The truncation logic and error handling added in e1d3c2b7 (Dec 4, 2025)
was lost when message handling was refactored from slack.ts to main.ts.
Changes:
- Add try/catch to all Slack message functions (respond, replaceMessage,
respondInThread, setTyping, setWorking)
- Truncate main messages at 35K chars (Slack limit 40K)
- Truncate thread messages at 20K chars
- Add helpful user-facing notes when truncation occurs
- Log Slack API errors as warnings instead of crashing
Fixes: msg_too_long API errors causing process crash
Co-authored-by: mom bot <mom-bot@living-bio.iam.gserviceaccount.com>
Koffi is only used on Windows for VT input support and fails to build
on Termux/Android and Linux systems without build tools. Moving it to
optionalDependencies allows installation to succeed on all platforms
while maintaining Windows functionality.
strip-ansi is imported in bash-executor.ts and two interactive mode
components but is not declared in package.json dependencies. This
causes ERR_MODULE_NOT_FOUND at runtime in strict package managers
like pnpm that do not hoist undeclared dependencies.
Co-authored-by: Graadient <graadient@users.noreply.github.com>
createBranchedSession() wrote the file and set flushed=true even when the
branched path had no assistant message. The next _persist() call saw no
assistant, reset flushed=false, and the subsequent flush appended all
in-memory entries to the already-populated file, duplicating the header
and entries.
Fix: defer file creation when the branched path has no assistant message,
matching the newSession() contract. _persist() creates the file on the
first assistant response.
closes#1672
- Map redacted_thinking to ThinkingContent with redacted: true instead of
adding a new content type. The opaque payload goes in thinkingSignature,
thinking text is set to "[Reasoning redacted]" so it renders naturally
everywhere. Cross-model transform drops redacted blocks.
- Skip interleaved-thinking-2025-05-14 beta header for Opus 4.6 / Sonnet 4.6
where adaptive thinking makes it deprecated/redundant.
- Do not send temperature when thinkingEnabled is true (incompatible with
both adaptive and budget-based thinking).
Based on #1665 by @tctev
Z.ai uses the same enable_thinking: boolean parameter as Qwen to control reasoning, not thinking: { type: "enabled" | "disabled" }.
The wrong parameter name means Z.ai ignores the disable request and always runs with thinking enabled, wasting tokens and adding latency.
Merge the Z.ai and Qwen branches since they use the same format.
PR by @okuyam2y
