It can be hard to track down failures, especially when they happen in multiple levels of submessages. Getting the full backtrace can be helpful, but it seems I am the only one who can really parse those to see what the issue was.
The main info I pull out is what the messages are and which depth of the submessages we are at. We can add some dbg!() statements to the framework... maybe when each message is executed, when each submessage is executed, each query is called, etc. basically all contract -> contract interactions could be traced.
Since multi-test only compiles for test environments (not into wasm), we can just dump everything with debug logs. Maybe we have some way to configure how many/which logs to dump, but first let's make sure we have all the needed info there, then make it optional.
It can be hard to track down failures, especially when they happen in multiple levels of submessages. Getting the full backtrace can be helpful, but it seems I am the only one who can really parse those to see what the issue was.
The main info I pull out is what the messages are and which depth of the submessages we are at. We can add some
dbg!()statements to the framework... maybe when each message is executed, when each submessage is executed, each query is called, etc. basically all contract -> contract interactions could be traced.Since multi-test only compiles for test environments (not into wasm), we can just dump everything with debug logs. Maybe we have some way to configure how many/which logs to dump, but first let's make sure we have all the needed info there, then make it optional.