Files
ersatztv/scripts/decisions_validate.py
T
timothy 0f565b1f7e fix(621): make an unparseable decision record loud instead of silently invisible
A file under the record wings that the dependency-free frontmatter reader cannot
parse returned [] and simply vanished from the corpus: decisions_validate.py printed
OK, build_decisions_catalog.py --check said "up to date", the record was absent, and
nothing anywhere errored. Reproduced end-to-end with a YAML block scalar (`rule: >-`),
which is the natural thing to reach for on this corpus's very long rule: values and
which parsed fine under PyYAML before #610's dependency-free reader.

An EXISTING record disappearing was already loud (the no-vanish diff check). The hole
was a NEWLY ADDED record, which that check structurally cannot see — no base state to
diff against — so the author's own PR looks clean. Hence a per-PATH check, not a
per-construct or diff-driven one: asserting "this path yields exactly one keyed
record" turns any present or future reader limitation from silent to loud in one move.

Adds record_wing_files/record_wing_faults to decisions_validate.py, surfaced through
validate() as ERRORS (a file in the record wings that is not a record is a mistake by
definition) and reported first, since a file that failed to parse leaves every
downstream check silently evaluating an incomplete corpus. The five top-level stripped
legacy archive files are exempt — they are generated "Records formerly in this file"
indexes, keyless by construction, and are what keeps older date-based pointers
resolvable.

_read_frontmatter is deliberately NOT extended to accept block scalars; rationale in
the new docs.record-wing-parse-guard record.

8 tests, mutation-verified: with the check neutered 5 go red, restored all 119 pass.
A live-corpus positive control asserts the wings are non-empty so a clean result can
never be vacuous.

Refs #621
2026-07-26 12:03:43 +02:00

594 lines
28 KiB
Python

#!/usr/bin/env python3
"""Validate ErsatzTV decision-record lifecycle invariants.
Replaces the line-level append-only guard (ersatztv#303 H9) with lifecycle checks that preserve its
spirit — rationale is never silently rewritten or deleted; every history touch is deliberate and
reviewable. A `Decisions-Edit: yes` commit trailer is kept ONLY for rationale-prose edits; routine
lifecycle metadata writes are marker-free. Fail-open on tooling trouble (missing refs, git errors,
parse issues), matching the old guard — with ONE deliberate exception: resolving the edit marker
itself fails CLOSED (`_edit_trailer_armed`), because a fail-open marker lookup is just #609 again.
"""
from __future__ import annotations
import argparse
import re
import subprocess
import sys
from datetime import date
from pathlib import Path
import scripts.decisions_lib as dl # noqa: E402 (run with PYTHONPATH=. or as module)
SKIP_HEADINGS = dl.SKIP_HEADINGS # single source of truth
# `signals` is required alongside the lifecycle fields: the `**Signals:**` line (plus `key:`) is what
# MemPalace's keyword recall matches on when surfacing a decision, so a record without it ingests with
# weak recall metadata and produces confident false-negatives for the "MemPalace to find, file to
# confirm" workflow (ersatztv#545).
REQUIRED_META = ("key", "status", "since", "supersedes", "superseded_by", "signals")
