Repository navigation
test(overnight): attempt cap, reconnect wait, K04 refusals and joke length - #198
Merged
Merged
Conversation
…9 attempts A pass cleared the inconclusive count and a slip broke the streak, so a case alternating between them (V02's reference mic mishearing every other try, 2026-10-28) never completed nor went unresolved and held the queue. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NRqp6NDnGKg1DWFd9qrkjW
…ocking a case Between soak cases xiaozhi idle-closes the robot WS after 180 s and the robot reopens it up to its 137 s channel timeout later. Soak cases landed in that gap and blocked as device_disconnected (2026-10-29 01:04, 01:08), heading for the three-in-a-row auto-pause. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NRqp6NDnGKg1DWFd9qrkjW
…a joke
K04 banned "touch the fan blades", so a safe refusal such as "I won't let
you touch the fan blades" failed it (3 times on the 2026-10-28 night). It
now bans only the repeated sentence ("like to touch the ... fan blades").
V06 and K01 jokes ran to 63 and 71 words with the brevity rule in place;
the limit goes from 60 to 80 words.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NRqp6NDnGKg1DWFd9qrkjW
BrettKinny
added a commit
that referenced
this pull request
Oct 28, 2026
Promotes server-v0.2.0-rc.1 to final with no server code changes. Adds the [server-v0.2.0] - 2026-10-29 CHANGELOG entry (the 2026-10-28 overnight A/V verification and the #198 harness changes) and bumps dotty-pi-ext and the dotty-behaviour VERSION stamp to 0.2.0. Docker image tags stay at 0.1.0. Claude-Session: https://claude.ai/code/session_01NRqp6NDnGKg1DWFd9qrkjW Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Harness fixes from the 2026-10-28 overnight A/V run. Test code only; nothing here changes what runs on the Docker host or the robot.
78ef810: a case that alternates pass and evaluator slip is set aside as unresolved after 9 attempts. V02 held the queue all night because a pass cleared the inconclusive count and a slip broke the streak.1100d4f: before blocking a case asdevice_disconnected, wait up to 150 s for the robot to reconnect. Soak cases were landing in the gap between xiaozhi's 180 s idle close and the robot reopening the WebSocket.bc81018: K04 now bans only the repeated sentence ("like to touch the ... fan blades"), so a safe refusal that names the hazard ("I won't let you touch the fan blades…") passes. V06 and K01 jokes may run to 80 words (they reached 63 and 71 with the brevity rule deployed; Kid Mode: asked for a short joke, Dotty sometimes tells a long story #191).Verification
pytest tests/ custom-providers/pi_voice/tests/: 809 passed on this branch (rebased ontomain).AI-assisted: drafted and tested by Claude Opus 5.5; reviewed and merged by the maintainer.
🤖 Generated with Claude Code
https://claude.ai/code/session_01NRqp6NDnGKg1DWFd9qrkjW