Specify the genesis variant for the `EvmSketch`

[hadjiszs]

Description

The sp1-contract-call crates are used in the aggchain proof for contract calls verification.

Recent update requires one genesis value (cf. here)

This genesis value must be properly specified according to the L1 for which the prover is instantiated.

Example in a test where the L1 is Sepolia:

let new = EvmSketch::builder()
    .at_block(BlockNumberOrTag::Number(final_block_number))
    .with_genesis(Genesis::Sepolia)
    .el_rpc_url(rpc_url_l2.clone())
    .build()
    .await?;

Solution

This with_genesis must be set properly in the production code to set whether the L1 is mainnet, sepolia, etc.

Several options:

• From a config file entry specifying the chain id
• The genesis variant directly
• ...

[hadjiszs]

@invocamanman do you think it is enough if we specify only Genesis::Sepolia or Genesis::Mainnet, or do you think there is some cases where we would need to go for Genesis::Custom(json)?

cf. https://github.com/succinctlabs/rsp/blob/main/crates/primitives/src/genesis.rs#L40-L78

cc @praetoriansentry

[jesusholoo]

i think if we use Genesis::Mainnet will be enough, for both chains.
The static calls that we do are extremely simple, do not involve new EIPs or precompileds, therefore, genesisMainnet will be enough ( this is just for "know" which hardfork they should execute)*

Even tho as we discuss in the last meeting maybe we can consider to make storage proofs ourselves in the future, but for now Mainnet is enough ^^

[jesusholoo]

Also, we already had this warning, see:

/// WARN: The static call must not use the `chainID` opcode, as it will
    /// return 1 (mainnet). The EVM version used by the Solidity compiler
    /// must be compatible with the version used in the static call. No
    /// special precompiled contracts are supported.
    /// Even though the current example satisfies these constraints, it's
    /// important to keep them in mind when updating the code.

Which basically means that we have been using "genesis:mainnet" implicitly

[hadjiszs]

i think if we use Genesis::Mainnet will be enough, for both chains.

For the bridge constraints test which interacts with Sepolia, I had to specify Genesis::Sepolia, otherwise we're failing in this reth error, from here in sp1-contract-cal

[hadjiszs]

Solved through #233 by @invocamanman : needed to specify Genesis::Sepolia because of the re-use of old deployed test data on old hardfork

So Genesis::Mainnet should be sufficient in all cases