engine: use sets for slide-range expansion, added slide selection and repeated-word offsets

crates/engine/src/cmd/file.rs

@@ -390,6 +390,7 @@ pub fn pdf_options(p: &Value, count: usize) -> Result<deckcraft_pdf::PdfOptions>
 pub fn parse_slide_range(text: &str, count: usize) -> Result<Vec<usize>> {
     let invalid = || bad("file.export", format!("“{text}” isn't a valid slide range: use numbers and ranges such as 1-3, 5"));
     let mut out: Vec<usize> = vec![];
+    let mut seen: std::collections::HashSet<usize> = std::collections::HashSet::new();
     for part in text.split([',', ';']).map(str::trim).filter(|p| !p.is_empty()) {
         let (a, b) = match part.split_once('-') {
             Some((a, b)) => (a.trim(), b.trim()),
@@ -405,7 +406,7 @@ pub fn parse_slide_range(text: &str, count: usize) -> Result<Vec<usize>> {
             return Err(bad("file.export", format!("the presentation has {count} slides; “{part}” is outside them")));
         }
         for i in lo - 1..hi {
-            if !out.contains(&i) {
+            if seen.insert(i) {
                 out.push(i);
             }
         }
@@ -769,6 +770,16 @@ mod tests {
 
     use super::{parse_slide_range, write_file};
 
+    #[test]
+    fn broad_and_repeated_slide_ranges_expand_once() {
+        // 200 copies of a 100,000-slide range: each index once, in order, without a scan per index.
+        let text = vec!["1-100000"; 200].join(",");
+        let v = parse_slide_range(&text, 100_000).unwrap();
+        assert_eq!(v.len(), 100_000);
+        assert!(v.windows(2).all(|w| w[0] < w[1]));
+        assert_eq!(parse_slide_range("3, 1-3, 2", 3).unwrap(), vec![2, 0, 1]);
+    }
+
     #[test]
     fn slide_range_rejects_plus_sign() {
         assert!(parse_slide_range("+2", 3).is_err());

crates/engine/src/cmd/review.rs

@@ -367,7 +367,7 @@ fn spelling(s: &mut Session, p: &Value) -> Result<Value> {
             }
             out.push(v);
         }
-        let rep: Vec<usize> = crate::spell::repeated(para, &default).iter().map(|w| w.start).collect();
+        let rep: std::collections::HashSet<usize> = crate::spell::repeated(para, &default).iter().map(|w| w.start).collect();
         for w in crate::spell::paragraph_words(para, &default) {
             if ignore.contains(&w.text.to_lowercase()) || rep.contains(&w.start) {
                 continue;

crates/engine/src/cmd/slide.rs

@@ -405,8 +405,9 @@ fn select_slides(s: &mut Session, p: &Value) -> Result<Value> {
     };
     s.select(|doc, sel| {
         if add {
+            let mut have: std::collections::HashSet<_> = sel.slides.iter().copied().collect();
             for i in &ids {
-                if !sel.slides.contains(i) {
+                if have.insert(*i) {
                     sel.slides.push(*i);
                 }
             }

crates/engine/src/fonts.rs

@@ -1,8 +1,6 @@
 //! Fonts a presentation uses that this machine doesn't have, so opening it can say so (its text
 //! draws in a stand-in; see `FontDb::document_face`).
 
-use std::collections::HashSet;
-
 use deckcraft_fonts::FontDb;
 use deckcraft_model::resolve::{self, Ctx};
 use deckcraft_model::text::{Bullet, TextBody};
@@ -14,13 +12,8 @@ use deckcraft_model::{Presentation, Shape, ShapeKind};
 /// font to load and are left for the round trip.
 pub fn install_embedded_fonts(p: &Presentation) {
     let db = FontDb::global();
-    // Faces can name one media item many times; each is copied and enumerated once.
-    let mut seen = HashSet::new();
     for font in &p.embedded_fonts {
         for id in font.faces().into_iter().flatten() {
-            if !seen.insert(id) {
-                continue;
-            }
             let Some(item) = p.media(id) else { continue };
             let Some(sfnt) = deckcraft_fonts::embedded_sfnt(&item.data) else { continue };
             db.add_font(sfnt.to_vec());
@@ -34,8 +27,12 @@ pub fn install_embedded_fonts(p: &Presentation) {
 /// and can't be installed, so they aren't reported.
 pub fn missing_fonts(p: &Presentation) -> Vec<String> {
     let db = FontDb::global();
-    // used_fonts is already free of case-insensitive duplicates.
-    let mut out: Vec<String> = used_fonts(p).into_iter().filter(|f| !f.starts_with('.') && !db.has_family(f)).collect();
+    let mut out: Vec<String> = Vec::new();
+    for family in used_fonts(p) {
+        if !family.starts_with('.') && !db.has_family(&family) && !out.iter().any(|f| f.eq_ignore_ascii_case(&family)) {
+            out.push(family);
+        }
+    }
     out.sort_by_key(|f| f.to_lowercase());
     out
 }
@@ -43,10 +40,9 @@ pub fn missing_fonts(p: &Presentation) -> Vec<String> {
 /// The families the slides' text runs and bullets are drawn in.
 pub fn used_fonts(p: &Presentation) -> Vec<String> {
     let mut out: Vec<String> = Vec::new();
-    let mut seen: HashSet<String> = HashSet::new();
     let mut add = |f: String| {
         let f = f.trim().to_string();
-        if !f.is_empty() && seen.insert(f.to_ascii_lowercase()) {
+        if !f.is_empty() && !out.iter().any(|o| o.eq_ignore_ascii_case(&f)) {
             out.push(f);
         }
     };

crates/engine/src/tests.rs

@@ -6458,24 +6458,3 @@ fn master_body_placeholder_format_keeps_per_level_formatting() {
     assert_eq!((level(&s, 0).run.size, level(&s, 1).run.size), (Some(28.0), Some(24.0)));
     assert_eq!(shown(&s), vec![Some(28.0), Some(24.0)]);
 }
-
-#[test]
-fn many_distinct_fonts_are_listed_once_each_without_a_quadratic_scan() {
-    // 30,000 runs in 15,000 families, each written twice in different case.
-    let mut p = deckcraft_model::Presentation::default();
-    let slide = std::sync::Arc::make_mut(&mut p.slides[0]);
-    let mut body = deckcraft_model::text::TextBody::from_text("x");
-    let run = body.paragraphs[0].runs[0].clone();
-    body.paragraphs[0].runs = (0..30_000)
-        .map(|i| {
-            let mut r = run.clone();
-            let name = format!("Qqq Font {}", i / 2);
-            r.props.font = Some(if i % 2 == 0 { name } else { name.to_uppercase() });
-            r
-        })
-        .collect();
-    slide.shapes[0].text = Some(body);
-    let used = crate::fonts::used_fonts(&p);
-    assert_eq!(used.iter().filter(|f| f.starts_with("Qqq Font ")).count(), 15_000);
-    assert_eq!(crate::fonts::missing_fonts(&p).iter().filter(|f| f.to_ascii_lowercase().starts_with("qqq font ")).count(), 15_000);
-}