First exploration: Jörmungandr NetworkLayer #366
Conversation
Sounds like a good idea. Extend the |
Why though ? |
initWallet = Wallet mempty mempty (SlotId 0 0, error "todo: genesisHash") |
|
Good point. I guess we have our first configuration parameter then 😅 |
| blocks `shouldBe` Right [] | ||
|
|
||
| it "should work for the first epoch" $ do | ||
| Right blocks <- runExceptT $ nextBlocks network (SlotId 0 0) | ||
| Right blocks <- runExceptT $ nextBlocks network (SlotId 0 0, undefined) |
There was a problem hiding this comment.
I think the idea was more like.. to update the network layer to play nicely with this. Which implies also an update in the http bridge implementation.
There was a problem hiding this comment.
The idea was to have:
data NetworkLayer t m = NetworkLayer
{ nextBlocks :: BlockHeader -> ExceptT ErrNetworkUnreachable m [Block]
, networkTip :: ExceptT ErrNetworkTip m BlockHeaderAnd these tests use the NetworkLayer, not the HttpBridgeLayer.
Without parametrising over tip, I don't see how we can avoid supplying undefined in place of a Hash "BlockHeader" here.
|
🤔 Oops, it would be the Originally I believed we would have the |
Well, we don't have the "current" block header either when we initially instantiate the wallet. I wouldn't go for a |
|
Starting from a clean slate -> #384 |
| @@ -461,34 +462,35 @@ newWalletLayer db nw tl = do | |||
| restoreSleep wid slot | |||
| Right blocks -> do | |||
| let next = view #slotId . header . last $ blocks | |||
| let nextBlock = BlockId $ view #prevBlockHash. header . last $ blocks -- TODO: WRONG!! THIS IS PARENT!!! | |||
There was a problem hiding this comment.
I think nextBlocks nw should be responsible for returning a new cursor
Issue Number
#219
Overview
Just exploring. WIP.
(SlotId, BlockId). Should maybe haveBlockHeaderinstead.HasGenesisHash targetor something)Comments
undefineds we have to pass in the tests are a bit unpleasant.newtype BlockId