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>
3.5 KiB
| id | status | status_updated | status_note | component | severity | timestamp | last_seen | occurrences | files | tags | related | legacy | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| bug-1593 | fixed | 2026-09-24 | scheduler | unknown | 2026-08-19T08:00:00Z | 2026-09-19T09:00:00Z | 1 |
|
|
|
true |
bug-1593: User-reported (live REPL testing): info resource all and info resource <name> aggregate every…
Description
User-reported (live REPL testing): info resource all and info resource <name> aggregate every worker's per-host resource values (total=sum(available), available=sum) for EVERY resource, including host-based ones like cpus/mem -- a fabricated, meaningless number for a per-host consumable already correctly shown by info worker <name>. User asked that info resource instead show resource metadata and be restricted to globally-shared resources only.
Environment
Not recorded (migrated from buglog.json).
Reproduction
Not recorded (migrated from buglog.json).
Root cause
xce_resource_t (xce/types.h) carries a flags bitmask (XCE_RESF_BUILTIN/DYNAMIC/GLOBAL/SHARED/EXTERNAL/RELEASE/DEFAULT_HOST/CONSUMABLE/LIMIT) that distinguishes a genuinely cluster-wide resource (GLOBAL/SHARED, stored once in core->sharedInst, indexed by r->idx -- the same pool RESVALS_FULL/RESVALS_DELTA event handling and snapshot save/load already treat as the defining store) from a host-based/per-worker-distributed resource (stored per-worker in w->resvals/w->resInUse, already correctly surfaced by repl_info_worker_one()). repl_info_resource_one() never consulted r->flags or core->sharedInst at all -- it iterated sim->core.workers unconditionally for every resource regardless of kind, summing w->resvals[r->idx] across the whole cluster. For a host-based resource (the common case -- cpus, mem, etc.) this total/available is not a meaningful quantity (no single cluster-wide capacity exists for a per-host resource); for a genuinely shared resource it also happened to work by accident (since a shared resource's value is typically replicated identically into every worker's resvals array too), but even then never reported the resource's own metadata (desc, valtype, flags, interval) the user asked for.
Fix
Fixed as part of task 12.5 (scheduler-core spec, Requirement 5 AC 22 amendment). repl_info_resource_one() rewritten to print resource DEFINITION metadata (name, desc, valtype, human-readable flags list, interval) plus, only for a resource flagged XCE_RESF_GLOBAL or XCE_RESF_SHARED, its current value read directly from core->sharedInst[r->idx] (never aggregated across workers). A host-based resource's metadata is still printed but with an explicit note directing the operator to info worker <name> for its actual per-host values -- no fabricated total/available. repl_info_resource() (info resource all) now skips any resource whose flags do not include GLOBAL or SHARED, per the user's explicit request to keep this command scoped to cluster-wide resources. New tests in test_repl_breakpoints_inspection.c construct a real GLOBAL/SHARED resource via xce_core_push_events() and a plain host-based one, asserting the metadata line and shared value appear for the former, the host-based one is excluded from info resource all, and both appear correctly for info resource <name> with the host-based one omitting a fabricated total/available pair.
Verification
Not recorded (migrated from buglog.json).
References
- Related: bug-1590