Conversation
…007,008,009 mainstream adapter + deterministic tests Coordinator-approval: required workspace wiring outside allowed dirs for integration/build validation (package.json, tsconfig.json, packages/bench/tsconfig.json).
… gaps (IKINV-005 IKINV-006 IKINV-007 IKINV-008)
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: bac99270bb
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| const inkOptions: InkRenderOptions = {}; | ||
|
|
||
| if (options.stdout) { | ||
| inkOptions.stdout = options.stdout as NodeJS.WriteStream; | ||
| } |
There was a problem hiding this comment.
toInkOptions rebuilds the options object from a small whitelist, so non-Ink fields are silently dropped before calling inkRender; this breaks the benchmark integration that passes { internal_backend: backend } and then waits for backend.waitForFrame() (for example packages/bench/src/scenarios/construction.ts lines 85-88, with the same pattern in other scenarios). With internal_backend stripped here, no frame is ever emitted to that backend and the benchmark can block on its first await instead of producing results.
Useful? React with 👍 / 👎.
1 participant
Work on Ink compatibility layer