test(e2e): consolidate more tests into multi-node and single-node categories - #24310
Merged
Conversation
spalladino
changed the title
spalladino
changed the title
spalladino
added
S-do-not-merge
spalladino
force-pushed
the
spl/e2e-consolidation
branch
from
June 26, 2026 12:13
df1c814 to
8733749
Compare
spalladino
changed the title
spalladino
force-pushed
the
spl/e2e-consolidation-2
branch
from
June 26, 2026 13:11
9fe3255 to
a218ac8
Compare
Relocate the four offense-detection slashing tests that already ran on the in-memory mock-gossip bus from e2e_p2p/ and e2e_slashing/ into multi-node/slashing/, swapping P2PNetworkTest for MultiNodeTestContext. Move the slashing-specific shared helpers (advanceToEpochBeforeProposer, awaitCommitteeExists, awaitOffenseDetected, awaitCommitteeKicked, findUpcomingProposerSlot, awaitProposalExecution) from e2e_p2p/shared.ts into multi-node/slashing/setup.ts and repoint the duplicate_* imports. Offense and assertion semantics are kept verbatim. Validator keys align because buildMockGossipValidators derives from getPrivateKeyFromIndex(i + 3), matching the ATTESTER_PRIVATE_KEYS_START_INDEX convention.
…MultiNodeTestContext Relocate data_withholding_slash and the two inactivity-slash tests from e2e_p2p/ into multi-node/slashing/, converting them from the real-libp2p P2PNetworkTest harness to the mock-gossip MultiNodeTestContext. Replace the standalone P2PInactivityTest base with an InactivityTest fixture over the context's fake-prover topology, and add a submitTxsThroughNode helper that repoints the context wallet at a node (the mock-gossip equivalent of submitTransactions). Offense and assertion semantics are preserved verbatim.
…ontext Relocate slash_veto_demo and the three sentinel tests (validators_sentinel, multiple_validators_sentinel, sentinel_status_slash) from e2e_p2p/ into multi-node/slashing/, converting them from P2PNetworkTest to the mock-gossip MultiNodeTestContext. validators_sentinel gains the .parallel suffix (4 top-level its). The late-joining fresh node becomes a createNonValidatorNode sentinel. Trim e2e_p2p/shared.ts down to the p2p tx helpers now that the slashing helpers live in multi-node/slashing/setup.ts. Offense/assertion semantics preserved.
Relocate the two L1-governance tests out of e2e_p2p/ into a new multi-node/governance/ folder. upgrade_governance_proposer converts to the mock-gossip MultiNodeTestContext (MOCK_GOSSIP_MULTI_VALIDATOR_OPTS + GOVERNANCE_TIMING). add_rollup stays on the P2PNetworkTest harness for now (its bootstrap-node restart and new-rollup node migration depend on real libp2p); only its imports and describe name are repointed. Mock conversion of add_rollup is deferred. Add the src/multi-node/governance/*.test.ts CI globs in bootstrap.sh, carry add_rollup's 20m timeout case to the new path, repoint the attested_invalid_proposal .test_patterns.yml entry, drop the stale valid_epoch_pruned_slash entry, and document the slashing/governance folders in the multi-node README.
Now that the slashing/sentinel/governance tests have moved to multi-node/, e2e_p2p/ holds only real-libp2p transport tests, so rename the folder to p2p/ for category consistency and delete reex.test.ts (pays setup cost for zero tests). The src/e2e_!(prover)/*.test.ts auto-glob stops matching after the rename, so add explicit src/p2p/*.test.ts globs (and repoint the reqresp glob) in both bootstrap.sh test arrays, making the compat name derivation robust to the slash-bearing p2p path. Repoint the blanket src/e2e_p2p/.* flake pattern to src/p2p/.*, the add_rollup harness import, and the AUTOMINE preset doc comment. reqresp_no_handshake stays until its flake is fixed.
The converted test relies on test.teardown() to stop nodes, so the outer nodes variable (and the AztecNodeService import) are no longer read.
