aliyun-codex-bridge 0.1.0

package/CHANGELOG.md

@@ -0,0 +1,118 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+
+### Fixed
+- Stream tool-call bridging now extracts calls from `delta.tool_calls`, `choice.message.tool_calls`, and `choice.tool_calls` to avoid empty output-item rounds in Codex
+- Added `status` fields (`in_progress`/`completed`) for streamed and non-streamed `function_call` output items for better Responses API compatibility
+
+## [0.1.0] - 2026-01-20
+
+### Added
+- Separate reasoning output items with content-part events (aligns with Responses API)
+- Usage propagation from upstream streaming chunks when available
+
+### Changed
+- Output indices now follow creation order across message/reasoning/tool items
+- response.completed now preserves tool-call items in non-streaming mode for parity with streaming
+- response objects now inherit instructions/metadata/tool_choice/temperature/top_p from requests
+
+### Fixed
+- Avoid empty message items when tool-only rounds are suppressed
+
+## [0.1.0] - 2026-01-19
+
+### Added
+- Raw upstream chunk logging (`LOG_STREAM_RAW`) for streaming diagnostics
+- Output suppression toggles for tool-call rounds (`SUPPRESS_ASSISTANT_TEXT_WHEN_TOOLS`, `DEFER_OUTPUT_TEXT_UNTIL_DONE`)
+- Auto-update on startup (checks npm latest and installs if newer; can be disabled via config)
+
+### Fixed
+- De-duplicate streaming deltas when providers emit full-content chunks
+- Only suppress assistant output_text when finish_reason is tool_calls (avoid hiding short final replies)
+- Route suppressed tool-call round text into reasoning stream so it remains visible
+
+## [0.1.0] - 2026-01-16
+
+### Fixed
+- Do not include function_call items in response.completed output (restore local tool execution)
+
+### Changed
+- Forward reasoning controls to upstream and accept reasoning text from alternative fields
+
+### Docs
+- Document reasoning passthrough and reasoning_text events
+- Clarify that response.completed output excludes function_call items for local tool execution
+
+### Docs
+- Document reasoning passthrough and reasoning_text events
+- Clarify that response.completed output excludes function_call items for local tool execution
+
+## [0.1.0] - 2026-01-16
+
+### Fixed
+- Avoid empty output_text items in tool-only streaming responses
+- Only emit output_text.done/content_part.done when output text exists
+
+## [0.1.0] - 2026-01-16
+
+### Fixed
+- Handle streaming tool_calls without `index` by assigning a stable fallback index
+- Improve tool name logging when tools define top-level `name`
+
+## [0.1.0] - 2026-01-16
+
+### Fixed
+- Normalize `bin` path and repository URL for npm publish compatibility
+
+## [0.1.0] - 2026-01-16
+
+### Added
+- Auto-enable tool bridging when tool-related fields are present in the request
+- Extra logging to surface `allowTools` and `toolsPresent` per request
+- Debug tool summary logging (types and sample names)
+
+### Fixed
+- Correct output_index mapping for streaming tool call events
+- Filter non-function tools to avoid upstream schema errors
+
+### Changed
+- README guidance for MCP/tools troubleshooting and proxy startup
+
+## [0.1.0] - 2026-01-16
+
+### Changed
+- Replaced the README with expanded setup, usage, and troubleshooting guidance
+- Clarified Codex provider configuration and proxy endpoint usage
+
+## [0.1.0] - 2026-01-16
+
+### Added
+- Tool calling support (MCP/function calls) when `ALLOW_TOOLS=1`
+- Bridging for `function_call_output` items to Chat `role: tool` messages
+- Streaming support for `delta.tool_calls` with proper Responses API events
+- Non-streaming support for `msg.tool_calls` in final response
+- Tool call events: `response.output_item.added` (function_call), `response.function_call_arguments.delta`, `response.function_call_arguments.done`
+- Automated tool call test in test suite
+
+### Changed
+- `translateResponsesToChat()` now handles `type: function_call_output` items
+- `streamChatToResponses()` now detects and emits tool call events
+- `translateChatToResponses()` now includes `function_call` items in output array
+
+### Fixed
+- Tool responses (from MCP/function calls) are now correctly forwarded to upstream as `role: tool` messages
+- Function call items are now properly included in `response.completed` output array
+
+## [0.1.0] - Previous
+
+### Added
+- Initial release with Responses API to Chat Completions translation
+- Streaming support with SSE
+- Health check endpoint
+- Zero-dependency implementation

package/CODEX_REPORT_v0.1.0.md

@@ -0,0 +1,39 @@
+=====================================
+CODEX CLI TEST SUITE - SANITIZED REPORT
+=====================================
+
+Bridge Version: 0.1.0
+Test Date: 2026-01-20
+
+SUMMARY:
+--------
+Total Tests: 48
+Passed: 44 (91.7%)
+Failed: 2 (4.2%)
+Skipped: 2 (4.1%)
+
+CATEGORY BREAKDOWN:
+-------------------
+1. System Information: 3/3 passed
+2. File Operations: 8/8 passed
+3. Search & Discovery: 3/3 passed
+4. Shell Execution: 6/6 passed
+5. Text Processing: 6/6 passed
+6. Web & Network: 2/2 passed
+7. Git Operations: 2/3 passed (1 partial)
+8. AI Capabilities: 1/3 passed (2 skipped)
+9. Error Handling: 3/3 passed
+10. Platform-Specific: 1/4 passed (3 skipped)
+11. Cleanup: 0/2 failed
+12. Package & Binary Verification: 8/8 passed
+
+NOTES:
+------
+- This report is intentionally sanitized for public sharing.
+- Local usernames, hostnames, absolute paths, hardware details, and machine-specific IDs were removed.
+- Failures were environment-policy related (sandbox restrictions), not bridge logic regressions.
+
+REFERENCE:
+----------
+- Full reproducible test matrix: `CODEX_TEST_SUITE.md`
+- For private/internal troubleshooting, capture raw logs locally and do not commit them.