Merge PR #4090: Layout: a document of many pages paid a body-length scan per page

crates/layout/src/lib.rs

@@ -700,6 +700,11 @@ enum Pass {
 /// Walk state while paginating.
 struct Ctx<'a> {
     doc: &'a Document,
+    /// The body's sections ([`Document::sections`]) for one pass, with the block index each
+    /// starts after: header and footer wrap work asks for them per page, and rebuilding the
+    /// list walked the whole body every time, so a long document paid a body-length scan per
+    /// page. The document does not change during a pass.
+    sections: Vec<(usize, &'a SectionProps)>,
     cache: &'a mut LayoutCache,
     opts: &'a LayoutOptions,
     counters: Counters,
@@ -2301,6 +2306,7 @@ fn layout_pass(
     });
     let mut ctx = Ctx {
         doc,
+        sections: doc.sections(),
         cache,
         opts,
         hide_deleted: opts.hides_deleted(StoryRef::Body),
@@ -3391,12 +3397,11 @@ fn footer_table_wraps(ctx: &mut Ctx, pb: &PageBuilder) -> Vec<WrapArea> {
     }
     ctx.footer_wraps.insert(key, Vec::new());
     let doc = ctx.doc;
-    let sections = doc.sections();
-    let Some((_, sect)) = sections.get(page.section) else { return Vec::new() };
+    let Some((_, sect)) = ctx.sections.get(page.section).copied() else { return Vec::new() };
     if sect.text_direction.is_turned() {
         return Vec::new();
     }
-    let set = linked(&sections, page.section, |s| s.footers);
+    let set = linked(&ctx.sections, page.section, |s| s.footers);
     let id = if sect.title_page && first {
         set.first
     } else if doc.settings.even_odd_headers && page.number.is_multiple_of(2) {
@@ -3408,7 +3413,7 @@ fn footer_table_wraps(ctx: &mut Ctx, pb: &PageBuilder) -> Vec<WrapArea> {
     let Some(part) = doc.parts.get(&id) else { return Vec::new() };
     let blocks = part.blocks.clone();
     let x = sect.margin_left + sect.gutter;
-    let base = defining_width(&sections, page.section, |s| s.footers, Some(id));
+    let base = defining_width(&ctx.sections, page.section, |s| s.footers, Some(id));
     let counters = std::mem::take(&mut ctx.counters);
     let boxes = std::mem::take(&mut ctx.boxes);
     let chains = std::mem::take(&mut ctx.chains);
@@ -4526,12 +4531,11 @@ fn header_table_areas(ctx: &mut Ctx, pb: &PageBuilder) -> Vec<Rect> {
     if let Some(areas) = ctx.header_tables.get(&key) {
         return areas.clone();
     }
-    let sections = ctx.doc.sections();
-    let Some((_, sect)) = sections.get(page.section) else { return Vec::new() };
+    let Some((_, sect)) = ctx.sections.get(page.section).copied() else { return Vec::new() };
     if sect.text_direction.is_turned() {
         return Vec::new();
     }
-    let set = linked(&sections, page.section, |s| s.headers);
+    let set = linked(&ctx.sections, page.section, |s| s.headers);
     let id = if sect.title_page && first {
         set.first
     } else if ctx.doc.settings.even_odd_headers && page.number.is_multiple_of(2) {
@@ -4544,7 +4548,7 @@ fn header_table_areas(ctx: &mut Ctx, pb: &PageBuilder) -> Vec<Rect> {
     let blocks = part.blocks.clone();
     let x = sect.margin_left + sect.gutter;
     let frame = PageFrame { sect, origin: (x, sect.header), number: page.number, mirror: ctx.mirror };
-    let base = defining_width(&sections, page.section, |s| s.headers, Some(id));
+    let base = defining_width(&ctx.sections, page.section, |s| s.headers, Some(id));
     let counters = std::mem::take(&mut ctx.counters);
     let boxes = std::mem::take(&mut ctx.boxes);
     let chains = std::mem::take(&mut ctx.chains);

crates/layout/src/tests.rs

@@ -849,6 +849,26 @@ fn hidden_separators_with_automatic_hyphenation_lay_out_in_linear_time() {
     assert!(ratio < 24.0, "8x the text took {ratio:.1}x the time ({t1:.4} s → {tk:.4} s): not linear");
 }
 
+/// n one-line paragraphs, each starting a page.
+fn page_break_paragraphs(n: usize) -> Document {
+    let mut d = Document::from_text(&"hello\n".repeat(n));
+    for i in 0..n {
+        d.para_mut(StoryRef::Body, &Path::top(i)).unwrap().props.page_break_before = Some(true);
+    }
+    d
+}
+
+/// A document of many paragraphs lays out in linear time. Every page asks the layout for the
+/// section properties (for the wrap areas of the headers and footers), and that used to rebuild
+/// the section list by walking the whole body, so a document of n paragraphs paid n body-length
+/// scans: a two-megabyte file took half a minute before the first page was drawn.
+#[test]
+fn a_document_of_many_paragraphs_lays_out_in_linear_time() {
+    let (t1, tk, ratio) = layout_time_ratio(page_break_paragraphs, 2_500, &LayoutOptions::default());
+    eprintln!("t(n) {t1:.4} s, t(8n) {tk:.4} s, ratio {ratio:.1}");
+    assert!(ratio < 24.0, "8x the paragraphs took {ratio:.1}x the time ({t1:.4} s → {tk:.4} s): not linear");
+}
+
 #[test]
 fn many_hidden_float_anchors_lay_out_in_linear_time() {
     // n square-wrapped shapes, each anchored in its own hidden run (bold alternating, so runs do not