diff --git a/crates/engine/src/js.rs b/crates/engine/src/js.rs
index b797bc8d83..6d344573a3 100644
--- a/crates/engine/src/js.rs
+++ b/crates/engine/src/js.rs
@@ -70,11 +70,15 @@ fn display(doc: &pdfcraft_cos::Document, f: &Field) -> i32 {
 
 /// A form field as scripts see it.
 pub fn field_state(doc: &pdfcraft_cos::Document, f: &Field) -> pdfcraft_js::FieldState {
+    field_state_with_display(f, display(doc, f))
+}
+
+fn field_state_with_display(f: &Field, display: i32) -> pdfcraft_js::FieldState {
     let mut s = pdfcraft_js::FieldState::new(f.name.clone(), kind(f.kind), f.value.clone());
     s.default = f.default.clone();
     s.readonly = f.read_only();
     s.required = f.has(pdfcraft_forms::flags::REQUIRED);
-    s.display = display(doc, f);
+    s.display = display;
     s.char_limit = f.max_len;
     s.options = match f.kind {
         FieldKind::CheckBox | FieldKind::Radio => f.widgets.iter().filter_map(|w| w.on_state.clone()).map(|o| (o.clone(), o)).collect(),
@@ -85,10 +89,19 @@ pub fn field_state(doc: &pdfcraft_cos::Document, f: &Field) -> pdfcraft_js::Fiel
 
 /// The changes a script made, against the fields as they were.
 fn changes(before: &[pdfcraft_js::FieldState], after: &[pdfcraft_js::FieldState]) -> Vec<FieldChange> {
+    if after.is_empty() {
+        return Vec::new();
+    }
+    // Outcomes contain only fields touched by the script. Index the original snapshot once,
+    // keeping the first match as the old linear lookup did for malformed duplicate names.
+    let mut by_name = std::collections::HashMap::with_capacity(before.len());
+    for field in before {
+        by_name.entry(field.name.as_str()).or_insert(field);
+    }
     after
         .iter()
         .filter_map(|a| {
-            let b = before.iter().find(|b| b.name == a.name)?;
+            let b = by_name.get(a.name.as_str())?;
             let c = FieldChange {
                 name: a.name.clone(),
                 value: (a.value != b.value).then(|| a.value.clone()),
@@ -106,14 +119,25 @@ pub struct JsRunner {
     pub doc: pdfcraft_js::DocInfo,
     pub doc_scripts: Vec<String>,
     pub output: JsOutput,
-    /// The fields in the document as the forms code last read them (for `display`).
+    /// Immutable document snapshot from the start of the edit (for widget `display`).
     cos: pdfcraft_cos::Document,
+    // `cos` is the same immutable pre-edit snapshot for the runner's lifetime. Cache only
+    // its scalar display readings, not COS objects: large forms exceed the object cache and
+    // otherwise reparse every widget for every calculation. Values/flags still come from
+    // the live `fields` snapshot on each call. A new edit gets a new runner and cache.
+    displays: std::collections::HashMap<pdfcraft_cos::ObjRef, i32>,
 }
 
 impl JsRunner {
     pub fn new(cos: &pdfcraft_cos::Document, file_name: &str) -> JsRunner {
         let doc = pdfcraft_js::DocInfo { file_name: file_name.to_string(), num_pages: pdfcraft_model::pages(cos).len(), page: 0, info: info(cos) };
-        JsRunner { doc, doc_scripts: pdfcraft_forms::document_scripts(cos), output: JsOutput::default(), cos: cos.clone() }
+        JsRunner {
+            doc,
+            doc_scripts: pdfcraft_forms::document_scripts(cos),
+            output: JsOutput::default(),
+            cos: cos.clone(),
+            displays: std::collections::HashMap::new(),
+        }
     }
 }
 
@@ -130,19 +154,28 @@ fn info(cos: &pdfcraft_cos::Document) -> Vec<(String, String)> {
 
 impl Scripts for JsRunner {
     fn run(&mut self, event: FieldEvent, script: &str, target: &Field, value: &str, fields: &[Field]) -> ScriptResult {
-        let states: Vec<_> = fields.iter().map(|f| field_state(&self.cos, f)).collect();
+        let states: Vec<_> = fields
+            .iter()
+            .map(|f| {
+                let display = match f.widgets.first() {
+                    Some(w) => *self.displays.entry(w.obj).or_insert_with(|| display(&self.cos, f)),
+                    None => pdfcraft_js::DISPLAY_VISIBLE,
+                };
+                field_state_with_display(f, display)
+            })
+            .collect();
+        let states = std::sync::Arc::new(states);
         let ev = pdfcraft_js::Event::field(event.name(), &target.name, value);
-        let o = pdfcraft_js::run(script, &ev, &self.doc, &states, &self.doc_scripts, Limits::default());
+        let o = pdfcraft_js::run_shared(script, &ev, &self.doc, std::sync::Arc::clone(&states), &self.doc_scripts, Limits::default());
         self.output.absorb(&o);
         if o.error.is_some() {
             // As in Acrobat, a failing script leaves the value alone (the error goes to the console).
