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
2.6 KiB
Markdown
38 lines
2.6 KiB
Markdown
---
|
|
id: "bug-1682"
|
|
status: "fixed"
|
|
status_updated: "2026-09-24"
|
|
status_note: ""
|
|
component: "scheduler"
|
|
severity: "unknown"
|
|
timestamp: "2026-08-24T21:40:00Z"
|
|
last_seen: "2026-09-19T09:30:00Z"
|
|
occurrences: 1
|
|
files: ["src/scheduler/src/plugin.c"]
|
|
tags: ["scheduler", "plugin", "session-hooks", "guard-bug", "latent", "scheduler-driver-equivalence"]
|
|
related: ["bug-1681", "bug-1492"]
|
|
legacy: true
|
|
---
|
|
|
|
# bug-1682: Found while fixing bug-1681 (source read of plugin_end_session(), src/scheduler/src/plugin.c): the…
|
|
|
|
## Description
|
|
Found while fixing bug-1681 (source read of plugin_end_session(), src/scheduler/src/plugin.c): the per-plugin iteration guard checks the WRONG field. plugin_end_session()'s loop tests `if (plugin->begin_session_fn != NULL) { plugin->end_session_fn(plugin); }` -- guards on begin_session_fn but calls end_session_fn. Pre-existing, not introduced by bug-1681's fix (untouched by that diff). Previously documented (not fixed) in testsuite/test_driver_session_ordering.c's own file header comment as a known 'CONCERN surfaced during this test's construction' -- flagged there for the reviewer, deliberately left unfixed as out of that task's scope. Confirmed still present in current code.
|
|
|
|
## Environment
|
|
_Not recorded (migrated from buglog.json)._
|
|
|
|
## Reproduction
|
|
_Not recorded (migrated from buglog.json)._
|
|
|
|
## Root cause
|
|
Copy-paste from plugin_begin_session()'s loop body (which correctly guards on begin_session_fn before calling it) without updating the guard field for the end_session variant.
|
|
|
|
## Fix
|
|
FIXED (2026-09-19, /xce-audit-invariants scheduler dimension-1 audit). Prior tag-only "fixed"/"closed-by-scheduler-sim-live-allocation" stamps on this entry were premature/incorrect -- confirmed by this audit's independent re-read of plugin.c that the guard-field bug was still genuinely present in code as of the scheduler-sim-live-allocation feature's completion (that feature touched plugin_job_election/plugin_prepare_allocation/plugin_commit_decision only, never plugin_begin_session/plugin_end_session's bodies). Now actually fixed: changed plugin_end_session()'s guard from plugin->begin_session_fn != NULL to plugin->end_session_fn != NULL (src/scheduler/src/plugin.c). Confirmed via TDD (RED: new test_end_session_invokes_end_session_fn_when_only_it_is_set() fails 84/85 against the pre-fix guard; GREEN: 85/85 after). Duplicate of bug-1492 (same defect, originally filed 2026-08-04, discovered again independently during this bug's own 2026-08-24 investigation without cross-referencing the earlier entry) -- see bug-1492 for the original filing.
|
|
|
|
## Verification
|
|
_Not recorded (migrated from buglog.json)._
|
|
|
|
## References
|
|
- Related: bug-1681, bug-1492
|