Skip to content

Release v0.2.2

Choose a tag to compare

View all tags
@github-actions github-actions released this 07 Sep 23:14
v0.2.2
75f9c93

Release v0.2.2

Installation & Upgrade

To install or upgrade to this version, use npm:

npm install -g llxprt-code@v0.2.2

Major Features

  1. Local AI server support with ephemeral socket configuration - improvements by @ahundt
  2. Comprehensive OAuth authentication improvements across all providers - Logout and log in with a different account
  3. Shared context is compatible between all providers. Switch at will in the same session without clearing!
  4. Zed editor integration - connect LLxprt Code to Zed using any model or provider you like!
  5. Debug logging improvements and standardization - mainly for diagnostics (DEBUG=llxprt:* llpxrt ...)

This release should be very stable with Cerebras but running with /set streaming disabled is highly recommended. Streaming provides no real improvement on this provider, and they do weird things.

Changes

  • chore(release): v0.2.2 (75f9c93)
  • refactor: remove meaningless mock theater test (fb8099b)
  • fix: resolve remaining CLI test failures (83af42b)
  • chore: update package-lock.json for v0.2.2 (ec66715)
  • fix: resolve failing tests and update to v0.2.2 (0144404)
  • fix: standardize imports in GeminiProvider to use relative paths (5cb86a4)
  • fix: resolve TypeScript build errors in GeminiProvider (482da9b)
  • fix: restore system prompt injection and tool parameter fixes in providers (0f8fc1c)
  • fix: update ToolFormatter to handle parametersJsonSchema field (656a8e2)
  • test: add comprehensive unit tests for ToolFormatter schema conversion (803196c)
  • fix: resolve Anthropic provider tool call formatting issues when switching providers (7afddbb)
  • fix: allow all providers to work without authentication for local endpoints (1027bc9)
  • fix: allow OpenAI provider to work without authentication for local endpoints (f7f5d58)
  • fix: OAuth support detection and authentication precedence for OpenAI provider (ad07e9d)
  • fix: update tests for new tool format conversion methods (aeb4846)
  • fix: normalize tool IDs across providers for seamless switching (b9d7bf8)
  • fix: add dual parameter support to file tools for model compatibility (c3b9517)
  • fix: provider switching and tool format conversion issues (b271cf6)
  • fix: centralize settings through SettingsService and fix circular dependencies (74ddc40)
  • Reworded architecture documentation and fixed formatting (32e1122)
  • fix: prevent synthetic tool call pairs from being created during Zed integration cancellation (f4973b3)
  • feat(core,docs): add local AI server support with ephemeral socket configuration (6c10458)
  • feat(core): implement ephemeral socket configuration for local AI servers (66737d5)
  • upgraded to modern tokenizer as fallback (109fdab)
  • feat: add folder structure to startup prompts (ee24757)
  • fix: comprehensive gemini provider improvements (38187b9)
  • Zed cancellation fixed (eac1aff)
  • restore really long prompt for gemini (86eda4d)
  • feat: add shared double-escape utility for GLM-4.5 and Qwen tool handling (ef0bffa)
  • fix: allow OpenAIProvider to work without API key for local/self-hosted endpoints (96006b2)
  • fix: consolidate multiple DebugLogger instances to single instance per function (c4432d9)
  • fix: migrate prompt config loaders and core client logging to DebugLogger (49b3f47)
  • fix(zed): Fix tool call/response pairing to match TUI behavior (07e996d)
  • fix: replace excessive console.log() calls with proper DebugLogger usage (2191433)
  • Improve Zed editor integration documentation (6573260)
  • fix footer test for windows (beff71d)
  • fix: correct OpenAIProvider integration test to use IContent speaker field (67589d1)
  • fix: use OPENAI_BASE_URL from environment in integration tests (ea34812)
  • fix: ensure OpenAIProvider always has a default baseURL (cd43253)
  • fix: update integration tests to use IContent format (d4623bc)
  • fix: revert OpenAIProvider to remove incorrect Responses API integration (1d023be)