             return ScriptResult { rc: true, value: value.to_string(), changes: Vec::new(), message: None };
         }
-        let after: Vec<_> = states.iter().map(|s| o.changed.iter().find(|c| c.name == s.name).unwrap_or(s).clone()).collect();
         ScriptResult {
             rc: o.rc,
             value: o.value.clone(),
-            changes: changes(&states, &after),
+            changes: changes(&states, &o.changed),
             message: (!o.rc).then(|| o.alerts.last().cloned()).flatten(),
         }
     }
@@ -213,8 +246,7 @@ impl Session {
             None => pdfcraft_js::Event::doc("Console"),
         };
         let mut o = pdfcraft_js::run(script, &event, &runner.doc, &states, &runner.doc_scripts, Limits::default());
-        let after: Vec<_> = states.iter().map(|s| o.changed.iter().find(|c| c.name == s.name).unwrap_or(s).clone()).collect();
-        let changes = changes(&states, &after);
+        let changes = changes(&states, &o.changed);
         let resets: Vec<Vec<String>> = o
             .requests
             .iter()
@@ -368,3 +400,47 @@ impl Session {
         Ok(names)
     }
 }
+
+#[cfg(test)]
+mod review_tests {
+    use super::*;
+
+    #[test]
+    fn duplicate_field_names_diff_against_first_original() {
+        let first = pdfcraft_js::FieldState::new("duplicate", pdfcraft_js::FieldType::Text, vec!["first".into()]);
+        let mut second = first.clone();
+        second.value = vec!["second".into()];
+        second.display = pdfcraft_js::DISPLAY_HIDDEN;
+        let before = vec![first.clone(), second.clone()];
+        assert!(changes(&before, std::slice::from_ref(&first)).is_empty());
+        let actual = changes(&before, std::slice::from_ref(&second));
+        assert_eq!(
+            actual,
+            vec![FieldChange {
+                name: "duplicate".into(),
+                value: Some(second.value),
+                display: Some(pdfcraft_js::DISPLAY_HIDDEN),
+                ..Default::default()
+            }]
+        );
+        let outcome = pdfcraft_js::run_shared(
+            "event.value = getField('duplicate').value; getField('duplicate').value = 'updated';",
+            &pdfcraft_js::Event::field("Calculate", "duplicate", ""),
+            &pdfcraft_js::DocInfo::default(),
+            std::sync::Arc::new(before.clone()),
+            &[],
+            Limits::default(),
+        );
+        assert_eq!(outcome.error, None);
+        assert_eq!(outcome.value, "first");
+        assert_eq!(outcome.changed.len(), 1);
+        assert_eq!(
+            changes(&before, &outcome.changed),
+            vec![FieldChange { name: "duplicate".into(), value: Some(vec!["updated".into()]), ..Default::default() }]
+        );
+        assert!(changes(&before, &[]).is_empty());
+        let mut missing = first;
+        missing.name = "missing".into();
+        assert!(changes(&before, &[missing]).is_empty());
+    }
+}
diff --git a/crates/engine/src/tests.rs b/crates/engine/src/tests.rs
index f125c41dac..67af975966 100644
--- a/crates/engine/src/tests.rs
+++ b/crates/engine/src/tests.rs
@@ -6106,3 +6106,185 @@ fn undo_never_makes_an_unsaved_document_clean() {
     s.undo(new).unwrap();
     assert!(!s.get(new).unwrap().dirty, "once saved, undo back to the saved state is clean");
 }
+/// Reproducible workload with 3,129 fields and 190 ordered custom calculations.
+/// No external PDF assets are needed; only in-memory contributor-original objects.
+fn synthetic_acroform(field_count: usize, calculation_count: usize) -> Vec<u8> {
+    use pdfcraft_cos::{Dict, Object, PdfString};
+    let mut doc = pdfcraft_cos::Document::open(Arc::new(fixture(6))).unwrap();
+    let mut refs = Vec::new();
+    let mut order = Vec::new();
+    for i in 0..field_count {
+        let name = match i {
+            0 => "ranks".to_string(),
+            1 => "trained".to_string(),
+            2 => "bonus".to_string(),
+            _ => format!("field{i}"),
+        };
+        let mut d = Dict::new();
+        d.set(b"Type".to_vec(), Object::name("Annot"));
+        d.set(b"Subtype".to_vec(), Object::name("Widget"));
+        d.set(b"FT".to_vec(), Object::name(if i == 1 { "Btn" } else { "Tx" }));
+        d.set(b"T".to_vec(), PdfString::text(&name));
+        d.set(b"V".to_vec(), if i == 1 { Object::name("Off") } else { Object::String(PdfString::text("0")) });
+        d.set(b"Rect".to_vec(), Object::Array([20, 20, 100, 40].map(Object::Int).to_vec()));
+        d.set(b"P".to_vec(), Object::Ref(pdfcraft_cos::ObjRef::new(4, 0)));
+        if (2..calculation_count + 2).contains(&i) {
+            let js = if i == 2 {
+                "var r = Number(getField('ranks').value); event.value = r + (r > 0 && getField('trained').isBoxChecked(0) ? 3 : 0);".to_string()
+            } else {
+                format!("event.value = Number(getField('{}').value) + 1;", if i == 3 { "bonus".to_string() } else { format!("field{}", i - 1) })
+            };
+            let mut a = Dict::new();
+            a.set(b"S".to_vec(), Object::name("JavaScript"));
+            a.set(b"JS".to_vec(), PdfString::text(&js));
+            let mut aa = Dict::new();
+            aa.set(b"C".to_vec(), Object::Dict(a));
+            d.set(b"AA".to_vec(), Object::Dict(aa));
+        }
+        let r = doc.add(Object::Dict(d));
+        refs.push(Object::Ref(r));
+        if (2..calculation_count + 2).contains(&i) {
+            order.push(Object::Ref(r));
+        }
+    }
+    doc.update_dict(pdfcraft_cos::ObjRef::new(4, 0), |d| d.set(b"Annots".to_vec(), Object::Array(refs.clone()))).unwrap();
+    let mut af = Dict::new();
+    af.set(b"Fields".to_vec(), Object::Array(refs));
+    af.set(b"CO".to_vec(), Object::Array(order));
+    doc.update_dict(doc.root().unwrap(), |d| d.set(b"AcroForm".to_vec(), Object::Dict(af))).unwrap();
+    pdfcraft_cos::write_incremental(&doc, &Default::default()).unwrap()
+}
+
+#[test]
+#[ignore = "manual performance measurement: 3,129 fields, 190 scripts"]
+fn acroform_interaction_benchmark() {
+    let bytes = Arc::new(synthetic_acroform(3129, 190));
+    // Optional local export for comparing CLI build profiles. Never overwrite an input.
