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>
3.7 KiB
| id | status | status_updated | status_note | component | severity | timestamp | last_seen | occurrences | files | tags | related | legacy | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bug-1567 | fixed | 2026-09-24 | scheduler | unknown | 2026-08-17T09:00:00Z | 2026-09-19T09:00:00Z | 2 |
|
|
true |
bug-1567: xce_scheduler -i -s : WARN 'failed to resolve job picker plugin (rc=-2)' on load; all…
Description
xce_scheduler -i -s : WARN 'failed to resolve job picker plugin (rc=-2)' on load; all pre-existing jobs stay PEND forever, nothing ever dispatches in simulation/interactive mode
Environment
Not recorded (migrated from buglog.json).
Reproduction
Not recorded (migrated from buglog.json).
Root cause
Two compounding, both explicitly documented 'Known scope boundary' gaps in the scheduler-core sim driver (task 11.1/11.2, .kiro/specs/scheduler-core): (1) sched_sim_driver_start() (sim.c) loads the initial snapshot via xce_sched_core_load_snapshot() BEFORE any picker plugin is registered into core->plugins -- only job_pick_plugin's NAME is threaded through from the ReplayHeader (sim.h:206-216's own doc comment). sched_resolve_job_picker() therefore returns -ENOENT (=-2), producing the WARN, and the picker-rebuild walk (on_queue_update/on_job_push) inside xce_sched_core_load_snapshot() is silently skipped for every job/queue in the initial snapshot. The real picker only gets registered moments later, in sim_register_default_picker() (sim_loop.c:381, called from sched_sim_driver_run() right after start() returns) -- but nothing re-pushes the already-loaded jobs/queues into it, so any job present in the replay file's INITIAL snapshot is permanently invisible to the fcfs picker's next_job(). (2) Independently, sim_picker_thread_fn() (sim_loop.c:448) has its own documented scope boundary: real resource-matching allocation (sched_make_allocation()) needs a live xce_scheduler_t* that the sim driver deliberately doesn't have, and no core-only equivalent exists yet (deferred to the separate, not-yet-designed .kiro/specs/scheduler-allocation spec) -- so EVERY candidate the picker selects is unconditionally turned into a PEND decision; dispatch is not implemented in simulation mode at all today, regardless of resource availability.
Fix
FIXED (TDD, RED->GREEN). New file src/scheduler/src/sim_picker.c: sim_bootstrap_picker() (replaces sim_loop.c's sim_register_default_picker(), moved out because sim_loop.c was already at its documented 800-line 'zero line-budget headroom' cap) registers the real builtin_fcfs picker AND then replays the picker-rebuild walk (on_queue_update per queue, on_job_push per pending job) that xce_sched_core_load_snapshot() had to skip at initial-snapshot-load time. Wired into scheduler_srcs, sched_sim_isolation_lib, and all 9 white-box test TUs that #include sim_loop.c directly (test_sim_isolation.c mechanically re-verified). New regression test: test_sim_event_loop.c's test_job_present_only_in_initial_snapshot_still_reaches_the_picker -- builds a fixture with a job baked into the INITIAL snapshot (not a ReplayEvent), asserts it gets a live PEND decision. RED confirmed by temporarily no-op'ing the rebuild walk (test failed as expected), GREEN with the real fix restored. Full scheduler suite green (41/41, only pre-existing unrelated bug-1560 skipped/excluded from that count). Note: finding (2) from the original diagnosis (sim mode never dispatches, only PENDs -- scheduler-allocation spec gap) is NOT fixed here, by design -- out of scope, a whole separate not-yet-designed spec.
Verification
Not recorded (migrated from buglog.json).
References
Not recorded (migrated from buglog.json).