…llers Introduce single-node/setup.ts with setupWithProver, a thin wrapper over SingleNodeTestContext.setup carrying today's fake-prover defaults, and migrate every single-node call site (proving, partial-proofs, l1-reorgs, recovery, misc) from SingleNodeTestContext.setup(...) to setupWithProver(...). Pure rename with no behavior change; the named factory establishes the single-node setup surface over the prover axis (the no-prover setupBlockProducer and real-BB setupWithRealProver factories, plus relocating the prod-seq and prover test files, follow separately). MultiNodeTestContext.setup is left untouched.
Moves multi-node/governance/add_rollup off P2PNetworkTest (real libp2p) onto the in-memory MockGossipSubNetwork, matching the sibling governance and slashing conversions. On the mock bus there is no bootstrap node or peer discovery, so the real-libp2p restart dance (stop/clear/re-add the bootstrap node, let new peers rediscover) collapses to: stop the old nodes/prover, warp, then spawn new nodes/prover on the new rollup version — they simply re-attach to the shared bus. The new rollup is deployed with a genesis archive root derived from the same funded-account set and timestamp the context used, so its genesis matches context.genesis. That lets the fake prover node (which always runs on context.genesis) prove the new rollup's checkpoints. Nodes repoint at the new rollup via rollupVersion alone; the aztec-node factory re-derives the L1 contract addresses from the registry, so the redundant l1Contracts override is dropped. Every governance assertion (rollup-add, migration, post-migration state, cross-chain bridging on both rollups) is preserved verbatim.
Move e2e_block_building, e2e_debug_trace, and e2e_multiple_blobs into single-node/block-building/ and convert them from raw setup() onto SingleNodeTestContext via the setupBlockProducer (prover-less) factory. Preserve each describe's behavior: block_building keeps aztecProofSubmissionEpochs=1 in its reorg describe so the prune/reinclude still fires, and multiple_blobs swaps its hand-rolled IDLE poll for the context's waitForSequencerState helper.
Move the four e2e_sequencer tests (escape_hatch_vote_only, gov_proposal,
reload_keystore, slasher_config) into single-node/sequencer/ and convert
them onto SingleNodeTestContext via setupBlockProducer. Each preserves its
explicit opts (committee size, slot/epoch durations, proof-submission
window, governance round/quorum) and its PXE syncChainTip semantics
('proposed', matching the original raw setup() default). gov_proposal keeps
its .parallel suffix; the now-empty e2e_sequencer/ directory is removed.
Move e2e_synching into single-node/sync/ and convert its raw setup() calls
onto SingleNodeTestContext via setupBlockProducer; its prune/reorg coverage
is L1-reorg- and manual-markAsProven-driven (no prover), and it pins
aztecProofSubmissionEpochs=1 to keep the prune-timeliness window the
fixture math was tuned against.
Repoint the bootstrap.sh test-list array: add the
single-node/{block-building,sequencer,sync}/*.test.ts globs, drop the
hardcoded e2e_block_building line and its exclusion from the generic
e2e_!(...) pattern, and carry its 25m timeout + AVM-input dump to
single-node/block-building/block_building via the case mechanism (the
loop's set_dump_avm already applies per test). Extend the
avm_check_circuit glob one level deeper for the 3-segment dump path.
Move e2e_prover/{full,client}.test.ts into single-node/prover/ and drop
the redundant describe-title prefixes (full_prover/client_prover ->
single-node/prover/{full,client}). Repoint the two special CI jobs for
the heavy real/fake full-prover run at the new path, preserving the
CI_FULL real (CPUS=16:MEM=96g:TIMEOUT=20m, e2e_prover_full_real) vs fake
(FAKE_PROOFS=1, e2e_prover_full_fake) split. The prover/ subfolder is
intentionally left out of the generic test glob, matching today: full
runs only via its dedicated jobs and client runs in no CI job.
