test(evm): migrate response-reading lib.js helpers off -b block (CON-256) by wen-coding · Pull Request #3406 · sei-protocol/sei-chain

wen-coding · 2026-05-07T05:23:43Z

Summary

Picks up where #3363 left off. Migrates the lib.js helpers that read post-execution events from -b block to -b sync + per-helper side-effect waits.

Pattern (introduced in #3363, extended here): submit with -b sync, then poll the actual on-chain side effect rather than seid q tx (which doesn't return under Autobahn). When the side effect is observable, return the post-state value; when it isn't, fall back to admin-sequence-advance + block-walk to recover the inclusion height.

Migrated:

storeWasm — side-effect: max wasm code_id grows (new getMaxWasmCodeId)
instantiateWasm — side-effect: new contract appears under codeId (new listContractsByCode)
registerPointerForERC20/721/1155 — side-effect: seid query evm pointer returns non-empty (factored into shared registerPointerForCw)
proposeCW20toERC20Upgrade, proposeParamChange — side-effect: new gov proposal with matching title (factored into findProposalByTitle)
executeWasm, associateWasm — fallback path: waitForAdminTxCommit waits on admin sequence advance, then block-walks to recover the actual inclusion height (no easy single-query side effect for arbitrary wasm execute / contract association)

Wait-timeout cases are wrapped with the same error prefixes ("storeWasm failed", "instantiateWasm failed", "contract deployment failed") that callers already match on, so existing assertions in DisableWasmTest and similar continue to work.

Other:

Removed associateSigner + hardcoded seilocalSignerPrivateKeys — setupSigners now uniformly funds the EVM address and reads back the sei address. The conditional branch became redundant after test(evm): wipe stale .openzeppelin manifest before upgradeability test #3431.
Dropped 3 res.code assertions in CW1155toERC1155PointerTest.js — res.code is now CheckTx (always 0 for mempool admission), not DeliverTx. Outcome verification stays via the existing post-state balance checks.

github-actions · 2026-05-07T05:24:38Z

The latest Buf updates on your PR. Results from workflow Buf / buf (pull_request).

Build	Format	Lint	Breaking	Updated (UTC)
`✅ passed`	`✅ passed`	`✅ passed`	`✅ passed`	May 19, 2026, 9:27 PM

wen-coding · 2026-05-07T05:50:47Z

This doesn't work until we link in block db which supports tx lookup.

codecov · 2026-05-07T20:04:51Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 59.04%. Comparing base (80a4364) to head (9f05f86).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3406      +/-   ##
==========================================
- Coverage   59.05%   59.04%   -0.01%     
==========================================
  Files        2188     2188              
  Lines      182088   182084       -4     
==========================================
- Hits       107530   107520      -10     
  Misses      64925    64925              
- Partials     9633     9639       +6

Flag	Coverage Δ
sei-db	`70.41% <ø> (ø)`

Flags with carried forward coverage won't be shown. Click here to find out more.
see 33 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

… waits The remaining -b block helpers in lib.js previously needed their post- execution events to extract the returned identifier (code_id, contract address, proposal_id, pointer_address). Under Autobahn -b block hangs, and the seid tx indexer isn't wired yet, so neither -b block nor a generic poll-by-txhash works. Replace each with a per-caller side-effect wait that goes through a query path that does work under Autobahn: - storeWasm: poll max code_id in `seid q wasm list-code` - instantiateWasm: diff `seid q wasm list-contract-by-code` before/after - proposeCW20toERC20Upgrade: scan gov proposals for matching title (extracted findProposalByTitle helper, also used by proposeParamChange) - registerPointerForERC20/721/1155: factored into registerPointerForCw; poll `seid q evm pointer` until non-empty - executeWasm, associateWasm: new waitForAdminTxCommit helper, waits for sender sequence to advance (the natural commit signal when the helper itself doesn't have a specific side effect to poll) The four EVM pointer deploy helpers (deployErc{20,721,1155}PointerFor*, deployErc20PointerNative) already polled via `provider.getTransactionReceipt` so only the broadcast mode needs to change. executeWasm now returns only the CheckTx response (not DeliverTx), so callers that asserted `res.code` to verify failure have to verify via state instead. CW1155toERC1155PointerTest already followed each such check with the state assertion that actually proves the outcome, so the res.code lines are removed as redundant. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

…merge The merge of #3431 (origin/wen/wipe_env_before_contract_upgradeability_test) silently kept our side of setupSigners because #3363 had touched the same region. Apply #3431's intent here: drop associateSigner and the seilocalSignerPrivateKeys lookup, and have setupSigners always use the fundAddress + self-send-tx flow. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

