Meeting notes: Development, Sep 18 2018

Meeting notes

Development meeting held Sep 18 @ 3PM UTC in grin/dev on Gitter, full chat transcript here.

Attendance:

antioch
hashmap
igno.peverell
tromp
quentinlesceller
yeastplume
garyyu
kargakis

Agenda & Actions

1. General update

  • Hard Fork went well, but had some couldn’t mine afterwards (fix went out)
  • Soft Fork went well.
  • Automated testing for mining has been difficult in past due to cuckoo memory requirements and travis CI.
  • @kargakis made a suggestion for “a suite that deploys a networks of a couple of nodes, runs for some time, then redeploys some/all nodes using a different version”.
  • Aggregate bulletproofs and a Schnorr sig improvement in the secpk256zkp fork via @jaspervdm, @Yeastplume has reviewed and bundled it into BIP32 mega PR: mimblewimble/grin#1501
  • @gary nd @antioch have a plan to fix the “stale sync_head” issue which was the cause of the “stuck in header sync loop”
  • @tromp has written an exhaustive cycle finder for cuckatoo and a working simple miner

2. Testnet 4

Focus on next two weeks is preparing for Testnet 4, includes

  • dual PoW
  • BIP32 mega PR (as above)
  • mean Cuckatoo CPU miner,
  • replace current prevhash with pow mmr root to the header (for FlyClient compatibility)

3. PoW

4. Other

  • @hashmap is playing with libp2p and TLS support for Hyper to research different peer discovery mechanisms.

  • Discussion on how to save space in headers. @gary ended up compiling this neat wiki page.

–END–

Feel free to raise suggestions / thoughts / comments and requests for future meeting agenda topics below!

@tromp

You should probably change the name of cuckatoo before it gets to concrete; I can see it being confusingly similar

cuckoo-lite?

It’s confusingly similar on purpose, as to most people the difference shouldn’t matter one bit.