Make FullProverTest extend SingleNodeTestContext so the real-prover suite
joins the single-node category over the none/fake/real prover axis. Extract
the context-hydration tail of SingleNodeTestContext.setup into a protected
hydrateFromContext(context) that reads slot/epoch durations from the resolved
context.config; FullProverTest reuses it (and super.teardown) for the rollup /
epoch-cache / chain-monitor / node-tracking / teardown machinery.
The environment build stays byte-for-byte today's: FullProverTest still calls
the raw setup(0, {...}) with its own opts (realVerifier L1, fundSponsoredFPC,
2 funded schnorr accounts, no top-level realProofs) rather than routing through
the base's default node config, since the base hardcodes aztecEpochDuration:6
whereas the real-BB job runs on the L1 env default. Everything prover-specific
is kept verbatim: the real BB prover-node config, the TokenSimulator harness,
the snapshot/account setup, the epoch warp + markAsProven, and the proven PXE.
The prover-node-before-BB-destroy teardown order is preserved.
The e2e_prover/ directory moved to single-node/prover/ in the Group B consolidation, so the !(prover) extglob exclusion on src/e2e_*/*.test.ts no longer excludes anything. Simplify to src/e2e_*/*.test.ts; the matched set is identical (153 files before and after).
…ropose revert The block-building conversion moved debug_trace onto setupBlockProducer / SingleNodeTestContext.setup, which defaults aztecEpochDuration to 6. The original raw setup() used the production default of 32. With epoch length 6 an epoch boundary lands at slot 6; the proposer-corruption test intercepts every propose and polls long enough to cross it, where proposer selection changes for the new epoch and the propose silently reverts (allowFailure:true inside the multicall -> empty logs -> propose_action_not_successful), so blocks stop being checkpointed and the 60s block-number wait times out. Pin aztecEpochDuration:32 to restore the single-epoch run the test had before the move.
The prod-seq tests were written against the raw setup() helper's 32-slot production epoch. After folding them onto setupBlockProducer the context's own aztecEpochDuration default of 6 applied, landing an epoch boundary mid-test where proposer selection changes and the boundary-slot propose silently reverts (no checkpoint lands, the chain stops). debug_trace caught it in CI; block_building, multiple_blobs, synching, gov_proposal, reload_keystore and slasher_config relied on the same default and were masked by fail-fast. Restore the 32-slot epoch in the factory; tests that need a shorter epoch still override it.
…ion collision The MultiNodeTestContext conversion deployed the new rollup with a genesisArchiveRoot equal to the context's primary rollup (same funded set + timestamp). Since the on-chain rollup version is uint32(keccak256(abi.encode(config, genesisState))), the new rollup's version collided with the primary's, so executing the governance proposal reverted with Governance__CallFailed -> Registry__RollupAlreadyRegistered. Offset the new rollup's genesis timestamp by 1s so its archive root (and thus version) differs, and repoint context.genesis at that genesis before spawning the migrated nodes/prover so the fake prover can still prove the new rollup's checkpoints. This restores the original test's semantics, where the new rollup ran on its own genesis.
…g tests The conversion onto setupBlockProducer dropped the funded-account count (setup(1, ...)) for the 'can simulate public txs while building a block', 'clears up all nullifiers if tx processing fails', and reorgs tests, while they still destructure accounts: [ownerAddress] from the context. With numberOfAccounts defaulting to 0, ownerAddress was undefined, producing 'Undefined argument admin/owner' and 'Account not found in wallet'. Pass numberOfAccounts: 1 to restore the single funded account these tests rely on.
FullProverTest builds its own real-BB environment rather than routing through the factory, so setupWithRealProver had no callers.
Move e2e_fees/ into single-node/fees/ (joining the fee-asset price-oracle test) and reconcile FeesTest onto SingleNodeTestContext. FeesTest now extends the base and builds its prover-backed environment with its own setup() opts, then calls hydrateFromContext to reuse the base rollup / epoch-cache / chain-monitor / node-tracking / teardown machinery, keeping only the fee/gas domain setup (FPC, fee juice, banana token) on the harness. The block-based catchUpProvenChain is preserved verbatim. Adds the .parallel suffix to the multi-it files and repoints the .test_patterns.yml fees entries.
