search

Near-One / rainbow-token-connector

ERC-20/NEP-141 Token Connector for Rainbow Bridge
62 stars 18 forks source link

fix: return `mint` promise from `deposit` #51

Closed chadoh closed 3 years ago

chadoh commented 3 years ago

This transaction, https://explorer.near.org/transactions/erDQKaeYvmjKaPQz7deNKqLxZy8RFB3tFeHNut3yTRa, is marked as Successful. But you can see that the mint call failed.

The attached deposit is less than the mimimum storage balance

The change in the first commit here will correctly mark the whole deposit as failing, in such cases.

However, this contract still has a bug. Given that the mint step failed, the factory contract should roll back the record_proof step, so that I can try the full deposit again. Right now when I retry the transaction, it fails a different way, https://explorer.near.org/transactions/FVtfxcz3SFZ5hhtbgnassHK454aAuBKfQ8yXofwRRAiT:

panicked at 'Event cannot be reused for depositing.', src/lib.rs:290:9

Also, I lost 100 DAI in the process of finding this bug. How do I get it back?

chadoh commented 3 years ago

@mfornet when I build locally I do not see changes to any wasm files, but buildkite says we failed to check in these altered wasm files (why are build artifacts part of the git repository?).

Do you see these files when you build? Can you push the wasm files to fix the build?

mfornet commented 3 years ago

@mfornet when I build locally I do not see changes to any wasm files, but buildkite says we failed to check in these altered wasm files (why are build artifacts part of the git repository?).

Do you see these files when you build? Can you push the wasm files to fix the build?

We are adding this artifacts to the repository to guarantee deterministic compilation, and having access to actually deployed code. It is in our plan to update this workflow, and generate artifacts automatically in CI, but unfortunately it is not done yet.

I'm re-compiling contracts now again. The issue was with a dependency that got outdated (not related to this PR). Not sure why previous commit failed, I will see if the output changes after this recompilation.