test: add Volume mount payload unit and live mount e2e coverage (3/4) by devin-ai-integration[bot] · Pull Request #1744 · e2b-dev/E2B

devin-ai-integration · 2026-08-21T12:29:47Z

Summary

3/4 of the stack replacing #1739 (base: #1743). Volume needed no migration — the existing suites (volume/volume.test.ts, volume/file.test.ts, volume_sync, volume_async) are already msw/monkeypatch based and stay in the default tier. This PR only fills the two gaps the tier split exposed.

Mocked, default tier — how a volume reaches the create-sandbox request:

// volume/mountPayload.test.ts, python: tests/shared/volume/test_mount_payload.py
Sandbox.create()                                   // no volumeMounts key
Sandbox.create({ volumes: { '/mnt/data': 'my-volume' } })
Sandbox.create({ volumes: { '/mnt/data': new Volume('vol-1', 'my-volume', 'volume-token') } })
→ body.volumeMounts === [{ volumeName: 'my-volume', mountPath: '/mnt/data' }]

E2E tier — the one Volume behavior a mock can't assert, that content actually survives on the real mount across sandboxes:

// volume/mount.test.ts, python: {sync,async}/volume_{sync,async}/test_mount.py
writer = create({ volumes: { '/mnt/data': volume } }); write('/mnt/data/hello.txt'); kill()
reader = create({ volumes: { '/mnt/data': volume } }); read('/mnt/data/hello.txt') === written

Not yet run against live infra.

Link to Devin session: https://app.devin.ai/sessions/e237ae2e3d8043fbabca7017400e6e57
Requested by: @mishushakov

devin-ai-integration · 2026-08-21T12:29:50Z

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

Disable automatic comment, CI, and merge conflict monitoring

chatgpt-codex-connector · 2026-08-21T12:29:53Z

Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits.
Credits must be used to enable repository wide code reviews.

changeset-bot · 2026-08-21T12:29:56Z

⚠️ No Changeset found

Latest commit: 0c481b0

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

github-actions · 2026-08-21T12:30:52Z

Package Artifacts

Built from 82bfaa2. Download artifacts from this workflow run.

JS SDK (e2b@2.44.2-devin-1787315092-test-tiers-volume.0):

npm install ./e2b-2.44.2-devin-1787315092-test-tiers-volume.0.tgz

CLI (@e2b/cli@2.16.4-devin-1787315092-test-tiers-volume.0):

npm install ./e2b-cli-2.16.4-devin-1787315092-test-tiers-volume.0.tgz

Python SDK (e2b==2.44.0+devin.1787315092.test.tiers.volume):

pip install ./e2b-2.44.0+devin.1787315092.test.tiers.volume-py3-none-any.whl

devin-ai-integration

TASTE.md review — test-only PR, so I checked the rules that bind on how tests exercise the public surface: parity across JS / sync Python / async Python (T-1, T-2), factory-vs-constructor acquisition (T-4), the positional/optional argument rule (T-3), option naming (T-10), and the timeout/config rules (T-44…T-50, nothing relevant here). The tier split itself and the msw / monkeypatch shapes match the convention the base PR established, and the e2e tier correctly holds only the one behavior a mock can't assert.

3 findings, all about the Volume-instance mount source:

JS payload test builds a Volume through its constructor (T-4).
Python does the same, and positionally, unlike every existing Volume test (T-4, T-3).
The async payload mirror covers only one of the three sync cases (T-1, T-2).

Not tied to a changed line:

The PR description documents the option as volumes: { '/mnt/data': 'my-volume' }, but the actual surface (and the code in this PR) is volumeMounts / volume_mounts. Worth fixing in the description so it doesn't read as a rename — option names must mirror exactly across SDKs (T-10).
Non-TASTE nit: the e2e volume names use Date.now() in JS and uuid4() in Python. uuid4() in both would avoid collisions between concurrent runs against the same account.

devin-ai-integration · 2026-08-21T12:32:49Z

+})
+
+test('Sandbox.create accepts a Volume instance as the mount source', async () => {
+  const volume = new Volume('vol-1', 'my-volume', 'volume-token')


T-4 — objects with remote lifecycles are obtained through async static factories (Volume.create, Volume.list); constructors are internal wiring. This is the first new Volume(...) call site in the JS SDK, and a test is documentation: it cements a construction path TASTE says users never take, and it also spells out the optional token positionally, which is exactly the chain T-3 rules out.

msw is already intercepting here, so the compliant form is to get the volume the way a caller would — add a http.post(apiUrl('/volumes'), …) handler returning { volumeID: 'vol-1', name: 'my-volume', token: 'volume-token' } and then const volume = await Volume.create('my-volume', { apiKey: TEST_API_KEY }). That keeps the test asserting the public path end to end instead of a hand-assembled object.

devin-ai-integration · 2026-08-21T12:32:50Z

+def test_create_accepts_a_volume_instance_as_the_mount_source(
+    monkeypatch, test_api_key
+):
+    volume = Volume("vol-1", "my-volume", "volume-token")


T-4 / T-3 — same constructor point as the JS side, plus a local inconsistency: token, domain, debug and proxy are optional on Volume.__init__, and every existing Volume test constructs by keyword (tests/sync/volume_sync/test_volume_content.py, tests/test_volume_client.py). Passing them positionally is the argument order TASTE rejects.