Move e2e_cross_chain_messaging/ into the new single-node/cross-chain/ folder and reconcile CrossChainMessagingTest onto SingleNodeTestContext. The harness now extends the base and builds its environment with its own setup() opts (optional prover node, EpochTestSettler auto-proving otherwise), then calls hydrateFromContext to reuse the base rollup / epoch-cache / chain-monitor / node-tracking / teardown machinery, keeping the CrossChainTestHarness domain object and the L1 inbox/outbox handles on the harness. The shared cross_chain_test_harness.ts stays in src/shared/ (it is also used by bench/client_flows and shared/uniswap_l1_l2); the moved files' import paths are repointed accordingly. Adds the single-node/cross-chain/ CI glob to both bootstrap arrays, repoints the per-test timeout case and the .test_patterns.yml entries, and applies the .parallel suffix to the multi-it files.
spalladino
force-pushed
the
spl/e2e-consolidation-2
branch
from
June 26, 2026 20:55
f5a11ea to
94e41a4
Compare
The clean-lite step run on every yarn-project bootstrap build deletes git-ignored files (build artifacts). It already spared node_modules and .yarn; also spare the top-level tmp/ scratch directory so local scratch files survive a build.
These three suites were converted to .parallel.test.ts in the e2e consolidation, which makes the CI runner extract each it() and run it in its own isolated container via run_test.sh "<test name>". That breaks them: - token_bridge_private: the second test relies on the L1 tokens minted by the first test, so in isolation the portal deposit reverts with ERC20InsufficientBalance. - failures (fees): "includes transaction that error in teardown" snapshots maxFeesPerGas at setup, but in isolation the chain reaches a higher feePerL2Gas by the time the tx executes, so the FPC setup phase reverts and the tx is dropped instead of mined-reverted. - token_bridge_failure_cases: two of its test names start with "Can't", and the runner's extract_test_names regex truncates at the apostrophe to "Can", producing an ambiguous -t filter that never runs the intended test. These suites share state via beforeAll and were only ever validated as sequential whole files (their bodies are byte-identical to the base branch). Rename them back to .test.ts so the runner executes the whole file in order. Suites with per-test beforeEach setup (token_bridge_public, etc.) remain parallel-safe and are left as .parallel.test.ts.
PhilWindle
approved these changes
Jun 29, 2026
spalladino
added a commit
that referenced
this pull request
Jul 1, 2026
## Motivation A-1178 (#24281) measures each e2e test's wall-clock as four coarse buckets (`setupFnMs`, `beforeHooksMs`, `bodyMs`, `afterHooksMs`). That tells us a test is slow but not *where* it loses time. Most e2e tests spend their wall-clock in the same handful of operations — standing up the environment, waiting for a tx to be mined/checkpointed/proven, waiting for a committee to form or an offense to be detected, client-side proving of seed txs, and warp scans hunting for a proposer slot. This decomposes that time into **named spans** so a single aggregate query over a full CI run answers *"across the suite, how much total wall-clock goes into proving / spinning up nodes / waiting for checkpoints?"