feat(specs,tests): deploy the EIP-8250 nonce manager as an ordinary contract

[spencer-tb]

Description

Implements ethereum/EIPs#12440 for EIP-8250: the nonce manager is an ordinary contract deployed by the EIP's keyless transaction, like the EIP-8141 expiry verifier in #3729. NONCE_MANAGER moves to 0x8250968C12e01A19d6F667b9B2F3b3A4d0e51cB7, apply_fork no longer installs it, and the framework pre-allocates it in genesis instead of using the ActivationInstall hook extension from #3683. The new test_contract_deployment.py deploys it with the EIP's transaction before, at and after the fork block, and the transition tests check that activation leaves it unchanged and installs nothing when it is missing.

The reference spec pin moves to the merged EIP text. Replaces #3722 and #3725.

Related Issues or PRs

• Update EIP-8250: small clarifications for alignment EIPs#12440
• Follows feat(specs,tests): implement EIP-8250 keyed nonces for frame transactions #3683, feat(specs, tests): deploy the EIP-8141 expiry verifier as an ordinary contract #3729 and refactor(test-forks,test-specs): remove the unused activation code install hook in EIP-8141 #3735, replaces feat(specs,tests): record the EIP-8250 activation in the fork block's access list #3722 and feat(specs,tests): reject an occupied EIP-8250 nonce manager address at the fork block #3725

Checklist

• Ran fast static checks to avoid CI fails, see Code Standards & Verifying Changes: just static
• PR title has the form <type>(<area>): <title>, where <type> and <area> come from an appropriate C-<type>, respectively A-<area>, label. The title should match the target squash commit message.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (eips/bogota/eip-8250@87e520e). Learn more about missing BASE report.

Additional details and impacted files

@@                   Coverage Diff                   @@
##             eips/bogota/eip-8250    #3736   +/-   ##
=======================================================
  Coverage                        ?   94.93%           
=======================================================
  Files                           ?      677           
  Lines                           ?    41903           
  Branches                        ?     3960           
=======================================================
  Hits                            ?    39779           
  Misses                          ?     1470           
  Partials                        ?      654           

Flag	Coverage Δ
unittests	94.93% <100.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[danceratopz]

🚀 The agents have comments and I'm doing my best to keep up 😆 In the meantime, they seemed worth sharing!

The agent checked the deployment independently: recovering the signature gives sender 0x7A57...8198, CREATE at nonce 0 gives 0x8250968C...1cB7, the hash matches, and the init code returns 60006000fd. The install is fully gone from the spec and the framework, and the genesis pre-allocation is how the account exists in tests. This also fixes the EIP-8250 fill on the base: since #3735 removed the activation install hook, blockchain-test genesis has no nonce manager, and 116 of 251 fixtures fail at 1efbafd. With this PR they all fill (261 passed).

One more question on the EIP text: why must NONCE_MANAGER have a nonzero nonce at activation? Nothing reads it, and the deployment transaction sets it to 1 anyway. If it is only there to match what the deployment produces, saying so in the EIP would stop implementers from adding a check for it.

just static passes and the EIP-8250 fill is green. I'd land this before #3724; #3724's own commits apply cleanly on top once it is rebased onto the branch.

[spencer-tb]

Nothing reads the nonce. The wording matches EIP-7997 and only mirrors what the deployment produces. I'll say so in the next EIPs PR.

fill-pypy fails the same way on every PR in the 8141 stack (#3735, #3739 too), not from this PR.