  • fix: complete test suite migration from IMessage to IContent (3227f65)
  • fix: resolve CI test failures after IContent migration (07e430b)
  • fix: resolve test failures after merge (7809055)
  • Merge branch 'unificationredeux' into main (312b39f)
  • fix: resolve final lint errors in logging test files (d0df492)
  • Fix ALL failing tests (100% pass rate achieved) (c819642)
  • Fix orphaned tool calls and Anthropic provider schema validation (411ffec)
  • fix: auto-patch orphaned tool calls with synthetic responses (23084c2)
  • fixed tool format and clear command (4c8ee81)
  • fix: fixed responses provider to work with IContent (470b588)
  • fix: resolve linting errors without using eslint-disable (226328c)
  • feat: implement IContent-only generation for all providers (9ee2521)
  • fix: resolve qwen model compatibility issues with OpenAI provider (fcc6c60)
  • fix: add JSON sanitization for qwen models and fix all TypeScript lint errors (8f0fda9)
  • fix: resolve all TypeScript lint errors and enforce strict type safety (a46d68e)
  • doc updates #212 and #214 (9790652)
  • Fix Qwen OAuth documentation (closes #212) (5340762)
  • Merge upstream gemini-cli up to v0.2.2 (commit 7598ef3) (dd8b089)
  • Fix multi-provider integration test to accept 404 errors (c3d3686)
  • Fix workflow variables to use correct names (bd0d6c4)
  • Fix e2e.yml to use KEY_VAR_NAME variable for secret indirection (0ea6008)
  • Fix: Use OPENAI_API_KEY without LLXPRT_ prefix (c638b4e)
  • Fix GitHub Actions to use repository variables instead of manual inputs (19ff0ea)
  • Fix Footer component test failures (d295b70)
  • test fixes, temporary removal of responses tests pending refactoring of responses provider, removal of mock theater tests from upstream (672a111)
  • Combined history support, fix for footer token count (tiktoken), seperation of OpenAIProvider from OpenAIResponsesProvider, refactoring and simplification of OpenAIProvider, compression fixes (9c2f183)
  • Fix history token count in the footer, fix compression (ad94b5f)
  • Initial shared history service (9aecd3b)
  • test: add comprehensive behavioral tests for GeminiChat (82ed311)
  • feat: add HistoryService with comprehensive test coverage (9c3a1cf)
  • Fix /keyfile command to properly set API keys (bdeef04)
  • refactor: reduce complexity of generateChatCompletion in OpenAIProvider (6ad2f82)
  • refactor: extract Qwen-specific logic into dedicated methods (cf2c81c)
  • refactor: simplify OpenAIProvider by removing unnecessary message variables (f5aeade)
  • refactor: extract Qwen double stringification fix to dedicated method (1c7af0a)
  • refactor: remove synthetic tool handling and malformed stream detection from OpenAIProvider (34daa3a)
  • Fixed OAuth fallback and Zed integration tools (1438358)
  • refactor: extract responses API functionality to dedicated OpenAIResponsesProvider (23ac66c)
  • fixed zed integration tools (39b89a2)
  • fix initialization issue after zed integration (a428202)
  • chore: restore OpenAIProvider with responses API support (a9638e6)
  • fix: prevent infinite retry loops with Qwen by improving stream validation (82237db)
  • fix: resolve ACP integration issue with provider profiles (5c5ddce)
  • feat: add comprehensive debug logging to Zed integration (ea24263)
  • fix: handle provider-based auth in Zed integration (2b2b5e3)
  • fix: properly initialize provider manager in Zed integration (56dc2c1)
  • feat: add configurable provider support to release workflow (95d056b)
  • test: skip flaky performance tests in CI (6bacfcb)
  • fix: remove Groq fallback - fail fast when secrets not configured (3df1672)
  • fix: use correct workflow_dispatch inputs syntax with Groq fallback (049a771)
  • feat: add configurable provider support to CI workflows (fdec23f)
  • Update README.md - no homebrew yet (0705a39)
  • chore(release): v0.2.2 (7598ef3)
  • chore: fix formatting in authCommand.ts (9dff311)
  • fix: remove failing property-based tests for edge cases (9b44f87)
  • fix: resolve remaining test failures from cherry-picks (2a1c2f0)
  • fix: resolve test failures from cherry-picked upstream changes (2fe501c)