* — a ranked list of where to invest in speedups. It extends A-1178 and builds on the consolidation (#24201/#24310) and helpers (#24404) work that turned each repeated operation into one shared definition, so wrapping it once instruments every test with no per-test edits. This is **instrumentation / data-gathering only** — no test behavior or timing changes. ## Approach - A generic `testSpan(name, fn)` / `testSpanSync(name, fn)` wrapper (`fixtures/timing.ts`) records `{ owner, name, start, end }` into the shared collector installed by the timing environment. When `TEST_TIMING_FILE` is unset there is no collector, so `testSpan()` calls `fn()` directly — exactly zero-cost, with no clock reads. - At flush, the timing environment groups spans by owner then tag and computes four numbers per `(owner, tag)`: - `count` — multiplicity is itself a signal (waited for a checkpoint 14× points at a loop to batch). - `totalMs` — naive sum (correct for serial repeats). - `busyMs` — duration of the **union** of the spans' intervals; concurrency-correct, so a `Promise.all` of 12 concurrent 3s spans reads ~3s instead of ~36s. The `busyMs ≪ totalMs` signature flags work that is run serially where it could be parallel. - `maxMs` — longest single occurrence, to catch one pathological wait hiding in a cheap average. - The aggregates attach as an additive `spans` map on each `type:"test"` / `type:"suite"` JSONL line. `setupFnMs` / `teardownFnMs` are kept for back-compat, now derived from the `setup:env:<mode>` / `teardown:env` tags. - Tags follow a stable `category:label` taxonomy (`setup:`, `wait:`, `tx:`, `warp:`, `wallet:`, `deploy:`, `other:`), tagged **by concept, not by function** — e.g. every checkpoint waiter maps to `wait:checkpoint` — so a label is a forever aggregation key regardless of which helper a test happened to call. - **One clock.** All spans use `performance.now()`. I verified Node's `perf_hooks` clock is process-wide: a `performance.now()` taken inside a fresh `vm` context (the jest sandbox realm) shares the same monotonic origin as the host realm (delta ~0.002ms; a fresh `perf_hooks.performance` inside a separate context anchors to the same `timeOrigin`). Interval-merging for `busyMs` across the sandbox and host realms is therefore valid, so **`busyMs` is included** (no fallback to `totalMs`-only needed). - **Background loop.** `startMempoolFeeder` runs interleaved with arbitrary tests; without care its prove/send spans would smear onto whichever test was current when each round fired. Its production is run inside an `AsyncLocalStorage`-scoped owner override (`other:mempool-feeder`), which pins every span in its async call tree to a fixed owner — isolated from concurrent test-body spans. That owner matches no test/suite record, so the feeder's spans are cleanly excluded from the per-test view. ## Changes by phase - **Phase 0 — collector + `testSpan()` + flush aggregation.** New `fixtures/timing.ts` (`testSpan`/`testSpanSync`/`withTestSpanOwner`). `timing_env.mjs` collector generalized from `fnSpans` to a `spans` array; `finalizeAndFlush` groups by owner/tag, merges intervals for `busyMs`, attaches the `spans` map, and derives the back-compat fields. - **Phase 1 — shared waits + `setup:node`.** Wrapped the waiters in `fixtures/wait_helpers.ts` (`wait:proposed` / `wait:proven`, `wait:checkpoint` / `wait:proven-checkpoint`, `wait:tx-mined`, `wait:l2-to-l1-witness`, `wait:pending-tx`, `wait:sequencer-state`) and the wait methods on `SingleNodeTestContext` / `MultiNodeTestContext` (`wait:epoch`, `wait:slot`, `wait:proof-window`, `wait:node-sync`, `wait:proof-submitted`, `wait:offense`, the multi-node `wait:checkpoint`/`wait:proven-checkpoint`/`wait:block` convergence waiters). Wrapped `createNode`/`createProverNode` with `setup:node`. - **Phase 2 — decompose `setup:env`.** Cracked the opaque setup in `fixtures/setup.ts` into `setup:env:anvil`, `:l1-deploy`, `:sequencer-start`, `:prover-node`, `:pxe`, and `wallet:create`. The top-level `setup:env` is tagged with the prover mode (`setup:env:none` / `:fake` / `:real`) so the three factories are comparable. - **Phase 3 — tx leaf wraps.