feat(specs,tests): Implement EIP-2780 by gurukamath · Pull Request #2175 · ethereum/execution-specs

gurukamath · 2026-02-09T16:15:04Z

🗒️ Description

Implement EIP-2780 and some basic tests.

Note: This PR contains some testing framework changes which are not compatible with other Amsterdam tests. This PR however, does not focus on fixing those. Those should be fixed in a future PR.
As a result, the CI is not likely to pass for Amsterdam

🔗 Related Issues or PRs

issue #1940

✅ Checklist

All: Ran fast tox checks to avoid unnecessary CI fails, see also Code Standards and Enabling Pre-commit Checks:
```
uvx tox -e static
```
All: PR title adheres to the repo standard - it will be used as the squash commit message and should start type(scope):.

Cute Animal Picture

TBD

…#2175) * refactor(bal): absence validators -> BalAccountAbsentValues * feat: raise for empty lists on BAL absence checks * refactor: changes from comments on PR ethereum#2175

SamWilsn

Quick, incomplete review

src/ethereum/forks/amsterdam/vm/gas.py

src/ethereum/forks/amsterdam/fork.py

src/ethereum/forks/amsterdam/vm/instructions/environment.py

marioevz · 2026-02-17T14:46:09Z

There's a possibility this becomes a blocker for Devnet-3 since it's listed as requirement for https://eips.ethereum.org/EIPS/eip-8037 which is included.

Spencer checked with Maria, and it seems like it's not a hard requirement but it might be a good idea to get this ready to merge.

gurukamath · 2026-02-17T15:19:31Z

Spencer checked with Maria, and it seems like it's not a hard requirement but it might be a good idea to get this ready to merge.

To my mind, this needs further clarity, especially when it comes to the EIP-7702 interactions. @jochem-brouwer has nicely summarised some of the open questions in this eth magicians thread

gurukamath · 2026-02-18T13:18:00Z

Raised a PR on the EIPs repo to clarify some interactions with EIP-7702 and EIP-7928. See here

We can proceed to write more tests specific to these cases once we agree on these points.

codecov · 2026-02-24T10:57:39Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 86.05%. Comparing base (8b6f6ea) to head (ce7c14c).

Additional details and impacted files

@@                     Coverage Diff                     @@
##           eips/amsterdam/eip-2780    #2175      +/-   ##
===========================================================
+ Coverage                    86.01%   86.05%   +0.03%     
===========================================================
  Files                          599      599              
  Lines                        36904    36993      +89     
  Branches                      3771     3796      +25     
===========================================================
+ Hits                         31744    31833      +89     
  Misses                        4551     4551              
  Partials                       609      609

Flag	Coverage Δ
unittests	`86.05% <100.00%> (+0.03%)`	⬆️

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

☔ View full report in Codecov by Sentry.
📢 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.

gurukamath · 2026-02-24T14:10:11Z

Converting back to draft so I can re-base this on the latest forks/amsterdam

gurukamath · 2026-02-25T10:35:01Z

@danceratopz This got a bit out of hand in terms of PR size. If this is too much, I'd be happy to split this up into the following 2 PRs.

Implementation + Framework updates and EIP-specific tests (in this PR)
Fix broken legacy tests including BALs and other older fork tests

Do let me know what you think

SamWilsn

From the spec side, my review is complete.

src/ethereum/forks/amsterdam/vm/gas.py

src/ethereum/forks/amsterdam/vm/instructions/environment.py

src/ethereum/forks/amsterdam/vm/instructions/system.py

src/ethereum/forks/amsterdam/fork.py

Rename G_* gas cost references to match upstream GAS_* naming convention in GasCosts dataclass after rebase: - G_WARM_ACCOUNT_ACCESS -> GAS_WARM_ACCOUNT_ACCESS - G_NEW_ACCOUNT -> GAS_NEW_ACCOUNT - G_COLD_ACCOUNT_ACCESS -> GAS_COLD_ACCOUNT_ACCESS - G_AUTHORIZATION -> GAS_AUTH_PER_EMPTY_ACCOUNT - G_VERY_LOW -> GAS_VERY_LOW - G_CODE_DEPOSIT_BYTE -> GAS_CODE_DEPOSIT_PER_BYTE Also fix: - Merge duplicate Amsterdam gas_costs() methods in forks.py to preserve GAS_BLOCK_ACCESS_LIST_ITEM=2000 from EIP-7928 - Replace pre.empty_account() with Address(pre.fund_eoa(amount=0)) - Add address_has_code=False to BALANCE ops on non-existent accounts - Rename TX_BASE_COST -> GAS_TX_BASE in amsterdam/transactions.py

Address PR review feedback from SamWilsn: - Rename GAS_COLD_ACCOUNT_COST_NOCODE to GAS_COLD_ACCOUNT_COST_NO_CODE across spec, testing framework, and test files - Add docstrings to EIP-2780 gas constants in vm/gas.py

danceratopz

Hey @gurukamath, here's some initial feedback. I must admit, I need to go deeper and improve my understanding to do this justice in a second round. I haven't looked at the tests updated for interactions yet. Will come back to this.