  • fix(ci): allow release branches to run (#7060) (#7061) (4a7860d)
  • cherrypick workflow fixes into preview release branch (#7052) (7c4e92a)
  • Merge upstream gemini-cli up to v0.2.1 (7f1a11f)
  • Fix sandbox npm command (#7176) (a31857d)
  • fix: resolve conflicts and fix build errors from cherry-picks (e4bb033)
  • fix(keyboard): Implement Tab and Backspace handling for Kitty Protocol (#7006) (Cherry-pick) (#7051) (882e218)
  • Zed preview patches (#7036) (62fb08f)
  • Hotfix/retry stream 6777 (#6881) (fc7d739)
  • Merge #6677 into release branch for hot fix. (#6850) (643439c)
  • chore(release): v0.2.1 (1a278f8)
  • Revert "Hotfix/retry stream 6777 (#6881)" (32204dc)
  • Revert "[cherrypic][hotfix] Do not call nextSpeakerCheck in case of API error (#7140)" (d2a4abe)
  • Revert "Hotfix/metrics stream error (#7156)" (2f6cf27)
  • chore(release): v0.2.0 (3a8ec2d)
  • Fix sandbox npm command (#7176) (9dd36d6)
  • Hotfix/metrics stream error (#7156) (0910b35)
  • [cherrypic][hotfix] Do not call nextSpeakerCheck in case of API error (#7140) (dc21a06)
  • fix(ci): allow release branches to run (#7060) (#7061) (8e4b9da)
  • fix(keyboard): Implement Tab and Backspace handling for Kitty Protocol (#7006) (Cherry-pick) (#7051) (4811eef)
  • cherrypick workflow fixes into preview release branch (#7052) (d354dc1)
  • Zed preview patches (#7036) (daea8f3)
  • fix: skip OAuth tests in CI environment (6cba900)
  • Remove legacy migration code and fix OAuth test suite (a5a4554)
  • Remove obsolete snapshot file (1a5b3f2)
  • Fix OAuth test skipping and test failures in CLI package (1f1e0b7)
  • Fix remaining OAuth test failures in CI (38eb8c3)
  • Fix formatting in oauth-integration.spec.ts (e8391a5)
  • Skip OAuth tests in CI environment (668bdf9)
  • Fix shellcheck warnings in verify-oauth-integration.sh (31d4a20)
  • Fix authentication storage and keyfile clearing issues (72ed4af)
  • Fix OAuth test failures after merge (7b705f3)
  • Fix GeminiProvider clearAuthCache to properly clear cached tokens (a64d2ea)
  • Skip OAuth browser tests in CI environment (2cf5fde)
  • Fix OAuth authentication issues - remove forced fallback and error suppression (1f92242)
  • fix: clear all Google OAuth files on Gemini logout (ddb0d19)
  • fix: properly clear Gemini OAuth credentials on logout (6e0fa89)
  • fix: correct logout functionality to properly disable OAuth (ce5af0d)
  • fix: restore USE_LOGIN_WITH_GOOGLE handling in GeminiProvider (9e81b09)
  • fix: handle USE_EXISTING_GEMINI_OAUTH signal properly in OAuth manager (acbccf3)
  • fix: revert Gemini OAuth to use existing LOGIN_WITH_GOOGLE flow (f2cd2d8)
  • fix: pass TokenStore to OAuth providers to enable token persistence (8047811)
  • feat: implement OAuth token persistence and logout functionality (96a42da)
  • Add deep analysis of OAuth authentication for all providers (b217021)
  • Add authentication fixes overview and project plan (5c7d0ba)
  • fix: properly handle stream_options ephemeral setting for Mistral compatibility (52dcfe3)
  • fix: resolve OAuth dialog and synthetic tool response issues (6c6e900)
  • Hotfix/retry stream 6777 (#6881) (86fd641)
  • chore(release): v0.2.0-preview.2 (#6868) (dda33bb)
  • Merge #6677 into release branch for hot fix. (#6850) (07b7df5)
  • feat: make read_file tool accept both file_path and absolute_path parameters (0a13867)
  • feat(openai): forward modelParams to Responses API (incl. retry); enable gpt-5 reasoning/text
  • fix: resolve Qwen provider stream hanging and parameter stringification issues (9a6cd63)
  • Add Cerebras Code Max/Pro plan configuration instructions to README (4e884e4)
  • chore: update package-lock.json for version 0.1.23 (a6827b7)
  • fix: use valid GitHub Action SHAs with ratchet comments (4a5ee62)
  • fix: pin GitHub Actions to specific SHAs for security (5f813bf)
  • chore: bump all package versions to 0.1.23 (44e3e90)
  • fix: update GitHub Actions to use version tags instead of commit SHAs (9ceb866)
  • chore(release): v0.2.0-preview.0 (b06e8db)

Full Changelog: v0.1.23...v0.2.2

Contributors

ahundt
Assets 3
Loading