Variables: a data set with no value for any bound variable shows a status message

Choosing a data set in the Variables panel, stepping through data sets
with the panel's arrows, and running `dataset.select`, `dataset.next`
and `dataset.prev` applied nothing and showed no message when no object
was bound to a variable that the data set has a value for. This
happens, for example, after Import Data Sets… defines new variables
without binding them.

The app now shows "nothing in data set A was applied: no object is
bound to a variable it has a value for" in the status bar, and the three
commands add it to the reply as `warning`. A data set with a value for
a bound variable gives no warning even when every object bound to it is
locked; `skipped` counts those objects. What applying a data set does
is unchanged.

Docs: docs/mcp.md, docs/target-app-parity.md.

crates/engine/src/cmd/variablecmds.rs

@@ -459,7 +459,7 @@ fn dataset_list(s: &mut Session, _p: &Value) -> Result<Value> {
 
 fn apply_dataset(s: &mut Session, name: &str) -> Result<Value> {
     const C: &str = "dataset.select";
-    let (applied, skipped, mut typed) = s.edit("Apply Data Set", |d, sel| {
+    let (applied, skipped, mut typed, bound) = s.edit("Apply Data Set", |d, sel| {
         let ds = d.variables.dataset(name).cloned().ok_or_else(|| bad(C, format!("no dataset named `{name}`")))?;
         // One object may be driven by several variables; walk them in definition order, so two
         // text variables on the same object land the later one's value and both are counted.
@@ -467,6 +467,8 @@ fn apply_dataset(s: &mut Session, name: &str) -> Result<Value> {
         for v in &d.variables.variables {
             pairs.extend(d.variables.objects_of(&v.name).into_iter().map(|id| (id, v.name.clone())));
         }
+        // Whether an object is bound to a variable that this row has a value for (the reply's `warning`).
+        let bound = pairs.iter().any(|(_, var)| ds.values.contains_key(var));
         let mut applied = 0usize;
         let mut skipped = 0usize;
         // Collects the type objects whose text the row rewrote, for the reply's `overflow`.
@@ -514,12 +516,18 @@ fn apply_dataset(s: &mut Session, name: &str) -> Result<Value> {
         d.variables.active_dataset = Some(name.to_string());
         // Every selected object left hidden or locked, such as one the row hid, is deselected.
         sel.deselect_uneditable(d);
-        Ok((applied, skipped, typed))
+        Ok((applied, skipped, typed, bound))
     })?;
     // Two text variables on one object rewrite it twice; it is named once.
     typed.sort_unstable();
     typed.dedup();
-    Ok(super::typecmd::with_overflow(s, &typed, json!({"name": name, "applied": applied, "skipped": skipped})))
+    let mut out = super::typecmd::with_overflow(s, &typed, json!({"name": name, "applied": applied, "skipped": skipped}));
+    // A row with no value for any bound variable applies nothing. The reply then carries a
+    // `warning`, which the app shows in its status bar.
+    if !bound {
+        out["warning"] = json!(format!("nothing in data set {name} was applied: no object is bound to a variable it has a value for"));
+    }
+    Ok(out)
 }
 
 fn dataset_select(s: &mut Session, p: &Value) -> Result<Value> {
@@ -1037,7 +1045,7 @@ pub fn specs() -> Vec<CommandSpec> {
             "Select Data Set",
             ["Window", "Variables"],
             None,
-            "{name} apply a dataset in one undo step, deselecting every selected object left hidden or locked (the key object too) → {applied, skipped, overflow?: [id…] (area type and type on a path whose text it rewrote and that no longer fits; left out when all fits)}",
+            "{name} apply a dataset in one undo step, deselecting every selected object left hidden or locked (the key object too) → {applied, skipped, overflow?: [id…] (area type and type on a path whose text it rewrote and that no longer fits; left out when all fits), warning? (when no object is bound to a variable that the dataset has a value for, so nothing was applied)}",
             has_doc,
             dataset_select
         ),