** `proveInteraction` → `tx:prove`, `ProvenTx.send` → `tx:send`; everything else (`proveTxs`/`proveAndSendTxs`/the submit helpers) aggregates through those leaves. `startMempoolFeeder` handled as above. - **Phase 4 — slashing/governance waiters.** Wrapped `awaitCommitteeExists` (`wait:committee`), `awaitOffenseDetected` (`wait:offense`), `awaitCommitteeKicked` (`wait:committee-kicked`), `awaitProposalExecution` (`wait:slash-execution`), and the `findUpcomingProposerSlot` / `advanceToEpochBeforeProposer` / `findSlotsWithProposers` scans (`warp:find-proposer`). - **Phase 5 — reporting.** The JSONL schema now carries the per-line `spans` map, and `TEST_TIMING_SPANS=1` emits one `type:"span"` line per occurrence (owner, name, ms) for deep dives. The repo now also ships a `track-e2e-times` skill (`yarn-project/.claude/skills/track-e2e-times/`) covering how to **collect and aggregate** the span timings locally: run the suite with `TEST_TIMING_FILE` set, find the per-worker JSONL, and print per-test sums plus the ranked span leaderboard (via the bundled `row.sh`). Only the step of publishing aggregate numbers to a running tracking log remains a separate out-of-band workflow. The leaderboard rollup is a small additive jq over the new `spans` maps: ``` jq -rs '[ .[] | select(.type=="test" or .type=="suite") | (.spans // {}) | to_entries[] ] | group_by(.key) | map({ tag: .[0].key, count: (map(.value.count) | add), busyMs: (map(.value.busyMs) | add), maxMs: (map(.value.maxMs) | max) }) | sort_by(-.busyMs) | .[] | "\(.busyMs)\t\(.count)\t\(.maxMs)\t\(.tag)"' ``` The existing `row.sh` is unaffected — it filters `type=="test"` and reads the unchanged `setupFnMs`/`bodyMs`/etc., ignoring the new `spans` key and `type:"span"` lines. ## Notes - `waitForEpoch`/`waitForSlot` (added to `@aztec/ethereum`'s `rollup_cheat_codes.ts` by #24404) are deliberately **not** instrumented in-package: that package should not depend on the e2e timing collector, and the `wait:epoch`/`wait:slot` concepts they cover are already captured at the e2e-context wrapper layer (`waitUntilEpochStarts`, the slot waiters). They have no e2e callers yet; a future call site picks them up via its context wrapper. - Spans do not partition `bodyMs` — a parent span includes its children, so `sum(spans) ≠ bodyMs`. Tagging is at the leaf wait/setup/tx level where additivity holds; the few intentional nests (e.g. `wait:committee-kicked` contains `wait:slash-execution`) carry distinct tags. Verified: `yarn build`, `yarn lint end-to-end`, `yarn format --check end-to-end` all pass. The flush aggregation (busyMs interval merge, back-compat derivation, feeder-owner exclusion, opt-in raw lines) is covered by unit tests in `end-to-end/src/shared/timing_env.test.ts`, which exercise the extracted `aggregateSpans` / `foldSpansInto` pure functions directly. Fixes A-1179 ## Example timing output Each e2e test run in CI emits one JSONL record per test and per suite, carrying a `spans` map keyed by `category:label` tag, where each value is `{count, totalMs, busyMs, maxMs}` (`busyMs` is the de-overlapped union duration, so concurrent occurrences of a span are counted once). An illustrative `type:"test"` line from CI run 28469687883 (commit de3635e): ```json { "suite": "optimistic.parallel", "type": "test", "name": "single-node/proving/optimistic happy path proves multiple epochs via checkpoint-driven flow", "status": "passed", "setupFnMs": 4707, "beforeHooksMs": 4717, "bodyMs": 327324, "teardownFnMs": 615, "afterHooksMs": 624, "totalMs": 332668, "startedAt": "2026-06-30T19:20:58.279Z", "spans": { "setup:env:anvil": { "count": 1, "totalMs": 82, "busyMs": 82, "maxMs": 82 }, "setup:env:l1-deploy": { "count": 1, "totalMs": 470, "busyMs": 470, "maxMs": 470 }, "setup:env:sequencer-start": { "count": 1, "totalMs": 3141, "busyMs": 3141, "maxMs": 3141 }, "setup:env:prover-node": { "count": 