# The rationale-edit marker is a git TRAILER, not a bare substring anywhere in the commit range.
# ersatztv#609: the original `[decisions-edit]` substring test armed on ANY commit message containing
# the literal string — including prose *about* the mechanism, and including a message explaining why
# no token was needed — which silently disabled all three body-diff comparisons while the job still
# reported green. Git recognises a trailer only in a message's final trailer block, so discussing the
# marker mid-message — which is what actually happened — can no longer arm anything. It is NOT a total
# immunity: a quoted example that IS the final paragraph of a NON-MERGE commit message parses as a
# real trailer and does arm (merge commits are excluded — see `_edit_trailer_armed`). See
# `ci.decisions-edit-trailer` for that residual, stated rather than papered over.
EDIT_TRAILER = "Decisions-Edit"
_EDIT_TRAILER_AFFIRMATIVE = frozenset({"yes", "true", "1"})
_LEGACY_EDIT_TOKEN = "[decisions-edit]" # noqa: S105 (a commit-message marker, not a credential)
# `date.fromisoformat` is NOT a YYYY-MM-DD validator: on Python >= 3.11 it also accepts ISO basic
# format ("20270101") and week dates ("2027-W01-1"), and which of those parse depends on the
# interpreter version — so the same corpus could validate differently on a dev machine and on the
# runner (pr-checks.yml pins only `python-version: '3.x'`). Gate on the extended calendar form first.
# The catalog's Review-due section additionally sorts on the raw STRING, which is only equivalent to
# sorting by date because this regex forces a fixed-width zero-padded form.
_STALE_AFTER_RE = re.compile(r"^\d{4}-\d{2}-\d{2}$")
def _parse_stale_after(value: str | None) -> date | None:
"""`stale-after` as a date, or None if absent, empty, or malformed.
Callers must NOT use a None return to mean "absent" — an empty or malformed value returns None
too. Distinguish on the raw field with `is not None` (absent stays None; present-but-empty is "").
"""
if not value or not _STALE_AFTER_RE.match(value.strip()):
return None
try:
return date.fromisoformat(value.strip())
except ValueError:
return None # e.g. 2027-02-30 — well-shaped but not a real date
def stale_records(records, today: date) -> list[tuple[str, str]]:
"""(heading, stale-after) for ACTIVE records that are past due — today >= stale-after.
OKF v0.2 semantics: an absolute date, no TTL. Deliberately NOT part of `validate()` — a record
going stale is the passage of time, not a defect in the commit under test, so it must never fail
a PR that didn't touch it. main() reports these as a non-blocking notice.
"""
out = []
for r in records:
if r.status != "active":
continue
d = _parse_stale_after(r.stale_after)
if d is not None and today >= d:
out.append((r.heading, r.stale_after))
return sorted(out)
def _key_of(ref: str | None) -> str | None:
if not ref or ref == "none":
return None
return ref.split("@", 1)[0].strip()
def validate(
records,
*,
archive_keys,
catalog_ok,
budget_ok,
removed,
rewritten,
archive_records=None,
demoted=(),
wing_faults=(),
) -> list[str]:
errs: list[str] = []
archive_records = archive_records or []
# #621: structural "one keyed record per file". ERRORS, not warnings — a file under the record
# wings that is not a record is a mistake by definition. Reported first because a file that
# failed to parse is absent from `records`, so every downstream check below is silently
# evaluating an incomplete corpus.
errs += list(wing_faults)
decision_recs = [r for r in records if r.heading not in SKIP_HEADINGS and r.status != "legacy-unmigrated"]
active_by_key: dict[str, int] = {}
known_keys = set(archive_keys)
for r in decision_recs:
if r.status not in dl.STATUSES:
errs.append(f"{r.heading}: invalid status {r.status!r}")
if not r.key or not dl.KEY_RE.match(r.key):
errs.append(f"{r.heading}: bad key format {r.key!r}")
continue
known_keys.add(r.key)
for f in REQUIRED_META:
if getattr(r, f) in (None, ""):
errs.append(f"{r.heading}: missing required metadata {f}")
# `stale-after` is OPTIONAL, but a malformed one is a defect in the commit that wrote it —
# it would silently never fire. Absent is fine; unparseable is not.
if r.stale_after is not None and _parse_stale_after(r.stale_after) is None:
errs.append(f"{r.heading}: stale-after {r.stale_after!r} is not a YYYY-MM-DD date")
if r.status == "active":
active_by_key[r.key] = active_by_key.get(r.key, 0) + 1
if r.status in ("superseded", "retired"):
errs.append(
f"{r.heading}: status {r.status} but still in the active set — relocate to docs/decisions/archive/"
)
n_meta = dl.metadata_line_count(r)
if n_meta > 1:
errs.append(f"{r.heading}: {n_meta} metadata blocks found (expected 1) — duplicate metadata block")
# PATH <-> KEY (#610). In the split layout the filename is derived from the key, which is what
# makes one-active-record-per-key a filesystem property rather than a check. That only holds if
# the two cannot drift, so enforce the correspondence directly.
for r in decision_recs + list(archive_records):
if not r.key:
continue
src = Path(r.source)
if dl.RECORDS_DIR not in src.parents and dl.ARCHIVE_DIR not in src.parents:
continue # legacy multi-record file; the rule applies to the split layout only
area, _, topic = r.key.partition(".")
