Phase 2: generate lineage-swap paired intent tapes (#103) #652
No reviewers
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!652
Loading…
Reference in a new issue
No description provided.
Delete branch "phase2/p2-11-a-lineage-swap"
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?
Task
Closes #103 (P2-11-A).
Summary
Generate complete sealed lineage-swap tapes with prospective fixed roles,
recorded assignment/rights draws and the identical128-epoch continuation.
Actual executed cuts have matching authority/access and the specified origins.
Scope
Sizes8/32/128, all seven unchanged mechanisms,84 smoke cells and14all/zero-target
boundary cells. Independent Python checks every record and role mapping for
30tapes. A_clone and saved minimum witnesses remain #105; deleted bridges are #104.
Design intent
Generation has only seed and size inputs. Both tapes and roles exist before
execution; logical handles remain distinct from physical IDs. Full execution
checks every actual suffix use, source/event binding, Read-only policy target,
clock and first difference. lineage_basic retains its ancestor-walk semantics.
Phase discipline
Protocol2.2/backlog2.2.1; #100/#101 and the separate CI fix #653 are closed.
Development fixtures establish correctness; production measurement, formal claims
and manuscript conclusions retain their separate gates. Complete csaI includes
the requested Phase5 recovery supplement.
Verification
a353fbc32ccef7e9355f74f7fb7d596beba3cb4epassed all192 tests in GCC Debug,GCC Release and Clang unoptimized Debug ASan/UBSan:98pairs and25,088actual uses
per build. The independent reference checks10,704records and2,784variates.
eded66b72c7f95fb54e40bd9d9f4df2f2203acbetimed out at120s on four largecells. The separate fix #654 passed all eight checks and merged at
862f8a4caf198a931c32cd1f7e03097a9eab7154.42678fc531c634933473fbb55603524d8e2ad390changes only workflow/docs from the failed head;every compiled source and test byte is unchanged. An additional full192-test
Debug
-g -O1ASan/UBSan run with one worker passed in538.57s on the samecompiled source tree, with no skips and unchanged reference fingerprints.
and CI integration evidence preserve exact provenance. The latter's
SHA-256 is
0e969d89e15c00b657585dcd4ee478235fb9330516eeceed94b0785780e680a5. Trackedci-integration.jsonrecords the supersededworkflow/doc hashes explicitly; original implementation evidence remains intact.
remote checks must pass on the final integrated head before the one-step issue
checklist update and merge. Both CI suites run one worker; sanitizers retain
assertions, instrumentation and120s per-test limits. The current overall budgets are90min under candidate #656.
Expanded-suite budget regression
The previous integrated head
f9b114fa0d5d3d9ce1530d7941af61083bd9c0fbpassed119/120tests before bothsanitizer jobs hit their30-minute overall deadline. #655 records this separate
failure. Candidate #656 changes only the two job budgets to90minutes.
Integration source
0c0bcdd1bb99918f75e7276d72efb39a3427157chas exactly the same compiled source/testtree as the preceding consumer; only workflow/docs differ and match the fix
candidate byte-for-byte. Earlier source and integration records remain intact.
Budget integration evidence has SHA-256
2999e9cec323fda71d31f8884bd6fa28826365c7c122694c585b502a801b34ab.Both #656 and this192-test consumer must pass all eight exact-head checks before
checking or merging #655/#656. Then this issue's native dependencies must be
closed and its own checks successful before checking #103 once and merging.
Baseline92-test capacity is not substituted for this expanded-suite gate.