Extensive testing for all examples #365
Conversation
alexandrosfilios
force-pushed
the
afilios/extend-examples
branch
from
bec6b71 to
5fb8216
Compare
alexandrosfilios
requested review from
diegoximenes,
gligneul and
rory-ocl
and removed request for
gligneul
stylus-tools/src/utils/testing.rs
Outdated
| use alloy::primitives::{Address, TxHash}; | ||
| use eyre::Result; | ||
|
|
||
| pub trait ExampleContractTester { |
There was a problem hiding this comment.
What is the advantage of defining a trait for that?
How about moving those fn out of the trait, they don't rely on self, and change init to something like this?
pub fn init(expected_abi: String, expected_constructor: String) -> Result<(Node, Address)>
There was a problem hiding this comment.
I would rename this from init to something like init_and_test_all too.
Maybe there will be a future test scenario that we will only want to init, and do those other tests in the specific test implementation.
There was a problem hiding this comment.
The main reason was because I wanted to override some methods, e.g. deployer(), but I can have both.
|
|
||
| #[tokio::test] | ||
| async fn call() -> Result<()> { | ||
| let exporter = stylus_tools::Exporter::builder().build(); |
There was a problem hiding this comment.
I suppose that you didn't rely on stylus_tools::utils::testing::ExampleContractTester in this test, and other tests, since you want to confirm that this stylus_tools::utils::testing::ExampleContractTester strategy is better than replicating the code in every test, right?
I think not replicating is better in this case, so we can move forward with a strategy like stylus_tools::utils::testing::ExampleContractTester 🙂
alexandrosfilios
force-pushed
the
afilios/extend-examples
branch
from
0ae4d46 to
9b9cf2f
Compare
Description
All examples now test more stylus commands. Enabled the tuple example.
Checklist