if src.parent.name != area or src.stem != topic:
errs.append(f"{r.heading}: key {r.key!r} does not match its path {src.parent.name}/{src.name}")
for r in archive_records:
if r.status == "active":
errs.append(f"{r.heading}: active record must not live under docs/decisions/archive/")
# An archived record's staleness is moot, but a malformed date there is still a typo worth
# catching — and checking both wings keeps the field's contract symmetric.
if r.stale_after is not None and _parse_stale_after(r.stale_after) is None:
errs.append(f"{r.heading}: stale-after {r.stale_after!r} is not a YYYY-MM-DD date")
for key, n in active_by_key.items():
if n > 1:
errs.append(f"key {key!r}: more than one active record ({n})")
# by_key map across BOTH wings, for reciprocity checks (existence-only checks still use
# known_keys, which additionally includes keys the caller only knows about via archive_keys).
by_key: dict[str, dl.Record] = {}
for r in decision_recs + list(archive_records):
if r.key:
by_key[r.key] = r
# reciprocal supersession (successor/predecessor may live in active OR archive). Runs over the
# union of active + archive records so a twice-reversed decision (an archived record whose
# superseded-by points to ANOTHER archived record) is checked from both sides too.
for r in decision_recs:
if r.status == "active" and r.superseded_by not in (None, "", "none"):
errs.append(f"{r.heading}: active record cannot already be superseded (superseded-by set)")
for r in decision_recs + list(archive_records):
sk = _key_of(r.superseded_by)
if sk:
if sk not in known_keys:
errs.append(f"{r.heading}: superseded-by points to unknown key {sk!r}")
else:
b = by_key.get(sk)
if b is not None and r.key and _key_of(b.supersedes) != r.key:
errs.append(f"{r.heading}: superseded-by {sk} but {sk} does not point back (supersedes)")
pk = _key_of(r.supersedes)
if pk:
if pk not in known_keys:
errs.append(f"{r.heading}: supersedes points to unknown key {pk!r}")
else:
a = by_key.get(pk)
if a is not None and r.key and _key_of(a.superseded_by) != r.key:
errs.append(f"{r.heading}: supersedes {pk} but {pk} does not point back (superseded-by)")
for h in removed:
errs.append(f"record removed from the active set without an archive copy: {h!r}")
for h in rewritten:
errs.append(f"rationale prose of {h!r} changed without a '{EDIT_TRAILER}: yes' commit trailer")
for h in demoted:
errs.append(f"{h}: migrated record demoted to legacy-unmigrated (metadata block removed)")
if not catalog_ok:
errs.append("docs/decisions/README.md active catalog is stale — run build_decisions_catalog.py")
if not budget_ok:
# non-blocking: reported as a warning by the caller (main()), never added here.
pass
return errs
# ---- git helpers (all fail-open: return neutral values on any error) ----
def _run(args: list[str]) -> str | None:
try:
r = subprocess.run(args, capture_output=True, text=True, timeout=30)
except (OSError, subprocess.SubprocessError):
return None
return r.stdout if r.returncode == 0 else None
def _merge_base(base: str, head: str) -> str | None:
out = _run(["git", "merge-base", base, head])
return out.strip() if out else None
def _active_paths_at(ref: str) -> list[str]:
"""Active decision files at a ref: decisions.md + docs/decisions/*.md minus README/archive/migration-map."""
out = _run(["git", "ls-tree", "-r", "--name-only", ref, "docs/decisions.md", "docs/decisions/"])
if out is None:
return []
paths = []
for p in out.splitlines():
if not p.endswith(".md"):
continue
if p.startswith("docs/decisions/archive/") or Path(p).name in dl._NON_DECISION_FILES:
continue
paths.append(p)
return paths
def _records_at(ref: str, paths: list[str]):
"""{key: Record} across the given paths at a ref.
