Conversation
|
We require contributors to sign our Contributor License Agreement, and we don't have @pei711 on file. You can sign our CLA at https://e2b.dev/docs/cla . Once you've signed, post a comment here that says '@cla-bot check' |
🦋 Changeset detectedLatest commit: 1549d0e The changes in this PR will be included in the next version bump. This PR includes changesets to release 2 packages
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
There was a problem hiding this comment.
TASTE.md review: no violations. The diff touches only Result serialization; checked parity (T-1), empty-object-as-present (T-13), undefined for absence (T-20), and docstrings (T-69). With this change, JS and Python both report data in formats() and serialize it, including {}, and both leave it out when it is missing or null.
|
@cla-bot check |
|
The cla-bot has been summoned, and re-checked this pull request! |
Fixes #1938.
Problem and change
Serializing Code Interpreter executions drops
Result.datain JavaScript and drops empty data objects in Python. Preserve that field when supplied, including{}, while keeping absent data out of the serialized result. Python sync and async code paths share the corrected result model.The production fix is one line in each SDK. Added JS and sync/async Python parser-to-serialization regressions and a patch changeset for both Code Interpreter packages.
Usage
Validation
Sandbox.connect/runCode, then JSON serialization preserves empty and populated data. This allocates no cloud sandbox and performs no real code execution.ty check e2b_code_interpreterpass.git diff --checkpasses.Tested on Windows with Node 24.13.1, Python 3.13.12 and pinned pnpm 10.34.5. The root
pnpm run format,lintandtypecheckwrappers were attempted; each stops when a Python package invokesmake, which is unavailable here. The affected Python package's equivalent Ruff and ty commands passed directly through uv. Cloud sandbox tests and other runtimes were not run; these checks make no sandbox or provider calls.This PR was fully generated with OpenAI Codex, including code, tests and description. Codex reviewed the change and ran the checks above; no human review is claimed.