tests/amsterdam/eip2780_reduce_intrinsic_tx_gas/__init__.py

packages/testing/src/execution_testing/forks/gas_costs.py

tests/amsterdam/eip2780_reduce_intrinsic_tx_gas/spec.py

src/ethereum/forks/amsterdam/vm/gas.py

packages/testing/src/execution_testing/forks/forks/forks.py

tests/amsterdam/eip2780_reduce_intrinsic_tx_gas/test_value_moving_transactions.py

src/ethereum/forks/amsterdam/vm/gas.py

src/ethereum/forks/amsterdam/vm/instructions/system.py

src/ethereum/forks/amsterdam/vm/gas.py

src/ethereum/forks/amsterdam/fork.py

danceratopz · 2026-03-18T08:15:36Z

@danceratopz This got a bit out of hand in terms of PR size. If this is too much, I'd be happy to split this up into the following 2 PRs.

Implementation + Framework updates and EIP-specific tests (in this PR)

Fix broken legacy tests including BALs and other older fork tests

That's a good idea, @gurukamath, I think splitting the older fork tests out would be a great idea to give more targeted reviews.

danceratopz

Hey @gurukamath, here's something that stood out really quickly upon taking another look at these.

tests/amsterdam/eip2780_reduce_intrinsic_tx_gas/test_value_moving_transactions.py

tests/amsterdam/eip2780_reduce_intrinsic_tx_gas/test_value_moving_calls.py

tests/amsterdam/eip2780_reduce_intrinsic_tx_gas/test_value_move_to_precompiles.py

danceratopz

Hey @gurukamath, here's a second round after reviewing tests/amsterdam/eip7928_block_level_access_lists/.

tests/amsterdam/eip7928_block_level_access_lists/test_block_access_lists.py

tests/amsterdam/eip7928_block_level_access_lists/test_block_access_lists_invalid.py

danceratopz · 2026-03-26T12:37:05Z

tests/amsterdam/eip7928_block_level_access_lists/test_block_access_lists_opcodes.py

 @pytest.mark.parametrize(
    "fails_at_extcodesize",
    [True, False],
    ids=["oog_at_extcodesize", "successful_extcodesize"],


I think "successful_extcodesize" as-is deviates from the original test's intent; it should be renamed to oog_phase2_at_extcodesize, in which case, how about?

Suggested change

ids=["oog_phase1_at_extcodesize", "oog_phase2_at_extcodesize"],

The question is do we need to add a new test case here "successful_extcodesize", or is it covered elsewhere?

I did not include a case here since it is covered specifically in tests/amsterdam/eip2780_reduce_intrinsic_tx_gas/test_account_access_opcodes.py::test_account_access_opcode. The original idea was to keep the implementation (and tests) of EIP-7928 and EIP-2780 as independent as possible. Not sure how essential this separation is anymore after the debates we've been having about merging stuff to forks/amsterdam

danceratopz · 2026-03-26T12:39:19Z

tests/amsterdam/eip7928_block_level_access_lists/test_block_access_lists_opcodes.py

        Op.PUSH20(target_contract)  # Target contract address
-        + Op.EXTCODESIZE(address_warm=False)  # Check code size (cold access)
+        + Op.EXTCODESIZE(
+            address_warm=False, address_has_code=False


We only have an Op.STOP, but unless this changes too, we should rename "successful_extcodesize" to "phase2_oog_at_extcodesize". We could consider adding a third successful case but I think the happy case is covered here https://github.com/gurukamath/execution-specs/blob/eip-2780/implement-eip-final/tests/amsterdam/eip7928_block_level_access_lists/test_block_access_lists.py#L246.

Suggested change

address_warm=False, address_has_code=False

address_warm=False, address_has_code=True

I think the test ids are a bit misleading and might need to change. The test essentially looks at the target being in BAL vs target not being in BAL. Pre 2780 these two scenarios exactly coincide with Success vs OOG of the opcode. Post 2780, they don't. Which is also why there is a disconnect between the target having code and us using address_has_code=False.

danceratopz

I reviewed all other non-specific Amsterdam tests for hard-coded constants. Didn't add inline comments here, just Clauded these up.Each addresses a few missing hard-coded gas constants from a single fork (Cancun, Prague, Osaka). I still need to self-review them:

gurukamath marked this pull request as draft February 9, 2026 16:15

SamWilsn reviewed Feb 11, 2026

View reviewed changes

LouisTsai-Csie mentioned this pull request Feb 13, 2026

[DEPRECATED] - feat(spec-specs): implement EIP-8037 state creation gas cost increase #2181