+    if let Some(path) = std::env::var_os("PDFCRAFT_BENCH_EXPORT") {
+        use std::io::Write;
+        let mut file = std::fs::OpenOptions::new().write(true).create_new(true).open(path).unwrap();
+        file.write_all(&bytes).unwrap();
+    }
+    let mut s = Session::new();
+    let id = s.open("synthetic-large-form.pdf", None, bytes, None).unwrap();
+    let value = |s: &Session, name: &str| s.get(id).unwrap().form.iter().find(|f| f.name == name).unwrap().value.clone();
+    for (name, v) in [
+        ("ranks", FieldValue::Text("1".into())),
+        ("trained", FieldValue::Check(true)),
+        ("ranks", FieldValue::Text("2".into())),
+        ("trained", FieldValue::Check(false)),
+        ("ranks", FieldValue::Text("0".into())),
+    ] {
+        let started = std::time::Instant::now();
+        s.apply(id, Edit::SetFieldValue { name: name.into(), value: v }).unwrap();
+        eprintln!("BENCH {name}: {:.3} ms", started.elapsed().as_secs_f64() * 1000.0);
+        assert!(s.take_js_output(id).errors.is_empty());
+    }
+    assert_eq!(value(&s, "bonus"), ["0"]);
+    assert_eq!(value(&s, "field191"), ["189"]);
+    assert_eq!(value(&s, "field3128"), ["0"]);
+}
+
+#[test]
+fn checkbox_and_rank_changes_recalculate_ordered_dependents() {
+    let mut s = Session::new();
+    let id = s.open("skills.pdf", None, Arc::new(synthetic_acroform(8, 4)), None).unwrap();
+    for (name, input, bonus) in [
+        ("trained", FieldValue::Check(true), "0"),
+        ("ranks", FieldValue::Text("1".into()), "4"),
+        ("trained", FieldValue::Check(false), "1"),
+        ("ranks", FieldValue::Text("2".into()), "2"),
+        ("trained", FieldValue::Check(true), "5"),
+        ("ranks", FieldValue::Text("0".into()), "0"),
+    ] {
+        s.apply(id, Edit::SetFieldValue { name: name.into(), value: input }).unwrap();
+        let fields = &s.get(id).unwrap().form;
+        let value = |name: &str| fields.iter().find(|f| f.name == name).unwrap().value.clone();
+        assert_eq!(value("bonus"), [bonus]);
+        assert_eq!(value("field5"), [(bonus.parse::<u32>().unwrap() + 3).to_string()]);
+        assert_eq!(value("field7"), ["0"]);
+        assert!(s.take_js_output(id).errors.is_empty());
+    }
+    s.undo(id).unwrap();
+    assert_eq!(s.get(id).unwrap().form.iter().find(|f| f.name == "bonus").unwrap().value, ["5"]);
+    let saved = s.save_bytes(id).unwrap();
+    let reopened = s.open("saved.pdf", None, saved, None).unwrap();
+    assert_eq!(s.get(reopened).unwrap().form.iter().find(|f| f.name == "bonus").unwrap().value, ["5"]);
+    let appearance = |s: &Session| {
+        let doc = s.get(id).unwrap();
+        let widget = doc.form.iter().find(|f| f.name == "bonus").unwrap().widgets.first().unwrap().obj;
+        doc.editor.as_ref().unwrap().cos.get(widget).as_dict().unwrap().get(b"AP").cloned()
+    };
+    let before = appearance(&s);
+    assert!(before.is_some());
+    s.apply(id, Edit::SetFieldValue { name: "field7".into(), value: FieldValue::Text("9".into()) }).unwrap();
+    assert_eq!(appearance(&s), before, "unchanged calculated values keep their appearance streams");
+}
+
+#[test]
+fn dynamic_calculation_reads_and_side_effects_reach_later_fields() {
+    use pdfcraft_cos::{Dict, Object, PdfString};
+    let mut cos = pdfcraft_cos::Document::open(Arc::new(synthetic_acroform(8, 4))).unwrap();
+    let fields = pdfcraft_forms::fields(&cos);
+    let bonus = fields.iter().find(|f| f.name == "bonus").unwrap();
+    let selector = fields.iter().find(|f| f.name == "field7").unwrap();
+    cos.update_dict(selector.obj, |d| d.set(b"V".to_vec(), PdfString::text("ranks"))).unwrap();
+    let mut a = Dict::new();
+    a.set(b"S".to_vec(), Object::name("JavaScript"));
+    a.set(b"JS".to_vec(), PdfString::text(
+        "var key = getField('field7').value; var v = Number(getField(key).value); getField('field6').value = v * 2; getField('field6').readonly = true; event.value = v + 1;"
+    ));
+    let mut aa = Dict::new();
+    aa.set(b"C".to_vec(), Object::Dict(a));
+    cos.update_dict(bonus.obj, |d| d.set(b"AA".to_vec(), Object::Dict(aa))).unwrap();
+    let bytes = pdfcraft_cos::write_incremental(&cos, &Default::default()).unwrap();
+    let mut s = Session::new();
+    let id = s.open("dynamic.pdf", None, Arc::new(bytes), None).unwrap();
+    s.apply(id, Edit::SetFieldValue { name: "ranks".into(), value: FieldValue::Text("4".into()) }).unwrap();
+    let fields = &s.get(id).unwrap().form;
+    let field = |name: &str| fields.iter().find(|f| f.name == name).unwrap();
