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>
3.2 KiB
| id | status | status_updated | status_note | component | severity | timestamp | last_seen | occurrences | files | tags | related | legacy | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bug-1756 | fixed | 2026-09-24 | worker | unknown | 2026-09-02T02:40:00Z | 2026-09-19T09:00:00Z | 1 |
|
|
|
true |
bug-1756: worker-rejoin-resync task 4.4 (Requirement 3.8, resource_sync.c): AddressSanitizer abort "Joining…
Description
worker-rejoin-resync task 4.4 (Requirement 3.8, resource_sync.c): AddressSanitizer abort "Joining already joined thread, aborting." while building the task's own new resync_pending-gate tests for resource_sync_thread(). Discovered live: the new tests are the first callers ever to exercise resource_sync_create()/resource_sync_start()/resource_sync_stop()/resource_sync_destroy() as a full lifecycle from a test binary (the two pre-existing tests only ever called collect_load_indices() directly). Each test calls resource_sync_stop(sync) then resource_sync_destroy(sync) back-to-back -- the exact same sequence xcecd.c's own real shutdown path (xcecd_cleanup(), src/worker/src/xcecd.c:580-582) already uses in production.
Environment
Not recorded (migrated from buglog.json).
Reproduction
Not recorded (migrated from buglog.json).
Root cause
resource_sync_stop() calls pthread_join(sync->thread, NULL) but never resets sync->thread to 0 (or any invalid sentinel) afterward. resource_sync_destroy() unconditionally calls resource_sync_stop() again as its first action, with no guard on whether the module was already stopped -- so a caller pattern of stop() then destroy() (used verbatim by both xcecd_cleanup() in production and the new task 4.4 tests) pthread_join()s the same already-joined thread handle a second time. This is undefined behavior; under ASAN it aborts hard instead of silently misbehaving. connection_handler.c's analogous network_handler_destroy() already guards its own equivalent second stop call with if (handler->running), showing this exact hazard is a known pattern elsewhere in the same daemon that resource_sync.c's destroy simply never adopted.
Fix
FIXED (2026-09-02, worker-rejoin-resync task 4.4, in-boundary companion fix -- resource_sync.c was already in this task's file boundary and the double-join blocked the task's own new tests from ever completing). resource_sync_destroy() now guards its internal resource_sync_stop() call with if (atomic_load(&sync->running)), mirroring network_handler_destroy()'s established idiom exactly -- if the caller already stopped the module explicitly, destroy() no longer re-joins the same thread handle. Verified via 5 consecutive full runs of the rebuilt resource_sync_test binary under ASAN: 5 Tests, 0 Failures each run, no abort (previously deterministic abort on the very first run reaching a stop()+destroy() pair). Full worker suite (meson test --suite worker) also green: 7/7 OK, no regressions, including controller_connector_test and xcecd_resync_test (task 4.3's own suites).
Verification
Not recorded (migrated from buglog.json).
References
- Related: bug-1519, bug-1757