Repository navigation
Honor the compute first-pair floor in release fanout checks - #183
luisleo526 wants to merge 1 commit into
Conversation
There was a problem hiding this comment.
Graphify reviewed this change.
No issues found by static checks — no coupling regressions or blocking issues in the code graph. That is not the same as safe to merge: see what was not checked.
Not checked: tests were not run; formal verification proved 0 of 2 changed function(s) (1 sampled, not proven, 1 not verified).
Formal verification. PR-changed functions: 1/2 verified (0 proven, 1 may-equivalent, 0 distinguished) · 1 not verified (1 vacuous).
Not verified on this run: check\_release (vacuous: never exercised).
Graphify review — findings
Makes the release fan-out check honor the compute first-pair floor: a selected compute role is only treated as enabled (delivery and receiver run required) when the release is at or above RELEASE_COMPUTE_FIRST_PAIR. Below the floor, or with the variable unset, compute shows up as a "disabled" row that passes, makes no receiver API calls and carries a first-pair policy message, while a malformed floor value now raises a ConfigurationError rather than being guessed at. The human table gets a new disabled state, distinct from excluded and unaccounted. Docs note that historical all-consumer probes start at v1.8.0, and new tests cover the below, at-or-above and malformed-floor cases, with test fixtures defaulting to an eligible floor.
No blocking issues surfaced.
Analysis details — impact, health, verification
Impact & health
Graphify review
Impact — 71 functions depend on the 71 functions this change touches.
Health — grade A; 4 existing hotspot(s) in the area this change touches (pre-existing, not introduced here):
find_receiver_run()— 2 callers, 6 callees (medium)check_release()— 1 callers, 9 callees (medium)main()— 1 callers, 9 callees (medium)wait_for_receiver()— 1 callers, 7 callees (medium)
Verification — 71 functions in the blast radius were not formally verified this run (proofs are advisory here).
Gate & verification
graphify gate
PASS — objectively clean (no health regressions, tests not run — proofs not run this pass (advisory)). Grounded, not self-assessed.
Advisory (not blocking):
- verification_scope: 71 function(s) in the blast radius were not formally verified this run
Test selection
Test selection
34 of 34 test file(s) selected (100%) via static blast radius.
Escalated to a full run for safety (
noncode-change,uncovered-code) — the selection is not trustworthy on its own (see below). CI should run the whole suite.
tests/test_facts_cards.py— full-run-safetytests/test_facts_descriptions.py— full-run-safetytests/test_facts_fanout.py— full-run-safetytests/test_facts_validate.py— full-run-safetytests/test_fingerprint_canonical.py— full-run-safetytests/test_parity_bundle.py— full-run-safetytests/test_parity_source_proof.py— full-run-safetytests/test_release_action_pins.py— full-run-safetytests/test_release_asset_fetch.py— full-run-safetytests/test_release_binding.py— full-run-safetytests/test_release_binding_workflows.py— full-run-safetytests/test_release_coordinator_receipts.py— full-run-safetytests/test_release_docs_runtime.py— full-run-safetytests/test_release_fanout_check.py— impact, changed-test, full-run-safetytests/test_release_guard_policy.py— full-run-safetytests/test_release_hosted_postmerge_receipts.py— full-run-safetytests/test_release_hpo_image.py— full-run-safetytests/test_release_main_push_inventory.py— full-run-safetytests/test_release_pair.py— full-run-safetytests/test_release_parity.py— full-run-safetytests/test_release_parity_image.py— full-run-safetytests/test_release_parity_runtime.py— full-run-safetytests/test_release_parity_wire.py— full-run-safetytests/test_release_receipt_collect_sweep.py— full-run-safetytests/test_release_receipt_legacy_previous.py— full-run-safetytests/test_release_receipt_shards.py— full-run-safetytests/test_release_receipts.py— full-run-safetytests/test_release_redeliver.py— full-run-safetytests/test_release_runtime_smoke.py— full-run-safetytests/test_release_sink_image.py— full-run-safetytests/test_release_sources.py— full-run-safetytests/test_release_workflows.py— full-run-safetytests/test_repo_descriptions.py— full-run-safetytools/release-dry-run/test_wfrun.py— full-run-safety
non-code file(s) changed (
docs/release-chain.md) → running the full suite for safety (a code graph can't see config/fixture/data deps)
changed code file(s) with no mapped test (
docs/release-chain.md) — a coverage gap or a missing link — running the full suite rather than only the selected tests
Selection is safe under the controlled-regression assumption; always-run tests + a periodic full run are the backstops. Advisory — it never changes the check verdict.
Formal verification
Could not verify: Could not verify check\_release.
The verifier did not have enough to check check\_release, so it is saying so rather than guessing. No false assurance is the whole point.
Guarantee: No guarantee either way, this is an honest abstention, not a pass.
Note: Reason: not verifiable: all 264 sampled inputs raised on both versions — the function never executed, so 'no divergence' would be vacuous (mostly NameError — names the real obstacle, not a sampling gap)
No difference found (not proven): No behavior difference found in render\_human (not a proof).
The verifier ran both versions of render\_human on many inputs and saw identical behavior every time. Strong evidence the change is safe, but evidence, not a proof.
Guarantee: Empirical: concolic exploration (CrossHair). A divergence on an untested input remains possible, so this is 'no counterexample found', not 'proven equivalent'.
Note: An input the sampler did not try could still differ.
|
Closing: no update for more than 24 hours. The branch is kept; the change returns inside one combined pull request after it is verified on current main. |
When compute is selected but a release precedes its configured first pair, the fanout checker currently reports a missing delivery that the planner deliberately omits. Reuse the planner's version policy and show compute as disabled for that pair; at or above the floor, delivery and receiver evidence remain required. Malformed selected-compute floors fail explicitly, and other roles retain their checks.
Validation: the updated 21-test module produced five expected failures against the unchanged checker, then passed all 21 tests with the fix. The four new policy targets cover below-floor, eligible, unset-floor and unselected compute cases through the real checker with fake HTTP clients. Required PR CI is still pending.
🤖 Generated with Claude Code