Types cleanup #485
Types cleanup #485
Conversation
g-r-a-n-t
force-pushed
the
type-traits-cleanup
branch
from
6edbd31
to
7849b1e
Compare
| if let Type::Struct(val) = &error_attributes.typ { | ||
| context.revert_errors.insert(val.clone()); | ||
| if let Type::Struct(_struct) = &error_attributes.typ { | ||
| context.revert_errors.insert(_struct.clone()); |
There was a problem hiding this comment.
note how we can get rid of this
There was a problem hiding this comment.
Instead of adding this information to the context and then building a list of Yul functions based on it, I think we should accumulate functions as operations are used. This should make things more manageable.
Basically, inside of the operation function, we would have a line like context.add_deps(operation_deps), where operation_deps is a list functions needed in the Yul runtime for the operation to work.
There was a problem hiding this comment.
Yes, that sounds like a good idea 👍
|
|
||
| let revert_data = expressions::expr(context, error_expr); | ||
| let size = abi_operations::encoding_size(&[val.clone()], vec![revert_data.clone()]); |
There was a problem hiding this comment.
encoding size is found in the revert function
g-r-a-n-t
force-pushed
the
type-traits-cleanup
branch
from
49dc4bc
to
1e53e88
Compare
Codecov Report
@@ Coverage Diff @@
## master #485 +/- ##
==========================================
+ Coverage 87.70% 87.79% +0.08%
==========================================
Files 78 80 +2
Lines 5263 5268 +5
==========================================
+ Hits 4616 4625 +9
+ Misses 647 643 -4
Continue to review full report at Codecov.
|
| if (iszero([test])) { | ||
| ([revert_fn]([msg], [size])) | ||
| [revert_operations::error_revert(&AbiType::from(string), msg)] |
There was a problem hiding this comment.
This looks a whole lot cleaner indeed :)
g-r-a-n-t
force-pushed
the
type-traits-cleanup
branch
from
1e53e88
to
3d2fb1a
Compare
What was wrong?
All type traits were defined in the analyzer crate, when they should have been split up among the crates where they're actually used.
How was it fixed?
AbiEncodingtrait and used a new trait calledJsonAbiin the abi crate.From<..>onAbiTypefor each Fe type.FeSizedtrait was renamed toEvmSizedand moved to the yulgen crate.There were a few other things done in this PR:
To-Do
OPTIONAL: Update Spec if applicable
Add entry to the release notes (may forgo for trivial changes)
Clean up commit history