At minimum match the existing convention:

Suggested change

volume = Volume("vol-1", "my-volume", "volume-token")

volume = Volume(volume_id="vol-1", name="my-volume", token="volume-token")

Better, mirror whatever the JS test ends up doing and obtain the instance from Volume.create against the mocked transport.

devin-ai-integration · 2026-08-21T12:32:50Z

+async def test_async_create_maps_mount_paths_to_named_volume_mounts(
+    monkeypatch, test_api_key
+):
+    body = await _async_request_body(
+        monkeypatch, test_api_key, {"/mnt/data": "my-volume"}
+    )
+
+    assert body["volumeMounts"] == [{"name": "my-volume", "path": "/mnt/data"}]


T-1 / T-2 — sync and async are full mirrors with the same method names and semantics, so their coverage shouldn't diverge either. Sync gets three cases (omitted, string name, Volume instance); async gets only the string-name one. The gap that matters is the instance case: AsyncSandbox.create maps AsyncVolume instances through its own transform in sandbox_async/main.py, and that mapping is asserted nowhere in the mocked tier — the only place an AsyncVolume mount source appears is the e2e test, which won't run by default.

Add the two missing async cases (None → no volumeMounts key, and {"/mnt/data": AsyncVolume(...)} → [{"name": …, "path": …}]) so the async mirror is complete. _async_request_body already takes the mounts argument, so each is a three-line test.

claude

LGTM — test-only PR, no production code changes.

Verified the mocked payload assertions ({name, path} shape, Volume instance vs string mount source, omission when no volumes) against the actual JS (sandboxApi.ts) and Python (sandbox_api.py) implementations — they match exactly. E2E mount tests follow the established pattern (create sandbox, write via mount, kill, recreate, read) and are correctly gated behind hostedTest/@pytest.mark.e2e.

Extended reasoning...

Overview

This PR adds test-only coverage for Volume mounts: a mocked unit test verifying the shape of the volumeMounts payload sent in the sandbox create request (JS and Python), and an e2e test verifying that content written to a volume-mounted path in one sandbox is readable from a second sandbox mounting the same volume. No production/SDK source files are touched.

Security risks

None. These are additive test files; no new code paths, auth logic, or data handling are introduced in shipped code.

Level of scrutiny

Low. Test-only change with no runtime impact. I still verified the assertions against the real implementation (packages/js-sdk/src/sandbox/sandboxApi.ts and packages/python-sdk/e2b/sandbox/sandbox_api.py) since the PR description quotes a stale payload shape (volumeName/mountPath) that doesn't match either the tests or the actual code (both use name/path) — the tests themselves are correct, only the description text is outdated.

Other factors

The e2e tests are properly gated behind hostedTest (JS) and @pytest.mark.e2e + skip_debug (Python) so they won't run in the default/mocked tier, consistent with the tiering effort described across this stack of PRs. Cleanup (kill sandboxes, destroy volume) is handled in try/finally blocks in both languages.

test: add Volume mount payload unit and live mount e2e coverage

0c481b0

devin-ai-integration Bot requested a review from mishushakov as a code owner August 21, 2026 12:29

devin-ai-integration Bot assigned mishushakov Aug 21, 2026

cla-bot Bot added the cla-signed label Aug 21, 2026

This was referenced Aug 21, 2026

test: split Template suites into mocked unit and e2e tiers (4/4) #1745

Open

test: split SDK suites into a mocked unit tier and an opt-in e2e tier #1739

Closed

devin-ai-integration Bot commented Aug 21, 2026

View reviewed changes

claude Bot reviewed Aug 21, 2026

View reviewed changes

Navigation Menu

Uh oh!

test: add Volume mount payload unit and live mount e2e coverage (3/4) - #1744

test: add Volume mount payload unit and live mount e2e coverage (3/4)#1744
devin-ai-integration[bot] wants to merge 1 commit into
devin/1787315092-test-tiers-sandboxfrom
devin/1787315092-test-tiers-volume

devin-ai-integration Bot commented Aug 21, 2026

Uh oh!

devin-ai-integration Bot commented Aug 21, 2026

Uh oh!

chatgpt-codex-connector Bot commented Aug 21, 2026

Uh oh!

changeset-bot Bot commented Aug 21, 2026

Uh oh!

github-actions Bot commented Aug 21, 2026

Uh oh!

devin-ai-integration Bot left a comment

Uh oh!

devin-ai-integration Bot Aug 21, 2026

Uh oh!

devin-ai-integration Bot Aug 21, 2026

Uh oh!

devin-ai-integration Bot Aug 21, 2026

Uh oh!

claude Bot left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

	volume = Volume("vol-1", "my-volume", "volume-token")
	volume = Volume(volume_id="vol-1", name="my-volume", token="volume-token")

Uh oh!

Conversation

devin-ai-integration Bot commented Aug 21, 2026

Summary

Uh oh!

devin-ai-integration Bot commented Aug 21, 2026

🤖 Devin AI Engineer

Uh oh!

chatgpt-codex-connector Bot commented Aug 21, 2026

Uh oh!

changeset-bot Bot commented Aug 21, 2026

⚠️ No Changeset found

Uh oh!

github-actions Bot commented Aug 21, 2026

Package Artifacts

Uh oh!

devin-ai-integration Bot left a comment

Choose a reason for hiding this comment

Uh oh!

devin-ai-integration Bot Aug 21, 2026

Choose a reason for hiding this comment

Uh oh!

devin-ai-integration Bot Aug 21, 2026

Choose a reason for hiding this comment

Uh oh!

devin-ai-integration Bot Aug 21, 2026

Choose a reason for hiding this comment

Uh oh!

claude Bot left a comment

Choose a reason for hiding this comment

Overview

Security risks

Level of scrutiny

Other factors

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant