Skip to content

fix(exec.sh): restore names only the errors the script added #179

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
MendixMau merged 1 commit into from
Sep 30, 2026
Merged

fix(exec.sh): restore names only the errors the script added #179

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,7 @@ Sections dated before 2026-09-19 predate the cycle and stay as they are.
- fix(project-bin/exec.sh): **the delta gate keys an error on code + message + location, counted, instead of the message text alone.** On a real 22-error baseline (Mendix 11.12.1), 20 errors shared two "Could not find widget" messages and CE0117 always reads "Error(s) in expression.", so a script adding another CE0117 in a new microflow, or another copy of a baseline error, was judged "identical" and kept. Now each Error is a hash of errorCode, message and every location's module / document / element, and the subset test is a multiset test. Names, not GUIDs, so a CREATE OR REPLACE that leaves a pre-existing error in place still reads as pre-existing. Field-run on a scratch copy of a greenfield PoC model — MendixMau
- learn(workflow-structure-rules, learned-workflow-patterns, module-folder-convention, learned-mdl-preflight, bug-logs): **five workflow rules re-probed on mxcli v0.24.0 / Mendix 11.13.0.** A forward `JUMP TO` builds clean (the CE6681 row taught direction as a platform rule; it was the jump-named-after-its-target defect, fixed upstream in v0.21.0); an interrupting boundary timer ending in `jump to` or `end workflow` builds clean (BUG-109 stamped no-longer-reproduces); BUG-76 splits — the bare-enum `DECISION` is now refused by `MDL-WF03` and still corrupts when forced, a Boolean `true`/`false` decision builds at 0 errors; `create or modify workflow … folder` places the workflow. New: one interrupting boundary per activity (`MDL-WF15`/CE6697) and no `end workflow` on a non-interrupting path (`MDL-WF08`/CE1844); `SET TASK OUTCOME` needs a signed-in named user ("Only named users can complete user task"), with the run-as-session Java action pattern. Evidence: a six-probe, four-control construct run (check, exec, native `mx check`, describe read-back) and a live proof test — card-disbursement requirements-driven build
- fix(project-tests/e2e/otel.js): **`capture()` pages back to t0 instead of reading only the newest 400 traces.** Jaeger answers the newest `limit` traces and says nothing about the rest; with timers running, 400 traces were two minutes, so a capture with t0 15 minutes back reported 0 errors over 24 real ERROR spans. It now pages by `end` while a page is full and still after t0, dedupes by trace, caps at `OTEL_MAX_PAGES` (25) and marks the result `.truncated` when the cap stops it short. Field run against the live Jaeger: t0 −15 min, 462 spans / 0 errors → 15,439 spans / 24 errors in 11 pages; a step's own capture stays one page — card-disbursement requirements-driven build
- fix(exec.sh): **when `exec.sh` undoes a script on a model that already had errors, it now names only the errors that script added.** Before, the restore printed every error on the model (22 on the field model, the one new one unmarked), and then said the errors were "PRE-EXISTING … This script is NOT the cause. Nothing will exec until that is cleared" and logged "blocked". Both were false: the script did add the error, and the delta gate keeps any later script that adds nothing new. Now the headline reads "this script added 1 new error(s) [CE0109] (22 in total, 21 were already there)", the new errors are listed in full, the old ones are summarised one line per code, and the build log records "1 new error(s): CE0109 — script rolled back; 22 pre-existing remain". New and old are split with the same code + message + location key the gate uses, so the two cannot disagree. With no measured baseline (`SKIP_BASELINE=1`) the output is unchanged. — MendixMau
- fix(project-bin/constants-audit.sh): **a `__SET_ME__`-style sentinel default reports `SENTINEL`, not `MODEL-SECRET`.** The audit only knew SET vs EMPTY, so the placeholder `learned-constants-and-secrets.md` Step 3 prescribes for a must-override secret got the same verdict as a real password in git and needed a waiver. The test reads the value's shape (`__[A-Za-z0-9_]+__`) and still prints nothing; the skill's verdict table gains the row. Field run on a copy of the model: 4 findings → 3, the hub password constant SENTINEL — card-disbursement requirements-driven build
- fix(review-report.js, coverage-check-all.sh): **a module with one coverage ledger per BRD no longer carries a permanent coverage FAULT over a clean count.** `coverage-check-all.sh` keeps `leaves:` on each per-BRD line, and `review-report.js` sums the counters across those lines (null, never zero, when a BRD line is FAULT / NO LEDGER, lacks a counter, or the lines miss the stated BRD count); an old installed copy without `leaves:` is named in the fault. Field run on two real ledgers: FAULT → pass, leaves 337 / 610, 947 summed over both — card-disbursement requirements-driven build
- fix(project-bin/check-page-shell.sh): **the page-body scan skips quoted strings and block comments.** An unbalanced brace inside a string literal (a JSON payload preview `Content: '{{ … }'`) left the depth up, the scan ran on into the next pages and a page with one H1 was reported as declaring 3; a `--` inside a string cut the line and did the same. Field run: 18 pages across 87 scripts, 1 violation (the false positive) → 0 — card-disbursement requirements-driven build
Expand Down
109 changes: 84 additions & 25 deletions project-bin/exec.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -563,22 +563,67 @@ err_codes() {
# Names, not elementId GUIDs: a CREATE OR REPLACE re-mints the GUIDs of a microflow whose
# pre-existing error it leaves in place, and that must still read as pre-existing. A rename
# does read as new, and restores — the safe direction. Hex tokens: no "|" or glob characters,
# whatever the message says.
# whatever the message says. ERRKEY_PY is the one definition of that key: err_set and
# err_report both prepend it, so "new" in the verdict and "NEW" in the report cannot drift.
ERRKEY_PY='import hashlib
def errkey(x):
locs = sorted("\x1e".join(str(l.get(k) or "") for k in ("module", "document", "element"))
for l in (x.get("locations") or []))
raw = "\x1f".join([x.get("errorCode") or "", x.get("message") or ""] + locs)
return hashlib.sha1(raw.encode("utf-8")).hexdigest()[:16]'
err_set() {
[ -n "$PY" ] || { echo ""; return; }
"$PY" - "$(native_path "$1")" <<'ERRSETPY' 2>/dev/null || echo ""
import hashlib, json, sys
{ printf '%s\n' "$ERRKEY_PY"; cat <<'ERRSETPY'; } | "$PY" - "$(native_path "$1")" 2>/dev/null || echo ""
import json, sys
d = json.load(open(sys.argv[1]))
keys = []
print('|'.join(sorted(errkey(x) for x in d.get('problems', []) if x.get('severity') == 'Error')))
ERRSETPY
}

