[Bug] Set from via filler for eth_call and eth_estimateGas

[kassandraoftroy]

Component

provider, pubsub

What version of Alloy are you on?

0.9.2

Operating System

Linux

Describe the bug

hey there i am having an issue with transaction simulation failing (where if i just skip simulation with the exact same payload and parameters it works)

seems like it has something to do with doing native transfers during my call (simulation) because I get back that I got this error execution reverted, data: "0x9fb7cfe2" which corresponds to this spot in my SC

https://github.com/kassandraoftroy/stealth-gas-contracts/blob/main/src/StealthGasStation.sol#L63

tx payload:

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

on contract address: 0x943285f1a29281e59514fF35Dc16E5a14E123a27

network: holesky testnet, using an alchemy rpc wss://eth-holesky.g.alchemy.com/v2/xxxxxxx

block number: ~3196935 (or current)

source code doing simulation: https://github.com/kassandraoftroy/stealth-gas-backend/blob/main/src/http_server.rs#L164-L186

caller/signer of tx being simulated: 0x4B5BaD436CcA8df3bD39A095b84991fAc9A226F1

here you can see that on tenderly simulation the same tx succeeds! https://www.tdly.co/shared/simulation/1efe08e3-a571-4511-82e1-46df3d3bbf05

my suspicion: it may be a bug in simulation where native transfers fail for some reason (maybe forwarding gas? maybe thinking the EOA can't receive eth? idk)

let me know what you think!

[mattsse]

could you perhaps provide a repro for this?

[DaniPopes]

Also not entirely sure how this relates to alloy, I would recommend you inspect the RPC request with RUST_LOG=alloy=trace and see if it could be an error in the parameters

[kassandraoftroy]

hey there I did a full repro and you can see it here: https://github.com/kassandraoftroy/repro-alloy-sim-err

but in the process of doing the repro i think i spotted something! when i add .with_from() to the simulation it works as expected...

i am leaving this open bc the error reported in the execution still seems extremely weird to me (NativeTransferFailed()) and it's what threw me off. if the error is actually downstream of not having .with_from why would i get error that execution fails at this line ? also why do i need to add .with_from to simulations but not when sending transaction? (if i skip simulation it succeeds even without .with_from)

[yash-atreya]

Adding more context from TG: https://t.me/ethers_rs/43253

if you pass this Tx struct to send_transaction and your provider has a signer and recommended fillers etc, then everything will go smooth, as i suppose the provider will fill in from for you under the hood. But if you estimate_gas on this same tx struct i suppose it won't necessarily fill in from and this was ultimately what made the behavior different in simulation on my call.

Changed title to reflect action item

[yash-atreya]

Resolved by #2011

[mattsse]

actually I think this isn't the case already, if I look at fillprovider

we actually never call the new prepare request call

[yash-atreya]

Looking into it