… waitForAdminTxCommit Three cleanups after CI surfaced regressions from the -b sync refactor: 1. storeWasm / instantiateWasm / registerPointerForCw wrap the waitForCondition timeout in a `<helper> failed: ...` prefix so callers that match on the prefix string (DisableWasmTest's wasm- disabled cases, ERC20toCW20PointerTest's pointer-of-a-pointer case) keep working. The chain rejects at DeliverTx in these cases (CheckTx accepts), so we surface the failure as a timeout on the side effect — the prefix preserves the existing error contract. 2. executeWasm / associateWasm now rewrite `response.height` to the actual inclusion block. The CheckTx response from -b sync carries `height: "0"`, but SeiEndpointsTest reads `res.height` to bound its sei_getLogs range. Find the inclusion block by scanning blocks between submit and observed commit for the tx hash (hex-uppercase SHA-256 of base64-decoded `block.data.txs` entries). 3. Add `getCurrentBlockHeight` and `findInclusionBlock` to support (2). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

- storeWasm: capture new code_id inside the wait. A second getMaxWasmCodeId() after the wait raced with any concurrent store landing in the same window. Serial test execution made the race dormant; capture-during-wait closes it. - instantiateWasm: same. The wait used .some() (boolean) and the return used .find() (specific entry) — two separate queries, each of which could see different state. Now captured inside the wait. - waitForAdminTxCommit: when sender sequence advanced (tx landed somewhere) but findInclusionBlock returned null, response.height was silently left at the CheckTx default. Log when this happens so the anomaly is visible in test output rather than producing a misleading height value. - CW1155toERC1155PointerTest: drop the unused res from three executeWasm calls (the res.code assertions that read it are gone), and on the negative-case test, add a comment pointing at the unchanged balances as the assertion that the unauthorized transfer was rejected at DeliverTx. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

philipsu522 · 2026-05-19T23:10:22Z

-    const output = await execute(command);
-    return JSON.parse(output);
+    const command = `seid tx wasm execute ${contractAddress} "${jsonString}" --amount ${coins} --from ${adminKeyName} --gas=5000000 --fees=1000000usei -y -b sync -o json`;
+    return await waitForAdminTxCommit(command)


philipsu522 · 2026-05-19T23:17:02Z

+    const seqBefore = await getAccountSequence(senderAddr)
+    const response = JSON.parse(await execute(command))
+    if (response.code !== 0) return response  // CheckTx rejection — surface immediately
+    await waitForCondition(


do we timeout in case there's some sort of bug here?

Yes, waitForCondition has a default 30s timeout, adding a note.

cursor · 2026-05-19T23:44:19Z

PR Summary

Medium Risk
Medium risk because it changes core test harness transaction submission/wait logic (wasm store/instantiate, pointer registration, gov proposals), which can introduce flakiness or false positives/negatives across many E2E tests if the new side-effect detection is incomplete or racy.

Overview
Migrates key contracts/test/lib.js helpers from -b block to -b sync and makes them wait for observable chain side effects before returning (e.g., new wasm code_id, new contract under a codeId, new gov proposal by title, and newly-registered pointer addresses), including shared polling utilities and improved error wrapping.

Adds a fallback waitForAdminTxCommit path for cases without an easy side-effect query (executeWasm, associateWasm) by waiting for the admin account sequence to advance and then scanning blocks to recover the inclusion height.

Simplifies signer setup by removing hardcoded private-key association logic and updates CW1155toERC1155PointerTest to stop asserting res.code (now CheckTx-only), relying on post-state balance checks instead.

^{Reviewed by Cursor Bugbot for commit 9f05f86. Bugbot is set up for automated code reviews on this repo. Configure here.}