+    assert_eq!(field("bonus").value, ["5"]);
+    assert_eq!(field("field5").value, ["8"], "later calculations see the ordered result");
+    assert_eq!(field("field6").value, ["8"], "side-effect value applied");
+    assert!(field("field6").read_only(), "side-effect property applied");
+    assert_eq!(field("field7").value, ["ranks"]);
+    assert!(s.take_js_output(id).errors.is_empty());
+}
+
+#[test]
+fn scripted_display_changes_survive_warm_cache_and_new_edit() {
+    use pdfcraft_forms::{FieldEvent, Scripts};
+    let mut cos = pdfcraft_cos::Document::open(Arc::new(synthetic_acroform(2, 0))).unwrap();
+    for (display, flags) in [(1, 6), (3, 36), (2, 0), (0, 4)] {
+        let fields = pdfcraft_forms::fields(&cos);
+        let target = fields.iter().find(|f| f.name == "ranks").unwrap();
+        let other = fields.iter().find(|f| f.name == "trained").unwrap();
+        let old_display = crate::js::field_state(&cos, other).display;
+        let mut runner = crate::js::JsRunner::new(&cos, "synthetic.pdf");
+        // Populate the display cache before a script writes the same field.
+        let warm = runner.run(FieldEvent::Calculate, "event.value = getField('trained').display;", target, "", &fields);
+        assert_eq!(warm.value, old_display.to_string());
+        let result = runner.run(
+            FieldEvent::Calculate,
+            &format!("getField('trained').display = {display}; event.value = getField('trained').display;"),
+            target,
+            "",
+            &fields,
+        );
+        assert_eq!(result.value, display.to_string());
+        assert_eq!(result.changes.len(), 1);
+        assert_eq!(result.changes.first().unwrap().display, Some(display));
+        pdfcraft_forms::apply_script_changes(&mut cos, &result.changes, &mut pdfcraft_forms::NoScripts).unwrap();
+        assert_eq!(cos.get(other.widgets.first().unwrap().obj).as_dict().unwrap().int(b"F").unwrap() & 38, flags);
+        // Within an edit the original implementation reads an immutable pre-edit COS snapshot.
+        let same_edit = runner.run(FieldEvent::Calculate, "event.value = getField('trained').display;", target, "", &fields);
+        assert_eq!(same_edit.value, old_display.to_string());
+        // The next edit must rebuild its cache from the newly written annotation flags.
+        let mut next = crate::js::JsRunner::new(&cos, "synthetic.pdf");
+        let read = next.run(FieldEvent::Calculate, "event.value = getField('trained').display;", target, "", &pdfcraft_forms::fields(&cos));
+        assert_eq!(read.value, display.to_string());
+        assert!(read.changes.is_empty());
+        assert!(runner.output.errors.is_empty());
+        assert!(next.output.errors.is_empty());
+    }
+}
diff --git a/crates/js/src/lib.rs b/crates/js/src/lib.rs
index d8d9fa8bac..785345c672 100644
--- a/crates/js/src/lib.rs
+++ b/crates/js/src/lib.rs
@@ -25,6 +25,7 @@
 
 use std::cell::RefCell;
 use std::collections::BTreeSet;
+use std::sync::Arc;
 
 use boa_engine::object::builtins::{JsArray, JsFunction};
 use boa_engine::object::{FunctionObjectBuilder, ObjectInitializer};
@@ -194,7 +195,7 @@ impl Default for Limits {
 }
 
 struct Host {
-    fields: Vec<FieldState>,
+    fields: Arc<Vec<FieldState>>,
     changed: BTreeSet<usize>,
     alerts: Vec<String>,
     console: Vec<String>,
@@ -286,7 +287,8 @@ fn read(ctx: &Context, name: &JsString) -> JsResult<FieldState> {
 fn write(ctx: &Context, name: &JsString, change: impl FnOnce(&mut FieldState)) -> JsResult<()> {
     let i = field_index(ctx, name)?;
     let mut h = host(ctx)?.borrow_mut();
-    change(&mut h.fields[i]);
+    let field = Arc::make_mut(&mut h.fields).get_mut(i).ok_or_else(|| error("the field is no longer available"))?;
+    change(field);
     h.changed.insert(i);
     Ok(())
 }
@@ -1098,6 +1100,13 @@ fn refuse(source: &str) -> Option<String> {
 /// Run `script` for `event`, after the document-level scripts (`doc_scripts`, which define the
 /// functions field scripts call), with the form's `fields`.
 pub fn run(script: &str, event: &Event, doc: &DocInfo, fields: &[FieldState], doc_scripts: &[String], limits: Limits) -> Outcome {
+    run_shared(script, event, doc, Arc::new(fields.to_vec()), doc_scripts, limits)
+}
+
+/// Like [`run`], sharing an immutable initial field snapshot. Scripts still get a fresh
+/// context; the first field write copies the snapshot, so neither the caller nor another
+/// script can observe uncommitted writes. Read-only calculations avoid copying the whole form.
