Merge pull request #21136 from calixteman/bug2033908

Avoid to add outlines having a deleted page which leads to clone a useless page (bug 2033908)

src/core/editor/pdf_editor.js

@@ -1452,9 +1452,12 @@ class PDFEditor {
         result.push({
           ...item,
           // When the item's own destination is invalid (but it has surviving
-          // children), clear the destination so the output item is a plain
-          // container rather than a broken link.
+          // children), clear the destination and rawDict so the output item is
+          // a plain container rather than a broken link. Clearing rawDict
+          // prevents #setOutlineItemDest from cloning a GoTo action that
+          // references a deleted page via its D array.
           dest: hasValidOwnDest ? item.dest : null,
+          rawDict: hasValidOwnDest ? item.rawDict : null,
           items: filteredChildren,
           _documentData: documentData,
         });

test/pdfs/.gitignore

@@ -886,6 +886,7 @@
 !outlines_se.pdf
 !radial_gradients.pdf
 !outlines_for_editor.pdf
+!outline_goto_action.pdf
 !mesh_shading_empty.pdf
 !acroform_calculation_order.pdf
 !extractPages_null_in_array.pdf

test/unit/api_spec.js

@@ -6971,6 +6971,41 @@ small scripts as well as for`);
 
         await newLoadingTask.destroy();
       });
+
+      it("should handle parent items whose dest was cleared but rawDict still holds a GoTo action (bug 2033603)", async function () {
+        // outline_goto_action.pdf has 2 pages and this outline:
+        //   "Parent"  /A << /S /GoTo /D [page1Ref /FitH 842] >>
+        //     "Child"  /Dest [page2Ref /FitH 0]
+        const loadingTask = getDocument(
+          buildGetDocumentParams("outline_goto_action.pdf")
+        );
+        const pdfDoc = await loadingTask.promise;
+        const data = await pdfDoc.extractPages([
+          { document: null, includePages: [1] },
+        ]);
+        await loadingTask.destroy();
+
+        const newLoadingTask = getDocument(data);
+        const newPdfDoc = await newLoadingTask.promise;
+        expect(newPdfDoc.numPages).toEqual(1);
+
+        const outline = await newPdfDoc.getOutline();
+        expect(Array.isArray(outline)).toEqual(true);
+        expect(outline.length).toEqual(1);
+
+        // "Parent" is kept as a plain container (dest cleared, no action).
+        const parent = outline[0];
+        expect(parent.title).toEqual("Parent");
+        expect(parent.dest).toEqual(null);
+        expect(parent.items.length).toEqual(1);
+
+        // "Child" keeps its explicit dest pointing to the (only) kept page.
+        const child = parent.items[0];
+        expect(child.title).toEqual("Child");
+        expect(Array.isArray(child.dest)).toEqual(true);
+
+        await newLoadingTask.destroy();
+      });
     });
 
     describe("extract pages with null values in arrays", function () {