You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The cross-SDK quality harness currently contains the Python-specific implementation of its customer-health lifecycle. That separates the executable assertions from the b2sdk revision they validate and duplicates ownership between repositories.
Add a small, versioned sdkharness contract in this repository:
.sdkharness/tests.tsv
a tracked executable for health/golden-path at target simulator
The executable will receive a loopback simulator URL, fixed fake credentials, and a pre-created bucket from sdkharness. It will prove authorize, upload, downloaded-byte equality, list, delete, and post-delete absence, then emit exactly one tab-separated SDKHARNESS_RESULT record.
This does not add production credentials, a live-B2 workflow, or a new scheduled job here. The harness continues to own simulator provisioning, canonical scenario semantics, centralized execution, evidence, reporting, and notification.
Rollout
The executable will first run as non-counting shadow evidence. The existing harness implementation remains authoritative until this change is merged and the merged revision passes side by side. Promotion/removal is a separate reviewed harness change.
Acceptance criteria
Runs against loopback simulator only and rejects unsafe/non-loopback targets.
Uses this checkout's b2sdk code rather than an unrelated installed release.
Cleans up its uploaded object on failures where possible.
Emits one deterministic PASS/FAIL result record without exposing credentials.
Focused tests cover success and configuration/failure boundaries.
Required changelog fragment and repository checks pass.
Activity