[shemnon]

Stacked on #104: review only this PR's commit. GitHub retargets it to main once #104 merges.

Sei adopts four Amsterdam EIPs without the rest of the fork (CON-528). SeiAmsterdamTime opts them in on top of Osaka. AmsterdamTime stays unset, so BAL, state gas, SLOTNUM and the other Amsterdam changes stay off.

• EIP-7708: ETH transfer and SELFDESTRUCT transfer logs
• EIP-7954: upstream 64 KiB code / 128 KiB initcode limits
• EIP-8024: DUPN, SWAPN, EXCHANGE via an Osaka + 8024 jump table
• EIP-8246: SELFDESTRUCT to self keeps the balance; Finalise uses the Amsterdam path, whose BAL recording stays off without IsAmsterdam
• IsSeiAmsterdam requires Osaka; the fork-order check enforces it

Tests

• core/vm/eips_sei_test.go: rules and fork order, the instruction set (Osaka plus 8024 only; SLOTNUM undefined), code size limits, SELFDESTRUCT balance and transfer log
• core/evm_sei_test.go: transfer logs

[cursor]

PR Summary

Medium Risk
Touches consensus-critical EVM transfers, SELFDESTRUCT finalization, and opcode tables, but only when SeiAmsterdamTime is set and with broad unit coverage.

Overview
Adds SeiAmsterdamTime and IsSeiAmsterdam so Sei can turn on a subset of Amsterdam EIPs on top of Osaka without enabling full Amsterdam (AmsterdamTime stays off—no BAL in Prepare, no SLOTNUM, no EIP-8037/8038 state gas).

Fork wiring: ChainConfig gains SeiAmsterdamTime (requires Osaka at or before it), Rules.IsSeiAmsterdam, and config fork-order / compatibility checks. The EVM selects a new seiAmsterdamInstructionSet (Osaka + EIP-8024 only) ahead of the Osaka table.

Behavior gated like Amsterdam where intended:

• EIP-7708: ETH transfer logs in Transfer and SELFDESTRUCT-to-other (IsAmsterdam || IsSeiAmsterdam).
• EIP-7954: Amsterdam 64 KiB / 128 KiB code and initcode limits in CheckMaxCodeSize / CheckMaxInitCodeSize.
• EIP-8246: No balance burn on SELFDESTRUCT-to-self; Finalise uses the Amsterdam finalization path (balance-only accounts after self-destruct); hooked state skips selfdestruct-burn tracing when Sei Amsterdam is active.

Tests cover rules/fork order, instruction set diff vs Osaka, code limits, transfer logs, and SELFDESTRUCT balance/logs.

^{Reviewed by Cursor Bugbot for commit d7e9860. Bugbot is set up for automated code reviews on this repo. Configure here.}

[seidroid]

This PR gates EIP-7708, 7954, 8024 and 8246 behind the new IsSeiAmsterdam flag (Osaka plus the subset). Every gated site checks the flag consistently. The fork-order check makes Osaka a requirement. BAL recording stays off because Prepare still only allocates stateAccessList under IsAmsterdam. The tests cover rules, the jump table, code-size limits, the SELFDESTRUCT balance and the transfer logs. I found no issues introduced by the PR.

Findings: 0 blocking | 0 non-blocking | 0 posted inline

Blockers

• None at the file/PR level.

Non-blocking

• None at the file/PR level.