address review: keep modern-comments precedence for already-owned parts

crates/pdf/src/lib.rs

@@ -498,9 +498,6 @@ impl Backdrops {
 /// Most pixels one slide raster may have; it is RGBA, and the backdrop and the content are two rasters.
 const MAX_RASTER_PIXELS: f64 = 64_000_000.0;
 
-/// Most ruled note lines one handout area gets (a letter page holds about 30).
-const MAX_NOTE_LINES: usize = 500;
-
 /// Render scale for a slide of `slide` units drawn `box_width` points wide at `dpi`, with the total pixel count capped.
 fn raster_scale(box_width: f64, slide: (f64, f64), dpi: f64) -> f64 {
     let scale = (box_width / 72.0 * dpi.clamp(36.0, 600.0)) / slide.0;
@@ -576,17 +573,20 @@ impl Exporter<'_> {
 
     /// Invisible glyphs exactly over the rasterised text.
     fn text_layer(&mut self, s: &mut Surface, pt: &PlacedText, view: Affine, visible: bool) -> Result<()> {
-        // Each paragraph's characters, gathered once: a long paragraph wraps into many runs.
-        let mut para_chars: std::collections::HashMap<usize, Vec<char>> = std::collections::HashMap::new();
         for run in &pt.layout.runs {
             if run.glyphs.is_empty() || run.size <= 0.0 {
                 continue;
             }
-            let chars = para_chars.entry(run.para).or_insert_with(|| {
-                pt.body.paragraphs.get(run.para).map(|p| p.runs.iter().flat_map(|r| r.text.chars()).collect()).unwrap_or_default()
-            });
-            let (a, b) = (run.chars.0.min(chars.len()), run.chars.1.min(chars.len()));
-            let text: String = chars.get(a..b.max(a)).unwrap_or_default().iter().collect();
+            let text: String = pt
+                .body
+                .paragraphs
+                .get(run.para)
+                .map(|p| p.runs.iter().map(|r| r.text.as_str()).collect::<String>())
+                .unwrap_or_default()
+                .chars()
+                .skip(run.chars.0)
+                .take(run.chars.1.saturating_sub(run.chars.0))
+                .collect();
             if text.trim().is_empty() {
                 continue;
             }
@@ -666,28 +666,19 @@ impl Exporter<'_> {
         let lh = size * 1.3;
         let mut y = area.y0 + size;
         s.set_fill(Some(grey(0x20, 1.0)));
-        let space = width(" ");
         'outer: for para in text.split('\n') {
-            // The line's width is kept as words are added, not measured again for each word.
             let mut line = String::new();
-            let mut line_w = 0.0;
             for word in para.split(' ') {
-                let word_w = width(word);
-                let cand_w = if line.is_empty() { word_w } else { line_w + space + word_w };
-                if cand_w > area.width() && !line.is_empty() {
+                let cand = if line.is_empty() { word.to_string() } else { format!("{line} {word}") };
+                if width(&cand) > area.width() && !line.is_empty() {
                     if y > area.y1 {
                         break 'outer;
                     }
                     s.draw_text(Point::from_xy(area.x0 as f32, y as f32), font.clone(), size as f32, &line, false, TextDirection::Auto);
                     y += lh;
                     line = word.to_string();
-                    line_w = word_w;
                 } else {
-                    if !line.is_empty() {
-                        line.push(' ');
-                    }
-                    line.push_str(word);
-                    line_w = cand_w;
+                    line = cand;
                 }
             }
             if y > area.y1 {
@@ -704,26 +695,20 @@ impl Exporter<'_> {
 
     /// Ruled lines for handwritten notes (3-per-page handouts).
     fn note_lines(&self, s: &mut Surface, r: Rect) {
-        // Counted, not stepped: at huge coordinates adding 22 leaves y unchanged, and a huge page
-        // would want millions of rules.
-        let first = r.y0 + 18.0;
-        let count = ((r.y1 - first) / 22.0).floor() + 1.0;
-        if count.is_nan() || count < 1.0 {
-            return;
-        }
+        let mut y = r.y0 + 18.0;
         s.set_stroke(Some(krilla::paint::Stroke {
             paint: krilla::color::rgb::Color::new(0xB0, 0xB0, 0xB0).into(),
             width: 0.5,
             ..Default::default()
         }));
-        for i in 0..(count.min(MAX_NOTE_LINES as f64) as usize) {
-            let y = first + 22.0 * i as f64;
+        while y <= r.y1 {
             let mut p = kurbo::BezPath::new();
             p.move_to((r.x0, y));
             p.line_to((r.x1, y));
             if let Some(path) = to_path(&p) {
                 s.draw_path(&path);
             }
+            y += 22.0;
         }
         s.set_stroke(None);
     }

