PDFBOX-5225: consider the field list parameter when there are widgets…

… with missing page references; adjust / add test

git-svn-id: https://svn.apache.org/repos/asf/pdfbox/trunk@1922019 13f79535-47bb-0310-9956-ffa450edef68

pdfbox/src/main/java/org/apache/pdfbox/pdmodel/interactive/form/PDAcroForm.java

@@ -759,6 +759,13 @@ private Rectangle2D getTransformedAppearanceBBox(PDAppearanceStream appearanceSt
         return transformedAppearanceBox.getBounds2D();
     }
 
+    /**
+     * Build a map of pages => widgets
+     * @param fields a list of fields to be flattened
+     * @param pages the page tree
+     * @return
+     * @throws IOException 
+     */
     private Map<COSDictionary,Set<COSDictionary>> buildPagesWidgetsMap(
             List<PDField> fields, PDPageTree pages) throws IOException
     {
@@ -777,6 +784,7 @@ private Map<COSDictionary,Set<COSDictionary>> buildPagesWidgetsMap(
                 }
                 else
                 {
+                    LOG.warn("missing /P entry (page reference) in a widget for field: " + field);
                     hasMissingPageRef = true;
                 }
             }
@@ -790,11 +798,21 @@ private Map<COSDictionary,Set<COSDictionary>> buildPagesWidgetsMap(
         // If there is a widget with a missing page reference we need to build the map reverse i.e. 
         // from the annotations to the widget.
         LOG.warn("There has been a widget with a missing page reference, will check all page annotations");
+        Set<COSDictionary> widgetDictionarySet = new HashSet<>();
+        for (PDField field : fields)
+        {
+            List<PDAnnotationWidget> widgets = field.getWidgets();
+            for (PDAnnotationWidget widget : widgets)
+            {
+                widgetDictionarySet.add(widget.getCOSObject());
+            }
+        }
+
         for (PDPage page : pages)
         {
             for (PDAnnotation annotation : page.getAnnotations())
             {
-                if (annotation instanceof PDAnnotationWidget)
+                if (widgetDictionarySet.contains(annotation.getCOSObject()))
                 {
                     fillPagesAnnotationMap(pagesAnnotationsMap, page, (PDAnnotationWidget) annotation);
                 }

pdfbox/src/test/java/org/apache/pdfbox/pdmodel/interactive/form/PDAcroFormFlattenTest.java

@@ -30,14 +30,14 @@
 import java.nio.file.Files;
 import java.nio.file.StandardCopyOption;
 import java.util.ArrayList;
+import java.util.Iterator;
 import java.util.List;
 import java.util.stream.Stream;
 
 import javax.imageio.ImageIO;
 
 import org.apache.pdfbox.pdmodel.PDDocument;
 import org.apache.pdfbox.Loader;
-import org.apache.pdfbox.pdfwriter.compress.CompressParameters;
 import org.apache.pdfbox.rendering.PDFRenderer;
 import org.apache.pdfbox.rendering.TestPDFToImage;
 import org.junit.jupiter.api.BeforeAll;
@@ -130,9 +130,6 @@ static void setUp()
         // annotations.
         "https://issues.apache.org/jira/secure/attachment/12994791/flatten.pdf,PDFBOX-4788.pdf",
 
-        // PDFBOX-4889: appearance streams with empty /BBox.
-        "https://issues.apache.org/jira/secure/attachment/13005793/f1040sb%20test.pdf,PDFBOX-4889.pdf",
-
         // PDFBOX-4955: appearance streams with forms that are not used.
         "https://issues.apache.org/jira/secure/attachment/13011410/PDFBOX-4955.pdf,PDFBOX-4955.pdf",
 
@@ -183,8 +180,9 @@ void flattenTestPDFBOX5254() throws IOException, URISyntaxException
         {
             testPdf.getDocumentCatalog().getAcroForm().flatten();
             testPdf.setAllSecurityToBeRemoved(true);
-            assertTrue(testPdf.getDocumentCatalog().getAcroForm().getFields().isEmpty());
-            testPdf.save(outputFile, CompressParameters.NO_COMPRESSION);
+            testPdf.save(outputFile);
+            assertTrue(testPdf.getDocumentCatalog().getAcroForm(null).getFields().isEmpty());
+            assertEquals(72, testPdf.getPage(0).getAnnotations().size());
         }
 
         // compare rendering
@@ -204,6 +202,61 @@ void flattenTestPDFBOX5254() throws IOException, URISyntaxException
         }
     }
 
+    /**
+     * Check that only VN_Name is removed in the field tree and in the annotations list. That field
+     * has an "orphan" widget that belongs to no page.
+     *
+     * @throws IOException
+     * @throws URISyntaxException
+     */
+    @Test
+    void flattenTestPDFBOX5225() throws IOException, URISyntaxException
+    {
+        String sourceUrl = "https://issues.apache.org/jira/secure/attachment/13027311/SourceFailure.pdf";
+        String targetFileName = "PDFBOX-5225.pdf";
+
+        generateSamples(sourceUrl, targetFileName);
+
+        File inputFile = new File(IN_DIR, targetFileName);
+        File outputFile = new File(OUT_DIR, targetFileName);
+
+        try (PDDocument testPdf = Loader.loadPDF(inputFile))
+        {
+            PDAcroForm acroForm = testPdf.getDocumentCatalog().getAcroForm();
+            List<PDField> list = new ArrayList<>();
+            list.add(acroForm.getField("VN_NAME"));
+            acroForm.flatten(list, false);
+            testPdf.setAllSecurityToBeRemoved(true);
+            testPdf.save(outputFile);
+            int count = 0;
+            Iterator<PDField> iterator = acroForm.getFieldTree().iterator();
+            while (iterator.hasNext())
+            {
+                iterator.next();
+                ++count;
+            }
+            assertEquals(76, count);
+            assertEquals(59, testPdf.getPage(0).getAnnotations().size());
+        }
+
+        // compare rendering
+        if (!TestPDFToImage.doTestFile(outputFile, IN_DIR.getAbsolutePath(),
+                OUT_DIR.getAbsolutePath()))
+        {
+            // check manually
+            System.err.println("Rendering of " + outputFile
+                    + " failed or is not identical to expected rendering in " + IN_DIR
+                    + " directory");
+        }
+        else
+        {
+            // cleanup input and output directory for matching files.
+            removeAllRenditions(inputFile);
+            inputFile.delete();
+            outputFile.delete();
+        }
+    }
+
     /*
      * Flatten and compare with generated image samples.
      *