Merged

marioevz assigned SamWilsn Feb 18, 2026

gurukamath force-pushed the eip-2780/implement-eip-final branch 3 times, most recently from 3f816ea to 0d197cc Compare February 19, 2026 15:51

gurukamath marked this pull request as ready for review February 19, 2026 15:52

gurukamath requested a review from danceratopz February 19, 2026 15:53

gurukamath marked this pull request as draft February 19, 2026 17:49

gurukamath force-pushed the eip-2780/implement-eip-final branch from 044f9b3 to 4c9e3e4 Compare February 20, 2026 13:45

gurukamath mentioned this pull request Feb 24, 2026

EIP-2780 Implementation Tracker: Reduce intrinsic transaction gas #1940

Open

25 tasks

gurukamath marked this pull request as ready for review February 24, 2026 10:59

gurukamath marked this pull request as draft February 24, 2026 14:10

gurukamath force-pushed the eip-2780/implement-eip-final branch 2 times, most recently from aa8045c to f04a782 Compare February 24, 2026 15:44

gurukamath marked this pull request as ready for review February 24, 2026 15:45

SamWilsn reviewed Mar 5, 2026

View reviewed changes

SamWilsn assigned danceratopz Mar 5, 2026

gurukamath added 3 commits March 13, 2026 11:07

feat(spec): Implement EIP-2780

161e0ac

feat(test): Update testing framework for EIP-2780

bd720ce

feat(test): Write basic EIP-2780 tests

92b3524

gurukamath added 16 commits March 13, 2026 11:07

feat(tests): add precompile tests

5daec8e

feat(tests): add value moving call tests

23338b9

fix(specs): update callcode value gas

90c3702

fix(spec): add recipient cost for 7702 delegations

6a7c08a

fix(tests): fix mkdocs issue

e99ef7a

feat(tests): update other opcode metadata

f3ff853

fix(tests): update BAL tests

9afd4f4

fix(tests): fix osaka tests for amsterdam

4b12452

refactor(tests): use RecipientType instead of multiple bools

9dd687f

fix(tests): fix prague tests for amsterdam

3141cd7

fix(tests): fix prague tests for amsterdam

288c9c4

fix(tests): fix cancun tests for amsterdam

0c98245

fix(tests): fix tests for older forks

5438ea3

fix(tests): minor fixes

7f32d51

feat(tests): add tests for account access opcodes

ae0b7d5

gurukamath force-pushed the eip-2780/implement-eip-final branch from 33a093e to 354ecc8 Compare March 13, 2026 10:25

gurukamath closed this Mar 13, 2026

gurukamath reopened this Mar 13, 2026

refactor: rename NOCODE -> NO_CODE and add docstrings

b590833

Address PR review feedback from SamWilsn: - Rename GAS_COLD_ACCOUNT_COST_NOCODE to GAS_COLD_ACCOUNT_COST_NO_CODE across spec, testing framework, and test files - Add docstrings to EIP-2780 gas constants in vm/gas.py

danceratopz reviewed Mar 17, 2026

View reviewed changes

SamWilsn approved these changes Mar 17, 2026

View reviewed changes

src/ethereum/forks/amsterdam/vm/gas.py Outdated Show resolved Hide resolved

src/ethereum/forks/amsterdam/fork.py Show resolved Hide resolved

gurukamath mentioned this pull request Mar 17, 2026

re-factor check_transaction #2522

Open

post review updates

1ae2024

gurukamath force-pushed the eip-2780/implement-eip-final branch from 8698153 to 1ae2024 Compare March 17, 2026 15:21

danceratopz reviewed Mar 26, 2026

View reviewed changes

post review updates 2

ce7c14c

danceratopz reviewed Mar 26, 2026

View reviewed changes


	ids=["oog_phase1_at_extcodesize", "oog_phase2_at_extcodesize"],

	address_warm=False, address_has_code=False
	address_warm=False, address_has_code=True

Conversation

gurukamath commented Feb 9, 2026 • edited by LouisTsai-Csie Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🗒️ Description

🔗 Related Issues or PRs

✅ Checklist

Cute Animal Picture

Uh oh!

SamWilsn left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

marioevz commented Feb 17, 2026

Uh oh!

gurukamath commented Feb 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gurukamath commented Feb 18, 2026

Uh oh!

codecov bot commented Feb 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

gurukamath commented Feb 24, 2026

Uh oh!

gurukamath commented Feb 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

SamWilsn left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

danceratopz left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

danceratopz commented Mar 18, 2026

Uh oh!

danceratopz left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

danceratopz left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

danceratopz Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

gurukamath Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

danceratopz Mar 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

gurukamath commented Feb 9, 2026 •

edited by LouisTsai-Csie

Loading

gurukamath commented Feb 17, 2026 •

edited

Loading

codecov bot commented Feb 24, 2026 •

edited

Loading

gurukamath commented Feb 25, 2026 •

edited

Loading

danceratopz Mar 26, 2026 •

edited

Loading