Find and Replace: bound the text a replacement builds

Each occurrence copied the whole replacement string, so a cell full of
short matches and a long replacement expanded tens of kilobytes of
input into about a gigabyte, in both the case-sensitive and the
case-insensitive paths.

Replacement now builds the new text only while it stays within
Excel's 32,767 characters per cell (counted as UTF-8 bytes, 4 each at
most) and within 256 MiB for the whole command, checking before each
copy. Past either limit the command fails and changes nothing.

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, HashSet};
+use std::collections::HashMap;
 
 use gridcraft_core::{CellRef, RangeRef, Value, compare};
 use gridcraft_model::{CfOperator, CfRule, CfValueKind, Color, CondFormat, PatternType, Sheet, Style, Underline, Workbook, default_icon_thresholds};
@@ -70,31 +70,15 @@ 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 {
-        if budget == 0 {
-            break;
-        }
-        let r = match used.and_then(|u| r.intersection(&u)) {
+        let r = match sh.used_range().and_then(|u| r.intersection(&u)) {
             Some(x) => x,
             None => continue,
         };
-        // 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;
+        for c in r.iter().take(2_000_000) {
             #[cfg(test)]
             SCANNED.with(|n| n.set(n.get() + 1));
             let v = sh.value(c);
@@ -294,8 +278,7 @@ impl CfCache {
                     _ => None,
                 },
                 CfRule::ColorScale { stops } => {
-                    // 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)) {
+                    if let Some(n) = v.as_f64() {
                         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])))
@@ -577,50 +560,6 @@ 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/edit/find.rs

@@ -104,6 +104,13 @@ pub(super) fn find(s: &mut Session, p: &Json) -> Result<Json> {
     Ok(json!({"cell": c.a1(), "sheet": si, "value": t, "count": hits.len()}))
 }
 
+/// Longest text one replacement may give a cell, in bytes: Excel's 32,767 characters, each up to
+/// four bytes. Every occurrence copies the replacement text, so a short cell full of matches and a
+/// long replacement would otherwise expand to gigabytes.
+const REPLACE_CELL_MAX_BYTES: usize = 32_767 * 4;
+/// Most bytes of replaced text one Replace All builds across all its cells.
+const REPLACE_TOTAL_MAX_BYTES: usize = 256 * 1024 * 1024;
+
 pub(super) fn replace(s: &mut Session, p: &Json) -> Result<Json> {
     let what = str_param(p, "what").unwrap_or("").to_string();
     let with = str_param(p, "with").unwrap_or("").to_string();
@@ -119,6 +126,7 @@ pub(super) fn replace(s: &mut Session, p: &Json) -> Result<Json> {
     let n = edit(s, |cx| {
         let sheets: Vec<usize> = if workbook { (0..cx.wb.sheets.len()).collect() } else { vec![cx.wb.active_sheet()] };
         let mut n = 0;
+        let mut total = 0usize;
         for si in sheets {
             let cells: Vec<(CellRef, String)> =
                 cx.wb.sheet(si).map(|sh| sh.cells.iter().map(|(c, _)| (c, sh.input_text(c))).collect()).unwrap_or_default();
@@ -130,12 +138,19 @@ pub(super) fn replace(s: &mut Session, p: &Json) -> Result<Json> {
                     continue;
                 }
                 let new = if whole {
-                    with.clone()
+                    Some(with.clone()).filter(|w| w.len() <= REPLACE_CELL_MAX_BYTES)
                 } else if case {
-                    text.replace(&what, &with)
+                    replace_cs_within(&text, &what, &with, REPLACE_CELL_MAX_BYTES)
                 } else {
-                    replace_ci(&text, &what, &with)
+                    replace_ci_within(&text, &what, &with, REPLACE_CELL_MAX_BYTES)
                 };
+                let Some(new) = new else {
+                    return Err(EngineError::Other(format!("The replacement makes the text of {} too long.", c.a1())));
+                };
+                total = total.saturating_add(new.len());
+                if total > REPLACE_TOTAL_MAX_BYTES {
+                    return Err(EngineError::Other("The replacements make too much text. Replace in a smaller range.".into()));
+                }
                 // A Python cell's text is its code: the replacement edits the code and the cell
                 // stays a Python cell. Code that would no longer fit in a PY formula is left as it was.
                 let old_py = cx.wb.sheet(si).and_then(|sh| sh.cell(c)).and_then(|x| x.py.clone());
@@ -190,7 +205,37 @@ fn folds_in_place(s: &str) -> bool {
     })
 }
 
