Integration: PDF crop box is part of an asset's identity; clippings keep anchored graphics (#391)

Asset::same_file compares pdf_crop (#362) with type, page and bytes: one
PDF page shown with two crop boxes keeps two assets, so add_asset,
adopt_asset, insert_asset (IDML/INDD import) and the load/save merge
never point a placement at another crop. Test in
identical_files_share_one_asset.

The copy clipping (edit.rs clip_doc) keeps the assets used_assets finds
in it, so graphics anchored in copied text, graphic table cells (#379),
notes and media posters keep their files; it collected only graphics on
the copied items before. Test:
copied_text_frames_keep_the_graphics_anchored_in_their_text.

crates/doc/src/assets.rs

@@ -15,10 +15,11 @@ use crate::{Asset, AssetId, Content, Document, Item};
 const MAX_DEPTH: usize = 64;
 
 impl Asset {
-    /// Whether `other` holds the same file: the same type, page and bytes. Assets without bytes
-    /// (a link that couldn't be read) are the same only when they link to the same path.
+    /// Whether `other` holds the same file: the same type, page, PDF crop box and bytes. Assets
+    /// without bytes (a link that couldn't be read) are the same only when they link to the same
+    /// path.
     pub fn same_file(&self, other: &Asset) -> bool {
-        if self.mime != other.mime || self.page != other.page {
+        if self.mime != other.mime || self.page != other.page || self.pdf_crop != other.pdf_crop {
             return false;
         }
         match (self.data.is_empty(), other.data.is_empty()) {
@@ -306,6 +307,13 @@ mod tests {
         let b = d.add_asset(Asset { link: Some("/x/p.png".into()), ..asset(2, b"same") });
         let c = d.add_asset(asset(3, b"diff"));
         let p = d.add_asset(Asset { page: 1, ..asset(4, b"same") });
+        // One PDF page shown with two crop boxes: two assets (the box is the asset's).
+        let pdf = |id, pdf_crop| Asset { mime: "application/pdf".into(), pdf_crop, ..asset(id, b"%PDF") };
+        let crop = d.add_asset(pdf(9, crate::PdfCrop::Crop));
+        let art = d.add_asset(pdf(10, crate::PdfCrop::Art));
+        assert_ne!(crop, art);
+        assert_eq!(d.add_asset(pdf(11, crate::PdfCrop::Art)), art);
+        d.assets.retain(|id, _| ![crop, art].contains(id));
         assert_eq!(a, b);
         assert_ne!(a, c);
         assert_ne!(a, p);

crates/engine/src/cmd/edit.rs

@@ -300,18 +300,13 @@ pub(crate) fn clip_doc(s: &Session) -> Result<designcraft_doc::Document> {
     for sp in d.spreads.iter_mut() {
         Arc::make_mut(sp).items.clear();
     }
-    // Kept frames and graphics (including inside groups).
+    // Kept frames (including inside groups).
     let mut frames = std::collections::HashSet::new();
-    let mut assets = std::collections::HashSet::new();
     for it in &items {
-        it.walk(&mut |i| match &i.content {
-            designcraft_doc::Content::Text(_) => {
+        it.walk(&mut |i| {
+            if let designcraft_doc::Content::Text(_) = &i.content {
                 frames.insert(i.id);
             }
-            designcraft_doc::Content::Graphic(g) => {
-                assets.insert(g.asset);
-            }
-            _ => {}
         });
     }
     if let Some(sp) = d.spreads.first_mut() {
@@ -325,6 +320,8 @@ pub(crate) fn clip_doc(s: &Session) -> Result<designcraft_doc::Document> {
     for st in d.stories.values_mut() {
         Arc::make_mut(st).frames.retain(|f| frames.contains(f));
     }
+    // Graphics on the items and in the kept stories (anchored objects, graphic cells, notes).
+    let assets = d.used_assets();
     d.assets.retain(|k, _| assets.contains(k));
     d.hyperlinks.clear();
     d.bookmarks.clear();
@@ -545,6 +542,28 @@ mod paste_tests {
         assert_eq!(pasted(&mut s, "edit.pasteInPlace", json!({"spread": 1})), (SpreadRef::Doc(1), Rect::new(w + 36.0, 36.0, w + 136.0, 136.0)));
     }
 
+    /// A copied text frame's clipping holds the files its anchored graphics show.
+    #[test]
+    fn copied_text_frames_keep_the_graphics_anchored_in_their_text() {
+        let mut s = Session::new();
+        s.execute("file.new", &json!({})).unwrap();
+        let png = designcraft_render::Rendered { width: 4, height: 4, pixels: vec![200u8; 64] }.to_png();
+        let placed = s.execute("file.place", &json!({"base64": super::super::file::base64_encode(&png), "name": "a.png", "x": 72, "y": 72})).unwrap();
+        let (g, aid) = (ItemId(placed["id"].as_u64().unwrap()), designcraft_doc::AssetId(placed["asset"].as_u64().unwrap()));
+        let f = ItemId(s.execute("frame.create", &json!({"rect": [36, 200, 300, 400], "content": "text"})).unwrap()["id"].as_u64().unwrap());
+        s.edit(|d, _| {
+            let it = d.item(g).cloned().ok_or(designcraft_doc::DocError::NoItem(g))?;
+            let sid = d.item(f).and_then(|i| i.text_frame()).map(|t| t.story).ok_or(designcraft_doc::DocError::NoItem(f))?;
+            super::super::anchored::anchor_items(d, sid, 0, vec![it], &Default::default())?;
+            Ok(())
+        })
+        .unwrap();
+        s.execute("selection.set", &json!({"ids": [f.0]})).unwrap();
+        s.execute("edit.copy", &json!({})).unwrap();
+        let clip = s.clipboard.clone().unwrap();
+        assert_eq!(clip.assets.keys().copied().collect::<Vec<_>>(), [aid]);
+    }
+
     /// Undo leaves the id of the copy it removed in the selection. Copy measures from the spread
     /// of the first selected object that is still there.
     #[test]