Compare commits

..
Author SHA1 Message Date
mos-dt-0andClaude Opus 5 1f4c7c0d37 fix(hygiene): .prettierignore must exclude Python build/test artifacts
ci/woodpecker/pr/ci Pipeline was successful
Prettier had no excludes for venv/__pycache__/.mypy_cache/.pytest_cache/htmlcov,
so any local Python virtualenv in the tree drops thousands of third-party files
into `pnpm format:check` and makes the gate unpassable in a working checkout
(observed: ~2400 files from an untracked apps/coordinator venv).

Same category as the existing node_modules/dist/.next entries. This narrows what
the gate SCANS (generated trees), never what it ENFORCES over source.

Found while making the three gates pass honestly for the remediation mission's
TASK-0, rather than reaching for --no-verify.

Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
2026-07-31 17:04:41 -05:00
mos-dt-0andMos 06e0d40352 feat(quality): CI test-membership guard — enumeration can no longer silently under-run the disk (#1017) (#1018)
ci/woodpecker/push/publish Pipeline was successful
ci/woodpecker/push/ci Pipeline was successful
Co-authored-by: mos-dt-0 <[email protected]>
2026-07-31 14:08:41 +00:00
Mos 166ee8c90f fix(wake): close fd 9 in the detector's sleep child so a dead detector's lock dies with it (#993)
ci/woodpecker/push/publish Pipeline was successful
ci/woodpecker/push/ci Pipeline was successful
2026-07-31 13:48:33 +00:00
11 changed files with 429 additions and 149 deletions
+8
View File
@@ -4,6 +4,14 @@ pnpm-lock.yaml
**/node_modules
**/drizzle
**/.next
# Python build/test artifacts — same category as node_modules/dist/.next above.
# Prettier must never scan generated trees; without these a local venv poisons
# `pnpm format:check` with thousands of third-party files.
**/venv
**/__pycache__
**/.mypy_cache
**/.pytest_cache
**/htmlcov
.claude/
docs/tess/TASKS.md
docs/scratchpads/
+5
View File
@@ -41,6 +41,11 @@ steps:
# (Constitution + dispatcher + each RUNTIME.md slice). See DESIGN §7 / R9.
- bash packages/mosaic/framework/tools/quality/scripts/check-resident-budget.sh --self-test
- bash packages/mosaic/framework/tools/quality/scripts/check-resident-budget.sh
# Test-membership guard (#1017): also first link of test:framework-shell.
# Invoked from BOTH surfaces it audits (F2, PR #1018) — the guard is link
# [0] of the pnpm chain, so severing that chain would silence it together
# with everything it guards; this direct line keeps one instrument running.
- bash packages/mosaic/framework/tools/quality/scripts/check-test-enumeration.sh
# Blocking gate (#791): a framework upgrade must never write or delete an
# operator-owned path. The HARD GATE proves an unanticipated operator sentinel
@@ -254,32 +254,15 @@ from urllib.parse import urlparse
def _origin_and_path(url):
# Normalize a URL to (scheme-class, host, distinguishing-port) + comment path.
#
# #991: http and https collapse into ONE scheme class ("web"). A Gitea whose
# ROOT_URL is configured http:// returns http:// object URLs even when every
# client reaches it over https://, so a scheme-strict comparison rejects the
# provider's own correct answer about a write that landed — a deterministic
# false negative on every comment posted against such a deployment. The
# scheme is also not what this check defends: the forgeries it exists to
# catch (look-alike host, decoy path prefix, wrong owner/repo/number) all
# vary the HOST or the PATH, both of which stay strict below. Any OTHER
# scheme (file:, ftp:, javascript:) remains distinguishing and is rejected.
#
# Port: an implicit port and its own scheme's default compare equal, so
# http://h == https://h. An EXPLICIT non-default port still distinguishes,
# because a different port is a different service on the same host.
# Normalize a URL to (scheme, host, effective-port) + comment path. The port
# defaults to the scheme's default (80 http / 443 otherwise) so an implicit
# port and its explicit default form compare equal.
parsed = urlparse(url or "")
scheme = (parsed.scheme or "").lower()
host = (parsed.hostname or "").lower()
if scheme in ("http", "https"):
scheme_class = "web"
default_port = 80 if scheme == "http" else 443
port = None if parsed.port in (None, default_port) else parsed.port
else:
scheme_class = scheme
port = parsed.port
return (scheme_class, host, port), parsed.path.rstrip("/")
default_port = 80 if scheme == "http" else 443
port = parsed.port if parsed.port is not None else default_port
return (scheme, host, port), parsed.path.rstrip("/")
try:
@@ -243,35 +243,15 @@ from urllib.parse import urlparse
def _origin_and_path(url):
