Repository navigation
test(subscriber): reconcile multi-tenant orphan batches exactly once - #56
Draft
echoomegaprime wants to merge 1 commit into
Conversation
This was referenced Nov 12, 2026
Open
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
The combined CertForge queue repair previously covered single-run orphan recovery, sequential idempotency, compensation rollback, competing worker attempts and immutable audit redaction. A multi-tenant batch reconciliation acceptance test was still missing.
Change
Adds only
tests/test_dispatch_orphan_batch_recovery.py: five synthetic tenants, five different failed legacy dispatches, one global sweep, and a second replay sweep. The test verifies for each tenant:The test deliberately stages failed-dispatch rows directly in a disposable fixture database rather than calling
claim_dispatchduring setup (which would reconcile earlier rows prematurely). Separate synthetic tenant principals avoid normal per-tenant submission rate limits. No production data or protected credentials are touched.Isolated execution
On FORGE, archived exact base commit
24277990381d376ab427e42851c62a03a3110a34into a disposable temp directory, added this precise 67-line test and ran:python3 -m pytest -q tests/test_dispatch_orphan_batch_recovery.py tests/test_dispatch_orphan_audit_redaction.pyResult: 4 passed in 4.99 seconds, exit 0. Initial scratch version failed during fixture setup at HTTP 429 due to a synthetic single-tenant submission rate limit; redesigned fixture with five tenant principals and independently reran green. No product issue inferred from fixture rate limiting.
Readback / release boundaries
9056dcae36a110e23e6b9721cc3524a65312620dverified. Compared to PR fix(subscriber): redact untrusted orphan-dispatch errors in immutable audit #55 base: 1 commit ahead, 0 behind, one new test file (+67 lines), no source code changed.