# err_report count|print <errors.json> — splits the post-exec errors against BASELINE_SET,
# counted the same way is_subset_of counts. Only when the baseline was measured
# (BASE_KNOWN=1); otherwise every error is listed as before, with no new/old split.
# count → "<new count> <new codes, comma-joined>", or "?" when the baseline is unknown.
# print → the NEW errors in full (code, message, where, expression detail), then the
# pre-existing ones as one line per code + message with a count.
# Why (field run, 2026-09-29): a restore on a 22-error model printed "22 error(s) found"
# and all 22 with locations; the one the script added was not marked, and the count was
# the same as before the script ran.
err_report() {
[ -n "$PY" ] || { [ "$1" = count ] && echo "?"; return; }
{ printf '%s\n' "$ERRKEY_PY"; cat <<'ERRREPPY'; } | "$PY" - "$1" "$(native_path "$2")" "$BASELINE_SET" "${BASE_KNOWN:-0}" 2>/dev/null || { [ "$1" = count ] && echo "?"; }
import json, sys
from collections import Counter
mode, known = sys.argv[1], sys.argv[4] == '1'
d = json.load(open(sys.argv[2]))
left = Counter(k for k in sys.argv[3].split('|') if k)
new, old = [], []
for x in d.get('problems', []):
if x.get('severity') != 'Error':
continue
locs = sorted('\x1e'.join(str(l.get(k) or '') for k in ('module', 'document', 'element'))
for l in (x.get('locations') or []))
raw = '\x1f'.join([x.get('errorCode') or '', x.get('message') or ''] + locs)
keys.append(hashlib.sha1(raw.encode('utf-8')).hexdigest()[:16])
print('|'.join(sorted(keys)))
ERRSETPY
k = errkey(x)
if known and left[k] > 0:
left[k] -= 1
old.append(x)
else:
new.append(x)
if mode == 'count':
print('%d %s' % (len(new), ','.join(sorted(set(x.get('errorCode') or '?' for x in new)))) if known else '?')
sys.exit(0)
if known:
print(' NEW, from this script (%d):' % len(new))
for e in new:
print(' ', e.get('errorCode', '?'), e.get('message', ''))
for loc in e.get('locations', []) or []:
print(' at', loc.get('module', '-'), '/', loc.get('document', '-'), '/', loc.get('element', '-'))
md = e.get('metadata') or {}
if md.get('expressionErrors'):
print(' expression:', md['expressionErrors'])
if old:
print(' Already there before this script, not why it was undone (%d):' % len(old))
for (code, msg), n in sorted(Counter((e.get('errorCode', '?'), e.get('message', '')) for e in old).items()):
print(' ', code, msg, ('x%d' % n) if n > 1 else '')
ERRREPPY
}