Mirrors the existing "EVM Module" CI job with AUTOBAHN=true plus the GIGA_EXECUTOR/GIGA_STORAGE/GIGA_OCC flag set used by the "Autobahn EVM GIGA Module" job. Runs evm_tests.sh under Autobahn: - EVMCompatabilityTest.js (broad EVM semantics: deployment, gas, variable types, assembly, trace, log filtering, edge cases — 78 cases) - EVMPrecompileTest.js (Sei precompiles: bank, addr, gov, distribution, staking, oracle, wasm) - SeiEndpointsTest.js (sei_*ExcludeTraceFail JSON-RPC paths) - AssociateTest.js (account association mechanics) Complements the narrower "Autobahn EVM GIGA Module" (smoke-tests EVMGigaTest.js on native+ERC20 transfers only) by adding broader EVM-compat + Sei-precompile + sei-endpoint + association coverage that aren't exercised elsewhere under Autobahn. Deliberately omits the flatkv post-steps from the CometBFT "EVM Module" job; flatkv-under-Autobahn is its own concern and would make per-job failure triage harder. Verified locally on a 4-node Autobahn docker cluster: 101 passing, 1 pending, 0 failing. Autobahn EVM Interoperability deferred until #3406 (lib.js response-reading helper migration) lands; without it, the suite's CW-deploy setup hooks depend on unmigrated -b block helpers and CI passing would be accidental rather than verified. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Mirrors the existing "EVM Module" CI job with AUTOBAHN=true plus the GIGA_EXECUTOR/GIGA_STORAGE/GIGA_OCC flag set used by the "Autobahn EVM GIGA Module" job. Same 4 steps as the CometBFT job: - evm_tests.sh (EVMCompatabilityTest, EVMPrecompileTest, SeiEndpointsTest, AssociateTest) - deploy_flatkv_evm_fixture.sh (seeds historical EVM state: balance, storage, code) - flatkv_evm_test.yaml (historical EVM RPC reads via cast: balance, storage, code at past block heights) - verify_flatkv_evm_store.sh (seidb dump-flatkv smoke check on the on-disk storage bucket) Complements the narrower "Autobahn EVM GIGA Module" (smoke-tests EVMGigaTest.js on native+ERC20 transfers only) by adding broader EVM-compat + Sei-precompile + sei-endpoint + association coverage plus historical-state correctness, none of which are exercised elsewhere under Autobahn. Verified locally on a 4-node Autobahn docker cluster: evm_tests.sh: 101 passing, 1 pending, 0 failing deploy_flatkv_evm_fixture.sh: fixture seeded at heights 17567/17571 (flatkv yaml + verify steps deferred to CI; the 3 flatkv components have zero -b block usages and use only EVM RPC + seid queries, so they're Autobahn-clean.) Autobahn EVM Interoperability deferred until #3406 (lib.js response-reading helper migration) lands; without it, the suite's CW-deploy setup hooks depend on unmigrated -b block helpers and CI passing would be accidental rather than verified. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

wen-coding marked this pull request as ready for review May 7, 2026 05:34

wen-coding added the non-app-hash-breaking label May 7, 2026

wen-coding marked this pull request as draft May 7, 2026 05:37

wen-coding mentioned this pull request May 17, 2026

test(autobahn): drop stale -b block TODO #3454

Open

wen-coding force-pushed the wen/migrate_response_reading_lib_helpers_off_b_block branch 2 times, most recently from 2e0fe83 to 9d86334 Compare May 18, 2026 22:57

wen-coding and others added 4 commits May 19, 2026 14:08

ci: retrigger after EVM Module hang (no code change)

68b8f86

wen-coding force-pushed the wen/migrate_response_reading_lib_helpers_off_b_block branch from cddbc2d to 68b8f86 Compare May 19, 2026 21:09

wen-coding changed the title ~~test(evm): introduce submitAndWaitForTx; migrate storeWasm + registerPointerForERC20 (CON-256)~~ test(evm): migrate response-reading lib.js helpers off -b block (CON-256) May 19, 2026

philipsu522 approved these changes May 19, 2026

View reviewed changes

wen-coding marked this pull request as ready for review May 19, 2026 23:43

wen-coding mentioned this pull request May 20, 2026

ci: add Autobahn EVM Module (CON-256) #3468

Draft

codchen approved these changes May 20, 2026

View reviewed changes

wen-coding added this pull request to the merge queue May 20, 2026

Merged via the queue into main with commit 1dd16d0 May 20, 2026
43 checks passed

wen-coding deleted the wen/migrate_response_reading_lib_helpers_off_b_block branch May 20, 2026 03:39

wen-coding mentioned this pull request May 20, 2026

test(evm): migrate cosmos deploy bash scripts off -b block (CON-256) #3469

Draft

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

test(evm): migrate response-reading lib.js helpers off -b block (CON-256)#3406

test(evm): migrate response-reading lib.js helpers off -b block (CON-256)#3406
wen-coding merged 5 commits into
mainfrom
wen/migrate_response_reading_lib_helpers_off_b_block

wen-coding commented May 7, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented May 7, 2026 •

edited

Loading

Uh oh!

wen-coding commented May 7, 2026

Uh oh!

codecov Bot commented May 7, 2026 •

edited

Loading

Uh oh!

philipsu522 May 19, 2026

Uh oh!

philipsu522 May 19, 2026

Uh oh!

wen-coding May 19, 2026

Uh oh!

cursor Bot commented May 19, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

wen-coding commented May 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Uh oh!

github-actions Bot commented May 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

wen-coding commented May 7, 2026

Uh oh!

codecov Bot commented May 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

philipsu522 May 19, 2026

Choose a reason for hiding this comment

Uh oh!

philipsu522 May 19, 2026

Choose a reason for hiding this comment

Uh oh!

wen-coding May 19, 2026

Choose a reason for hiding this comment

Uh oh!

cursor Bot commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Summary

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

wen-coding commented May 7, 2026 •

edited

Loading

github-actions Bot commented May 7, 2026 •

edited

Loading

codecov Bot commented May 7, 2026 •

edited

Loading

cursor Bot commented May 19, 2026 •

edited

Loading