+#[cfg(test)]
 pub(super) fn replace_ci(text: &str, what: &str, with: &str) -> String {
+    replace_ci_within(text, what, with, usize::MAX).unwrap_or_default()
+}
+
+/// `text` with every `what` replaced by `with`, or `None` once the result passes `limit` bytes
+/// (checked before each copy, so nothing larger is allocated).
+pub(super) fn replace_cs_within(text: &str, what: &str, with: &str, limit: usize) -> Option<String> {
+    if what.is_empty() {
+        return (text.len() <= limit).then(|| text.to_string());
+    }
+    let mut out = String::new();
+    let mut i = 0;
+    for (pos, m) in text.match_indices(what) {
+        if out.len() + (pos - i) + with.len() > limit {
+            return None;
+        }
+        out.push_str(text.get(i..pos).unwrap_or(""));
+        out.push_str(with);
+        i = pos + m.len();
+    }
+    let rest = text.get(i..).unwrap_or("");
+    if out.len() + rest.len() > limit {
+        return None;
+    }
+    out.push_str(rest);
+    Some(out)
+}
+
+/// Case-insensitive [`replace_cs_within`].
+pub(super) fn replace_ci_within(text: &str, what: &str, with: &str, limit: usize) -> Option<String> {
     let lower = text.to_lowercase();
     let w = what.to_lowercase();
     // Slicing `text` at positions found in `lower` is only sound when folding kept every byte
@@ -198,15 +243,22 @@ pub(super) fn replace_ci(text: &str, what: &str, with: &str) -> String {
     // boundaries (U+212A KELVIN SIGN shrinks 3→1 bytes, U+0130 grows 2→3), where the tail used
     // to be dropped. Those fall back to the case-sensitive replace, as length changes already do.
     if lower.len() != text.len() || w.is_empty() || !folds_in_place(text) || !folds_in_place(what) {
-        return text.replace(what, with);
+        return replace_cs_within(text, what, with, limit);
     }
     let mut out = String::new();
     let mut i = 0;
     while let Some(pos) = lower.get(i..).and_then(|s| s.find(&w)) {
+        if out.len() + pos + with.len() > limit {
+            return None;
+        }
         out.push_str(text.get(i..i + pos).unwrap_or(""));
         out.push_str(with);
         i += pos + w.len();
     }
-    out.push_str(text.get(i..).unwrap_or(""));
-    out
+    let rest = text.get(i..).unwrap_or("");
+    if out.len() + rest.len() > limit {
+        return None;
+    }
+    out.push_str(rest);
+    Some(out)
 }

crates/engine/src/cmd/edit/tests.rs

@@ -353,6 +353,36 @@ fn cut_paste_that_overflows_the_grid_moves_nothing() {
     assert_eq!(get(&mut s, "XFD2")["value"], json!(2.0));
 }
 
+#[test]
+fn replace_refuses_text_that_would_grow_past_a_cell() {
+    // 10,000 matches × a 1,000-character replacement would make a 10 MB cell (and a few such
+    // cells gigabytes); the replacement is refused before that text is built.
+    let mut s = Session::new();
+    s.new_workbook();
+    let long = "a".repeat(10_000);
+    s.execute("range.setValues", json!({"range": "A1", "values": [[long.clone()], ["abc"]]})).unwrap();
+    for case in [false, true] {
+        let err = s.execute("edit.replace", json!({"what": "a", "with": "x".repeat(1_000), "matchCase": case})).expect_err("too long");
+        assert!(err.to_string().contains("too long"), "{err}");
+        assert_eq!(s.execute("cell.get", json!({"cell": "A1"})).unwrap()["value"], json!(long));
+        assert_eq!(s.execute("cell.get", json!({"cell": "A2"})).unwrap()["value"], json!("abc"));
+    }
+    // A replacement that stays within a cell's limit still happens.
+    s.execute("edit.replace", json!({"what": "b", "with": "x".repeat(100)})).unwrap();
+    assert_eq!(s.execute("cell.get", json!({"cell": "A2"})).unwrap()["value"], json!(format!("a{}c", "x".repeat(100))));
+}
+
+#[test]
+fn bounded_replace_stops_at_its_limit() {
+    use super::find::{replace_ci_within, replace_cs_within};
+    assert_eq!(replace_cs_within("abab", "a", "xy", 6).as_deref(), Some("xybxyb"));
+    assert_eq!(replace_cs_within("abab", "a", "xy", 5), None);
+    assert_eq!(replace_ci_within("AbAb", "a", "xy", 6).as_deref(), Some("xybxyb"));
+    assert_eq!(replace_ci_within("AbAb", "a", "xy", 5), None);
+    assert_eq!(replace_cs_within("abc", "z", "", 3).as_deref(), Some("abc"));
+    assert_eq!(replace_cs_within("abc", "z", "", 2), None);
+}
+
 #[test]
 fn replace_case_insensitive() {
     assert_eq!(replace_ci("Apple apple APPLE", "apple", "pear"), "pear pear pear");