Skip to content

fix(harbor): handle absent pi streams and record audited rerun - #56

Merged
minixalpha merged 2 commits into
mainfrom
bench/pi-dev3-rerun-20261102
Nov 3, 2026
Merged

minixalpha merged 2 commits into
mainfrom
bench/pi-dev3-rerun-20261102

Conversation

@minixalpha

@minixalpha minixalpha commented Nov 2, 2026 •

Copy link
Copy Markdown
Owner

Handle absent stdout/stderr in Pi model-registration checks so valid installations no longer fail on None. Preserve rejection of diagnostics and missing models, and add three regression cases.

Record three fresh Pi reruns with the original agent, model configuration, task revisions, and time limits. Sampler scored 0/9 after confirmed timeout cancellation; WAL passed 7/7 and pipeline passed 4/4. Both passes read reference tests and solutions matching the registered task hashes and are explicitly marked as non-independent. Preserve the original 34-task results.

Include synchronized English/Chinese reports and benchmark READMEs, public configurations, and a 326-file freeze manifest. Raw logs and credential values are excluded.

Validation: all 304 Harbor tests and all three CI jobs passed. Report hashes and links validate; the recorded audit confirms all 1,449 original frozen files are unchanged.

@minixalpha
minixalpha merged commit 1085a74 into main Nov 3, 2026
3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant