MendixMau · Nov 1, 2026 · Oct 29, 2026
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -37,6 +37,7 @@ Sections dated before 2026-09-19 predate the cycle and stay as they are. Named m
 - fix(page-audit, design-audit): **the audits produce their own `.claude/loop/page-scope.json`**: when it is missing or older than the model they run `bin/page-scope.sh` (`scope-ensure.js`, 2 s) instead of FAULTing on 0 pages; a failed regeneration never blanks an existing file and names the failure. Field: scope deleted -> page-audit ran on 9 pages, `touch` of the .mpr -> regenerated once — speed experiment E2
 - new(look-ledger, journey-runner, module-review): **a journey run's screenshots are LOOK evidence**: `bin/look-ledger.sh shots` (run by journey-runner after every run) maps each reached step's PNG to the one page whose widgets (`DESCRIBE PAGE`) it landed on, a URL pattern breaking ties, and lists every owed page as `OPEN` (open that PNG; the Read hook then marks the page SEEN) or `UNREACHED` (short gap-fill journey, or `page-audit.js` as before) with the line `owed N = A OPEN + B UNREACHED`; a shot older than the page's last build is not evidence and nothing is claimed when unsure. Field: 8 owed pages = 5 OPEN + 3 UNREACHED from two journey runs, no separate screenshot pass; baseline words 64,004 -> 64,193 — speed experiment E2
 - fix(affected-tests): **a page, flow or navigation change no longer selects 0 journeys because the journeys name widgets, not `Module.Name`**: the affected scope now reads the widget names of every changed page/snippet and of each calling page (`DESCRIBE PAGE`, one extra warm-catalog `mxcli` call, 0.07 s) and selects each journey whose JSON holds one; a navigation or layout change selects every journey; when callers or widgets cannot be measured (`widgets NOT MEASURED` row) it selects every journey (unsure means include). Field: 067 (Request_Edit page) 0 -> J-REQ-01, 090 (navigation) 0 -> all 5 journeys — speed experiment E2
+- fix(canvas): **`mf-canvas-facts.py` reads a note line's sides by the model's Mendix version, so on Mendix 11 it no longer reports NOTE-SIDE for lines that face their object.** A note numbers its sides top 0, bottom 1, left 2, right 3 on Mendix 10, but like every other object (top 0, right 1, bottom 2, left 3) on Mendix 11; the script knew only the Mendix 10 table. It now reads `_ProductVersion` from the `.mpr`'s `_MetaData` table (major 11 or higher: the activity numbering; unreadable: Mendix 11) and prints `# note sides: …` as its first line. One app's 305 hand-drawn note lines, matched across its 10.24 and 11.15 versions: the upgrade renumbered the note's side on every line (0 to 0, 1 to 2, 2 to 3, 3 to 1) and left the object's side alone, and the script's fact counts for the whole app are now identical on both. mxcli v0.28.0 (mendixlabs/mxcli#1620) writes the activity numbering whatever the project's version. Fixture section V (renumbered copy of the golden model). — saingge
 - process(design-artifacts, conversion-runbook): **Stage 3 runs in parallel by default, from one brief and a mechanical task list**: the lead fixes modules, canonical names and design tokens, writes `docs/progress/stage3-brief.md`, and takes the tasks from the stage-3 lines of `bin/lib/artifact-check.sh` (so no artifact is forgotten); one subagent per 1–2 screens or artifact group, a shared opening line so the start is cached, mechanical tasks on the cheap tier, 5-line replies, the lead the only writer of `PROJECT.md`. The runbook's Stage 3 fan-out row points there, and its Stage 2 row names `facts-lock.sh resolve`. Measured on one project (single runs, Sonnet, 250k auto-compact): serial 13 min / $3.60, parallel 10 min / $4.44, parallel with a brief 6 min / $4.32; 0 name mismatches in the parallel runs (serial dropped 3 entities); both parallel runs forgot the process map and swimlanes, which the task list now covers. Baseline words 64,004 before and after — Stage 3 experiment 2026-10-28
 - fix(facts-lock, gate-check): **parallel BRDs keep one spelling per identifier**: `bin/facts-lock.sh <project> resolve` freezes each casing conflict (`WIPStatus`/`WipStatus`) whose majority spelling is used by strictly more BRD sources than any other and records the count in the lock (`resolvedBy`); a tie stays contested, and `build` still never picks on its own. Stage 2 of `gate-check.sh` now reads the lock once there are 2+ BRDs: a BRD that contradicts a frozen identifier FAILs and is quoted; no lock yet is a note, never a FAIL. Measured cause: 20 parallel BRDs on PROJECT-C carried 10 casing conflicts into Stage 3 as open questions, and facts-lock.sh, documented only as advice, reached 1 of 6 projects. Fixtures: `test-facts-lock.sh` resolve cases (4), `test-bug03-gates.sh` T19 (2 of 4 fail on the previous gate-check) — speed/parallel experiments 2026-10-28
 - fix(row-groups): **a build plan whose number and Kind cells are formatted as code (`` `BUILD` ``, `` `12` ``) is read like a plain one**: before, it read 0 BUILD rows, so with `Row groups: on` grouping silently did nothing (the fast-lane agent in the 2026-10-28 speed experiment had to strip the backticks by hand). Fixture: `test-row-groups.sh` G5 (both cases fail on the previous row-groups.sh) — speed experiment 2026-10-28

diff --git a/project-bin/mf-canvas-facts.py b/project-bin/mf-canvas-facts.py
@@ -27,16 +27,19 @@
 CALIBRATION, measured from Studio Pro 10.24 screenshots: a variable label is ~5.5 units per
 character, two lines (name, type) under the shape; an activity fits 3 caption lines of ~16
 characters and no word over 14; a decision diamond fits 2 lines of ~10 characters and no word
-over 10. Connection sides are stored as indexes: an annotation (note) numbers its sides top 0,
-bottom 1, left 2, right 3; every other object top 0, right 1, bottom 2, left 3. The note is
-always the origin of a note line. A flow line is the cubic Bezier through the two stored control
-vectors, sampled at 48 points. SENSE is a prompt to read the true branch, not a verdict: on one
-hand-drawn corpus 15 of its 18 hits were fine (`$Message = empty` really means "valid"), while all
-9 hits in the mxcli-written flows were real.
-
-OUTPUT: per flow `# Module.Name  (n)` then one indented fact line each; last line
-`# facts: <flows> flows, OVERLAP a, LINE b, ...`. Field run, measured 2026-10-21 on one Mendix 10.24 app (1,588 flows, read in one pass):
-on the 36 flows one mxcli-built branch added or changed in a single module, before any repair, OVERLAP 19,
+over 10. Connection sides are stored as indexes. Every object except a note numbers its sides top 0,
+right 1, bottom 2, left 3. A note (annotation) numbers its own sides by the model's Mendix version:
+on Mendix 10 top 0, bottom 1, left 2, right 3; on Mendix 11 like every other object. The version is
+the .mpr's `_MetaData._ProductVersion`; the first output line says which numbering was used, and a
+version that cannot be read counts as Mendix 11. The note is always the origin of a note line. A flow
+line is the cubic Bezier through the two stored control vectors, sampled at 48 points. SENSE is a
+prompt to read the true branch, not a verdict: on one hand-drawn corpus 15 of its 18 hits were fine
+(`$Message = empty` really means "valid"), while all 9 hits in the mxcli-written flows were real.
+
+OUTPUT: first `# note sides: <numbering used>`, then per flow `# Module.Name  (n)` then one indented
+fact line each; last line `# facts: <flows> flows, OVERLAP a, LINE b, ...`. Field run, measured
+2026-10-21 on one Mendix 10.24 app (1,588 flows, read in one pass): on the 36 flows one mxcli-built
+branch added or changed in a single module, before any repair, OVERLAP 19,
 LINE 27, NOTE-SIDE 17, LEFTWARD 3, TURN 4, SENSE 9, CAPTION 148; after the branch's later repair
 commits, OVERLAP 3, LINE 7, TURN 4, CAPTION 113 and the rest 0.
 
@@ -55,7 +58,8 @@
 KINDS = ("OVERLAP", "LINE", "NOTE-SIDE", "LEFTWARD", "TURN", "CAPTION", "SENSE")
 CH = 5.5
 ACT_SIDES = {0: (0, -1), 1: (1, 0), 2: (0, 1), 3: (-1, 0)}
-NOTE_SIDES = {0: (0, -1), 1: (0, 1), 2: (-1, 0), 3: (1, 0)}
+NOTE_SIDES_10 = {0: (0, -1), 1: (0, 1), 2: (-1, 0), 3: (1, 0)}   # a note on Mendix 10
+NOTE_SIDES = ACT_SIDES      # a note on Mendix 11 and up; main() sets it from the model's version
 SIDE_NAME = {(0, -1): "top", (0, 1): "bottom", (-1, 0): "left", (1, 0): "right"}
 NEG = re.compile(r"\b(not|no|non|none|never|missing|wrong|without|empty|invalid|inactive|incomplete|"
                  r"incorrect|disabled|absent|expired|refused|failed|lacks?)\b|n't\b|"
@@ -66,6 +70,23 @@
                  r"validated|verified)\b", re.I)
 
