier β
- [Johann/Jaynti] how can state be exited?
=> Johann will write design for
exitable state object
and share with others.
attendees: Jaynti, Vaibhav,Sami, Alex, Johba
- [kosta / johann] update on off-chain executor, see PR.
- [johann] hydrated state architecture
- tasks are independent, can be started on already
- Jaynti: will work on this till next week
- how likely can we use EIP726?
- needs to be implemented in all clients => like months/years away.
- how to design contract exit games? Contract Shell Exit Proposal
- open problem: when should exits be possible?
attendees: Matheus, Jaynti, Vaibhav, Ashish, Sami, Esteban, Alex, Kosta, Johba
-
Matheus β
- Voting projects, research into Plasma since 2017.
- Latest code: JS implementation for stealth addresses and ring signatures with Lotion.js as Plasma chain.
- Liveliness is a problem / looking at Swarm to fix availability problem.
-
Jaynti => Digested Computation π‘
- Integrate VM state hashing into execution after each op-code.
- Commit merkle tree of all steps with block to root chain.
=> Allows to do direct challenges without binary search. π
=> Downside of potentially big performance impact. π
-
Sami => Stack Proofs as Unbalanced Merkle Tree π
- The stack proof can be an unbalanced merkle tree.
=> Popping and pushing super easy. π
=> Probably max top 7 elements needed, rest in regular hash.
- The stack proof can be an unbalanced merkle tree.
-
Johba => General Updates
- Posted 2 bounties to Gitcoin: 1, 2
- More research needed into optimized proofs (hydrated state architecture).
- UTXO model allows limited contract interaction.