Keyed by `key`, NOT by heading (#610). Heading-keying made a pure RENAME look like a removal
plus an unrelated addition, so retitling a record failed CI as an "unlogged removal" — a trap
that has bitten before. `key` is the record's identity and is stable across a retitle, a move
between files, and the legacy->frontmatter format change, which is also what lets this compare
correctly straight through the migration commit.
A record with no key (legacy-unmigrated) falls back to its heading, prefixed so it cannot
collide with a real dotted key.
"""
by_key = {}
for p in paths:
blob = _run(["git", "show", f"{ref}:{p}"])
if blob is None:
continue
for rec in dl.parse_text(blob, Path(p)):
if rec.heading in SKIP_HEADINGS:
continue
by_key[rec.key or f"heading:{rec.heading}"] = rec
return by_key
def _archive_records_at(ref: str):
out = _run(["git", "ls-tree", "-r", "--name-only", ref, "docs/decisions/archive/"])
paths = [p for p in (out or "").splitlines() if p.endswith(".md")]
by_key = {}
for p in paths:
blob = _run(["git", "show", f"{ref}:{p}"])
if blob is None:
continue
for rec in dl.parse_text(blob, Path(p)):
# Filter SKIP_HEADINGS exactly as _records_at does. Without it the generated
# "Records formerly in this file" section in each stripped archive file parses as a
# pseudo-record; they then collide by heading across files, and any later PR that
# touches those generated bullets trips a false "rationale prose changed without a
# trailer" failure.
if rec.heading not in SKIP_HEADINGS:
by_key[rec.key or f"heading:{rec.heading}"] = rec
return by_key
def _rationale(rec) -> str:
"""Record body with the contiguous top metadata block stripped, whitespace-normalized.
Only the block from the first non-blank line (when it is the `key:` meta line) up to the first
following blank line is metadata. A **Rule:**/**Signals:**/**Mechanics:**/`key:` line appearing
later in rationale prose is prose, not metadata (mirrors decisions_lib's contiguous-block rule).
"""
lines = rec.body.splitlines()
start = 0
while start < len(lines) and not lines[start].strip():
start += 1
if start < len(lines) and lines[start].strip().startswith("`key:"):
end = start + 1
while end < len(lines) and lines[end].strip():
end += 1
rest = lines[:start] + lines[end:]
else:
rest = lines
return "\n".join(s for s in (ln.strip() for ln in rest) if s)
def _edit_trailer_armed(mb: str, head: str) -> bool:
"""True when some NON-MERGE commit in `mb..head` carries an affirmative `Decisions-Edit:` trailer.
A non-affirmative value (`no`, `not-needed`, …) deliberately does NOT arm the exemption: the
marker must be an explicit yes, so a trailer written to *record that no edit was intended* can
never disable the guard — the ersatztv#609 failure mode, one abstraction level up.
This is the one git helper here that deliberately does NOT fail open (see the module docstring):
on a git error it returns False, leaving the guard ON. A lost marker costs a contributor one
clear error message and a re-push; a lost guard costs a silent history rewrite in the PR that
most needed policing. Failing open here would reintroduce #609 through a different door.
`--no-merges` skips merge commits: on a `pull_request` event `actions/checkout` lands on a
synthetic merge commit whose body is composed by the forge from the PR description, not by the
author, so a trailer parsed out of it was never a deliberate marker.