 
+# -- which numbering the model's notes use ---------------------------------------------------------
+def note_numbering(model):
+    """(note side table, text for the output line), from the major version in the .mpr's _MetaData."""
+    try:
+        db = model._db()
+        try:
+            ver = db.execute("SELECT _ProductVersion FROM _MetaData").fetchone()[0]
+        finally:
+            db.close()
+        major = int(re.match(r"\d+", ver).group())
+    except Exception:                   # no table, no row, no number: say so, and take the newer numbering
+        return ACT_SIDES, "Mendix 11 numbering (model version unreadable, assumed)"
+    if major >= 11:
+        return ACT_SIDES, "Mendix 11 numbering (model version %s)" % ver
+    return NOTE_SIDES_10, "Mendix 10 numbering (model version %s)" % ver
+
+
 # -- the unit as plain data ----------------------------------------------------------------------
 def plain(t, v):
     """A mxpatch (type, key, value) value as dicts and lists; arrays keep their version marker."""
@@ -346,6 +367,8 @@ def main(argv):
     except mxpatch.PatchError as e:
         sys.stderr.write("mf-canvas-facts: %s\n" % e)
         return 2
+    global NOTE_SIDES
+    NOTE_SIDES, numbering = note_numbering(model)
     # model.units() would stop at the first unit that does not decode; read each candidate on
     # its own so one bad unit is a named fault and every other flow is still measured.
     needles = [mxpatch._bson_str(0x02, "$Type", t) for t in FLOW_TYPES]
@@ -375,6 +398,7 @@ def main(argv):
     found = {q for q, _ in picked} | {q.split(".", 1)[1] for q, _ in picked}
     missing = [n for n in names if n not in found]
     totals = dict.fromkeys(KINDS, 0)
+    print("# note sides: %s" % numbering)
     for q, mf in sorted(picked, key=lambda x: x[0]):
         res = check(mf, exposed)
         print("# %s  (%d)" % (q, len(res)))

diff --git a/tests/wave2/test-mf-canvas-facts.sh b/tests/wave2/test-mf-canvas-facts.sh
@@ -10,6 +10,10 @@
 # (see its CAPTURE.txt). Each fact has a negative control in the same model: a parameter column
 # (no OVERLAP), a decision captioned to match its true branch (no SENSE).
 #
+# The note numbering depends on the model's Mendix version (section V). The captured dump says
+# 11.14.0; the golden facts below were measured on a Mendix 10 reading of it, so model() stamps
+# 10.24.18 over it, and section V builds the Mendix 11 copies.
+#
 # Positive control: against a project-bin without the script every case fails -- it does not
 # exist. [guard] marks regression guards.
 set -uo pipefail
@@ -27,11 +31,31 @@ bad() { FAIL=$((FAIL+1)); printf '  FAIL %s\n' "$1"; [ -n "${2:-}" ] && printf '
 . "$TK/bin/lib/portable.sh"; require_py
 export MXPATCH_BACKUP_DIR="$WORK/bak"
 
-# model <dir> -- the base model's units, the captured five over them, the captured .mpr.
+# model <dir> [version] -- the base model's units, the captured five over them, the captured .mpr
+# stamped with the Mendix version (default 10.24.18.103793, a Mendix 10 model).
 model() {
   rm -rf "$1"; mkdir -p "$1"; cp -r "$BASE/mprcontents" "$1/"; cp -r "$FX/overlay/mprcontents" "$1/"
-  "$PY" -c 'import sqlite3,sys; c=sqlite3.connect(sys.argv[1]); c.executescript(open(sys.argv[2]).read()); c.commit()' \
-    "$1/TestApp.mpr" "$FX/overlay/TestApp.mpr.sql"
+  "$PY" -c 'import sqlite3,sys; c=sqlite3.connect(sys.argv[1]); c.executescript(open(sys.argv[2]).read())
+c.execute("UPDATE _MetaData SET _ProductVersion = ?", (sys.argv[3],)); c.commit()' \
+    "$1/TestApp.mpr" "$FX/overlay/TestApp.mpr.sql" "${2:-10.24.18.103793}"
+}
+# notes <dir> set <n> | remap -- the origin side of every note line: all set to n, or renumbered the way
+# the Mendix 10 -> 11 upgrade does (top 0 -> 0, bottom 1 -> 2, left 2 -> 3, right 3 -> 1). Prints how many.
+notes() {
+  "$PY" -I -c 'import sys; sys.path.insert(0, sys.argv[1]); import mxpatch
+m = mxpatch.Model(sys.argv[2]); n = 0
+for p in m.unit_files():
+    raw = open(p, "rb").read()
+    if b"AnnotationFlow" not in raw:
+        continue
+    u = mxpatch.Unit(p, raw)
+    for d in mxpatch.walk(u.doc):
+        if mxpatch.get(d, "$Type") == "Microflows$AnnotationFlow":
+            old = mxpatch.get(d, "OriginConnectionIndex")
+            mxpatch.set_value(d, "OriginConnectionIndex", int(sys.argv[4]) if sys.argv[3] == "set" else {0: 0, 1: 2, 2: 3, 3: 1}[old])
+            n += 1
+    m.write_unit(u, apply=True, log=lambda s: None)
+print(n)' "$BIN" "$1" "$2" "${3:-}"
 }
 FACTS() { "$PY" -I "$BIN/mf-canvas-facts.py" "$@"; }
 
