feat: exported in-process testing network by Thunkar · Pull Request #24629 · AztecProtocol/aztec-packages

Thunkar · 2026-07-09T10:20:14Z

Upstreaming the utilities build for aztec-kit so external projects can benefit from our e2e scaffolding. Essentially the same thing we already use, but pure ts rather than the unpublishable .sh script. Verified with ci-full-no-test-cache

The pure-Node signal/exit handlers in startAnvil only reap anvil when node runs its own handlers (explicit stop, Ctrl+C, normal exit, uncaught exception). On a parent hard-kill (SIGKILL/crash/OOM) they never run, and the detached anvil is orphaned — verified empirically. That is the exact failure the removed anvil_kill_wrapper.sh guarded against and a known source of CI flakiness. Spawn anvil under a small inlined bash watchdog that polls its parent (this node process) and reaps anvil via an EXIT/INT/TERM trap when the parent dies for any reason. Inlined (not a shipped .sh) so it works from the published npm tarball, with the resolved anvil binary passed via $ANVIL_BIN so it works without anvil on PATH.

The watchdog trapped cleanup on EXIT/INT/TERM, but a trapped TERM does not terminate bash — it ran cleanup and then resumed the parent-poll loop, so the supervisor never exited on stop() until killChild's 5s SIGKILL escalation. That 5s collided with jest's 5s hook timeout and hung afterEach teardown (forwarder_l1_tx_utils.test.ts). Trap the kill on EXIT only and have INT/TERM just exit (firing the EXIT trap); use 'sleep & wait' so the poll is interruptible. stop() now returns in ~1s; orphan-on-parent-death still holds.

nchamo

Some small comments

nchamo · 2026-07-13T14:47:13Z

+        await stopNode();
+        await stopAnvil();


Should we run them in parallel? So that if the first fails, the second one is also stopped?

nchamo · 2026-07-13T14:50:42Z

+  const envBin = process.env.ANVIL_BIN;
+  if (envBin && existsSync(envBin)) {
+    return envBin;
+  }


I understand having defaults if $ANVIL_BIN is not set. But if it is set and it doesn't exist, then we should probably fail, instead of using default version

mmm, yeah reworked this part

nchamo · 2026-07-13T14:51:25Z

+      () => {},
+    );
+
+    return {


Claude suggested we also implement asyncDispose

I did and I don't know if I'm a fan of the syntax, but removing the finally is neat

nchamo · 2026-07-13T14:53:00Z

+ * `maxFeesPerGas.feePerL2Gas must be >= gasFees.feePerL2Gas`. Apply via
+ * `wallet.setMinFeePadding(TEST_FEE_PADDING)` on every test wallet that sends txs.
+ */
+export const TEST_FEE_PADDING = 30;


Is this the same as PIPELINED_FEE_PADDING for the same reasons? Just checking if we could re-use it or if they are different things

It is, but for dependency reasons made PIPELINED_FEE_PADDING an alias

spalladino · 2026-07-13T20:03:47Z

+  /** Override the default per-address fee juice granted at genesis to funded addresses. */
+  initialAccountFeeJuice: Fr;


Should this be optional if it's a default override?

spalladino · 2026-07-13T20:05:15Z

+      expect(await a.node.getBlockNumber()).toBeGreaterThanOrEqual(0);
+      expect(await b.node.getBlockNumber()).toBeGreaterThanOrEqual(0);


Should we test strictly greater, to make sure block production works?

spalladino · 2026-07-13T20:08:19Z

+  if (envBin && existsSync(envBin)) {
+    return envBin;
+  }


Should we also test it's executable?

spalladino · 2026-07-13T20:19:56Z

+  /** anvil block time in seconds. Omit for automine (the default). */
+  l1BlockTime?: number;


Heads up that createLocalNetwork sets useAutomineSequencer to true by default, which sets anvil to automine during startup (see sequencer-client/src/sequencer/automine/automine_sequencer.ts start()). So a user setting eg l1BlockTime=12 will see their anvil still set to automine. I'd remove this option altogether for now, or add proper support for it if we really have a need for it.

spalladino · 2026-07-13T20:20:23Z

+ * Necessary because `deployAztecL1Contracts` shells out to bare `forge`/`solc`, and those inherit
+ * PATH from us. Since the aztec-up change that stopped polluting the user's interactive PATH,
+ * `forge`/`cast`/`anvil`/`nargo` are only reachable via `~/.aztec/current/internal-bin/`.


I'd rather modify deployAztecL1Contracts than meddling with PATH

Right, this is reworked now

…ec-packages into gj/upstream_testing_utils

spalladino

Looks good. One thing to look out for is whether the user has a local anvil version that works with Aztec. Problem is, I don't know what's the min version we need (I guess the one pinned in the docker image?).

Thunkar · 2026-07-14T13:29:35Z

Looks good. One thing to look out for is whether the user has a local anvil version that works with Aztec. Problem is, I don't know what's the min version we need (I guess the one pinned in the docker image?).