crates/pdf/src/tests.rs

@@ -677,16 +677,3 @@ fn raster_scale_caps_total_pixels_not_each_side() {
     let s = raster_scale(720.0, (960.0, 540.0), 200.0);
     assert!((s - 720.0 / 72.0 * 200.0 / 960.0).abs() < 1e-12);
 }
-
-#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
-#[cfg_attr(not(target_arch = "wasm32"), test)]
-fn huge_paper_and_long_notes_export_in_bounded_work() {
-    // At 1e19 pt adding a rule's 22 pt leaves y unchanged: the rules are counted, not stepped. A narrow page keeps the raster small.
-    let p = deck(1);
-    let _ = export(&p, &PdfOptions { paper: (612.0, 1.0e19), ..opts(PageLayout::Handouts { per_page: 3 }) });
-    // 20,000 short words of notes, wrapped with the line width kept as words are added.
-    let mut p = deck(1);
-    Arc::make_mut(&mut p.slides[0]).notes = TextBody::from_text(&"ab ".repeat(20_000));
-    let bytes = export(&p, &opts(PageLayout::Notes)).expect("export");
-    assert!(bytes.starts_with(b"%PDF-"));
-}

crates/pptx/src/read.rs

@@ -663,52 +663,55 @@ fn read_slides(
     let pkg = imp.pkg;
     let fallback_layout = p.masters.first().and_then(|m| m.layouts.first()).map(|l| l.id).unwrap_or_default();
     let mut file_slide_ids: HashMap<u32, SlideId> = HashMap::new();
-    let mut notes_by_target: HashMap<String, TextBody> = HashMap::new();
+    // A notes or comments part belongs to one slide; one shared by many must not be read (and
+    // copied) once per slide that names it.
+    let mut owned_parts: HashSet<String> = HashSet::new();
     for (file_id, path, id) in slide_ids {
         let Some(doc) = docs.get(&opc::norm(path)) else { continue };
         let part = Part::new(pkg, path, doc);
         let layout = part.rels.first_of("slideLayout").and_then(|r| layout_by_part.get(&opc::norm(&r.target)).copied()).unwrap_or(fallback_layout);
         let mut slide = read_slide(imp, &part, doc, *id, layout);
-        if let Some(notes) = slide_notes(imp, &part, &mut notes_by_target) {
+        if let Some(notes) = slide_notes(imp, &part, &mut owned_parts) {
             slide.notes = notes;
         }
-        read_slide_comments(pkg, pres_part, &part, &mut slide);
+        read_slide_comments(pkg, pres_part, &part, &mut slide, &mut owned_parts);
         file_slide_ids.insert(*file_id, *id);
         p.slides.push(Arc::new(slide));
     }
     file_slide_ids
 }
 
-/// The slide's notes text, read once per notes part.
-fn slide_notes(imp: &mut Imp, part: &Part, cache: &mut HashMap<String, TextBody>) -> Option<TextBody> {
+/// The slide's notes text. A notes part is the notes of the first slide that names it.
+fn slide_notes(imp: &mut Imp, part: &Part, owned: &mut HashSet<String>) -> Option<TextBody> {
     let n = part.rels.first_of("notesSlide")?;
-    let key = opc::norm(&n.target);
-    if let Some(cached) = cache.get(&key) {
-        return Some(cached.clone());
+    if !owned.insert(opc::norm(&n.target)) {
+        log::warn!("pptx: notes part {} is named by more than one slide; kept on the first", n.target);
+        return None;
     }
     let pkg = imp.pkg;
     let nd = load(pkg, &n.target)?;
     let np = Part::new(pkg, &n.target, &nd);
-    let notes = read_notes(imp, &np, &nd.root);
-    cache.insert(key, notes.clone());
-    Some(notes)
+    Some(read_notes(imp, &np, &nd.root))
 }
 
 /// Comments. A modern part and the legacy part can describe the same threads (PowerPoint
 /// writes both): the modern one wins, its flattened legacy copy is skipped.
-fn read_slide_comments(pkg: &Package, pres_part: &Part, part: &Part, slide: &mut Slide) {
-    let mut parts: Vec<(Doc, bool)> = vec![];
+/// Each comments part is read once, for the first slide that names it.
+fn read_slide_comments(pkg: &Package, pres_part: &Part, part: &Part, slide: &mut Slide, owned: &mut HashSet<String>) {
+    let mut parts: Vec<(Doc, bool, bool)> = vec![];
     for c in part.rels.all_of("comments") {
+        // An already-owned part is not re-read, but its kind still decides modern precedence.
+        let fresh = owned.insert(opc::norm(&c.target));
         if let Some(cd) = pkg.xml(&c.target) {
             // The content type names the modern part; a package that types it only through a
             // Default entry instead of an Override is still modern when its root is p188.
             let modern = pkg.content_type(&c.target).as_deref() == Some(crate::write::CT_COMMENTS_MODERN) || cd.root.prefix() == "p188";
-            parts.push((cd, modern));
+            parts.push((cd, modern, fresh));
         }
     }
-    let has_modern = parts.iter().any(|(_, m)| *m);
-    for (cd, modern) in parts {
-        if has_modern && !modern {
+    let has_modern = parts.iter().any(|(_, m, _)| *m);
+    for (cd, modern, fresh) in parts {
+        if !fresh || (has_modern && !modern) {
             continue;
         }
         slide.comments.extend(read_comments(pkg, pres_part, &cd.root, modern));

