add block hint #42
add block hint #42
Conversation
tcoratger
force-pushed
the
block-hint
branch
from
d501fab
to
0fc876a
Compare
Co-authored-by: Clément Walter <[email protected]>
There was a problem hiding this comment.
I think that we'd like instead to implement From<SealedBlock> for a KethBlock, and such a KethBlock would fit well with the already existing https://github.com/lambdaclass/cairo-vm/blob/main/vm/src/vm/vm_memory/memory_segments.rs#L132-L146 so that it's easier is we update the models
There was a problem hiding this comment.
KethBlock would be a Vec<MaybeRelocatable>?
There was a problem hiding this comment.
It'd have been better to import the test_block function from test_os.cairo but since #43 is supposed to remove this, I won't block the PR for this
| -> HintExecutionResult { | ||
| // This call will first add a new memory segment to the VM (the base) | ||
| // Then we load the block into the VM starting from the base | ||
| vm.gen_arg(&Into::<KethPayload>::into(block.clone()).0)?; |
There was a problem hiding this comment.
where do you actually assign the base to the block cairo variable here?
There was a problem hiding this comment.
rust question: why do we need to clone?
There was a problem hiding this comment.
- For the base, this is done here no? https://github.com/lambdaclass/cairo-vm/blob/7d19956781901fb05c76f7c02fa82bd907f0f9a2/vm/src/vm/vm_memory/memory_segments.rs#L136 we add a new segment which is automatically set as the base for the block, is there anything else to set?
- For the clone, this is a closure question, for more explicitness, the exact error without cloning is
cannot move out ofblock, a captured variable in anFnclosure move occurs becauseblockhas typeSealedBlock, which does not implement theCopytrait
| impl From<Option<u64>> for KethPayload { | ||
| fn from(value: Option<u64>) -> Self { | ||
| MaybeRelocatable::from(Felt252::from(value.unwrap_or_default())).into() | ||
| } | ||
| } |
There was a problem hiding this comment.
not sure where we have Option<u64> in blocks, but this would be converted into a model.Option in cairo, meaning a struct with 2 values {is_some: 0, value: 0}
There was a problem hiding this comment.
blob_gas_used for example is an option, at the moment the model.Block is constructed like that so maybe let us keep as is and fine tuned this after modification on Cairo side no?
| fn from(value: Bloom) -> Self { | ||
| vec![ | ||
| MaybeRelocatable::from(Felt252::from(value.len())), | ||
| MaybeRelocatable::from(Felt252::from_bytes_be_slice(&value.0 .0)), |
There was a problem hiding this comment.
bloom is a 256 long bytes array, so a vec of felt
| impl From<U256> for KethPayload { | ||
| fn from(value: U256) -> Self { | ||
| MaybeRelocatable::from(Felt252::from_bytes_be_slice( | ||
| &value.to_be_bytes::<{ U256::BYTES }>(), | ||
| )) | ||
| .into() | ||
| } | ||
| } |
There was a problem hiding this comment.
the u256 is not trivial as it can be either Uint256* (so a pointer to a new segment with two values) or directly inlined into the struct, and consequently split in _low and _high parts
There was a problem hiding this comment.
Here I put the case where this is always inlined into the struct
| impl From<Bytes> for KethPayload { | ||
| fn from(value: Bytes) -> Self { | ||
| vec![ | ||
| MaybeRelocatable::from(Felt252::from(value.len())), | ||
| MaybeRelocatable::from(Felt252::from_bytes_be_slice(&value.0)), | ||
| ] | ||
| .into() | ||
| } | ||
| } |
There was a problem hiding this comment.
bytes are an array of felt, not a single felt
| fn from(value: Vec<u8>) -> Self { | ||
| vec![ | ||
| MaybeRelocatable::from(Felt252::from(value.len())), | ||
| MaybeRelocatable::from(Felt252::from_bytes_be_slice(&value)), |
There was a problem hiding this comment.
same, needs to be an array
| impl From<B256> for KethPayload { | ||
| fn from(value: B256) -> Self { | ||
| vec![ | ||
| MaybeRelocatable::from(Felt252::from_bytes_be_slice(&value.0[16..])), | ||
| MaybeRelocatable::from(Felt252::from_bytes_be_slice(&value.0[0..16])), | ||
| ] | ||
| .into() | ||
| } | ||
| } |
There was a problem hiding this comment.
see comment, it depends on the target model
| impl From<Option<B256>> for KethPayload { | ||
| fn from(value: Option<B256>) -> Self { | ||
| value.unwrap_or_default().into() | ||
| } | ||
| } |
| // The length of the signature payload | ||
| MaybeRelocatable::from(Felt252::from(value.payload_len())), | ||
| // The signature payload | ||
| MaybeRelocatable::from(Felt252::from_bytes_be_slice(&value.to_bytes())), |
| // Run the cairo program with the hint processor and the block | ||
| // | ||
| // We unwrap the result to ensure that the program ran successfully | ||
| let _ = cairo_run(&program, &config, &mut hint_processor).unwrap(); |
There was a problem hiding this comment.
because the program does nothing, there is no chance to have this run failing. (you could remove the hint it'd be the same)
No description provided.