-
Notifications
You must be signed in to change notification settings - Fork 2.8k
fix: gemini maxOutputTokens and reasoning config #9375
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
Re-review complete. No new issues were identified in the latest commit; Gemini reasoning typings and tests remain valid.
Mention @roomote in a comment to request specific changes to this pull request or fix all unresolved issues. |
daniel-lxs
approved these changes
Nov 19, 2025
mrubens
approved these changes
Nov 19, 2025
mini2s
added a commit
to zgsm-ai/costrict
that referenced
this pull request
Nov 19, 2025
* fix: prevent MCP server restart when toggling tool permissions (RooCodeInc#8633) * fix: prevent MCP server restart when toggling tool permissions Add isProgrammaticUpdate flag to distinguish between programmatic config updates and user-initiated file changes. Skip file watcher processing during programmatic updates to prevent unnecessary server restarts. * fix(mcp): prevent server reconnection when toggling disabled state Fixed bug where MCP servers would reconnect instead of staying disabled when toggled off. The issue was that toggleServerDisabled() used stale in-memory config instead of reading the fresh config from disk after writing the disabled flag. Changes: Added readServerConfigFromFile() helper to read and validate server config from disk Updated disable path to read fresh config before calling connectToServer() Updated enable path to read fresh config before calling connectToServer() This ensures the disabled: true flag is properly read, causing connectToServer() to create a disabled placeholder connection instead of actually connecting the server. + refactor(mcp): use safeWriteJson for atomic config writes Replace JSON.stringify + fs.writeFile with safeWriteJson in McpHub.ts to prevent data corruption through atomic writes with file locking. * fix(mcp): prevent race condition in isProgrammaticUpdate flag Replace multiple independent reset timers with a single timer that gets cleared and rescheduled on each programmatic config update. This prevents the flag from being reset prematurely when multiple rapid updates occur, which could cause unwanted server restarts during the file watcher's debounce period. + fix(mcp): ensure isProgrammaticUpdate flag cleanup with try-finally Wrap safeWriteJson() calls in try-finally blocks to guarantee the isProgrammaticUpdate flag is always reset, even if the write operation fails. This prevents the flag from being stuck at true indefinitely, which would cause subsequent user-initiated config changes to be silently ignored. * docs(readme): update readme images and image compression * docs: replace inline base64 images with image file references * Merge remote-tracking branch 'upstream/main' into roo-to-main * feat(terminal): refactor execa command execution for shell handling * Feat: Add Minimax Provider (fixes RooCodeInc#8818) (RooCodeInc#8820) Co-authored-by: xiaose <[email protected]> Co-authored-by: Matt Rubens <[email protected]> * Release v3.29.4 (RooCodeInc#8906) * fix: Gate auth-driven Roo model refresh to active provider only (RooCodeInc#8915) * feat: add zai-glm-4.6 model to Cerebras and set gpt-oss-120b as default (RooCodeInc#8920) * feat: add zai-glm-4.6 model and update gpt-oss-120b for Cerebras - Add zai-glm-4.6 with 128K context window and 40K max tokens - Set zai-glm-4.6 as default Cerebras model - Update gpt-oss-120b to 128K context and 40K max tokens * feat: add zai-glm-4.6 model to Cerebras provider - Add zai-glm-4.6 with 128K context window and 40K max tokens - Set zai-glm-4.6 as default Cerebras model - Model provides ~2000 tokens/s for general-purpose tasks * add [SOON TO BE DEPRECATED] warning for Q3C * chore: set gpt-oss-120b as default Cerebras model * Fix cerebras test: update expected default model to gpt-oss-120b * Apply suggestion from @mrubens Co-authored-by: Matt Rubens <[email protected]> --------- Co-authored-by: kevint-cerebras <[email protected]> Co-authored-by: Matt Rubens <[email protected]> * roo provider: update session token on every request (RooCodeInc#8923) * roo provider: update session token on every request * Cleanup: remove unused imports * Also refresh token before completePrompt() * feat: add We're hiring link to announcement modal (RooCodeInc#8931) Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com> Co-authored-by: Roo Code <[email protected]> Co-authored-by: Matt Rubens <[email protected]> * Fix: Enhanced codebase index recovery and reuse ('Start Indexing' button now reuses existing Qdrant index) (RooCodeInc#8588) Co-authored-by: daniel-lxs <[email protected]> * fix: make code index initialization non-blocking at activation (RooCodeInc#8933) * fix(context): truncate type definition to match max read line (RooCodeInc#8509) Co-authored-by: daniel-lxs <[email protected]> * fix: prevent infinite loop when canceling during auto-retry (RooCodeInc#8902) * fix: prevent infinite loop when canceling during auto-retry - Add abort check after backoffAndAnnounce in first-chunk retry logic - Add abort check after backoffAndAnnounce in mid-stream retry logic - Properly handle task abortion to break retry loops Fixes RooCodeInc#8901 * docs: add critical comments explaining abort checks - Document the importance of abort checks after backoff - Explain how these checks prevent infinite loops - Add context for future maintainability --------- Co-authored-by: Roo Code <[email protected]> * feat: rename MCP Errors tab to Logs for mixed-level messages (RooCodeInc#8894) - Update McpView.tsx to use "logs" tab ID instead of "errors" - Rename translation key from tabs.errors to tabs.logs in all locales - Change empty state message from "No errors found" to "No logs yet" This better reflects that the tab shows all server messages (info, warnings, errors), not just errors. Fixes RooCodeInc#8893 Co-authored-by: Roo Code <[email protected]> * feat: improve @ file search for large projects (RooCodeInc#8805) * feat: improve @ file search for large projects - Increase default file limit from 5,000 to 10,000 (configurable up to 500,000) - Respect VSCode search settings (useIgnoreFiles, useGlobalIgnoreFiles, useParentIgnoreFiles) - Add 'maximumIndexedFilesForFileSearch' configuration setting - Add tests for new functionality Conservative default of 10k keeps memory usage low while still providing 2x improvement. Users with large projects can opt-in to higher limits (up to 500k). This is a simplified alternative to PR RooCodeInc#5723 that solves the same problem without the complexity of caching. Ripgrep is already fast enough for 10k+ files, and the benefit of caching doesn't justify 2,200+ lines of additional code and maintenance burden. Fixes RooCodeInc#5721 * fix: add missing translations for maximumIndexedFilesForFileSearch setting * test: improve file-search tests to verify configuration behavior * fix: remove search_and_replace tool from codebase (RooCodeInc#8892) Co-authored-by: Roo Code <[email protected]> Co-authored-by: daniel-lxs <[email protected]> Co-authored-by: Hannes Rudolph <[email protected]> * Release v3.29.5 (RooCodeInc#8942) * Changeset version bump (RooCodeInc#8907) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Matt Rubens <[email protected]> * Fix cost and token tracking b
Fixes reasoning not being set to high or low properly.