# Normalize a URL to (scheme-class, host, distinguishing-port) + comment path.
#
# #991: http and https collapse into ONE scheme class ("web"). A Gitea whose
# ROOT_URL is configured http:// returns http:// object URLs even when every
# client reaches it over https://, so a scheme-strict comparison rejects the
# provider's own correct answer about a comment that landed — a deterministic
# false negative on EVERY review comment posted against such a deployment.
# That matters more here than anywhere else: on a host where no seat can
# create a review OBJECT, the comment-form review record this path produces
# is the only gate-16 evidence available, and this check refuses all of it.
# The scheme is also not what the check defends: the forgeries it exists to
# catch (look-alike host, decoy path prefix, wrong owner/repo/kind/number)
# all vary the HOST or the PATH, both of which stay strict below. Any OTHER
# scheme (file:, ftp:, javascript:) remains distinguishing and is rejected.
#
# Port: an implicit port and its own scheme's default compare equal, so
# http://h == https://h. An EXPLICIT non-default port still distinguishes,
# because a different port is a different service on the same host.
# Normalize a URL to (scheme, host, effective-port) + comment path. The port
# defaults to the scheme's default (80 http / 443 otherwise) so an implicit
# port and its explicit default form compare equal.
parsed = urlparse(url or "")
scheme = (parsed.scheme or "").lower()
host = (parsed.hostname or "").lower()
if scheme in ("http", "https"):
scheme_class = "web"
default_port = 80 if scheme == "http" else 443
port = None if parsed.port in (None, default_port) else parsed.port
else:
scheme_class = scheme
port = parsed.port
return (scheme_class, host, port), parsed.path.rstrip("/")
default_port = 80 if scheme == "http" else 443
port = parsed.port if parsed.port is not None else default_port
return (scheme, host, port), parsed.path.rstrip("/")
try:
@@ -41,13 +41,7 @@
# that other host's credential cross-host;
# 10. leaves NO temp files behind (POST/GET bodies + metadata) on either the
# success or the failure path — nested function-scoped RETURN traps do not
# clobber each other and every scratch file is removed on all exit paths;
# 11. (#991) ACCEPTS a record whose provider-returned URL differs from the repo
# remote ONLY in http-vs-https — the standing state on a deployment whose
# Gitea ROOT_URL is misconfigured — while still REJECTING a different
# explicit port and a non-web scheme. Every prior fixture here was https://
# and every negative varied only host or path, so the one axis that fails
# in production had zero coverage: the fixtures encoded the assumption.
# clobber each other and every scratch file is removed on all exit paths.
set -euo pipefail
@@ -67,29 +61,15 @@ STATE_FILE="$WORK_DIR/comments.json"
# A dedicated scratch dir the wrapper is pointed at via TMPDIR, so the leak
# check can assert every POST/GET body + metadata temp file is cleaned up.
TMP_SCRATCH="$WORK_DIR/scratch"
# #1007: a sandboxed HOME. detect-platform.sh's step-0 per-agent identity lookup
# reads ~/.config/mosaic/gitea-tokens/<identity>, which is OUTSIDE both
# XDG_CONFIG_HOME and MOSAIC_CREDENTIALS_FILE — so on any seat that has a real
# per-agent token the suite resolves a PRODUCTION credential and dies at the
# authenticated-identity read (HTTP 401) before reaching case 1. Pointing HOME
# at the work dir keeps that lookup inside the sandbox.
HOME_DIR="$WORK_DIR/home"
cleanup() {
rm -rf "$WORK_DIR"
}
trap cleanup EXIT
mkdir -p "$REPO_DIR" "$BIN_DIR" "$XDG_DIR" "$TMP_SCRATCH" "$HOME_DIR"
mkdir -p "$REPO_DIR" "$BIN_DIR" "$XDG_DIR" "$TMP_SCRATCH"
git -C "$REPO_DIR" init -q
git -C "$REPO_DIR" remote add origin https://git.mosaicstack.dev/mosaicstack/stack.git
# #1007: shadow any GLOBAL `mosaic.gitIdentity` with an empty REPO-LOCAL value.
# A repo-local key shadows the global and reads back empty at rc=0, which is the
# only way to make step 0 fall through inside a sandbox. Note the env-var route
# (MOSAIC_GIT_IDENTITY="") does NOT work: detect-platform.sh reads it with
# `${MOSAIC_GIT_IDENTITY:-}`, and `:-` treats set-but-empty identically to
# unset, so setting it empty is silently the same as not setting it at all.
git -C "$REPO_DIR" config mosaic.gitIdentity ""
ISSUE_NUMBER=7
REPO_SLUG="mosaicstack/stack"
@@ -286,19 +266,6 @@ elif mode == "url-wrong-repo":
elif mode == "url-suffix-injection":
# Prefix-injected: a bare endswith("/<slug>/issues/7") test would ACCEPT this.
issue_url = f"https://git.mosaicstack.dev/deceptive/{repo}/issues/7"
elif mode == "url-scheme-downgrade":
# #991, and the ONLY one of these modes that must be ACCEPTED. A Gitea whose
# ROOT_URL is http:// returns http:// object URLs for a repo cloned over
# https://. Same host, same path, correct record — the provider is telling
# the truth about a write that landed.
issue_url = f"http://git.mosaicstack.dev/{repo}/issues/7"
elif mode == "url-wrong-port":
# An EXPLICIT non-default port is a different service on the same host and
# must stay distinguishing — collapsing the scheme must not collapse this.
issue_url = f"https://git.mosaicstack.dev:8443/{repo}/issues/7"
elif mode == "url-non-web-scheme":
# Only http/https collapse. Any other scheme stays distinguishing.
issue_url = f"ftp://git.mosaicstack.dev/{repo}/issues/7"
record = {
"id": new_id,
"body": body,
@@ -398,7 +365,6 @@ run_comment() {
(
cd "$REPO_DIR"
PATH="$BIN_DIR:$PATH" \
HOME="$HOME_DIR" \
TMPDIR="$TMP_SCRATCH" \
XDG_CONFIG_HOME="$XDG_DIR" \
MOSAIC_CREDENTIALS_FILE="$CREDENTIALS_FILE" \
@@ -579,17 +545,13 @@ if grep -q " $ACTING_LOGIN\$" "$AUTH_LOG"; then
fi
assert_no_temp_leak "cross-host"
# Cases 7-12 (#865 Blocker 3): the created record's id/author/body are all
# correct, but its provider-returned issue_url does not belong to this issue on
# this provider/repo. Verification pins the URL's ORIGIN (scheme-class + host +
# explicit non-default port) and its FULL path (deployment prefix + exact
# owner/repo + kind + number), so each must FAIL CLOSED. A bare endswith/suffix
# test would wrongly accept the look-alike-host and prefix-injection variants.
# url-wrong-port and url-non-web-scheme (#991) bound the scheme relaxation from
# the other side: collapsing http/https must not also collapse a different port
# or a different scheme family.
for bad_mode in url-wrong-host url-wrong-owner url-wrong-repo url-suffix-injection \
url-wrong-port url-non-web-scheme; do
# Cases 7-10 (#865 Blocker 3): the created record's id/author/body are all
# correct, but its provider-returned issue_url is forged. Verification pins the
# URL's ORIGIN (scheme+host+effective-port) and its FULL path (deployment prefix
# + exact owner/repo + kind + number), so each forgery must FAIL CLOSED. A bare
# endswith/suffix test would wrongly accept the look-alike-host and
# prefix-injection variants.
for bad_mode in url-wrong-host url-wrong-owner url-wrong-repo url-suffix-injection; do
if run_comment "$bad_mode"; then
echo "FAIL: forged comment URL ($bad_mode) was accepted" >&2
cat "$OUTPUT_FILE" >&2
@@ -606,19 +568,4 @@ done
# issue_url (already exercised by Case 1's fresh-success), so the tightened check
# is not rejecting genuine writes.
# Case 13 (#991): the deployment's ROOT_URL is http:// while the repo remote is
# https://, so the provider returns an http:// issue_url for a record that is
# otherwise entirely correct. This is not a forgery — it is the provider's own
# truthful answer about a write that landed — and a scheme-strict comparison
# rejects it deterministically, on EVERY comment, converting a successful write
# into a reported failure. It must be ACCEPTED. Host, path, owner, repo, kind
# and number all remain strict; only the http/https distinction is relaxed.
run_comment url-scheme-downgrade
grep -q 'Added and verified comment on Gitea issue #7 (comment ID 1)' "$OUTPUT_FILE"
# The write really happened and was read back by exact id — this case passes
# through the same POST/GET chain as case 1, not around it.
grep -q "^POST $API_BASE/issues/7/comments$" "$CURL_LOG"
grep -q "^GET $API_BASE/issues/comments/1$" "$CURL_LOG"
assert_no_temp_leak "url-scheme-downgrade"
echo "issue-comment.sh REST create + exact-id read-back regression passed"
@@ -436,19 +436,6 @@ elif mode == "comment-url-wrong-repo":
elif mode == "comment-url-suffix-injection":
# Prefix-injected: a bare endswith("/<slug>/pulls/123") test would ACCEPT it.
pr_url = f"{_origin}/deceptive{_slug}/pulls/123"
elif mode == "comment-url-wrong-port":
# #991 bound: an EXPLICIT non-default port is a different service on the same
# host. Relaxing http-vs-https must NOT relax this.
pr_url = f"{_p.scheme}://{_p.hostname}:8443{_slug}/pulls/123"
elif mode == "comment-url-non-web-scheme":
# #991 bound: ONLY http/https collapse; any other scheme stays distinguishing.
pr_url = f"ftp://{_p.netloc}{_slug}/pulls/123"
elif mode == "comment-url-scheme-downgrade":
# #991, and the only URL mode here that must be ACCEPTED. A Gitea whose
# ROOT_URL is http:// returns http:// object URLs for a repo reached over
# https://. Same host, same path, correct record — a truthful provider
# answer about a comment that landed, not a forgery.
pr_url = f"http://{_p.netloc}{_slug}/pulls/123"
elif mode == "comment-mixed-case-slug":
# #875: EXPECTED_REPO_SLUG is taken verbatim from GITEA_API_BASE and can be
# mixed-case (e.g. "USC/uconnect"), but Gitea canonicalizes the returned
@@ -903,16 +890,11 @@ fi
assert_no_temp_leak "review-body-reuse"
# Cases 12-15 (#865 Blocker 3): a PR comment whose id/author/body are all correct
# but whose provider-returned pull_request_url does not belong to this PR must
# FAIL CLOSED. Verification pins the URL's ORIGIN (scheme-class + host + explicit
# non-default port) and FULL path (deployment prefix + exact owner/repo + kind +
# number); a bare endswith/suffix test would wrongly accept the look-alike-host
# and prefix-injection variants. comment-url-wrong-port and
# comment-url-non-web-scheme (#991) bound the scheme relaxation from the other
# side: collapsing http/https must not also collapse a different port or a
# different scheme family.
for bad_mode in comment-url-wrong-host comment-url-wrong-owner comment-url-wrong-repo \
comment-url-suffix-injection comment-url-wrong-port comment-url-non-web-scheme; do
# but whose provider-returned pull_request_url is forged must FAIL CLOSED.
# Verification pins the URL's ORIGIN (scheme+host+effective-port) and FULL path
# (deployment prefix + exact owner/repo + kind + number); a bare endswith/suffix
# test would wrongly accept the look-alike-host and prefix-injection variants.
for bad_mode in comment-url-wrong-host comment-url-wrong-owner comment-url-wrong-repo comment-url-suffix-injection; do
if run_review "$bad_mode" comment durable-body; then
echo "FAIL: forged comment URL ($bad_mode) was accepted" >&2
cat "$OUTPUT_FILE" >&2
@@ -938,19 +920,6 @@ run_review comment-mixed-case-slug comment durable-body https://git.mosaicstack.
grep -q 'Added and verified comment on Gitea PR #123' "$OUTPUT_FILE"
assert_no_temp_leak "comment-mixed-case-slug"
# Case 15c (#991): the deployment's Gitea ROOT_URL is http:// while every client
# reaches it over https://, so the provider returns an http:// pull_request_url
# for a comment that is otherwise entirely correct. Same class as 15b — a
# legitimate provider response, not a spoof — and a scheme-strict compare
# rejects it on EVERY comment, deterministically. That is not a cosmetic false
# negative here: on a host where no seat can create a review OBJECT, this
# comment-form record is the only gate-16 evidence obtainable, and the wrapper
# refuses all of it while the comment sits durably on the PR. Host, path, owner,
# repo, kind and number stay strict; only http-vs-https is relaxed.
run_review comment-url-scheme-downgrade comment durable-body
grep -q 'Added and verified comment on Gitea PR #123' "$OUTPUT_FILE"
assert_no_temp_leak "comment-url-scheme-downgrade"
