- Introduced a new `indexing` group in the environment variable schema with fields for AST grep usage, static graph enablement, and relationship types.
- Updated the CodexLens configuration to support new indexing features.
feat: Enhance DashboardToolbar with session and fullscreen controls
- Added props for session sidebar visibility and fullscreen mode to the DashboardToolbar component.
- Implemented handlers for toggling session sidebar and fullscreen mode.
- Updated the toolbar layout to include session sidebar toggle and fullscreen button.
refactor: Improve TerminalGrid and TerminalPane components
- Refactored GridGroupRenderer to handle pane size changes directly via store.
- Enhanced TerminalPane to remove unused file browser logic and improve layout handling.
- Updated key generation for child panes to ensure stability.
feat: Extend CodexLens API for staged Stage-2 expansion modes
- Added support for `staged_stage2_mode` in the CodexLens API, allowing for different expansion strategies.
- Updated semantic search handlers to process new stage-2 mode parameter.
- Implemented validation and handling for new stage-2 modes in the backend.
test: Add benchmarks for staged Stage-2 modes comparison
- Created a benchmark script to compare performance and results of different staged Stage-2 modes.
- Included metrics for latency, overlap, and diversity across modes.
Wire FlowExecutor into orchestrator routes for actual flow execution with
pause/resume/stop lifecycle management. Add CLI session audit system with
audit-routes backend and Observability tab in IssueHub frontend. Introduce
cli-session-mux for cross-workspace session routing and QueueSendToOrchestrator
UI component. Normalize frontend API response handling for { data: ... }
wrapper format and propagate projectPath through flow hooks.
In codex-lens, add per-server opened-document cache in StandaloneLspManager
to avoid redundant didOpen notifications (using didChange for updates), and
skip warmup delay for already-warmed LSP server instances in ChainSearchEngine.
- Introduced a new benchmark results file for performance comparison on 2026-02-09.
- Added a test for LspGraphBuilder to ensure it does not expand nodes at maximum depth.
- Created a test for the staged search pipeline to validate fallback behavior when stage 1 returns empty results.
- Added new benchmark result files: compare_2026-02-09_score_fast3.json and compare_2026-02-09_score_fast4.json.
- Implemented KeepAliveLspBridge to maintain a persistent LSP connection across multiple queries, improving performance.
- Created unit tests for staged clustering strategies in test_staged_stage3_fast_strategies.py, ensuring correct behavior of score and dir_rr strategies.
- Created a new JSON file with benchmark results from the run on 2026-02-09.
- Added tests for the StandaloneLspManager to verify path normalization on Windows, including handling of percent-encoded URIs and ensuring plain Windows paths remain unchanged.
- Implemented final verification tests for contentPattern to validate behavior with empty strings, dangerous patterns, and normal patterns.
- Created glob pattern matching tests to verify regex conversion and matching functionality.
- Developed infinite loop risk tests using Worker threads to isolate potential blocking operations.
- Introduced optimized contentPattern tests to validate improvements in the findMatches function.
- Added verification tests to assess the effectiveness of contentPattern optimizations.
- Conducted safety tests for contentPattern to identify edge cases and potential vulnerabilities.
- Implemented unrestricted loop tests to analyze infinite loop risks without match limits.
- Developed tests for zero-width pattern detection logic to ensure proper handling of dangerous regex patterns.
- Added category support for programming and documentation languages in Config.
- Implemented category-based filtering in HybridSearchEngine to improve search relevance based on query intent.
- Introduced functions for filtering results by category and determining file categories based on extensions.
- Updated VectorStore to include a category column in the database schema and modified chunk addition methods to support category tagging.
- Enhanced the WatcherConfig to ignore additional common directories and files.
- Created a benchmark script to compare performance between Binary Cascade, SPLADE, and Vector semantic search methods, including detailed result analysis and overlap comparison.
- Implemented `get_all_chunks` method in `VectorStore` class to fetch all semantic chunks from the database.
- Added a new benchmark script `analyze_methods.py` for analyzing hybrid search methods and storage architecture.
- Included detailed analysis of method contributions, storage conflicts, and FTS + Rerank fusion experiments.
- Updated results JSON structure to reflect new analysis outputs and method performance metrics.
