internal error: entered unreachable code: We should never have decoded calldata for contract creations

[pcaversaccio]

Component

Forge

Have you ensured that all of these are up to date?

• Foundry
• Foundryup

What version of Foundry are you on?

forge 0.2.0 (93f64c7 2023-10-27T00:24:49.496418315Z)

What command(s) is the bug in?

forge test

Operating System

Linux

Describe the bug

Multiple times during CI runs I face the following error (example)

Running 1 test for test/invariants/CreateX_Invariants.t.sol:CreateX_Invariants
[PASS] statefulFuzz_EtherBalance() (runs: 5000, calls: 75000, reverts: 74998)
The application panicked (crashed).
Message:  internal error: entered unreachable code: We should never have decoded calldata for contract creations
Location: crates/evm/traces/src/lib.rs:262

I'm doing an invariant test on my create factory here. The issue as it seems is that for certain initCode inputs (i.e. decoded calldata) it panics. It's getting pretty annoying lately tbh. Any way to disable this in my test environment?

[mattsse]

@Evalir what changed with trace decoding?

foundry/crates/evm/traces/src/lib.rs

Line 262 in cc760aa

unreachable!("We should never have decoded calldata for contract creations");

this unreachable doesn't look sound to me

[Evalir]

@mattsse we'll have to triage—this unreachable was already there from the previous implementation. Possible that the issue might come from another place

[mattsse]

looks like we're decoding init code now, so I'm removing the unreachable for now

#6141

this shouldn't be there anyway

[mattsse]

any other issues we need to fix asap @pcaversaccio ? :)

sorry for the inconveniences lately, lots of moving parts atm -.-

[mattsse]

triggering new release once merged

[pcaversaccio]

any other issues we need to fix asap @pcaversaccio ? :)

sorry for the inconveniences lately, lots of moving parts atm -.-

Well you might look at a similar issue I reported some months ago: #5629. Very weird behaviour.

[pcaversaccio]

any other issues we need to fix asap @pcaversaccio ? :)
sorry for the inconveniences lately, lots of moving parts atm -.-

Well you might look at a similar issue I reported some months ago: #5629. Very weird behaviour.

See the debugging info I added there:

If I replace hex"01" with abi.encode(hex"01") it will run successfully (if I use abi.encodePacked instead it will revert; the function deployCreate uses bytes as input for the argument, which is the init code). So two questions @Evalir: why is this throwing only if I use the ci profile and why is there an issue with the hex/abi.encodePacked instruction?

[mattsse]

okay that's the same bug, yeah.

so the bug here is actually pretty simple that we just didn't handle the case where we were able to decode the constructor calldata

this panic is now removed with #5629, so no more crashes here

[pcaversaccio]

okay that's the same bug, yeah.

so the bug here is actually pretty simple that we just didn't handle the case where we were able to decode the constructor calldata

this panic is now removed with #5629, so no more crashes here

you mean #6141

[Evalir]

Amazing! Is this closeable @pcaversaccio ?

[pcaversaccio]

Amazing! Is this closeable @pcaversaccio ?

one sec -currently running my CI with the latest release

[pcaversaccio]

LGTM - thx for fixing.