This does cost a real (if narrow) false negative, and the trade is deliberate. Merging main into
a PR branch is discouraged by convention but NOT mechanically blocked — `prepush-rebase-check.sh`
only refuses a branch that is BEHIND origin/main, and a merge makes origin/main an ancestor, so
the push is allowed. An author who resolves a rationale conflict in such a merge and puts the
ONLY marker on that merge commit gets their legitimate rewrite rejected. That failure is LOUD and
costs one extra commit carrying the trailer; honoring forge-composed merge bodies would instead
disable the guard SILENTLY, which is the #609 failure itself. Prefer the loud one.
`unfold` joins RFC-822 continuation lines before the value is compared. Without it, a folded
`Decisions-Edit: no\\n yes` yields two lines and the continuation ` yes` arms the exemption on
its own — the value the author actually wrote is `no`.
"""
fmt = f"--format=%(trailers:key={EDIT_TRAILER},valueonly,unfold)"
out = _run(["git", "log", "--no-merges", fmt, f"{mb}..{head}"])
if out is not None and any(ln.strip().lower() in _EDIT_TRAILER_AFFIRMATIVE for ln in out.splitlines()):
return True
# Retired-form nudge: without it, a contributor using the pre-#609 habit gets a bare "prose
# changed" failure and no hint that the marker's FORM (not their intent) is what changed.
if _LEGACY_EDIT_TOKEN in (_run(["git", "log", "--no-merges", "--format=%B", f"{mb}..{head}"]) or "").lower():
print(
f"::warning::decisions-validate: the commit range carries the retired "
f"{_LEGACY_EDIT_TOKEN} marker but no '{EDIT_TRAILER}: yes' trailer — since ersatztv#609 "
"the rationale-edit marker is a git trailer, and the substring no longer exempts anything",
file=sys.stderr,
)
return False
def _diff_findings(base: str, head: str) -> tuple[list[str], list[str], list[str]]:
"""(removed, rewritten, demoted) between merge-base(base,head) and head. Fail-open → ([], [], []).
Records are MATCHED by `key` and REPORTED by heading. Matching by key means a retitle is a
retitle rather than a removal-plus-addition (#610) — heading-keying made renaming a heading fail
CI as an "unlogged removal", a trap that has bitten before — and it is what lets the comparison
work straight through the legacy→frontmatter migration, since `key` survives both the format
change and the move between files. Headings are still what appears in the error, because a key
alone is not enough for a human to find the record.
"""
mb = _merge_base(base, head)
if not mb:
print(
f"::warning::decisions-validate: could not resolve merge-base({base},{head}); no-vanish/body-diff skipped",
file=sys.stderr,
)
return [], [], []
token = _edit_trailer_armed(mb, head)
base_active = _records_at(mb, _active_paths_at(mb))
head_active = _records_at(head, _active_paths_at(head))
head_archive = _archive_records_at(head)
base_archive = _archive_records_at(mb)
def name(rec) -> str:
return rec.heading
gone = set(base_active) - set(head_active)
removed = [name(base_active[k]) for k in sorted(gone) if k not in head_archive]
# an archive record present at base must not vanish entirely (neither wing has it at head)
archive_gone = set(base_archive) - set(head_archive)
removed += [name(base_archive[k]) for k in sorted(archive_gone) if k not in head_active]
rewritten: list[str] = []
if not token:
# surviving records whose rationale prose changed
for k in set(base_active) & set(head_active):
if _rationale(base_active[k]) != _rationale(head_active[k]):
rewritten.append(name(head_active[k]))
# archived records must body-match their base active copy (no laundering rewrites via archive)
for k in gone & set(head_archive):
if _rationale(base_active[k]) != _rationale(head_archive[k]):
rewritten.append(name(head_archive[k]))
# archive records that survive in the archive wing: rationale must not be rewritten either
for k in set(base_archive) & set(head_archive):
if _rationale(base_archive[k]) != _rationale(head_archive[k]):
rewritten.append(name(head_archive[k]))
