Testnet V2 todo

[liamzebedee]

tinychain testnet 1 is running, and we've sent a couple txs and mined a bunch of blocks. lots of todo items that are spread out among different areas of the stack, nearly all of these are linearized tasks - ie. small in size, low risk, don't involve a lot of unknowns

https://x.com/liamzebedee/status/1816057946369057070

• bugs
  • sync panics process after downloader has no more peers for jobs
  • node doesn't sync to full tip yet (need to ingest bodies)
  • test is failing in CI for some reason, can't access logs
• refactorings
  • getlongestchainhaslist returns the headers tip, should return full tip
  • explorer accounts page returns accounts in random order
  • the explorer is horribly inefficient lol. needs to hot reload too for new blocks from the node
• security
  • setup txs so they have unique nonces
  • add transaction replay protection / build "seen" index or "nonce" index
• sync
  • test branches and reorgs (gut: this will be straightforward)
  • upon first start, scan for missing block bodies. add these to wantlist
  • wait until fully synced before mining and spamming the net
  • update full tip
  • interactive bissect with peers to get common ancestor. then sync from that point.
  • when we receive a new block and don't know parent, do "fast path" and ask for most recent 10 headers
• build mempool ([WIP] Mempool #18)
  • add fee mechanism
  • reingest old txs??
  • add to block body
• mining
  • restart miner whenever we have a new full tip
  • restart miner whenever we have a new tx bundle from mempool
• add ability to send transactions
  • design some interface for users / CLI
  • "add transaction" user endpoint
  • gossip transactions, if !seen then add to mempool
  • add "gettxstatus" to check where/if tx was confirmed
• saving state
  • save the miner public key to database under wallets, load on start
  • save the peers we've seen to database, load from db on start
• performance improvements
  • verify signatures in parallel (maybe using GPU?)
  • fetch blocks in parallel for syncgetblockdata
  • add proper indexes on sql database for various things
  • add a cache / memoize the longestchainhashlist function
  • add a cache / memoize the reconstructstate function
• simple block explorer on each node (Block explorer (tinyscan) #19)
  • take the sqlite database
  • get the longest chain hash list
  • then build a simple HTML interface around it
    • homepage - latest block, num transactions
    • /blocks/:hash
    • /txs/:hash
    • /accounts/:pubkey

Stretch:

• load peers from dns seed? or load peers from bittorrent or ethereum or something.
• move to using protobufs for rpc Protobufs + gRPC #13

Nice to have:

• add ecdsa recovery or use schnor signatures to reduce tx/pubkey size by 50%
• add the state root
  • maybe via something like https://twitter.com/fede_intern/status/1818273961148383361
• setup hole punching / NAT stuff
• size improvements
  • use hash(pubkey) for accounts instead of pubkey. could be smaller ([32] vs. [20]byte).
• build a simple terminal UI interface
  • list of blocks
  • list of accounts
  • after every update
• mining algorithm
  • use ethash