Replies: 1 comment
-
We spoke about this briefly offline. After some thought here is what I think:
With the short-circuiting approach, we'd have to retry the ones that are known to have succeeded, as well as the ones that are unknown (this follows from (1) and (2)). I think this means that we should return enough information in the error response for us to figure out which ones failed. For v1, I think that means the signer should just use the "single version" of the contract call. The "single version" approach makes things super simple on the signer side. It's more costly because of fees, so maybe we can do the ideal thing if we have time. Footnotes
|
Beta Was this translation helpful? Give feedback.
-
Surfaced in #243, there's a discrepancy in how we handle submitting batched transactions (deposits | withdrawals) on-chain.
Currently, in both user paths, a signer submits a list of either deposits | withdrawals & all individual transactions should pass. If a single one fails, then they all fail with an error message pointing to the failed entry in the list.
As mentioned by @djordon, this might not necessarily be the best way to handle this logic. Is it a better alternative to complete the transactions even if one or more of them fail?
Beta Was this translation helpful? Give feedback.
All reactions