@@ -69,6 +93,33 @@ for c in sys.argv[2:]: print("%s=%s" % (c, bool(f.NEG.search(c))))' "$BIN/mf-can
 printf '%s' "$OUT" | tail -1 | grep -q '^# facts: 15 flows, OVERLAP 1, LINE 1, NOTE-SIDE 1, LEFTWARD 0, TURN 0, CAPTION 0, SENSE 1$' \
   && ok "summary line counts every fact kind" || bad "summary" "$(printf '%s' "$OUT" | tail -1)"
 
+echo "== V: the note numbering follows the model's Mendix version =="
+printf '%s\n' "$OUT" | head -1 | grep -qx '# note sides: Mendix 10 numbering (model version 10.24.18.103793)' \
+  && ok "the first output line names the numbering used (Mendix 10)" || bad "numbering line" "$(printf '%s\n' "$OUT" | head -1)"
+M11="$WORK/m11"; model "$M11" 11.15.0
+OUT11="$(FACTS "$M11" --module MyFirstModule 2>&1)"
+printf '%s\n' "$OUT11" | head -1 | grep -qx '# note sides: Mendix 11 numbering (model version 11.15.0)' \
+  && ok "the first output line names the numbering used (Mendix 11)" || bad "numbering line 11" "$(printf '%s\n' "$OUT11" | head -1)"
+printf '%s' "$OUT" | grep -q '^  NOTE-SIDE note line Annotation@360,100 -> Action@360,200 leaves the note left' \
+  && ! printf '%s' "$OUT11" | grep -q '^  NOTE-SIDE' \
+  && ok "mxcli's (2,0) note above its activity: NOTE-SIDE on Mendix 10, none on Mendix 11 (bottom of the note)" \
+  || bad "(2,0) above the activity, 10 vs 11" "$OUT11"
+# Every origin side of that note, on a Mendix 10 model, against the same line renumbered as the upgrade does
+# on a Mendix 11 model: the same NOTE-SIDE result. The note sits above its activity, so only its bottom faces it.
+for n in 0 1 2 3; do
+  A="$WORK/n10-$n"; B="$WORK/n11-$n"; model "$A"; model "$B" 11.15.0
+  C1="$(notes "$A" set "$n")"; C2="$(notes "$B" set "$n")"; C3="$(notes "$B" remap)"
+  R10="$(FACTS "$A" --module MyFirstModule | grep -c '^  NOTE-SIDE')"; R11="$(FACTS "$B" --module MyFirstModule | grep -c '^  NOTE-SIDE')"
+  WANT=1; [ "$n" = 1 ] && WANT=0
+  [ "$C1/$C2/$C3" = "1/1/1" ] && [ "$R10" = "$WANT" ] && [ "$R11" = "$WANT" ] \
+    && ok "origin side $n on Mendix 10 and the upgraded $n on Mendix 11 give the same NOTE-SIDE result ($WANT)" \
+    || bad "origin side $n: lines $C1/$C2/$C3, NOTE-SIDE on 10: $R10, on 11: $R11, want $WANT"
+done
+MU="$WORK/mu"; model "$MU"; "$PY" -c 'import sqlite3,sys; c=sqlite3.connect(sys.argv[1]); c.execute("DROP TABLE _MetaData"); c.commit()' "$MU/TestApp.mpr"
+OUTU="$(FACTS "$MU" --module MyFirstModule 2>&1)"; RC=$?
+[ "$RC" -eq 0 ] && printf '%s\n' "$OUTU" | head -1 | grep -qx '# note sides: Mendix 11 numbering (model version unreadable, assumed)' \
+  && ok "[guard] a version that cannot be read: Mendix 11 numbering, said on the first line, still exit 0" || bad "unreadable version rc=$RC" "$(printf '%s\n' "$OUTU" | head -1)"
+
 echo "== S: selection and faults =="
 OUT="$(FACTS "$M/mprcontents" MyFirstModule.ACT_Customer_Check ACT_Customer_Merge 2>&1)"; RC=$?
 [ "$RC" -eq 0 ] && [ "$(printf '%s\n' "$OUT" | grep -c '^# MyFirstModule\.')" = 2 ] \