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>
3.3 KiB
| id | status | status_updated | status_note | component | severity | timestamp | last_seen | occurrences | files | tags | related | legacy | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bug-1598 | fixed | 2026-09-24 | scheduler | unknown | 2026-08-19T00:00:00Z | 2026-09-19T09:00:00Z | 1 |
|
|
true |
bug-1598: AddressSanitizer: stack-use-after-return on address inside sim (xce_sched_sim_driver_t) at…
Description
AddressSanitizer: stack-use-after-return on address inside sim (xce_sched_sim_driver_t) at cbuf_get, thread T2 (sim_picker_thread_fn), reading test_run_deactivates_tlog_virtual_clock_logging_on_completion's stack frame after it returned
Environment
Not recorded (migrated from buglog.json).
Reproduction
Not recorded (migrated from buglog.json).
Root cause
sched_sim_driver_run() (src/scheduler/src/sim_loop.c) spawns sim_picker_thread_fn on a real pthread that is, by this codebase's own documented design (that function's own header comment), never joined or cancelled -- no sim-driver teardown/fini function exists yet. It loops forever on cbuf_get(&sim->core.wbuf, ...), reading sim for the lifetime of the process. Every white-box test in test_sim_event_loop.c that calls sched_sim_driver_run()/sched_sim_driver_start() with a stack-local xce_sched_sim_driver_t sim; was therefore ALWAYS a latent stack-use-after-return once its test function returned -- it simply hadn't been caught yet, because no prior test ordering/stack-frame-reuse pattern had triggered ASAN's shadow-memory detector. Adding one new test (test_run_deactivates_tlog_virtual_clock_logging_on_completion, part of the scheduler-core design-review follow-up wiring tlog virtual-clock activation into sim.c/sim_loop.c) between two existing tests shifted the stack layout enough to make the pre-existing latent bug fire deterministically (reproduced on 2 independent runs).
Fix
Changed the fixture from stack-local to static xce_sched_sim_driver_t sim; storage in three places: test_run_processes_multiple_sessions_in_order_and_no_longer_returns_enosys(), test_run_deactivates_tlog_virtual_clock_logging_on_completion() (the test that first exposed it), and the run_sim_and_capture_stdout() helper (used once, by test_job_present_only_in_initial_snapshot_still_reaches_the_picker()). Safe specifically because: (a) sched_sim_driver_start() unconditionally memset()s the whole struct as its first action, so no stale-data leak across runs; (b) each affected function is called exactly once per test-binary process (verified via grep, no re-entrancy), so no second call's memset can ever stomp a still-live zombie thread's cbuf mutex/condvar state from a first call -- static would be UNSAFE on a helper called more than once while an earlier zombie thread might still be alive. Does NOT fix the underlying thread-lifecycle gap (still no join/cancel) -- that is a separate, larger, already-documented-as-deferred architectural item (a real sim-driver teardown/fini function). Verified via 2 independent direct runs of the rebuilt sched_sim_event_loop_test binary post-fix: 4 Tests, 0 Failures, clean exit code 0 both times, no ASAN abort.
Verification
Not recorded (migrated from buglog.json).
References
Not recorded (migrated from buglog.json).