# is_subset_of <candidate> <superset> — both are err_set's "|"-joined sorted key lists.
Expand Down Expand Up @@ -638,6 +683,7 @@ fi
# "this script broke it" from "it was already broken". Costs one extra mxbuild;
# skip with SKIP_BASELINE=1 when you know the tree is clean.
BASELINE_SET=""
BASE_KNOWN=0
if [ "${SKIP_BASELINE:-0}" != "1" ] && [ -x "$MXBUILD" ] && [ -x "$JAVA_EXE" ]; then
echo "→ Pre-flight: checking whether the model already has errors..."
_BF=$(mktemp /tmp/mxbuild-baseline.XXXXXX)
Expand All @@ -664,6 +710,7 @@ if [ "${SKIP_BASELINE:-0}" != "1" ] && [ -x "$MXBUILD" ] && [ -x "$JAVA_EXE" ];
[ "$_BC" = "0" ] && [ "$MXTK_MXBUILD_EXIT" -ne 0 ] && _BC="?"
fi
rm -f "$_BF"
case "$_BC" in ''|*[!0-9]*) ;; *) BASE_KNOWN=1 ;; esac
if [ "$_BC" = "?" ]; then
echo " ⚠ Baseline NOT measured: mxbuild exited $MXTK_MXBUILD_EXIT without checking the model."
[ -n "${MXTK_MXBUILD_WHY:-}" ] && echo " mxbuild says: $MXTK_MXBUILD_WHY"
Expand Down Expand Up @@ -818,8 +865,22 @@ if [ -x "$MXBUILD" ] && [ -x "$JAVA_EXE" ]; then
rm -f "$ERRORS_FILE"
else
GATE_STATE="fail"
echo " ✗ mxbuild: $CE_COUNT error(s) found — restoring snapshot to avoid loading a corrupt MPR."
CE_CODES=$(err_codes "$ERRORS_FILE")
read -r NEW_COUNT NEW_CODES <<EOF
$(err_report count "$ERRORS_FILE")
EOF
case "$NEW_COUNT" in
''|*[!0-9]*)
NEW_COUNT="?"
echo " ✗ mxbuild: $CE_COUNT error(s) found — restoring snapshot to avoid loading a corrupt MPR." ;;
*)
_OLD=$(( CE_COUNT - NEW_COUNT ))
if [ "$_OLD" -gt 0 ]; then
echo " ✗ mxbuild: this script added $NEW_COUNT new error(s) [$NEW_CODES] ($CE_COUNT in total, $_OLD were already there) — undoing it."
else
echo " ✗ mxbuild: this script added $NEW_COUNT new error(s) [$NEW_CODES] — undoing it."
fi ;;
esac
# restore_snapshot (above) is the one restore implementation, through restore-mpr.sh.
# Without it the attribution check below rebuilt the SAME broken model and blamed the
# error on "PRE-EXISTING" (v1 model, 2026-09-26). RESTORED gates that check.
Expand All @@ -829,19 +890,8 @@ if [ -x "$MXBUILD" ] && [ -x "$JAVA_EXE" ]; then
# old one-liner printed "CE0117 Error(s) in expression." and nothing else, so the reader
# had to re-run mxbuild by hand on a scratch copy to learn which activity (greenfield
# pilot, 2026-09-04: an unqualified enum value in a CHANGE, found only that way).
[ -n "$PY" ] && "$PY" - "$ERRORS_FILE" <<'PYEOF' 2>/dev/null || true
import json, sys
d = json.load(open(sys.argv[1]))
for e in d.get('problems', []):
if e.get('severity') != 'Error':
continue
print(' ', e.get('errorCode', '?'), e.get('message', ''))
for loc in e.get('locations', []) or []:
print(' at', loc.get('module', '-'), '/', loc.get('document', '-'), '/', loc.get('element', '-'))
md = e.get('metadata') or {}
if md.get('expressionErrors'):
print(' expression:', md['expressionErrors'])
PYEOF
# With a measured baseline, the NEW errors come first and the rest are summarised.
err_report print "$ERRORS_FILE" || true
cp "$ERRORS_FILE" "$LAST_ERRS"
echo " → Full error detail: .mpr-snapshots/last-mxbuild-errors.json"

Expand Down Expand Up @@ -870,7 +920,16 @@ PYEOF
rm -f "$BASE_ERRS"

echo ""
if [ "$BASE_COUNT" != "0" ] && [ "$BASE_COUNT" != "?" ]; then
if [ "$NEW_COUNT" != "?" ] && [ "$BASE_COUNT" != "0" ] && [ "$BASE_COUNT" != "?" ]; then
# The baseline was measured and this script added errors on top of it. The old
# branch below said "NOT the cause. Nothing will exec until that is cleared" here,
# which was false twice over: the script added the error, and the delta gate keeps
# any later script that adds nothing new (field run, 2026-09-29).
echo " → Undone: this script added $NEW_COUNT new error(s) [$NEW_CODES]. Fix those and re-run."
echo " The model is back to its $BASE_COUNT earlier error(s) [$BASE_CODES]. Those did not cause this; a script"
echo " that adds nothing new is kept even while they remain."
log_build "❌ gate failed" "$NEW_COUNT new error(s): $NEW_CODES — script rolled back; $BASE_COUNT pre-existing remain"
elif [ "$BASE_COUNT" != "0" ] && [ "$BASE_COUNT" != "?" ]; then
echo " ⚠ PRE-EXISTING: the restored model already fails with $BASE_COUNT error(s) [$BASE_CODES]."
echo " This script is NOT the cause. Nothing will exec until that is cleared."
case "$BASE_CODES" in
Expand Down
Loading