Open fainashalts opened 4 days ago
should this be split into two parts? the first being a static analysis and the second being the PR into anvil to support debug_traceCall
?
@tremarkley yeah I think that would be ok. The reason I didn't do so was I felt like we probably wouldn't release the tool until we can simulate, so I wondered whether we should spend engineering time on the static analysis part. What do you think?
@fainashalts yeah, I think the right call is to prioritize adding debug_traceCall
to Anvil. If for some reason we can't get that supported or it significantly increases the scope of this, then we can consider fallback options (e.g. static analysis)
In order to determine in all cases whether an interop message is sent, we must simulate
eth_sendRawTransaction
, likely usingdebug_traceCall
.debug_traceCall
is not currently supported by Anvil, though it seems the Anvil devs are open to supporting it. This task likely involves two parts:debug_traceCall
op-simulator
to identify RPCs that result in interop messages.