+pub fn run_shared(script: &str, event: &Event, doc: &DocInfo, fields: Arc<Vec<FieldState>>, doc_scripts: &[String], limits: Limits) -> Outcome {
     let failed = |why: String| Outcome { rc: true, value: event.value.clone(), change: event.change.clone(), error: Some(why), ..Default::default() };
     if let Some(why) = std::iter::once(script).chain(doc_scripts.iter().map(String::as_str)).find_map(refuse) {
         return failed(why);
@@ -1119,12 +1128,12 @@ pub fn run(script: &str, event: &Event, doc: &DocInfo, fields: &[FieldState], do
     run_here(script, event, doc, fields, doc_scripts, limits)
 }
 
-fn run_here(script: &str, event: &Event, doc: &DocInfo, fields: &[FieldState], doc_scripts: &[String], limits: Limits) -> Outcome {
+fn run_here(script: &str, event: &Event, doc: &DocInfo, fields: Arc<Vec<FieldState>>, doc_scripts: &[String], limits: Limits) -> Outcome {
     let mut ctx = Context::default();
     ctx.runtime_limits_mut().set_loop_iteration_limit(limits.loop_iterations);
     ctx.runtime_limits_mut().set_recursion_limit(limits.recursion);
     ctx.insert_data(RefCell::new(Host {
-        fields: fields.to_vec(),
+        fields,
         changed: BTreeSet::new(),
         alerts: Vec::new(),
         console: Vec::new(),
diff --git a/crates/js/src/tests.rs b/crates/js/src/tests.rs
index 61d3a71e51..1d2856b8bd 100644
--- a/crates/js/src/tests.rs
+++ b/crates/js/src/tests.rs
@@ -1132,3 +1132,37 @@ mod session {
         assert!(o.abandoned || o.error.as_deref().is_some_and(|e| e.contains("too long")), "{o:?}");
     }
 }
+
+#[test]
+fn shared_field_snapshots_isolate_writes_errors_and_script_globals() {
+    let initial = Arc::new(fields());
+    let before = initial.as_ref().clone();
+    let event = Event::field("Calculate", "total", "");
+    let write = run_shared(
+        "var scriptLocal = 1; getField('qty').value = 9; getField('agree').checkThisBox(0, true); event.value = getField('qty').value;",
+        &event,
+        &doc(),
+        Arc::clone(&initial),
+        &[],
+        Limits::default(),
+    );
+    assert_eq!(write.error, None);
+    assert_eq!(write.value, "9");
+    assert_eq!(write.changed.iter().find(|f| f.name == "qty").unwrap().value, ["9"]);
+    assert_eq!(write.changed.iter().find(|f| f.name == "agree").unwrap().value, ["Yes"]);
+    assert_eq!(*initial, before, "script writes cannot mutate the caller's snapshot");
+    let failed = run_shared("getField('qty').value = 99; throw new Error('rejected');", &event, &doc(), Arc::clone(&initial), &[], Limits::default());
+    assert!(failed.error.is_some());
+    assert_eq!(*initial, before, "an exception cannot leak partial writes into the initial snapshot");
+    let read = run_shared(
+        "event.value = typeof scriptLocal + ':' + getField('qty').value + ':' + getField('agree').isBoxChecked(0);",
+        &event,
+        &doc(),
+        initial,
+        &[],
+        Limits::default(),
+    );
+    assert_eq!(read.error, None);
+    assert_eq!(read.value, "undefined:4:false");
+    assert!(read.changed.is_empty());
+}
diff --git a/docs/form-performance.md b/docs/form-performance.md
new file mode 100644
index 0000000000..c00adf0913
--- /dev/null
+++ b/docs/form-performance.md
@@ -0,0 +1,278 @@
+# AcroForm JavaScript performance
+
+Status: implemented; proposed as an independent performance PR. Last reviewed and updated:
+2026-10-19. Measurements below are historical synthetic evidence collected before this split,
+not fresh measurements on upstream main. No private document is needed for the benchmark.
+
+## Architecture
+
+- `crates/ui-egui/src/forms_ui.rs` queues `Edit::SetFieldValue` for a text commit or checkbox
+  toggle. `crates/automation/src/forms.rs::form_fill` uses the same edit API.
+- `crates/engine/src/lib.rs::Session::apply` runs edits on a copy, preserving rollback and undo.
+  `run_edit` calls `pdfcraft_forms::set_value_with` with an engine `JsRunner`.
+- `crates/forms/src/lib.rs` validates and writes the edited value, then runs one calculation
+  pass. `/CO` precedes calculated fields omitted from `/CO`. Each script sees prior writes.
+  There is no dependency graph or dependency-based filtering: dynamic references and script
+  side effects are evaluated on every pass.
