Conversation
…e error handling in chat history
- Add chat.runError.title strings for en, ja, ru, and zh (banner uses t('runError.title')).
- Assert the English run-error title in the chat execution graph E2E flow.
- Add a unit test so loadHistory clears stale runError when the latest turn ends in a successful assistant reply.
Made-with: Cursor
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c9dcbf4ab4
ℹ️ About Codex in GitHub
Codex has been enabled to automatically 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 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| "runError": { | ||
| "title": "Model API call error" | ||
| }, |
There was a problem hiding this comment.
Remove duplicate runError key that overrides new title
The newly added runError object is shadowed by an existing runError key later in the same JSON file, so the new title is never used at runtime. In JSON, duplicate keys keep only the last value, which means t('runError.title') still resolves to the older string ("Model call failed") and the new E2E assertion for "Model API call error" will fail. The same duplicate-key pattern is present in src/i18n/locales/ja/chat.json, ru/chat.json, and zh/chat.json as well.
Useful? React with 👍 / 👎.
Add data-testid=chat-run-error to the run-error banner and target it in the execution-graph E2E so CI passes when the UI locale is not English. Made-with: Cursor
Summary
Related Issue(s)
Type of Change
Validation
Checklist