# Case 16 (#865 ITEM 1, current-head TOCTOU): the PR head advances between the
# pre-submit head read (which pins the review) and the post-verify re-read. The
# review is genuinely created and verified as pinned to the OLD head, but the
@@ -0,0 +1,165 @@
#!/usr/bin/env bash
# check-test-enumeration.sh — CI test-membership guard (#1017).
#
# CI reaches shell suites through two hand-enumerated surfaces:
# S1 packages/mosaic/package.json scripts."test:framework-shell"
# S2 .woodpecker/ci.yml direct `bash packages/mosaic/framework/tools/...` commands
#
# A hand-enumerated allowlist re-arms its own gap: a new suite never auto-joins,
# so the list silently under-runs the disk (17 of 39 suites were invisible when
# #1017 was filed). This guard makes that under-run impossible to do silently:
#
# FAIL when a suite-shaped file exists on disk and is neither enumerated on
# the UNION of both surfaces nor listed in the exclusions file.
# ("Enumerated", deliberately — F1/F2 on PR #1018 proved this guard sees
# NAMING, not reachability, and its words must not claim otherwise.)
# FAIL when either surface names a path that does not exist on disk
# (a rename manufactures a stale entry silently — checked BOTH directions).
# FAIL when an exclusion entry has no reason, names a path that is gone,
# names a path that is also enumerated (contradiction), or names a path
# outside the population (dead weight that looks like coverage).
#
# POPULATION PATTERN — a deliberate decision, stated per #1017's record:
# basename matches *test*.sh (contains "test", ends ".sh"). Deliberately BROAD:
# the strict `test-*.sh` prefix cannot even name three real boundary files
# (tmux/agent-send.test.sh — CI-run; orchestrator/smoke-test.sh;
# wake/validate-973/microtest-wake-assert.sh), and three independent censuses
# handled that last file three different ways with no trace of the judgement.
# The broad pattern makes such files MEMBERS, so their disposition must be a
# signed exclusion, not an accident of the glob. The SAME pattern is applied to
# both sides of the comparison (disk and enumeration) — a comparison globbed two
# ways runs on two different populations. Scripts outside the pattern on both
# sides symmetrically (e.g. check-resident-budget.sh, verify-sanitized.sh) are
# check-scripts, not suites; their existence is still verified via the
# both-directions rule because every surface-named path must exist on disk.
#
# The surfaces are PARSED, never line-ranged: three seats independently
# mis-scoped hand-written line ranges against these files (#1017 thread). S1 is
# read via JSON + command-chain tokenization; S2 by extracting every
# packages/mosaic/framework/tools/ token wherever it appears in the file.
#
# Exclusions file format (framework/tools/quality/test-enumeration-exclusions.txt):
# <repo-relative-path> | <non-empty reason>
# Lines starting with # and blank lines are ignored. An exclusion is a recorded
# decision someone signed, not an omission nobody made.
set -uo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
ROOT="$(cd "$SCRIPT_DIR/../../../../../.." && pwd)"
while (( $# )); do
case "$1" in
--root) ROOT="$(cd "$2" && pwd)"; shift 2 ;;
*) echo "usage: check-test-enumeration.sh [--root <repo-root>]" >&2; exit 2 ;;
esac
done
PKG_JSON="$ROOT/packages/mosaic/package.json"
CI_YML="$ROOT/.woodpecker/ci.yml"
TOOLS_DIR="$ROOT/packages/mosaic/framework/tools"
EXCLUSIONS="$TOOLS_DIR/quality/test-enumeration-exclusions.txt"
for f in "$PKG_JSON" "$CI_YML"; do
[[ -f "$f" ]] || { echo "FAIL: required surface file missing: $f" >&2; exit 2; }
done
[[ -d "$TOOLS_DIR" ]] || { echo "FAIL: tools dir missing: $TOOLS_DIR" >&2; exit 2; }
fail_count=0
fail() { printf 'FAIL %s\n' "$1"; fail_count=$(( fail_count + 1 )); }
# in_population <repo-relative path> — the single pattern, used for BOTH sides.
in_population() {
local base; base="$(basename "$1")"
[[ "$base" == *test*.sh ]]
}
# --- Surface 1: package.json test:framework-shell, parsed, repo-relative -----
# Tokens are script paths iff they contain "/" and end .sh/.py; interpreter
# names and flags are skipped. Paths are relative to packages/mosaic/.
mapfile -t S1 < <(python3 - "$PKG_JSON" <<'PY'
import json, shlex, sys
cmd = json.load(open(sys.argv[1]))["scripts"].get("test:framework-shell", "")
seen = []
for seg in cmd.split("&&"):
for tok in shlex.split(seg):
if "/" in tok and (tok.endswith(".sh") or tok.endswith(".py")):
path = "packages/mosaic/" + tok
if path not in seen:
seen.append(path)
print("\n".join(seen))
PY
)