+- `crates/engine/src/js.rs::JsRunner::run` snapshots the fields, invokes `pdfcraft_js::run_shared`,
+  then converts changed JavaScript field states into form updates.
+- `crates/js/src/lib.rs::run_shared` starts a scoped worker per script. `run_here` builds a fresh Boa
+  context, installs the API, evaluates document scripts and evaluates the action. No script or
+  bytecode cache is present in this AcroForm path. XFA uses a separate session implementation.
+- `crates/forms/src/scripting.rs::apply_changes` applies script side effects before the next
+  calculation. It rereads the field snapshot when side effects occur. Calculation results
+  equal to the existing value already skip value writes and appearance generation.
+- Checkboxes write the field `/V` and widget `/AS`; they enter the same calculation pass as
+  text edits. JavaScript exposes the export value or `Off`, and `isBoxChecked`.
+- `Session::refresh_scoped` writes an incremental working revision, rereads fields and rebuilds
+  the render pool. `crates/ui-egui/src/editing.rs::apply_edit` compares old/new fields and
+  invalidates affected pages (including dependent fields on other pages). It does not blindly
+  invalidate every page for an ordinary AcroForm fill. Rendering uses page/tile requests;
+  widget appearance streams are separate from cached page rasters.
+
+## Findings and changes
+
+The dominant measured cost was the bridge, not executing calculation logic:
+
+1. `JsRunner` expanded a sparse `Outcome::changed` into a complete second snapshot. `changes`
+   then searched the first snapshot linearly for **each** field, even when JavaScript had not
+   changed any field directly. That repeated quadratic comparison ran after every calculation.
+   It now compares the sparse outcome only, with an indexed original snapshot when needed.
+   Empty outcomes return immediately; unchanged writes still produce no form updates. The
+   console/button path uses the same conversion.
+2. Once that was removed, field-state construction dominated. Reading every widget's display
+   flags on every script traversed the same immutable pre-edit COS snapshot. This workload
+   exceeds COS's 1,024-object editing-cache cap, repeatedly reparsing widget dictionaries.
+   The runner now memoizes only those scalar display readings by widget reference. Each edit
+   creates a new runner; current values, read-only/required flags and options still come from
+   the current field snapshot. No global cache or object-cache limit changes were needed.
+
+All calculation actions still execute in the same order. No dependency filtering, script
+removal, context reuse, bytecode caching or recursive scheduling changes were introduced.
+Equal calculated values already skipped appearance regeneration; this behavior remains.
+The Release investigation also found a full field-state copy inside the JavaScript host on
+**every** script invocation. `pdfcraft_js::run_shared` now accepts an immutable shared snapshot;
+`Arc::make_mut` detaches it on the first field write. The engine retains the initial snapshot
+for comparing effects. Read-only calculations no longer copy every field again. Each script
+still gets a fresh context and thread; exceptions, globals and uncommitted writes remain
+isolated. Scripts that write fields retain the previous copy isolation. The existing borrowed
+`run` API remains available. No new dependencies were added.
+
+## Measurements
+
+Historically measured under Rust 1.99.1 on Linux x86_64 with the default optimized development
+profile, two-job builds. Serial benchmark runs had no concurrent compilation. These are
+**synthetic engine measurements, not measurements of the original Pathfinder PDF or GUI input
+latency**. Five edits: ranks 1, checkbox on, ranks 2, checkbox off, ranks 0. Every edit runs all
+190 custom calculations among 3,129 fields on a six-page document. Widgets are colocated
+on the first page; the fixture is an engine calculation stress test, not a layout replica.
+
+| Synchronous edit | Before (ms) | After (ms) |
+|---|---:|---:|
+| ranks 1 | 4,425.274 | 424.465 |
+| checkbox on | 4,378.642 | 458.716 |
+| ranks 2 | 4,291.131 | 521.721 |
+| checkbox off | 4,391.942 | 492.851 |
+| ranks 0 | 4,322.570 | 453.392 |
+| Median | **4,378.642** | **458.716** |
+
+Median speedup: **9.55×**, approximately **89.5% less synchronous editing time**. The synthetic
+scripts are deliberately small; this does not predict the exact result for other documents.
+Raw logs and the baseline test executable are retained locally in `target/form-performance/`.
+
+Historical serial runs collected with optional instrumentation (mean per edit; inclusive
+nested spans). That instrumentation is not included in this PR:
+
+| Phase | Before (ms) | After (ms) |
+|---|---:|---:|
+| JavaScript, including thread/context setup | 225.934 | 201.584 |
+| Changed-field comparison | 2,814.805 | 0.024 |
+| Entire script dispatch, all 190 scripts | 4,320.643 | 352.733 |
+| Calculation pass, including changed appearances | 4,399.084 | 390.958 |
+| Direct field write | 0.201 | 0.196 |
+| Appearance generation, including Format | 61.225 | 22.981 |
+| Form event, including direct write and calculations | 4,411.767 | 402.981 |
+| Engine refresh/serialization | 23.499 | 21.826 |
+| Whole engine edit | 4,442.144 | 431.539 |
+
+An intermediate run after sparse comparison alone measured 1,193.419 ms/edit in field-state
+construction. After display memoization it measured 101.821 ms/edit. The intermediate run's
+whole engine edit averaged 1,592.101 ms. Appearance time also benefits from the changed cache
+working set; there was no change to which calculated appearances are generated.
+
+Rendering and compositor time are not included in these edit numbers. The original sheet and
+GUI acceptance run remain required to quantify them.