1, "totalMs": 139, "busyMs": 139, "maxMs": 139 }, "setup:env:pxe": { "count": 1, "totalMs": 391, "busyMs": 391, "maxMs": 391 }, "wallet:create": { "count": 1, "totalMs": 79, "busyMs": 79, "maxMs": 79 }, "setup:env:fake": { "count": 1, "totalMs": 4707, "busyMs": 4707, "maxMs": 4707 }, "wait:epoch": { "count": 5, "totalMs": 180210, "busyMs": 180210, "maxMs": 36067 }, "tx:prove": { "count": 4, "totalMs": 2123, "busyMs": 2123, "maxMs": 621 }, "tx:send": { "count": 4, "totalMs": 144305, "busyMs": 144305, "maxMs": 36124 }, "wait:proven-checkpoint": { "count": 4, "totalMs": 0, "busyMs": 0, "maxMs": 0 }, "wait:node-sync": { "count": 4, "totalMs": 401, "busyMs": 401, "maxMs": 101 }, "teardown:env": { "count": 1, "totalMs": 615, "busyMs": 615, "maxMs": 615 } }, "commit": "de3635e74eb89071b52939d1583d072ba2c3852c", "branch": "spl/a1179-track-common-spans", "runId": "28469687883" } ```
vezenovm
added a commit
that referenced
this pull request
Jul 1, 2026
…stant The reverted comment (copied verbatim from before #24429) still said the value must be larger than MAX_PRIVATE_LOGS_PER_TX while setting it to 20, and cited e2e_l1_with_wall_time, which no longer exists (deleted in #24310). Replace with an accurate description of the real bound and mark the value explicitly as an unsafe test-only setting.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Continues the e2e test consolidation building on the now-merged #24201. This PR completes the
single-nodeandmulti-nodecategories (no single-node- or multi-node-topology test remains outsidethem), settles
e2e_p2p/down to transport-only, and adds a smallinfra/category.Context
Leftovers from #24201's consolidation.
e2e_p2p/had become a catch-all: offense-detection (slashing),sentinel, and governance tests sat alongside the genuine transport tests, and
e2e_slashing/duplicated theslashing topology. Most of those aren't about real networking — they exercise proposals/attestations/offense
detection that the in-memory
MockGossipSubNetworkreproduces faithfully. Separately, a long tail ofproduction-sequencer, prover, fee, cross-chain, and validator tests still lived as loose
e2e_*filesoutside the categories the previous PR established.
A note on
mockGossipSubNetwork: the flag is inert on theP2PNetworkTestpath —setup_p2p_test.ts'screateNodenever wires the mockp2pServiceFactory, so tests that set it were in fact on real libp2pdespite their comments.
MultiNodeTestContextwires the mock bus genuinely, so each conversion moves thetest onto a real in-memory bus.
Approach
multi-node/slashing/and the governance tests intomulti-node/governance/, converting each fromP2PNetworkTesttoMultiNodeTestContext(mock gossip).e2e_p2p/→p2p/once only transport tests remain, adding explicit CI globs.single-node/, over named setup factories on anone/fake/real prover axis, and make
FullProverTestaSingleNodeTestContextsub-base.FeesTest,CrossChainMessagingTest, and the multi-validatornode test — onto the category bases, and move the public-testnet smoke into a new
infra/category. Afterthis,
single-nodeandmulti-nodeare complete.Hierarchy before → after
e2e_p2p/{broadcasted_invalid_block_proposal_slash,data_withholding_slash,slash_veto_demo,inactivity_slash,inactivity_slash_with_consecutive_epochs,validators_sentinel,multiple_validators_sentinel,sentinel_status_slash}ande2e_slashing/{attested_invalid_proposal,broadcasted_invalid_checkpoint_proposal_slash}→multi-node/slashing/(sentinel folds in here — it exists to drive slashing)e2e_p2p/{add_rollup,upgrade_governance_proposer}→multi-node/governance/e2e_multi_validator/e2e_multi_validator_node→multi-node/block-production/multi_validator_node.parallel(one node hosting multiple validator keys; converted toMultiNodeTestContext)e2e_p2p/→p2p/(transport