# --- Surface 2: ci.yml, every framework/tools token wherever it appears ------
# Comment lines (first non-whitespace char is #) are skipped BEFORE matching:
# commenting an invocation out is the most common way a suite actually gets
# disabled, and a raw-text regex would keep calling it enumerated (F1, 20155 on
# PR #1018 — demonstrated, not argued). Known residual limit: a path named only
# in a TRAILING comment on a live line still matches; no such line exists today
# and full fidelity would need a YAML parser the CI image does not ship.
mapfile -t S2 < <(grep -vE '^[[:space:]]*#' "$CI_YML" \
| grep -oE 'packages/mosaic/framework/tools/[A-Za-z0-9_./-]+\.(sh|py)' | sort -u)
# --- Union, and its population-restricted view -------------------------------
declare -A ENUM=() ENUM_POP=()
for p in "${S1[@]:-}" "${S2[@]:-}"; do
[[ -n "$p" ]] || continue
ENUM["$p"]=1
in_population "$p" && ENUM_POP["$p"]=1
done
# --- Direction B: every surface-named path must exist on disk ----------------
for p in "${!ENUM[@]}"; do
[[ -f "$ROOT/$p" ]] || fail "STALE ENUMERATION: surfaces name '$p' but it does not exist on disk"
done
# --- Exclusions: parsed with the same rigor the enumeration gets -------------
declare -A EXCLUDED=()
if [[ -f "$EXCLUSIONS" ]]; then
lineno=0
while IFS= read -r line; do
lineno=$(( lineno + 1 ))
[[ "$line" =~ ^[[:space:]]*(#|$) ]] && continue
path="${line%%|*}"; reason="${line#*|}"
path="$(echo "$path" | sed 's/^[[:space:]]*//; s/[[:space:]]*$//')"
reason="$(echo "$reason" | sed 's/^[[:space:]]*//; s/[[:space:]]*$//')"
if [[ "$line" != *"|"* || -z "$reason" ]]; then
fail "EXCLUSION MISSING REASON: line $lineno ('$path') — an exclusion is a recorded decision someone signed"
continue
fi
if [[ ! -f "$ROOT/$path" ]]; then
fail "STALE EXCLUSION: line $lineno excludes '$path' which does not exist on disk"
continue
fi
if ! in_population "$path"; then
fail "EXCLUSION OUTSIDE POPULATION: line $lineno excludes '$path' which the population pattern does not name — dead weight that reads as coverage"
continue
fi
if [[ -n "${ENUM[$path]:-}" ]]; then
fail "CONTRADICTORY EXCLUSION: line $lineno excludes '$path' which the surfaces already enumerate"
continue
fi
EXCLUDED["$path"]=1
done < "$EXCLUSIONS"
fi
# --- Direction A: disk population must be enumerated or signed-excluded ------
disk_total=0
unlisted=0
while IFS= read -r f; do
rel="${f#"$ROOT"/}"
in_population "$rel" || continue
disk_total=$(( disk_total + 1 ))
if [[ -z "${ENUM_POP[$rel]:-}" && -z "${EXCLUDED[$rel]:-}" ]]; then
fail "UNENUMERATED: '$rel' exists on disk but is neither enumerated on any CI surface nor signed in the exclusions file"
unlisted=$(( unlisted + 1 ))
fi
done < <(find "$TOOLS_DIR" -type f -name '*.sh' | sort)
if (( fail_count > 0 )); then
printf 'enumeration guard: %d failure(s) — population %d, enumerated (in-population) %d, excluded %d\n' \
"$fail_count" "$disk_total" "${#ENUM_POP[@]}" "${#EXCLUDED[@]}"
exit 1
fi
printf 'enumeration guard: OK — population %d, enumerated (in-population) %d, excluded (signed) %d, surfaces name %d path(s), all present on disk\n' \
"$disk_total" "${#ENUM_POP[@]}" "${#EXCLUDED[@]}" "${#ENUM[@]}"
@@ -0,0 +1,166 @@
#!/usr/bin/env bash
# test-check-test-enumeration.sh — needles for the enumeration guard (#1017).
#
# Every failure mode the guard promises gets BOTH polarities:
# NEEDLE a fixture that MUST trip the guard, asserted on the guard's OWN
# words (--out) — exit 1 alone cannot distinguish "caught the rogue
# file" from "choked on the fixture".
# CONTROL a fixture that MUST pass. A guard that failed unconditionally
# would satisfy every needle here — the null-case defect the guard's
# own subject matter (#1017) exists to make impossible.
#
# The needles encode the specific errors that produced #1017's thread:
# n6 is the 20124 boundary file (a suite the strict prefix cannot name);
# n2b proves surface 2 is PARSED, not line-ranged (three seats mis-scoped
# hand-written ranges against ci.yml);
# n5/n7 keep the exclusions file honest so it cannot become the next silent cap;
# n8/c4 are F1 (20155): a commented-out ci.yml line is NOT enumeration —
# commenting-out is the most common way a suite actually gets disabled,
# and it must fail loud in one direction without false-staling the other.
set -uo pipefail
HERE="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
GUARD="$HERE/check-test-enumeration.sh"
TMP="$(mktemp -d)"; trap 'rm -rf "$TMP"' EXIT
PASS=0; FAIL=0
# fixture <name> — a minimal repo root the guard accepts via --root:
# one suite enumerated on S1 (plus a naming-outlier suite, so the S1 parser's
# handling of non-prefix names is always exercised), one on S2, one check-script
# named on S2 that is outside the population, and an empty exclusions file.
fixture() {
local r="$TMP/$1"
mkdir -p "$r/packages/mosaic/framework/tools/git" \
"$r/packages/mosaic/framework/tools/tmux" \
"$r/packages/mosaic/framework/tools/quality/scripts" \
"$r/.woodpecker"
printf '#!/usr/bin/env bash\nexit 0\n' > "$r/packages/mosaic/framework/tools/git/test-a.sh"
printf '#!/usr/bin/env bash\nexit 0\n' > "$r/packages/mosaic/framework/tools/tmux/outlier.test.sh"
printf '#!/usr/bin/env bash\nexit 0\n' > "$r/packages/mosaic/framework/tools/quality/scripts/test-ci.sh"
printf '#!/usr/bin/env bash\nexit 0\n' > "$r/packages/mosaic/framework/tools/quality/scripts/verify-thing.sh"
cat > "$r/packages/mosaic/package.json" <<'JSON'
{"scripts": {"test:framework-shell": "bash framework/tools/git/test-a.sh && bash framework/tools/tmux/outlier.test.sh"}}
JSON
cat > "$r/.woodpecker/ci.yml" <<'YML'
steps:
sanitize:
commands:
- bash packages/mosaic/framework/tools/quality/scripts/verify-thing.sh
guard:
commands:
- bash packages/mosaic/framework/tools/quality/scripts/test-ci.sh
YML
: > "$r/packages/mosaic/framework/tools/quality/test-enumeration-exclusions.txt"
printf '%s' "$r"
}
# expect <kind> <want-exit> <desc> [--out <substring>] -- <root>
expect() {
local kind="$1" want="$2" desc="$3"; shift 3
local need_out=""
while (( $# )); do
case "$1" in
--out) need_out="$2"; shift 2 ;;
--) shift; break ;;
esac
done
local root="$1" got=0 out
out="$(bash "$GUARD" --root "$root" 2>&1)" || got=$?