+
+## Release comparison and remaining cost
+
+Historical comparison used the repository's Release profile (thin LTO, one code-generation
+unit) and an optional standalone profiler, which is not included in this PR.
+A fresh in-memory fixture was exported with create-new semantics to
+`target/form-performance/synthetic-large.pdf` (SHA-256
+`300431eeae117458d96e52877e12a9e63ea8dcbb8ad4500da717ba00ebb2f1e1`). It remained unchanged.
+The example alternated `ranks` between 1 and 2 for six samples per process, always executing
+all 190 scripts. Page rendering was timed separately.
+
+Initial matched comparison, before the additional shared-snapshot change:
+
+| Build | Samples | Median engine edit (ms) |
+|---|---:|---:|
+| Development | 6 | 460.697 |
+| Release | 6 | 408.358 |
+
+Release alone reduced the median by about 11%. A detailed Release run measured 106.947 ms/edit
+constructing field states, 174.696 ms in JavaScript setup/execution, and 19.808 ms refreshing
+engine view data. Calculation-order setup was only 0.319 ms. That supported investigating the
+host's extra copy, rather than changing calculation scheduling or disabling calculations.
+
+The pre-change Release profiler was retained, then compared with the shared-snapshot Release
+profiler in alternating runs (old, new, old, new). No compilation ran during measurement.
+
+| Release version | Samples | Median edit (ms) | Range (ms) |
+|---|---:|---:|---:|
+| Before shared snapshots | 12 | 381.380 | 353.468–447.833 |
+| With shared snapshots | 12 | 267.063 | 258.545–313.777 |
+
+The additional optimization reduced median engine time by **30.0%** (1.43× speedup).
+This is a separate comparison from the earlier five-edit benchmark; do not combine their
+speedup factors. The page-render/setup cost in the initial matched comparison was about
+6–9 ms for this deliberately overlapping synthetic layout, not for the original sheet.
+
+After shared snapshots, a historical Release run with optional instrumentation averaged:
+
+| Phase | Mean per edit (ms) |
+|---|---:|
+| Field-state construction | 75.754 |
+| JavaScript setup/execution | 83.198 |
+| Complete script dispatch | 200.480 |
+| Calculation pass | 237.780 |
+| Calculated-value writes, including appearances | 25.089 |
+| Appearances | 23.712 |
+| All complete field enumerations (three per edit) | 27.387 |
+| Calculation-order setup | 0.305 |
+| Engine refresh | 19.128 |
+| Whole engine edit | 272.807 |
+
+These inclusive timings show that scheduling is now negligible on the fixture. The main
+remaining costs are constructing/disposing field snapshots and creating/executing isolated
+JavaScript contexts. Further snapshot caching or context reuse would need additional semantic
+work and real-document evidence; no dependency-based skipping or global/context reuse was
+introduced. The current shared-snapshot change was retained because its paired measurement
+showed a substantial improvement and isolation regressions passed.
+
+**Remaining limit:** these numbers do not quantify real-document or compositor latency.
+A representative document and delayed field are still needed for a reproducible engine/render/UI
+comparison. Prior manual acceptance on the integrated release reported responsive field editing;
+that is not a timing measurement or independent validation of this split.
+
+## Reproduce
+
+Use the same compiler, build profile and machine for before/after runs. The default development
+profile optimizes workspace crates at level 1 and dependencies at level 2. Build with two jobs
+and a separate target directory, then run the benchmark without concurrent compilation:
+
+```sh
+CARGO_TARGET_DIR=target/acroform-benchmark cargo test -p pdfcraft-engine -j 2 --lib acroform_interaction_benchmark --no-run
+CARGO_TARGET_DIR=target/acroform-benchmark cargo test -p pdfcraft-engine -j 2 --lib acroform_interaction_benchmark -- --ignored --nocapture
+```
+
+The benchmark creates a six-page, 3,129-field/190-calculation workload from contributor-original
+PDF objects in memory and times five synchronous edits with `Instant`. It also checks the final
+calculated chain and an independent field. Widgets overlap on page one; this is an engine stress
+test, not a realistic page layout. There is no machine-dependent timing threshold in CI.
+
+To export a synthetic fixture for a standalone comparison, set `PDFCRAFT_BENCH_EXPORT` to a new
+filename when running the ignored benchmark. Its parent must exist; export refuses to overwrite
+any existing file. The ordinary benchmark writes no PDF.
+
+## Validation
+
+Regression coverage includes shared-snapshot write/error/global isolation, duplicate field names,
+scripted display changes with a warm cache and a fresh edit, ordered checkbox/rank calculations,
+dynamic field lookup, property/value side effects, undo/save/reopen, and unchanged appearances.
+The existing JavaScript and engine suites cover the unchanged public scripting behavior.
+
+Before the split, the ignored benchmark passed both before and after optimization. Earlier
+scoped native/WASM checks, Clippy, layering, asset and parity checks, UI form/JavaScript tests,
+and automation `javascript_through_tools` checks also passed on the mixed review history.
+Those historical checks do not certify these independent branches or hosted CI.
+
+## Independent branch validation
+
+Validated on upstream `a12fe8f4116984ebbc870eb514346ea922233316` using a fresh, isolated
+Cargo target directory, offline locked dependencies and two build jobs:
+
+- All five focused regressions passed separately: duplicate names, warm/new-edit display
+  caches, ordered checkbox/rank dependents, dynamic lookup/side effects and shared-snapshot
+  write/error/global isolation.
