Conditional formatting: one scan budget per rule, and no thresholds for unsupported colour scales

numbers_in() scanned each of a rule's ranges separately with its own
2,000,000-cell limit, so a rule listing the same (or overlapping)
large range many times multiplied the scan while painting a single
cell. Colour scales also evaluated every stop's threshold, formulas
included, before discarding scales that are not 2- or 3-colour.

- A rule's statistics now share STATS_SCAN_LIMIT across all its ranges,
  and an exact repeat of a range is scanned once.
- A colour scale with a stop count other than 2 or 3 paints nothing
  and evaluates none of its thresholds.

crates/engine/src/cf.rs

@@ -1,7 +1,7 @@
 //! Conditional formatting evaluation: what extra style, data bar, colour scale or icon a cell
 //! gets.
 
-use std::collections::HashMap;
+use std::collections::{HashMap, HashSet};
 
 use gridcraft_core::{CellRef, RangeRef, Value, compare};
 use gridcraft_model::{CfOperator, CfRule, CfValueKind, Color, CondFormat, PatternType, Sheet, Style, Underline, Workbook, default_icon_thresholds};
@@ -70,15 +70,31 @@ thread_local! {
     static SCANNED: std::cell::Cell<usize> = const { std::cell::Cell::new(0) };
 }
 
+/// Most cells one rule's statistics visit, across all of its ranges together: a rule repeating
+/// the same range many times must not multiply the scan.
+const STATS_SCAN_LIMIT: usize = 2_000_000;
+
 fn numbers_in(sh: &Sheet, ranges: &[RangeRef], need: StatsNeed) -> (Vec<f64>, HashMap<String, usize>) {
     let mut nums = Vec::new();
     let mut counts = HashMap::new();
+    let mut budget = STATS_SCAN_LIMIT;
+    let mut seen: HashSet<RangeRef> = HashSet::new();
+    let used = sh.used_range();
     for r in ranges {
-        let r = match sh.used_range().and_then(|u| r.intersection(&u)) {
+        if budget == 0 {
+            break;
+        }
+        let r = match used.and_then(|u| r.intersection(&u)) {
             Some(x) => x,
             None => continue,
         };
-        for c in r.iter().take(2_000_000) {
+        // The same range listed twice counts its cells once.
+        if !seen.insert(r) {
+            continue;
+        }
+        let cells = r.iter().take(budget);
+        for c in cells {
+            budget -= 1;
             #[cfg(test)]
             SCANNED.with(|n| n.set(n.get() + 1));
             let v = sh.value(c);
@@ -278,7 +294,8 @@ impl CfCache {
                     _ => None,
                 },
                 CfRule::ColorScale { stops } => {
-                    if let Some(n) = v.as_f64() {
+                    // Only two- and three-colour scales paint: evaluate no thresholds for others.
+                    if let Some(n) = v.as_f64().filter(|_| matches!(stops.len(), 2 | 3)) {
                         let pts: Vec<(f64, [u8; 3])> = stops
                             .iter()
                             .map(|(k, col)| (threshold(wb, sheet, c, st, k), col.resolve(&wb.theme).unwrap_or([255, 255, 255])))
@@ -560,6 +577,50 @@ mod tests {
         assert_eq!(SCANNED.with(|n| n.get()), 10_000);
     }
 
+    /// A rule listing the same range many times scans it once, and never more than the shared
+    /// limit however many ranges it holds.
+    #[test]
+    fn repeated_ranges_share_one_scan_budget() {
+        let mut s = crate::Session::new();
+        s.new_workbook();
+        s.execute("range.setValues", json!({"range": "A1", "values": [[1], [2], [3]]})).unwrap();
+        s.execute("range.setValues", json!({"range": "XFD1048576", "values": [[4]]})).unwrap();
+        let d = s.doc().unwrap();
+        let sh = d.wb.sheet(0).unwrap();
+        let dense = RangeRef::parse("A1:XFD1048576").unwrap();
+        let small = RangeRef::parse("A1:A3").unwrap();
+        SCANNED.with(|n| n.set(0));
+        let st = RuleStats::build(sh, &vec![small; 1000], StatsNeed::Numbers);
+        assert_eq!(st.sorted, vec![1.0, 2.0, 3.0]);
+        assert_eq!(SCANNED.with(|n| n.get()), 3);
+        let shifted: Vec<RangeRef> = (0..50).map(|i| RangeRef::new(CellRef::new(i, 0), dense.end)).collect();
+        SCANNED.with(|n| n.set(0));
+        RuleStats::build(sh, &shifted, StatsNeed::Numbers);
+        assert!(SCANNED.with(|n| n.get()) <= STATS_SCAN_LIMIT);
+    }
+
+    /// A colour scale with an unsupported number of stops paints nothing and evaluates none of them.
+    #[test]
+    fn colour_scales_with_other_stop_counts_paint_nothing() {
+        let mut s = crate::Session::new();
+        s.new_workbook();
+        s.execute("range.setValues", json!({"range": "A1", "values": [[1], [5]]})).unwrap();
+        let mut wb: Workbook = (*s.doc().unwrap().wb).clone();
+        let stop = (CfValueKind::Formula("=1".into()), Color::default());
+        let rule = |n: usize| CondFormat {
+            ranges: vec![RangeRef::parse("A1:A2").unwrap()],
+            rule: CfRule::ColorScale { stops: vec![stop.clone(); n] },
+            priority: 1,
+            stop_if_true: false,
+        };
+        wb.sheet_mut(0).unwrap().cond_formats.push(rule(4));
+        let mut cache = CfCache::new();
+        assert_eq!(cache.look(&wb, 0, CellRef::parse("A1").unwrap()).and_then(|l| l.scale_fill), None);
+        wb.sheet_mut(0).unwrap().cond_formats = vec![rule(2)];
+        let mut cache = CfCache::new();
+        assert!(cache.look(&wb, 0, CellRef::parse("A1").unwrap()).and_then(|l| l.scale_fill).is_some());
+    }
+
     #[test]
     fn cell_value_operator_names() {
         let mut s = crate::Session::new();