We can refine this, but I think ATM we can assume that if they're deep enough to have an e2e test that uses this external network, they also have the toolchain installed

Upstreaming the utilities build for `aztec-kit` so external projects can benefit from our e2e scaffolding. Essentially the same thing we already use, but pure ts rather than the unpublishable .sh script. Verified with `ci-full-no-test-cache` (cherry picked from commit 7e09008)

…o next (AztecProtocol#24933) Forward-ports the **prover** and **node / world-state / barretenberg** slices of the v5-next → next backlog (work merged to `v5-next` after the ~2026-07-08 cut that reshaped `next`). ## Applied (clean cherry-picks, chronological) - feat(prover): revive cancelled proving jobs from a persisted aborted state (AztecProtocol#24578) - fix(prover-node): rebuild pruned checkpoint provers and recover the epoch (AztecProtocol#24436) - feat: exported in-process testing network (AztecProtocol#24629) - fix(validator): sync world state before forking in checkpoint proposal validation (AztecProtocol#24694) ## ⚠️ Needs owner conflict-resolution (conflict against reshaped `next`; not included here) Cherry-pick onto this branch and resolve: - [ ] `git cherry-pick -x a4e3a44` — feat(world-state): support prefilled nullifiers in genesis state (AztecProtocol#24567) **[RESHAPE — v6 subsystem rewritten; re-implement, do not merge]** - [ ] `git cherry-pick -x c44e74e` — fix(prover-node): do not abort in-flight proving jobs on a clean shutdown (AztecProtocol#24579) - [ ] `git cherry-pick -x 1039d38` — feat: allow custom proof submission target address (AztecProtocol#24270) Part of the manual v5-next → next backlog sweep. Draft until conflicts are resolved and CI is green.

wip

1f11ffb

Thunkar self-assigned this Jul 9, 2026

Thunkar added ci-full Run all master checks. ci-draft Run CI on draft PRs. labels Jul 9, 2026

Thunkar changed the title ~~wip~~ feat: in-process testing network Jul 9, 2026

Thunkar added 2 commits July 9, 2026 12:32

Thunkar added ci-full-no-test-cache and removed ci-full Run all master checks. labels Jul 9, 2026

Thunkar added 2 commits July 9, 2026 14:18

Merge branch 'merge-train/fairies-v5' into gj/upstream_testing_utils

7150308

fixes

a365a6a

Thunkar changed the title ~~feat: in-process testing network~~ feat: exported in-process testing network Jul 9, 2026

Thunkar requested review from nchamo and spalladino July 9, 2026 13:41

Thunkar marked this pull request as ready for review July 9, 2026 13:41

nchamo reviewed Jul 13, 2026

View reviewed changes

spalladino reviewed Jul 13, 2026

View reviewed changes

Thunkar added 2 commits July 14, 2026 09:34

Merge branch 'merge-train/fairies-v5' of github.com:AztecProtocol/azt…

1f5c254

…ec-packages into gj/upstream_testing_utils

comments from review

1b3134b

Thunkar requested a review from nventuro as a code owner July 14, 2026 12:21

Merge branch 'merge-train/fairies-v5' into gj/upstream_testing_utils

cab4a30

Thunkar requested review from nchamo and spalladino and removed request for nventuro July 14, 2026 12:24

nchamo approved these changes Jul 14, 2026

View reviewed changes

spalladino approved these changes Jul 14, 2026

View reviewed changes

Thunkar merged commit 7e09008 into merge-train/fairies-v5 Jul 14, 2026
12 checks passed

Thunkar deleted the gj/upstream_testing_utils branch July 14, 2026 13:55

AztecBot mentioned this pull request Jul 14, 2026

feat: merge-train/fairies-v5 #24682

Merged

AztecBot mentioned this pull request Jul 21, 2026

chore: port v5-next to next #24825

Open

PhilWindle mentioned this pull request Jul 21, 2026

chore: forward-port clean v5-next backlog to next #24849

Closed

PhilWindle mentioned this pull request Jul 23, 2026

chore(port): forward-port v5-next prover + node/world-state backlog to next #24933

Merged

3 tasks

nchamo mentioned this pull request Jul 23, 2026

chore(port): forward-port remaining v5-next fairies backlog to next #24949

Merged

		/** Override the default per-address fee juice granted at genesis to funded addresses. */
		initialAccountFeeJuice: Fr;

		expect(await a.node.getBlockNumber()).toBeGreaterThanOrEqual(0);
		expect(await b.node.getBlockNumber()).toBeGreaterThanOrEqual(0);

		/** anvil block time in seconds. Omit for automine (the default). */
		l1BlockTime?: number;

Uh oh!

Conversation

Thunkar commented Jul 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nchamo left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

spalladino left a comment

Choose a reason for hiding this comment

Uh oh!

Thunkar commented Jul 14, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Thunkar commented Jul 9, 2026 •

edited

Loading