Multi-era exts #298
Multi-era exts #298
Conversation
multi-era/rust/src/utils.rs
Outdated
| pub fn hash(&self) -> BlockHeaderHash { | ||
| let bytes = match self { | ||
| Self::Byron(block) => match block { | ||
| // why there's no to_bytes or sth for byron block headers? |
There was a problem hiding this comment.
@gostkin There should be. I think it's the ToBytes trait? We don't implement the same traits as the other eras since we can't do encoding preservation/etc so it directly uses cbor_event's, but byron/cip25 should still have that serialization::ToBytes (in core/rust/src/serialization.rs). There's a TODO about regeneration there so we might want to think about if we want to keep this long-term or if there was a way to use the other traits for non-preserve types. For Byron it's not a big deal since everything is canonical anyway but for cip25 it's weird since some things are subsets but maybe those aren't implementing any of the traits.
There was a problem hiding this comment.
More context: we had these ToBytes/FromBytes traits from before we did preserve-encodings. (possibly named differently e.g.ToFromBytes or ToCBORBytes/FromCBORBytes - I don't remember). The new CML's Deserialize which now has those to/from bytes methods now has the ability to preserve encodings and/or force canonical encoding which byron/cip25 don't.
Byron could kind of fake it since it's all guaranteed to be canonical anyway so both methods would do the same thing. This would have to be changed in cddl-codegen though since when you're not preserving encodings in general we don't make any assumptions about all inputs being canonical so we don't implement these traits and instead directly implement cbor_event::Serialize like we did before. You couldn't blanket impl it like with ToBytes since it would blanket impl it for everything not just the byron types.
CIP25 is weird since it's not canonical but skips some data e.g. parses a subset of objects so round-tripping isn't preserved. We could maybe regen cip25 and have it impl those types with the exception of the outermost object (the one equivalent to a tx metadatum) which IIRC doesn't implement ToBytes anyway.
gostkin
force-pushed
the
egostkin/multiera-extensions
branch
6 times, most recently
from
e60f987
to
d8b2893
Compare
gostkin
force-pushed
the
egostkin/multiera-extensions
branch
from
d8b2893
to
c574039
Compare
multi-era/rust/src/utils.rs
Outdated
| pub fn hash(&self) -> [u8; 32] { | ||
| let bytes = match self { | ||
| Self::Byron(block) => match block { | ||
| // why there's no to_bytes or sth for byron block headers? |
There was a problem hiding this comment.
Was this comment meant to be removed?
| @@ -375,6 +376,10 @@ impl TransactionBody { | |||
| encodings: None, | |||
| } | |||
| } | |||
|
|
|||
| pub fn hash(&self) -> [u8; 32] { | |||
| blake2b256(&self.to_cbor_bytes()) | |||
There was a problem hiding this comment.
[u8; u32] is more general but less consistent than TransactionHash. I guess on the rust side there's an Into impl so it's okay.
Use `TransactionBody` instead of of [u8; 32] to be consistent with `hash_transaction()`. Moves all the functions that were added directly to `mod.rs` in #298 files to `utils.rs` files where they should be since they aren't auto-generated.
* TX hash mismatch for non-canonical TXs This only concerns the free-floating `hash_transaction()`. Fixes #314 * *TxBody.hash() for multi-era rework Use `TransactionBody` instead of of [u8; 32] to be consistent with `hash_transaction()`. Moves all the functions that were added directly to `mod.rs` in #298 files to `utils.rs` files where they should be since they aren't auto-generated.
Adds better support for multi-era. Mostly hash functions are added to tx bodies / blocks. Fixed pool naming as well