Rongsong Shen
d26ff8870d
The buglog was one 27K-line JSON file whose fields covered only symptom,
root cause and fix, as free text: 2% of entries named the code version a
bug was found on, 12% the build profile, 36% a way to reproduce, and 58% of
the file was OpenWolf hook auto-detections ("Missing guard clause",
"Function not marked async" on .md files).
- .claude/skills/buglog/bug-record.md: the record standard. Eight
sections (Summary as the title, Description, Environment, Reproduction;
Root cause, Fix, Verification, References once fixed), each with the
content it must contain -- e.g. Environment needs the OS, the libc the
failing binary loads (from its ELF interpreter: glibc 2.42 in the nix
shell vs 2.40 on the host), toolchain, build profile and commit.
- .wolf/bugs/<id>.md: 500 records migrated from buglog.json; every field
verified to round-trip. They carry legacy: true and an explicit "not
recorded" where the old log had no such field; component derived from
file paths, severity from keywords (else unknown), fix_commits from
hashes in status_note/fix that git resolves (91 of 463 fixed).
- .wolf/buglog-archive.json: the 710 hook auto-detections, verbatim.
- .wolf/buglog.json is now an index generated from the records, keeping
the fields OpenWolf's hooks and `openwolf bug search` read.
- scripts/buglog.py rewritten: new (template + Environment collected from
the machine and --binary), check, status (refuses a status the record
doesn't support; fixed needs Root cause, Fix, Verification and
--commit hashes git resolves), note, recur, set, show, list, triage,
env, index. 29 unittest cases in scripts/tests/test_buglog.py.
- scripts/git-hooks/pre-commit runs `check --staged`: incomplete staged
records, a legacy record hand-changed to fixed without the fix-time
sections, a new record claiming legacy: true, or a stale index refuse
the commit.
- scripts/gitea-sync.py reads the records: issue body = the record's
sections plus linked fix commits (origin remote, GITEA_COMMIT_REPO to
override), and a body hash so an issue's text is refreshed when the
record changes.
- bug-1866 rewritten to the full standard as the worked example.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
38 lines
3.8 KiB
Markdown
38 lines
3.8 KiB
Markdown
---
|
|
id: "bug-1623"
|
|
status: "fixed"
|
|
status_updated: "2026-09-24"
|
|
status_note: ""
|
|
component: "scheduler"
|
|
severity: "unknown"
|
|
timestamp: "2026-08-22T12:40:00Z"
|
|
last_seen: "2026-09-19T09:00:00Z"
|
|
occurrences: 1
|
|
files: ["src/scheduler/testsuite/test_sim_e2e_original_motivation.c"]
|
|
tags: ["scheduler", "sim", "replay", "reconciliation", "test-authoring", "scheduler-sim-decision-parity"]
|
|
related: ["bug-1611", "bug-1613"]
|
|
legacy: true
|
|
---
|
|
|
|
# bug-1623: Test-authoring bug while building scheduler-sim-decision-parity task 4.3…
|
|
|
|
## Description
|
|
Test-authoring bug while building scheduler-sim-decision-parity task 4.3 (test_sim_e2e_original_motivation.c): initial version asserted resInUse checkpoints of 17.00/18.00/20.00/12.00 (derived from the RECORDING driver own in-memory resInUse arithmetic) but the actual replay-side info worker output showed only 12.00 (ANCHOR own charge) at every checkpoint -- the first three assertions failed.
|
|
|
|
## Environment
|
|
_Not recorded (migrated from buglog.json)._
|
|
|
|
## Reproduction
|
|
_Not recorded (migrated from buglog.json)._
|
|
|
|
## Root cause
|
|
Recorded-Decision Reconciliation (sim_reconcile_unaccounted_jobs(), sim_loop.c, scheduler-sim-decision-parity tasks 2.2-2.4/3) matches a RUN-state, unaccounted job against a DISPATCH decision recorded in THAT SAME replayed round only -- never an earlier or later round (each round decoded_session is decoded fresh and freed immediately after that round own processing). A job that arrives PEND and is genuinely live-dispatched by a real allocator in round K, then confirmed RUN via a JOB_STATE_CHANGE event in a LATER round K+1, can never be reconciled during replay: round K+1 own decoded decisions do not contain a match (the real DISPATCH is in round K own data), so the job stays RUN+unaccounted forever (Requirement 4 graceful degradation, correctly, not a bug in production). This is structurally impossible to avoid for ANY organically-arriving job, since a job cannot simultaneously be a fresh PEND candidate for a live picker (required to ever get a real DISPATCH decision) AND already-RUN-before-the-picker-runs (required for reconciliation to even consider it) in the same round -- confirmed by direct source read of sim_loop.c/driver.c/builtin_fcfs.c before writing any test code, then independently reconfirmed empirically by running the real test binary. Only ONE job in the whole fixture (ANCHOR_ID) was deliberately engineered so its arrival+RUN-confirmation events AND its own injected DISPATCH decision land in the exact same round (round 1) -- every other churn job (J2-J12) arrives and is genuinely live-dispatched by the test own capacity-aware allocator thread in ITS OWN arrival round, so its dispatch can never be reconciled later, exactly like every other real production job whose confirmation lags its dispatch by more than zero rounds.
|
|
|
|
## Fix
|
|
Changed all four info-worker checkpoint assertions from the RECORDING-side arithmetic (17.00/18.00/20.00/12.00, i.e. what a normal-mode continuation of the same session would show) to the REPLAY-side reality: a constant used=12.00 (ANCHOR_ID own charge only) at all four checkpoints spanning the 24-round replay, since no churn job dispatch is ever reconciled. Added an explicit non-tautology guard (used=0.00 never appears) and an extensive header-comment/round-table-comment explanation of why churn-job charges never surface during replay, so a future reader does not mistake this for an oversight. Also independently reconfirmed via a live discriminating-power check: commenting out sim_loop.c own sim_reconcile_unaccounted_jobs() call site made the ANCHOR_ID DISPATCH-visibility assertion fail for the right reason; restoring it (git diff clean on production files afterward) made the whole suite green again.
|
|
|
|
## Verification
|
|
_Not recorded (migrated from buglog.json)._
|
|
|
|
## References
|
|
- Related: bug-1611, bug-1613
|