+- `cargo test -p pdfcraft-js -p pdfcraft-engine -p pdfcraft-forms -j 2`: engine 285 unit
+  and 9 integration tests, JavaScript 47 tests, and forms 80 tests passed. Two unit timing
+  probes and two private-file acceptance tests were ignored in the regular suite.
+- The ignored 3,129-field/190-calculation benchmark was then run explicitly and passed;
+  its final calculated chain and independent field remained correct. This is a diagnostic
+  timing run, not a new before/after comparison on this upstream base.
+- Scoped `cargo clippy --all-targets -- -D warnings` and WASM checks of those three
+  crates passed, as did formatting, whitespace, layering, asset and parity checks.
+- `javascript_through_tools` passed. A CLI built in the isolated target directory then
+  reproduced bonuses 1 → 4 → 5 → 2 → 0; save/reopen and `page_render` passed, and the
+  synthetic page image was inspected.
+
+These are local checks, not hosted CI. No GUI executable was rebuilt and no private PDF
+was used for this independent validation.
+
+## Rebase validation
+
+Revalidated on upstream `932bf9c6bfdd63e2a21bb831696a3447b9def9cf` (2026-10-18).
+The performance implementation and all five focused regressions are unchanged from the
+previously reviewed fix. Upstream's newer tests and CSV documentation are retained.
+
+- Engine: 286 unit and 9 integration tests passed; JavaScript: 48 passed; forms: 82 passed.
+  Two unit timing probes and two private-file acceptance tests remain ignored in the
+  regular suite.
+- All five focused regressions and the explicit 3,129-field/190-calculation benchmark passed.
+- Scoped all-targets Clippy with `-D warnings`, WASM, formatting, whitespace, layering,
+  asset and parity checks passed.
+- `javascript_through_tools` and a freshly built CLI's fill/save/reopen/render smoke check
+  passed. The bonus chain remained 1 → 4 → 5 → 2 → 0, and the synthetic render was inspected.
+
+Validation used a separate Cargo target directory. No GUI build or private PDF was needed;
+these are local checks, not hosted CI or a new before/after performance comparison.
+
+## PR #3318 refresh
+
+Rebased onto freshly fetched upstream main
+`e397efa85451be00f8d56df7649807fe31e019c7` on 2026-10-19. Both roadmap conflicts
+retain upstream's menu, button-label and content-editing notes alongside the performance
+notes. Every performance source, regression-test and parity hunk is exactly the original
+performance patch applied to current upstream; upstream's additional tests are retained.
+Main advanced after initial validation on `43550c44`; the non-overlapping redaction update
+was incorporated and all checks below were repeated on the final base.
+
+- All five focused regressions passed.
+- Engine: 287 unit and 15 integration tests passed. JavaScript: 49 tests passed. Forms:
+  84 tests passed. Doc tests passed. The regular suite still ignored two timing probes
+  and two private-file tests; the synthetic benchmark was then run explicitly and passed.
+- The 3,129-field/190-calculation benchmark preserved the expected calculated values. Its
+  five diagnostic edits took 317.347–328.488 ms; this is not a new before/after comparison.
+- Scoped all-targets Clippy with warnings denied, WASM, formatting, whitespace, layering,
+  assets and parity passed.
+- Automation JavaScript and CLI checks passed: bonuses 1 → 4 → 5 → 2 → 0, save/reopen,
+  and page rendering. The synthetic render was inspected.
+
+Builds used an isolated Cargo target outside the worktrees. No GUI executable was built,
+no private PDF was used, and the compatibility branch was not changed. These are local
+checks, not hosted CI or a new performance speedup claim.
+
+## Revision history
+
+- 2026-10-19: refresh PR #3318 onto upstream `e397efa8`, retain upstream roadmap additions
+  and all performance changes, and repeat focused, scoped, benchmark and headless validation.
+
+- 2026-10-18: rebase onto upstream `932bf9c6`; retain all performance source edits and regression
+  tests; preserve upstream CSV documentation and repeat scoped native/WASM and headless checks.
+- 2026-10-18: separate performance changes for review on current upstream main; add duplicate-name
+  and scripted-display coverage; keep historical optional instrumentation results as evidence only.
+- 2026-10-18: retain shared snapshots after a paired 30% synthetic Release improvement and
+  verify snapshot isolation.
+- 2026-10-17: remove quadratic outcome conversion and repeated display reads; record the
+  synthetic before/after benchmark.
diff --git a/parity/acrobat-features.toml b/parity/acrobat-features.toml
index 7aca7d5c8b..4415d7ac5d 100644
--- a/parity/acrobat-features.toml
+++ b/parity/acrobat-features.toml
@@ -4596,7 +4596,7 @@ milestone = "M6"
 status = "shipped"
 commands = ["form.field.properties"]
 tools = ["form_set_props", "form_fill"]
-evidence = ["crates/forms/src/tests.rs::formats_validation_and_calculations_run_like_acrobat"]
+evidence = ["crates/forms/src/tests.rs::formats_validation_and_calculations_run_like_acrobat", "crates/engine/src/tests.rs::checkbox_and_rank_changes_recalculate_ordered_dependents", "crates/engine/src/tests.rs::dynamic_calculation_reads_and_side_effects_reach_later_fields"]
 notes = "/CO order, then other calculated fields; recalculated after every fill and Clear form"
 
 [[feature]]