# Demotion: a record that HAD a key at base and lost it at head. Under key-matching that shows
# up as its key vanishing while a `heading:`-fallback entry appears for the same heading, so it
# is detected on the heading axis rather than the key axis.
head_by_heading = {r.heading: r for r in head_active.values()}
demoted: list[str] = []
for k in gone:
rec = base_active[k]
if not rec.key:
continue
successor = head_by_heading.get(rec.heading)
if successor is not None and successor.key != rec.key:
demoted.append(rec.heading)
# a demoted record is not also "removed" — report the specific fault, not both
removed = [h for h in removed if h not in set(demoted)]
return sorted(set(removed)), sorted(set(rewritten)), sorted(set(demoted))
def record_wing_files(records_dir: Path | None = None, archive_dir: Path | None = None) -> list[Path]:
"""Every file that MUST be exactly one keyed record.
Scope is the split-layout record wings only:
* `docs/decisions/records/**/*.md` — the active wing
* `docs/decisions/archive/<area>/*.md` — the archived wing, nested by area since #610
Deliberately NOT included: `docs/decisions/archive/*.md` at the top level. Those are the five
stripped legacy topic files left behind by the #610 split (`api.md`, `scan.md`, `spa.md`,
`startup.md`, `release-ci-governance.md`); each parses to a single KEYLESS pseudo-record from its
generated "Records formerly in this file" index, which is exactly what keeps older date-based
pointers resolvable. They are indexes, not records. `_NON_DECISION_FILES` (README.md and
friends) is filtered on top of that.
"""
records_dir = dl.RECORDS_DIR if records_dir is None else records_dir
archive_dir = dl.ARCHIVE_DIR if archive_dir is None else archive_dir
files: list[Path] = []
if records_dir.exists():
files += sorted(records_dir.rglob("*.md"))
if archive_dir.exists():
files += sorted(archive_dir.glob("*/*.md")) # area-nested only — see docstring
return [p for p in files if p.name not in dl._NON_DECISION_FILES]
def record_wing_faults(records_dir: Path | None = None, archive_dir: Path | None = None) -> list[str]:
"""Structural check (#621): every file in the record wings parses to exactly ONE keyed record.
Without this, a file the frontmatter reader cannot parse yields `[]` and simply VANISHES — no
error, no warning, validator OK, catalog "up to date", record absent from the corpus. That is
the corpus's own failure mode turned on itself: the one thing worse than a missing record is a
missing record that reports success (cf. the #609 marker that printed OK while doing nothing and
the #603 `stale-after` that silently never fired).
An EXISTING record disappearing was already loud — the no-vanish diff check catches it. This
closes the case the diff check structurally cannot see: a NEWLY ADDED record, where the author's
own PR looks clean because there is no prior state to diff against.
Making it path-driven rather than record-driven is the point: it converts ANY present or future
reader limitation from silent to loud, instead of enumerating the constructs we happen to know
about today.
"""
faults: list[str] = []
for p in record_wing_files(records_dir, archive_dir):
try:
recs = dl.parse_file(p)
except Exception as exc: # unreadable/undecodable file is a fault, not a crash
faults.append(f"{p}: could not be read as a decision record ({exc})")
continue
if len(recs) != 1:
faults.append(
f"{p}: parsed to {len(recs)} records, expected exactly 1 — a file under the record "
f"wings must be one record. Common cause: YAML the dependency-free frontmatter "
f"reader does not accept (a block scalar such as `rule: >-` or `rule: |`, any "
f"indented/nested structure, or unterminated `---` frontmatter). Put the value on "
f"ONE quoted line, as all other records do."
)
continue
if not recs[0].key:
faults.append(f"{p}: parsed to a record with no `key` — every record in the wings must be keyed.")
return faults
def _archive_keys() -> set[str]:
keys: set[str] = set()
if dl.ARCHIVE_DIR.exists():
for f in dl.ARCHIVE_DIR.rglob("*.md"): # rglob: archive is nested by area after #610
for r in dl.parse_file(f):
if r.key:
keys.add(r.key)
return keys
def _budget_total() -> int:
"""Lines of PROSE in the active corpus — YAML frontmatter excluded.
