Could not find node error

[0xalpharush]

$ evm-cfg evm -o cfg.dot --open
thread 'main' panicked at src/cfg_gen/cfg_graph.rs:214:9:
Could not find node for pc 62

6080604052348015600f57600080fd5b5060b580601d6000396000f3fe6080604052348015600f57600080fd5b506004361060285760003560e01c8063f42e8cdd14602d575b600080fd5b603c60383660046058565b603e565b005b60005b60ff811015605457816054576001016041565b5050565b600060208284031215606957600080fd5b81358015158114607857600080fd5b939250505056fea2646970667358221220a206d90c473b6930258d5789495c41b79941b5334c47a76b6e618d3571716d5164736f6c634300081c0033

    contract X {
        function Y(bool yes) external {
            for (uint256 i = 0; i < 255; i++) {
                if (yes) {
                    break;
                }
            }
        }
    }

[plotchy]

Did some looking into this. I get some weird results when debugging the analysis, like things dont make sense. The bytecode wants to jump to PCs that aren't jumpdests.

Can i ask how you compiled this? I'd like to dig in further. When i use either of these i get expected output.

solc X.sol --bin-runtime --no-cbor-metadata

solc X.sol --bin-runtime --no-cbor-metadata --optimize --via-ir

Version: 0.8.26+commit.8a97fa7a.Darwin.appleclang

[0xalpharush]

I believe I just ran forge build X.sol

[plotchy]

ok sounds good will check it out

[plotchy]

ah on second look I can recognize that this is the deploy code rather than the runtime.

The jumps on the runtime portion aren't intended to make sense in the deploy code, so they panic.

I could probably catch those and still output a graph. That would also help with factories and stuff they have contract code that is ready to deploy but not intended to run in the same context