Is there a canonical way to handle error states within a smart contract?
I am looking for a consistent method for guaranteed off-chain and on-chain validation and error handling. For example the ability to retry or defer execution in an asynchronous way until success or rollback without loss of funds or fees. Preferably without having to build a message queue, pub/sub pattern, or leverage other third party centralized services which defeats the purpose of using a “smart” contract …
The general idea is a smart contract that can mint tokens on demand without locking or losing any ADA regardless of whether it is used correctly or network behaves. The following diagram loosely illustrates what I am trying to achieve at a high level.
Can you pls elaborate on this? As all transactions are deterministic and you will only be charged fees or loose collaterals when your deterministic transaction fails like this:
Phase1 validation failure - no fee charged and no collateral loss
Phase1 validation success (no-script tx) - fee charged, collateral is not applicable
any Phase2 validation failure in a tx (only after Phase1 is valid) - no fee charged, but all collateral is lost
all Phase2 validation success in a tx (only after Phase1 is valid) - only the fee charged
Alright so to answer to my musings on recovery, retry, etc is probably “not supported” because it is presumed impossible to have indeterminate state between off-chain, on-chain, validation phases, settlement, etc. The node should be trusted to handle each phase and execution appropriately.
If that is correct then I assume adding logic that impacts phase 1 validity is the only real method to ensure success. In other words phase 2 should effectively never fail as long as phase 1 was valid and covered all criteria.
I guess my question could then be reworded as is there a method for creating a “self auditing contract” to ensure every possible condition in phase 1 is covered correctly?