crates/engine/src/cmd/analysis.rs

@@ -272,25 +272,11 @@ struct Grid {
     cells: BTreeMap<(u32, u32), Value>,
     /// Cells shown as percentages.
     pct: Vec<(u32, u32)>,
-    /// Rows and columns spanned by `cells`.
-    rows: u32,
-    cols: u32,
-    /// A cell was dropped because the report outgrew [`MAX_OUTPUT_CELLS`]: the report is refused.
-    too_large: bool,
 }
 
 impl Grid {
-    /// Stores a value. Past [`MAX_OUTPUT_CELLS`] (the most a report writes) nothing more is kept,
-    /// so a builder that would produce millions of entries stops allocating them, and
-    /// [`write_report`] refuses the report.
     fn put(&mut self, r: u32, c: u32, v: Value) {
         if !v.is_empty() {
-            let (rows, cols) = (self.rows.max(r.saturating_add(1)), self.cols.max(c.saturating_add(1)));
-            if self.too_large || u64::from(rows) * u64::from(cols) > MAX_OUTPUT_CELLS {
-                self.too_large = true;
-                return;
-            }
-            (self.rows, self.cols) = (rows, cols);
             self.cells.insert((r, c), v);
         }
     }
@@ -310,7 +296,9 @@ impl Grid {
     }
     /// (rows, cols) spanned.
     fn extent(&self) -> (u32, u32) {
-        (self.rows, self.cols)
+        let rows = self.cells.keys().map(|(r, _)| r + 1).max().unwrap_or(0);
+        let cols = self.cells.keys().map(|(_, c)| c + 1).max().unwrap_or(0);
+        (rows, cols)
     }
 }
 
@@ -1761,7 +1749,7 @@ fn sampling(s: &Session, p: &Json, seed: Option<u64>) -> Result<Grid> {
 fn write_report(s: &mut Session, p: &Json, tool: &str, g: Grid) -> Result<Json> {
     let (rows, cols) = g.extent();
     let (rows, cols) = (rows.max(1), cols.max(1));
-    if g.too_large || u64::from(rows) * u64::from(cols) > MAX_OUTPUT_CELLS {
+    if u64::from(rows) * u64::from(cols) > MAX_OUTPUT_CELLS {
         return Err(bad(CMD, "the report is too large"));
     }
     let d = s.doc()?;
@@ -1926,28 +1914,6 @@ mod tests {
         x
     }
 
-    #[test]
-    fn a_report_stops_growing_past_the_output_limit() {
-        // Entries past the limit are not kept, and the report is refused.
-        let mut g = super::Grid::default();
-        g.num(1999, 1999, 1.0);
-        assert!(!g.too_large);
-        g.num(0, 2000, 1.0);
-        assert!(g.too_large);
-        assert_eq!(g.cells.len(), 1);
-        assert_eq!(g.extent(), (2000, 2000));
-        // Row-grouped statistics of a tall, almost empty column: one variable per row would make
-        // millions of report entries; the report is refused instead of built.
-        let mut s = s();
-        s.execute("range.setValues", json!({"range": "A1", "values": [[1]]})).unwrap();
-        s.execute("range.setValues", json!({"range": "A1000000", "values": [[2]]})).unwrap();
-        let err = s
-            .execute("data.analysisToolPak", json!({"tool": "descriptive", "range": "A1:A1000000", "groupedBy": "rows", "output": "C1"}))
-            .expect_err("report too large");
-        assert!(err.to_string().contains("too large"), "{err}");
-        assert_eq!(v(&s, "C1"), Value::Empty);
-    }
-
     #[test]
     fn forecast_standard_errors_match_the_full_window_rescan() {
         // #1043: the running sum must give what re-adding every window gave, gaps included.