local why=""
[[ "$got" == "$want" ]] || why="wanted exit $want, got $got"
if [[ -z "$why" && -n "$need_out" && "$out" != *"$need_out"* ]]; then
why="exit $got as expected, but output never said: $need_out"
fi
if [[ -z "$why" ]]; then
printf ' PASS [%-7s] %s (exit %s)\n' "$kind" "$desc" "$got"
PASS=$(( PASS + 1 ))
else
printf ' FAIL [%-7s] %s — %s\n' "$kind" "$desc" "$why"
printf '%s\n' "$out" | sed 's/^/ | /'
FAIL=$(( FAIL + 1 ))
fi
}
excl() { printf '%s\n' "$2" >> "$1/packages/mosaic/framework/tools/quality/test-enumeration-exclusions.txt"; }
echo "=== c1: a fully consistent fixture passes ==="
R="$(fixture c1)"
expect CONTROL 0 "consistent tree: both surfaces enumerated, nothing unlisted" \
--out "enumeration guard: OK" -- "$R"
echo "=== n1: an on-disk suite reachable from no surface must fail ==="
R="$(fixture n1)"
printf '#!/usr/bin/env bash\nexit 0\n' > "$R/packages/mosaic/framework/tools/git/test-rogue.sh"
expect NEEDLE 1 "unlisted suite is named in the failure" \
--out "UNENUMERATED: 'packages/mosaic/framework/tools/git/test-rogue.sh'" -- "$R"
echo "=== n6: the 20124 boundary file — a suite the strict prefix cannot name ==="
R="$(fixture n6)"
printf '#!/usr/bin/env bash\nexit 0\n' > "$R/packages/mosaic/framework/tools/git/rogue.test.sh"
expect NEEDLE 1 "naming-outlier suite (*.test.sh) is a population member, not invisible" \
--out "UNENUMERATED: 'packages/mosaic/framework/tools/git/rogue.test.sh'" -- "$R"
echo "=== c3: a non-suite script outside the pattern is outside it on BOTH sides ==="
R="$(fixture c3)"
printf '#!/usr/bin/env bash\nexit 0\n' > "$R/packages/mosaic/framework/tools/git/check-unrelated.sh"
expect CONTROL 0 "check-script on disk, unlisted, outside population: not the guard's business" \
--out "enumeration guard: OK" -- "$R"
echo "=== n2/n2b: a surface naming a path absent from disk must fail — both surfaces ==="
R="$(fixture n2)"
rm "$R/packages/mosaic/framework/tools/git/test-a.sh"
expect NEEDLE 1 "S1 (package.json) stale entry" \
--out "STALE ENUMERATION: surfaces name 'packages/mosaic/framework/tools/git/test-a.sh'" -- "$R"
R="$(fixture n2b)"
rm "$R/packages/mosaic/framework/tools/quality/scripts/test-ci.sh"
expect NEEDLE 1 "S2 (ci.yml) stale entry — proves ci.yml is parsed, not line-ranged" \
--out "STALE ENUMERATION: surfaces name 'packages/mosaic/framework/tools/quality/scripts/test-ci.sh'" -- "$R"
echo "=== c2: a rogue suite with a SIGNED exclusion passes, and is counted ==="
R="$(fixture c2)"
printf '#!/usr/bin/env bash\nexit 0\n' > "$R/packages/mosaic/framework/tools/git/test-rogue.sh"
excl "$R" "packages/mosaic/framework/tools/git/test-rogue.sh | non-hermetic pending fixture work (needle-suite specimen)"
expect CONTROL 0 "signed exclusion is honoured and visible in the summary" \
--out "excluded (signed) 1" -- "$R"
echo "=== n3: an exclusion with no reason is not a decision ==="
R="$(fixture n3)"
printf '#!/usr/bin/env bash\nexit 0\n' > "$R/packages/mosaic/framework/tools/git/test-rogue.sh"
excl "$R" "packages/mosaic/framework/tools/git/test-rogue.sh | "
expect NEEDLE 1 "empty reason rejected" --out "EXCLUSION MISSING REASON" -- "$R"
R="$(fixture n3b)"
printf '#!/usr/bin/env bash\nexit 0\n' > "$R/packages/mosaic/framework/tools/git/test-rogue.sh"
excl "$R" "packages/mosaic/framework/tools/git/test-rogue.sh"
expect NEEDLE 1 "missing separator rejected (the path alone is not a signature)" \
--out "EXCLUSION MISSING REASON" -- "$R"
echo "=== n4: an exclusion whose path is gone is stale, not satisfied ==="
R="$(fixture n4)"
excl "$R" "packages/mosaic/framework/tools/git/test-vanished.sh | was excluded once, then deleted"
expect NEEDLE 1 "stale exclusion rejected" --out "STALE EXCLUSION" -- "$R"
echo "=== n5: excluding an enumerated suite is a contradiction, not belt-and-braces ==="
R="$(fixture n5)"
excl "$R" "packages/mosaic/framework/tools/git/test-a.sh | already in CI but excluded anyway"
expect NEEDLE 1 "contradictory exclusion rejected" --out "CONTRADICTORY EXCLUSION" -- "$R"
echo "=== n8/c4: a commented-out ci.yml line is not enumeration (F1, 20155) ==="
R="$(fixture n8)"
printf '#!/usr/bin/env bash\nexit 0\n' > "$R/packages/mosaic/framework/tools/git/test-disabled.sh"
printf ' # - bash packages/mosaic/framework/tools/git/test-disabled.sh\n' >> "$R/.woodpecker/ci.yml"
expect NEEDLE 1 "suite named only in a commented-out invocation is UNENUMERATED" \
--out "UNENUMERATED: 'packages/mosaic/framework/tools/git/test-disabled.sh'" -- "$R"
R="$(fixture c4)"
printf ' # - bash packages/mosaic/framework/tools/git/test-vanished.sh\n' >> "$R/.woodpecker/ci.yml"
expect CONTROL 0 "comment naming an absent path raises no false stale-enumeration" \
--out "enumeration guard: OK" -- "$R"
echo "=== n7: excluding a file outside the population is dead weight, not coverage ==="
R="$(fixture n7)"
excl "$R" "packages/mosaic/framework/tools/quality/scripts/verify-thing.sh | not a suite but signing it anyway"
expect NEEDLE 1 "out-of-population exclusion rejected" --out "EXCLUSION OUTSIDE POPULATION" -- "$R"
echo
printf 'enumeration-guard needles: %d passed, %d failed\n' "$PASS" "$FAIL"
(( FAIL == 0 ))
