Bind worker credentials to paired origin (#224) #236
No reviewers
Labels
No labels
accessibility
bug
documentation
duplicate
enhancement
good first issue
help wanted
invalid
P1
P2
P3
question
wontfix
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
Creatidy/scarcity-router!236
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "issue-224-origin-bound-worker"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Scope
Refs #224. Refuse worker runtime connections to an origin different from the stored paired origin before sending the worker credential. Host, port and scheme changes fail closed; paired-origin operation and origin-bound private CA trust remain intact. This is containment only, not verified origin migration. Tier B, certificate lifecycle and migration remain open.
CI remediation: isolate package-lifecycle unit tests from checkout
dist/and ambient baseline overrides using a temporary synthetic wheel. Production package behavior remains unchanged.Validation
uv sync --only-dev; fixed-seedPYTHONHASHSEED=0 make check: 3,181 passed, 18 skipped, basedpyright 0/0; standaloneuv run basedpyright,make package-check, and whitespace checks passed.Review
Review #7 approved the preceding HEAD but is invalidated by the fixture change. Fresh complete security-critical GPT-6.1 Sol High review and current-HEAD CI are pending; all historical #224 review ordinals remain preserved.
Reserving imminent independent whole-PR security/correctness/UX review #7/10 for issue #224 (six previous invocations retained under PR #233). Frozen HEAD
ce7be1b1241c24f4a785d6c7e8a928231a091b8e, base and merge-basef141a6998fb0b9588d4908bc22c286586d1efc1a. Requested nativepr-reviewer, GPT-6.1 Sol/openai/high as explicitly owner-directed. Scope is origin-mismatch containment only, not origin migration. Required CI pending; no approval is implied by this reservation.Independent whole-PR review #7/10: APPROVE, findings
[], limitations[], exact HEADce7be1b1241c24f4a785d6c7e8a928231a091b8e, base/merge-basef141a6998fb0b9588d4908bc22c286586d1efc1a. Fresh nativepr-reviewersessionses_e74174032ffeunZ7bEpmZkBCh4, dispatched GPT-6.1 Sol/openai/high; backend runtime identity independently unobserved. Reviewer inspected complete frozen diff and worker connection, credential, trust, TLS, service and UX context, verified clean frozen refs and diff whitespace; did not run test suites. UX: different-origin override fails before connection with actionable diagnostic; same-origin services unchanged. CI remains required, and origin migration/tier B/certificate lifecycle are not claimed.CI run #3376 failed at repository tests: 5 failures and 1 error in
test_package_lifecycle, all because the fake lifecycle unexpectedly reads a baseline wheel from checkoutdist/, absent on the clean runner. Earlier local validation had a pre-existing built wheel, so did not reproduce a clean checkout. Forgejo v15 Actions jobs/artifacts MCP endpoints returned 404; the public run UI exposed the failure log. Scoped test-fixture remediation is underway, supplying a synthetic temporary wheel rather than depending on checkout artifacts. This changes HEAD and invalidates review #7; a fresh complete review will be required under the same #224 counter.Final
make checkafter package-fixture remediation ran 3,181 tests and failed once intest_openai_codex_acquisition.MalformedAndBoundedOutput.test_oversized_line_with_secret_never_leaks: captured stderr contained an exception from_FakeAppServer._feedafter a closed pipe. The assertion truncates the underlying traceback, so the exact errno remains unverified. This resembles the historical excluded #166 feeder/close race, but is not proven identical and is not attributed to this PR. The immediately preceding full standalone suite passed 3,181 tests. A changed-condition fixed-hash-seed gate will be attempted without suppressing or skipping the test; this failure remains recorded.Reserving imminent fresh independent whole-PR review #8/10 for issue #224. Frozen HEAD
4b903874da53dc6ba06740b70d5aedb55054f252, base/merge-basef141a6998fb0b9588d4908bc22c286586d1efc1a. Review #7 APPROVE was invalidated by the CI-fixture commit; prior ordinals remain consumed. Scope includes credential-origin containment and the necessary clean-runner fixture repair, not verified migration. Requested nativepr-reviewer, GPT-6.1 Sol/openai/high. Fixed-seed full gate passed; current-head CI #3379 running.Independent whole-PR review #8/10: APPROVE, findings
[], limitations[], exact HEAD4b903874da53dc6ba06740b70d5aedb55054f252, base/merge-basef141a6998fb0b9588d4908bc22c286586d1efc1a. Fresh nativepr-reviewersessionses_e73e588c7ffeMVlkB9BVh3wQzy, dispatched GPT-6.1 Sol/openai/high; backend runtime identity independently unobserved. Complete frozen diff, credential-origin/TLS/CA/service paths, test fixture isolation and UX assessed; reviewer did not run suites. UX: mismatch gives actionable credential-free failure and paired-origin behavior is retained. Current-head CI #3379 still running; no migration, tier B or lifecycle completion claimed.