The budget exists to bound how much narrative a reader/agent must get through, so it counts
prose, not metadata. Under the #610 split each record carries ~11 frontmatter lines plus two
fences (1789 lines across 166 records), which are the structured restatement of what used to be
one dense backtick line — counting them would inflate the metric without any new knowledge
being added.
NOTE this is a change of METRIC, not a consolidation: it re-measures the same corpus, it does
not shrink it. At the time of the split, whole-file counting put the corpus a few hundred lines
over budget; prose-only puts the same content comfortably under. The consolidation work is still
worth doing — it is simply no longer being signalled by a warning that was partly measuring
punctuation. Pre-migration this function is inert: no legacy file has frontmatter.
"""
total = 0
for f in dl.active_files():
if not f.exists():
continue
text = f.read_text(encoding="utf-8")
lines = text.splitlines()
if dl.has_frontmatter(text):
end = next((i for i, ln in enumerate(lines[1:], start=1) if ln.rstrip() == "---"), None)
if end is not None:
lines = lines[end + 1 :]
total += len(lines)
cat = dl.TOPIC_DIR / "README.md"
if cat.exists():
total += len(cat.read_text(encoding="utf-8").splitlines())
return total
def _budget_ok(limit: int) -> bool:
return _budget_total() <= limit
def _catalog_ok() -> bool:
try:
import scripts.build_decisions_catalog as bc # pyright: ignore[reportMissingImports]
except Exception:
return True # fail-open; the CI job also runs an independent --check step
want = bc.render_catalog(dl.all_active_records())
cat = dl.TOPIC_DIR / "README.md"
have = cat.read_text(encoding="utf-8") if cat.exists() else ""
return want.strip() == have.strip()
def main(argv=None) -> int:
ap = argparse.ArgumentParser()
ap.add_argument("--base")
ap.add_argument("--head")
# Aggregate active-corpus budget. Re-baselined 2026-07-21 (#542): 4800 -> 5600. The corpus grew
# by ~530 lines because 31 workflow/CI/review rules were MOVED into it from
# docs/handoffs/chicorytv-issue-queue.md, which was their only copy. That is the corpus doing its
# job, not drift — the knowledge existed already, it just wasn't retrievable. Do not raise this
# again to accommodate genuinely new records; that is what the warning is for.
ap.add_argument("--budget", type=int, default=5600)
args = ap.parse_args(argv)
records = dl.all_active_records()
archive_records = []
if dl.ARCHIVE_DIR.exists():
for f in dl.ARCHIVE_DIR.rglob("*.md"): # rglob: archive is nested by area after #610
archive_records += dl.parse_file(f)
removed, rewritten, demoted = _diff_findings(args.base, args.head) if args.base and args.head else ([], [], [])
budget_ok = _budget_ok(args.budget)
errs = validate(
records,
archive_keys=_archive_keys(),
catalog_ok=_catalog_ok(),
budget_ok=budget_ok,
removed=removed,
rewritten=rewritten,
archive_records=archive_records,
demoted=demoted,
wing_faults=record_wing_faults(),
)
if not budget_ok:
print(
f"::warning::decisions-validate: aggregate active-corpus is {_budget_total()} lines "
f"(budget {args.budget}) — schedule a consolidation",
file=sys.stderr,
)
stale = stale_records(records, date.today())
if stale:
listed = "; ".join(f"{h} (stale-after {d})" for h, d in stale)
print(
f"::notice::decisions-validate: {len(stale)} active record(s) past their stale-after date "
f"— re-confirm or extend: {listed}",
file=sys.stderr,
)
unmigrated = [r for r in records if r.status == "legacy-unmigrated" and r.heading not in SKIP_HEADINGS]
if unmigrated:
print(f"::notice::{len(unmigrated)} legacy-unmigrated decision record(s) remain (must trend to 0).")
if errs:
for e in errs:
print(f"decisions-validate: {e}", file=sys.stderr)
return 1
print("decisions-validate: OK")
return 0
if __name__ == "__main__":
raise SystemExit(main())