@@ -0,0 +1,45 @@
# test-enumeration-exclusions.txt — signed exclusions for check-test-enumeration.sh (#1017).
#
# Every entry is a recorded decision: a suite-shaped file that exists on disk,
# is NOT reachable from any CI surface, and carries the reason someone signed
# for that. The guard FAILS on an entry with no reason, a stale path, or a path
# the surfaces already enumerate. Burning an entry down = making it CI-reachable
# (package.json test:framework-shell or a ci.yml step) and deleting its line.
#
# Format: <repo-relative path> | <reason>
# All entries below were signed at #1017's filing base (main 826a8b3b, 2026-07-31)
# by pepper (sb-it-1-dt); measurements cited are one-run assertions from that seat.
# --- tools/git: the #1007 five — non-hermetic, resolve real credentials ---
packages/mosaic/framework/tools/git/test-pr-merge-gitea-empty-uid.sh | resolves real credentials (#1007 census); joins CI after the wrapper-half hermeticity fix (git -C scoping)
packages/mosaic/framework/tools/git/test-issue-create-interactive-auth.sh | resolves real credentials (#1007 census); joins CI after the wrapper-half hermeticity fix
packages/mosaic/framework/tools/git/test-gitea-login-resolution.sh | resolves real credentials (#1007 census); joins CI after the wrapper-half hermeticity fix
packages/mosaic/framework/tools/git/test-pr-metadata-gitea.sh | resolves real credentials (#1007 census, fourth entry via family-grep); joins CI after the wrapper-half hermeticity fix
packages/mosaic/framework/tools/git/test-issue-comment-readback.sh | resolves real credentials (#1007 census, fifth entry); joins CI after the wrapper-half hermeticity fix
# --- tools/git: push guards — measured green locally, CI-image fitness unverified ---
packages/mosaic/framework/tools/git/test-push-guard.sh | measured green at 826a8b3b (46 passed / 0 failed, one run, 2026-07-31); CI-image fitness unverified; #1017 burndown
packages/mosaic/framework/tools/git/test-mutate-push-guard.sh | measured green at 826a8b3b (8/0, 13 mutants killed 0 survived, one run, 2026-07-31); requires setsid (util-linux), absent from the alpine base image; #1017 burndown
packages/mosaic/framework/tools/git/test-issue-create-body-safety.sh | hermeticity unaudited — the unprotected suite in #1007's protected/unprotected split; audit before CI; #1017 burndown
# --- tools/git: unmeasured ---
packages/mosaic/framework/tools/git/test-verify-clean-clone.sh | unmeasured in CI image; asserts git file-mode (100644/755) semantics that need verification on the CI filesystem first; #1017 burndown
packages/mosaic/framework/tools/git/test-help-exit-code.sh | unmeasured in CI image; stub-based (#701 regression harness), likely CI-fit; #1017 burndown
packages/mosaic/framework/tools/git/test-lane-brief-pr-linkage.sh | unmeasured in CI image; fixture-based (#546/#547 regression harness), likely CI-fit; #1017 burndown
# --- tools/tmux: require a live tmux server ---
packages/mosaic/framework/tools/tmux/test-send-message-socket.sh | requires a real tmux server on a throwaway socket; CI image ships no tmux; #1017 burndown (needs tmux in image or a signed permanent exclusion)
packages/mosaic/framework/tools/tmux/test-send-message-verdict.sh | requires real tmux-pane fixtures on a throwaway socket; CI image ships no tmux; #1017 burndown (same condition as its sibling)
# --- single-suite directories: unmeasured in CI ---
packages/mosaic/framework/tools/fleet/test-start-agent-session.sh | unmeasured in CI image; stubs tmux via a fake bin dir, likely CI-fit; #1017 burndown
packages/mosaic/framework/tools/glpi/test-list-http-status.sh | unmeasured in CI image; stub-based (#807 regression harness), likely CI-fit; #1017 burndown
packages/mosaic/framework/tools/orchestrator/test-board-roll.sh | unmeasured in CI image; file-fixture based, likely CI-fit; #1017 burndown
packages/mosaic/framework/tools/woodpecker/test-ci-wait-exit-matrix.sh | unmeasured in CI image; drives ci-wait.sh against a stub pipeline-status.sh, likely CI-fit; #1017 burndown