crates/pptx/tests/it/comments.rs

@@ -146,6 +146,33 @@ fn comments_on_several_slides_round_trip() {
     assert_eq!(q.slides[2].comments[0].text, "Slide three");
 }
 
+/// A slide that names a modern part another slide already owns, plus its own legacy part, must
+/// not import the legacy copy: modern precedence holds even when the modern part is not re-read.
+#[test]
+fn owned_modern_part_still_suppresses_later_legacy() {
+    let mut p = deck();
+    let layout = p.slides[0].layout;
+    let s2 = defaults::new_slide(&mut p, layout).expect("slide");
+    p.slides.push(Arc::new(s2));
+    Arc::make_mut(&mut p.slides[1]).comments = vec![Comment { author: "Cara".into(), text: "Slide two".into(), ..Default::default() }];
+    let bytes = deckcraft_pptx::export(&p).expect("export");
+    let rels2 = part(&bytes, "ppt/slides/_rels/slide2.xml.rels").replace("modernComment2.xml", "modernComment1.xml");
+    let mut z = zip::ZipArchive::new(std::io::Cursor::new(&bytes)).expect("zip");
+    let mut out = zip::ZipWriter::new(std::io::Cursor::new(Vec::new()));
+    for i in 0..z.len() {
+        let mut f = z.by_index(i).expect("entry");
+        out.start_file(f.name(), zip::write::SimpleFileOptions::default()).expect("start");
+        if f.name() == "ppt/slides/_rels/slide2.xml.rels" {
+            std::io::Write::write_all(&mut out, rels2.as_bytes()).expect("write");
+        } else {
+            std::io::copy(&mut f, &mut out).expect("copy");
+        }
+    }
+    let q = deckcraft_pptx::import(&out.finish().expect("finish").into_inner()).expect("import");
+    assert_eq!(q.slides[0].comments.len(), 2);
+    assert!(q.slides[1].comments.is_empty(), "{:?}", q.slides[1].comments);
+}
+
 /// A legacy-only package (what older tools read and write) still imports: 1/8 pt positions,
 /// flattened replies.
 #[test]

crates/pptx/tests/it/malformed.rs

@@ -287,3 +287,34 @@ fn unreadable_slide_is_reported() {
     assert_eq!(lost, vec!["ppt/slides/slide2.xml".to_string()]);
     assert_eq!(deckcraft_pptx::import(&broken).expect("import").slides.len(), 1);
 }
+
+/// Many slides naming one notes part: it is read for the first slide only, not copied onto every
+/// slide that names it.
+#[test]
+fn a_shared_notes_part_is_read_once() {
+    use std::sync::Arc;
+    let mut p = defaults::new_presentation(None);
+    let layout = p.slides[0].layout;
+    for _ in 0..3 {
+        let s = defaults::new_slide(&mut p, layout).expect("slide");
+        p.slides.push(Arc::new(s));
+    }
+    for (i, s) in p.slides.iter_mut().enumerate() {
+        Arc::make_mut(s).notes = deckcraft_model::text::TextBody::from_text(&format!("notes {i}"));
+    }
+    let mut es = entries(&deckcraft_pptx::export(&p).expect("export"));
+    for (n, d) in es.iter_mut() {
+        if n.starts_with("ppt/slides/_rels/") {
+            let x = String::from_utf8(d.clone()).expect("utf8");
+            let x = x
+                .replace("notesSlide2.xml", "notesSlide1.xml")
+                .replace("notesSlide3.xml", "notesSlide1.xml")
+                .replace("notesSlide4.xml", "notesSlide1.xml");
+            *d = x.into_bytes();
+        }
+    }
+    let q = deckcraft_pptx::import(&rezip(&es)).expect("import");
+    assert_eq!(q.slides.len(), 4);
+    assert_eq!(q.slides[0].notes.text(), "notes 0");
+    assert!(q.slides[1..].iter().all(|s| s.notes.text().is_empty()), "a shared notes part is not copied");
+}