• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

lsm / neokai / 26801008748
86%

Build:
DEFAULT BRANCH: dev
Ran 02 Jun 2026 05:46AM UTC
Jobs 0
Files 0
Run time –
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

Canceled at 02 Jun 2026 03:35PM UTC via web
26801008748

push

github

web-flow
fix(daemon): forward reasoning_content from OpenAI Chat Completions bridge (#2050)

* fix(daemon): forward reasoning_content from OpenAI Chat Completions bridge

The bridge silently dropped reasoning_content from SSE deltas, causing
custom endpoints (DeepSeek R1, o-series via Chat Completions, QwQ, etc.)
to show only content tokens while hiding the model's reasoning chain.

- Add reasoning_content to OpenAIChatStreamChoice.delta type
- Track thinking block state (thinkingOpen, thinkingBlockIndex) during
  streaming, mirroring the Responses bridge pattern
- Emit Anthropic thinking SSE events via existing translator helpers:
  contentBlockStartThinkingSSE, thinkingDeltaSSE, contentBlockStopSSE
- Close thinking block before opening text or tool_use blocks
- Replace heuristicOutputText with heuristicOutputTokens so both reasoning
  and content deltas count toward the fallback usage estimate
- Add unit tests verifying reasoning_content is translated and counted

* fix(daemon): close thinking block before tool_use and accumulate heuristic strings

- Add closeThinkingBlock() inside openToolCall so reasoning→tool_calls
  transitions stop the thinking block before the tool_use block starts.
- Accumulate heuristic reasoning/text strings and estimate tokens once at
  stream end instead of per-fragment, preventing inflated counts when
  backends stream single-character deltas.
- Add unit tests for reasoning→tool_calls and reasoning-only streams.

* fix(daemon): propagate thinking config to OpenAI Chat bridge via side-channel

Mirrors the pattern introduced in PR #2049 for the Responses bridge.

- Add setSessionThinkingConfig to OpenAIChatBridgeServer. The bridge stores
  per-session thinking configs and merges them into request bodies when the
  SDK omits the thinking field.
- Encode sessionId in ANTHROPIC_AUTH_TOKEN (custom-endpoint:${sessionId})
  so the bridge can identify which session config to apply.
- CustomEndpointProvider implements Provider.setSessionTh... (continued)
Source Files on build 26801008748
Detailed source file information is not available for this build.
  • Back to Repo
  • Github Actions Build #26801008748
  • b453c537 on github
  • Prev Build on dev (#26796026000)
  • Next Build on dev (#26801282614)
  • Delete
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc