Engine events mutate authority before detecting exhausted clocks or event IDs #629
Labels
No labels
bug
cli
core
docs
event
experiment
figure
invariant
metrics
oracle
phase-0
phase-1
phase-2
phase-3
phase-4
phase-5
phase-6
provenance
revocation
tests
workload
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
erikinkinen/AES#629
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
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?
Problem
Engine::stepandEngine::step_measuredapply the graph operation beforeallocating its event identity and incrementing the logical clock. At
time == UINT64_MAX, an accepted event wraps time to zero and publishes itsauthority change. An exhausted event-ID cursor also reaches the generator's
fatal overflow check after the operation instead of returning a hard error.
Reproduced on
5ef0cd37b1902300f6bbb0c2c24d524cd6a7fa0fwith a valid subject/object and a fresh proposedcapability, then setting the event clock to its maximum:
The probe exits 1 because the expected overflow error and unchanged authority
are absent. This violates the research overflow/no-mutation contract and can
produce a noncontiguous accepted history.
Required behavior
Preflight both the logical clock and the next event identity before applying
an event through either engine entry point. Return
Errc::Overflowwithoutchanging graph, history, clock, identity cursors or the measured work meter.
Keep the ID generator's existing direct-call overflow behavior and valid
legacy executions unchanged.
Acceptance