tests only;reex.test.tsdeleted)e2e_slashing/removed (now empty)e2e_block_building,e2e_debug_trace,e2e_multiple_blobs→single-node/block-building/e2e_synching→single-node/sync/e2e_sequencer/*, pluse2e_multi_eoa,e2e_publisher_funding_multi,e2e_sequencer_config→single-node/sequencer/e2e_prover/{full,client}→single-node/prover/e2e_fees/*(incl.e2e_fee_asset_price_oracle) →single-node/fees/e2e_cross_chain_messaging/*→single-node/cross-chain/e2e_bot→single-node/bot/e2e_public_testnet/*→infra/(new category)e2e_simple→ folded intosingle-node/misc/node_block_api(block-data-API check; the redundant deploy-and-proveitdropped)e2e_l1_with_wall_timedeleted (its multi-tx-block assertion is covered bysingle-node/block-building/block_building)Helpers introduced
single-node/setup.ts— two factories overSingleNodeTestContext.setup:setupWithProver(fakein-process prover — the single-node default) and
setupBlockProducer(no prover; raisesaztecProofSubmissionEpochsto1024so unproven blocks aren't pruned, PXE onsyncChainTip: 'proposed').FullProverTest,FeesTest,CrossChainMessagingTestnowextend SingleNodeTestContextvia aprotected hydrateFromContext(context)split out ofsetup, so each reuses the base'srollup/epoch-cache/chain-monitor/proven-chain-waiter/node-tracking/teardown machinery while keeping its own
domain setup (real-BB env / FPC + gas state /
CrossChainTestHarness).advanceToEpochBeforeProposer,awaitCommitteeExists,awaitOffenseDetected,awaitCommitteeKicked,findUpcomingProposerSlot,awaitProposalExecution) movefrom
e2e_p2p/shared.tsintomulti-node/slashing/setup.ts, plussubmitTxsThroughNode.multi-node/slashing/inactivity_setup.ts(InactivityTest) andmulti-node/governance/setup.ts(
GOVERNANCE_TIMING).Semantic changes
P2PNetworkTestharness to the genuine in-memory mock-gossip bus ofMultiNodeTestContext. Offense andassertion semantics are preserved verbatim; node creation maps
createNode(s)/createProverNodeontocreateValidatorNodeAt/createNonValidatorNode/the context's fake prover; real-libp2p concerns(
waitForP2PMeshConnectivity, bootstrap nodes,removeInitialNode) drop out.add_rollupconverts toMultiNodeTestContexttoo: its real-libp2p restart dance collapses to "stop oldnodes/prover, warp, spawn new nodes/prover on the new rollup version" re-attaching to the shared in-memory
bus; the new rollup is deployed with a genesis matching
context.genesisso the fake prover can prove it.setupBlockProducer); since the env default proof-submissionwindow is
1, the factory raises it to1024so blocks aren't pruned. Tests that depend on crossing thewindow (the
block_buildingreorg cases,synching's prune) pinaztecProofSubmissionEpochsexplicitly.FeesTest/CrossChainMessagingTestkeep their bespoke prover-backedsetup()opts byte-for-byte andpreserve
catchUpProvenChain/advanceToEpochProvenexactly; only the shared environment machinery movesto the base.
FullProverTestlikewise keeps its real-BB env (L1 env epoch default, not the base's6).Test fixes
.parallelsuffix where they have >1 top-levelit:attested_invalid_proposal,broadcasted_invalid_checkpoint_proposal_slash,validators_sentinel,multi_validator_node, and the fees/cross-chain files (failures,gas_estimation,private_payments;all five cross-chain files).
bootstrap.shper-test timeout case fore2e_cross_chain_messaging/l1_to_l2(which no longermatched the
.paralleltest name) is repointed tosingle-node/cross-chain/l1_to_l2.parallel.API changes
Test-infra only.
Open decisions (not blocking; for follow-up)
single-node/prover/client.test.tsruns in no CI job. Pre-existing — the olde2e_prover/dir wasnever globbed (only
fullhad dedicated real/fake jobs); the move preserved that. Decide whether to startrunning it (e.g. add
src/single-node/prover/!(full).test.tsto the CI globs). Tracked as A-1301.Part of A-1176