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-1507"
|
|
status: "fixed"
|
|
status_updated: "2026-09-24"
|
|
status_note: ""
|
|
component: "scheduler"
|
|
severity: "crash"
|
|
timestamp: "2026-08-07T09:10:00Z"
|
|
last_seen: "2026-09-19T09:00:00Z"
|
|
occurrences: 1
|
|
files: ["src/scheduler/src/repl_break.c"]
|
|
tags: ["scheduler-core", "task-12.2", "repl", "xstring", "heap-corruption", "free", "rtti"]
|
|
related: ["bug-1506"]
|
|
legacy: true
|
|
---
|
|
|
|
# bug-1507: sched_repl_breakpoints_inspection_test: 'free(): invalid pointer' SIGABRT inside…
|
|
|
|
## Description
|
|
sched_repl_breakpoints_inspection_test: 'free(): invalid pointer' SIGABRT inside repl_breakpoint_free(), reached from repl_cmd_break()'s malformed-filter rejection path ("break JOB_PUSH nofiltermark"). gdb backtrace: malloc_printerr <- repl_breakpoint_free (repl_break.c:375) <- repl_cmd_break <- repl_dispatch <- test_break_malformed_filter_rejected.
|
|
|
|
## Environment
|
|
_Not recorded (migrated from buglog.json)._
|
|
|
|
## Reproduction
|
|
_Not recorded (migrated from buglog.json)._
|
|
|
|
## Root cause
|
|
repl_breakpoint_free() called FREEUP() (a bare free()) on bp->token and each filter's key/value, all of which are makeString()-produced xstrings -- RTTI reference-counted objects (src/lib/xstring.c: xstring_type has custom .reference/.dereference interface functions; the returned pointer is the object's DATA region inside an RTTI-managed allocation, not a plain malloc() chunk). Releasing one requires going through the RTTI dereference path (xstring_dereference(), which only fires the real free when the refcount hits zero via the owning kvtab's key_del), never a bare free(). This codebase's own kill_sched_job()/kill_sched_queue() (schedobj.c) establish the identical precedent for every other makeString()'d field in the scheduler tree (xce_sched_job_t.submit fields, xce_sched_queue_t.name, etc.): none of them are ever explicitly released either -- they live for the process's lifetime, matching this whole test suite's 'detect_leaks=0 by design' convention. My new repl_break.c code was the first place in this file family to call free() directly on a makeString() result.
|
|
|
|
## Fix
|
|
Removed the FREEUP() calls on bp->token/filters[i].key/filters[i].value from repl_breakpoint_free() (repl_break.c) -- it now only frees the plain calloc'd bp->filters array and bp itself, matching kill_sched_job()/kill_sched_queue()'s established never-release-xstrings precedent. Updated repl_internal.h's xce_repl_filter_t doc comment to state this explicitly so a future maintainer doesn't reintroduce the same free() on an RTTI-managed pointer.
|
|
|
|
## Verification
|
|
_Not recorded (migrated from buglog.json)._
|
|
|
|
## References
|
|
- Related: bug-1506
|