update plutus to alonzo/rc-2 #2425
Conversation
JaredCorduan
force-pushed
the
jc/update-plutus
branch
2 times, most recently
from
148a0f0
to
b7e5068
Compare
JaredCorduan
changed the title
| @@ -264,7 +264,7 @@ plutus_data = ; New | |||
| / { * plutus_data => plutus_data } | |||
| / [ * plutus_data ] | |||
| / big_int | |||
There was a problem hiding this comment.
big_int can be any big int now for the same reason.
JaredCorduan
force-pushed
the
jc/update-plutus
branch
from
6d3e033
to
5af208b
Compare
JaredCorduan
force-pushed
the
jc/update-plutus
branch
2 times, most recently
from
71cc4d7
to
9c5cb98
Compare
JaredCorduan
force-pushed
the
jc/update-plutus
branch
from
9c5cb98
to
b17905d
Compare
| Left e -> fail e | ||
| Right cm -> pure cm | ||
|
|
||
| decodeArrayAsMap :: Ord a => Set a -> Decoder s b -> Decoder s (Map a b) |
There was a problem hiding this comment.
Okay, so we gave you the cost model as Map Text Integer to make it very homogenous, and so you could work with it somewhat independently of the API we give you. But if we're using defaultCostModelParams during deserialization, then that's kind of gone out the window.
In which case: it seems like this would be easier if it was Map ModelKey Integer where ModelKey has a Enum/Bounded instances? It also would make for a nicer, more generic function:
decodeArrayAsMap :: (Enum k, Bounded k, Ord k) => Decoder s v -> Decoder s (Map k v)
There was a problem hiding this comment.
Yep, I agree, that would probably be nicer to work with and maintain.
There was a problem hiding this comment.
There was a problem hiding this comment.
Thanks @michaelpj ! The enumeration will have to match what we are currently doing (ascending order by key) if we want to make this change sometime before the next hardfork after alonzo (since the cost model is a part of the script integrity hash).
As of IntersectMBO/cardano-ledger#2425, the ledger serialises only the values of a CostModel, rather than the keys. This is to allow for far more compact serialisation (the string names were taking a lot of space!) However, it means that we reassemble the names from the Plutus default cost model parameters, and hence generators must match these names. This commit alters the cost model generator to simply set values over the Plutus `defaultCostModelParams`.
This PR ended up being less trivial than I expected, so I will leave some notes here for posterity.
Datato 64 bytes, so our tests and CDDL have been adapted accordingly.102is now serialized as#6.102([uint, [* a]])instead of#6.102([uint, * a]).