[gretzke]

Linear: ROUTE-2217 (item 1).

The README has a new section for deployments on a chain that runs the Amsterdam (Glamsterdam) upgrade. EIP-8037 adds state gas for deployed code (1,530 gas per byte), so a large contract needs more than 2^24 gas. The forge script simulation uses the configured evm_version (cancun in this repo) and sets a gas limit that is too low. The section says to add --skip-simulation, so forge takes each gas limit from eth_estimateGas on the RPC node. It also says that --gas-limit sets the block gas limit of the local simulation, not the transaction gas limit, and that the deployer needs more ETH (37,435,950 gas with Amsterdam rules against 5,253,854 with Osaka rules for the example).

The numbers come from a v4 PoolManager deployment through forge script and the CREATE2 factory:

Path	Gas limit	Gas used	Result
--broadcast (local simulation, forge 1.5.0 and 1.8.6)	7,256,884	6,968,987	reverted
--broadcast --skip-simulation	49,427,194 (anvil estimate times 1.3)	37,435,950	success

I ran both on a mainnet fork with an anvil that applies the Amsterdam gas rules. Then I checked the same transaction with eth_simulateV1 on Sepolia (Geth v1.17.8, reth v2.7.0, v2.8.0 and v2.8.1). The real clients estimate 37,734,228 gas. The transaction succeeds at 1.3 times this value (49,054,496) and reverts at 7,256,884. I did not run the Universal Router deploy scripts themselves. --skip-simulation exists with the same behavior in Foundry 1.3.6 to 1.5.0 (crates/script/src/broadcast.rs).

AI-Generated Description

Linear: ROUTE-2217 (item 1).

Summary

Adds a README section documenting how to deploy contracts on Ethereum L1 after the Amsterdam (Glamsterdam) upgrade.

Changes

• README.md: New "Deploying on Ethereum L1 after the Amsterdam upgrade" section explaining that EIP-8037 adds state gas for deployed code (1,530 gas per byte), which pushes large contract deployments past the 2^24 gas threshold. Documents that forge script's local simulation uses the configured evm_version (older than Amsterdam), sets a gas limit that is too low, and causes the deployment to revert on chain. Recommends --skip-simulation so forge uses eth_estimateGas from the RPC node. Clarifies that --gas-limit sets the block gas limit of the local simulation, not the transaction gas limit.

Notes

• Documentation only — no contract or script changes
• Numbers come from a v4 PoolManager deployment through forge script and the CREATE2 factory: 37,435,950 gas with Amsterdam rules vs 5,253,854 gas with Osaka rules
• Verified against mainnet fork (anvil with Amsterdam gas rules) and eth_simulateV1 on Sepolia (Geth v1.17.8, reth v2.7.0–v2.8.1)

[linear-code]

ROUTE-2217