# --- naming-boundary files the strict test-*.sh prefix cannot even name ---
# (#1017: three independent censuses handled the microtest file three different
# ways — editorial drop, structural exclusion, accidental inclusion — with no
# recorded judgement. These lines ARE that judgement, signed.)
packages/mosaic/framework/tools/orchestrator/smoke-test.sh | behavior smoke checks for coord continue/run workflows, run manually by orchestrator seats; unmeasured in CI; #1017 burndown
packages/mosaic/framework/tools/wake/validate-973/microtest-wake-assert.sh | #973 instrument self-test, run as a precondition of the validate-973 evidence procedure rather than as a standing CI suite; #1017 burndown candidate
@@ -529,7 +529,19 @@ cmd_run() {
echo "detector.sh: WARN — off-host liveness beacon emit failed (see beacon.sh); the off-host absence check remains the authoritative dead-man." >&2
fi
[ "$once" -eq 1 ] && break
sleep "$interval"
# Close the detector lock fd in the sleep child; otherwise an orphaned sleep
# keeps the single-instance flock (fd 9, taken at exec 9> above) alive after
# the detector parent dies. The lock is non-blocking (`flock -n`, above), so
# for as long as that sleep survives a replacement instance is REFUSED and
# exits rather than queueing. This particular hold is BOUNDED by one poll
# interval (WAKE_DETECTOR_INTERVAL, default 30s): when the orphaned sleep
# exits its copy of fd 9 closes, ending this bounded sleep-child hold. It
# does NOT follow that the next start succeeds — other inheritors of fd 9
# (the M1 adapter, M2 sink grandchildren) are outside this patch's scope and
# can keep holding the flock. The cost this removes is a restart window in
# which every supervisor retry fails on the sleep child's account.
# `9>&-` closes ONLY the child's copy — the parent's lock is unaffected.
sleep "$interval" 9>&-
done
}
+1 -1
View File
@@ -25,7 +25,7 @@
"lint": "eslint src",
"typecheck": "tsc --noEmit",
"test": "vitest run --passWithNoTests && pnpm run test:framework-shell",
"test:framework-shell": "python3 src/lease-broker/daemon_deadline_unittest.py && python3 src/lease-broker/normative_fragments_unittest.py && python3 src/lease-broker/receipt_challenge_unittest.py && python3 src/lease-broker/context_recovery_unittest.py && python3 src/lease-broker/recovery_runtime_unittest.py && python3 src/lease-broker/recovery_b1_adversarial_unittest.py && python3 src/lease-broker/framework_skill_portability_unittest.py && python3 src/mutator-gate/runtime_tools_unittest.py && python3 src/mutator-gate/runtime_launch_guard_unittest.py && python3 src/mutator-gate/version_coupling_unittest.py && python3 framework/tools/lease-broker/check-runtime-launches.py --root ../.. && bash framework/tools/codex/test-pr-diff-context.sh && bash framework/tools/qa/test-deps-preflight.sh && bash framework/tools/git/test-pr-review-gitea-comment.sh && bash framework/tools/git/test-pr-review-repo-host-override.sh && bash framework/tools/git/test-ci-queue-wait-branch-absent.sh && bash framework/tools/git/test-git-credential-mosaic.sh && bash framework/tools/git/test-gitea-token-identity.sh && bash framework/tools/_scripts/test-install-ordering-guard.sh && bash framework/tools/tmux/agent-send.test.sh && bash framework/tools/wake/test-wake-store-ack.sh && bash framework/tools/wake/test-wake-store-enqueue-race.sh && bash framework/tools/wake/test-wake-digest-hmac.sh && bash framework/tools/wake/test-wake-digest-quarantine.sh && bash framework/tools/wake/test-wake-detector.sh && bash framework/tools/wake/test-wake-fn-oracle.sh && bash framework/tools/wake/test-wake-reconcile.sh && bash framework/tools/wake/test-wake-beacon.sh && bash framework/tools/wake/test-wake-preimage.sh && bash framework/tools/wake/test-wake-install.sh"
"test:framework-shell": "bash framework/tools/quality/scripts/check-test-enumeration.sh && bash framework/tools/quality/scripts/test-check-test-enumeration.sh && python3 src/lease-broker/daemon_deadline_unittest.py && python3 src/lease-broker/normative_fragments_unittest.py && python3 src/lease-broker/receipt_challenge_unittest.py && python3 src/lease-broker/context_recovery_unittest.py && python3 src/lease-broker/recovery_runtime_unittest.py && python3 src/lease-broker/recovery_b1_adversarial_unittest.py && python3 src/lease-broker/framework_skill_portability_unittest.py && python3 src/mutator-gate/runtime_tools_unittest.py && python3 src/mutator-gate/runtime_launch_guard_unittest.py && python3 src/mutator-gate/version_coupling_unittest.py && python3 framework/tools/lease-broker/check-runtime-launches.py --root ../.. && bash framework/tools/codex/test-pr-diff-context.sh && bash framework/tools/qa/test-deps-preflight.sh && bash framework/tools/git/test-pr-review-gitea-comment.sh && bash framework/tools/git/test-pr-review-repo-host-override.sh && bash framework/tools/git/test-ci-queue-wait-branch-absent.sh && bash framework/tools/git/test-git-credential-mosaic.sh && bash framework/tools/git/test-gitea-token-identity.sh && bash framework/tools/_scripts/test-install-ordering-guard.sh && bash framework/tools/tmux/agent-send.test.sh && bash framework/tools/wake/test-wake-store-ack.sh && bash framework/tools/wake/test-wake-store-enqueue-race.sh && bash framework/tools/wake/test-wake-digest-hmac.sh && bash framework/tools/wake/test-wake-digest-quarantine.sh && bash framework/tools/wake/test-wake-detector.sh && bash framework/tools/wake/test-wake-fn-oracle.sh && bash framework/tools/wake/test-wake-reconcile.sh && bash framework/tools/wake/test-wake-beacon.sh && bash framework/tools/wake/test-wake-preimage.sh && bash framework/tools/wake/test-wake-install.sh"
},
"dependencies": {
"@mosaicstack/brain": "workspace:*",