Merge pull request #359 from plaisted/master

link annotation fix for PdfDocumentBuilder

src/UglyToad.PdfPig.Tests/Writer/PdfDocumentBuilderTests.cs

@@ -101,6 +101,33 @@
             }
         }
 
+        [Fact]
+        public void CanFastAddPageAndStripLinkAnnots()
+        {
+            var first = IntegrationHelpers.GetDocumentPath("outline.pdf");
+            var contents = File.ReadAllBytes(first);
+
+            var annotCount = 0;
+            byte[] results = null;
+            using (var existing = PdfDocument.Open(contents, ParsingOptions.LenientParsingOff))
+            using (var output = new PdfDocumentBuilder())
+            {
+                output.AddPage(existing, 1);
+                results = output.Build();
+                var pg = existing.GetPage(1);
+                var annots = pg.ExperimentalAccess.GetAnnotations().ToList();
+                annotCount = annots.Count;
+                Assert.Contains(annots, x => x.Type == Annotations.AnnotationType.Link);
+            }
+
+            using (var rewritten = PdfDocument.Open(results, ParsingOptions.LenientParsingOff))
+            {
+                var pg = rewritten.GetPage(1);
+                var annots = pg.ExperimentalAccess.GetAnnotations().ToList();
+                Assert.Equal(annotCount - 1, annots.Count);
+                Assert.DoesNotContain(annots, x => x.Type == Annotations.AnnotationType.Link);
+            }
+        }
 
         [Fact]
         public void CanReadSingleBlankPage()

src/UglyToad.PdfPig/Writer/IPdfStreamWriter.cs

@@ -8,6 +8,13 @@
 
     internal interface IPdfStreamWriter : IDisposable
     {
+        /// <summary>
+        /// Sets if the stream writer should attempt to deduplicate objects.
+        /// May not have any affect if <see cref="IPdfStreamWriter"/> does not
+        /// support deduplication.
+        /// </summary>
+        bool AttemptDeduplication { get; set; }
+
         /// <summary>
         /// The underlying stream used by the writer.
         /// </summary>

src/UglyToad.PdfPig/Writer/PdfDedupStreamWriter.cs

@@ -23,13 +23,16 @@
             ms.SetLength(0);
             TokenWriter.WriteToken(token, ms);
             var contents = ms.ToArray();
-            if (hashes.TryGetValue(contents, out var value))
+            if (AttemptDeduplication && hashes.TryGetValue(contents, out var value))
             {
                 return value;
             }
 
             var ir = ReserveObjectNumber();
-            hashes.Add(contents, ir);
+            if (AttemptDeduplication)
+            {
+                hashes.Add(contents, ir);
+            }
 
             offsets.Add(ir.Data, Stream.Position);
             TokenWriter.WriteObject(ir.Data.ObjectNumber, ir.Data.Generation, contents, Stream);

src/UglyToad.PdfPig/Writer/PdfDocumentBuilder.cs

@@ -330,6 +330,10 @@ namespace UglyToad.PdfPig.Writer
             var streams = new List<PdfPageBuilder.CopiedContentStream>();
             if (pageInfo.Page.TryGet(NameToken.Contents, out IToken contentsToken))
             {
+                // Adobe Acrobat errors if content streams ref'd by multiple pages, turn off
+                // dedup if on to avoid issues
+                var prev = context.AttemptDeduplication;
+                context.AttemptDeduplication = false;
                 if (contentsToken is ArrayToken array)
                 {
                     foreach (var item in array.Data)
@@ -347,6 +351,7 @@ namespace UglyToad.PdfPig.Writer
                     streams.Add(new PdfPageBuilder.CopiedContentStream(
                         WriterUtil.CopyToken(context, ir, document.Structure.TokenScanner, refs) as IndirectReferenceToken));
                 }
+                context.AttemptDeduplication = prev;
             }
 
             // manually copy page dict / resources as we need to modify some
@@ -379,15 +384,55 @@ namespace UglyToad.PdfPig.Writer
             {
                 if (kvp.Key == NameToken.Contents || kvp.Key == NameToken.Parent || kvp.Key == NameToken.Type)
                 {
+                    // don't copy these as they'll be handled during page tree writing
                     continue;
                 }
 
                 if (kvp.Key == NameToken.Resources)
                 {
+                    // merge parent resources into child
                     CopyResourceDict(kvp.Value, resources);
                     continue;
                 }
 
+                if (kvp.Key == NameToken.Annots)
+                {
+                    var val = kvp.Value;
+                    if (kvp.Value is IndirectReferenceToken ir)
+                    {
+                        val = document.Structure.TokenScanner.Get(ir.Data).Data;
+                    }
+                                                    
+                    if (!(val is ArrayToken arr))
+                    {
+                        // should be array... ignore and remove bad dict
+                        continue;
+                    }
+
+                    // -> ignore links to resolve issues with refencing non-existing pages
+                    // at some point should add support for copying the links if the
+                    // pages are copied as well but for now just fix corruption
+                    var toAdd = new List<IToken>();
+                    foreach (var annot in arr.Data)
+                    {
+                        DictionaryToken tk = GetRemoteDict(annot);
+                        if (tk == null)
+                        {
+                            // malformed
+                            continue;
+                        }
+                        if (tk.TryGet(NameToken.Subtype, out var st) && st is NameToken nm && nm == NameToken.Link)
+                        {
+                            // link -> ignore
+                            continue;
+                        }
+                        toAdd.Add(WriterUtil.CopyToken(context, tk, document.Structure.TokenScanner, refs));
+                    }
+                    // copy rest
+                    copiedPageDict[NameToken.Annots] = new ArrayToken(toAdd);
+                    continue;
+                }
+
                 copiedPageDict[NameToken.Create(kvp.Key)] =
                     WriterUtil.CopyToken(context, kvp.Value, document.Structure.TokenScanner, refs);
             }
@@ -508,10 +553,14 @@ namespace UglyToad.PdfPig.Writer
                     pageDictionary[NameToken.MediaBox] = RectangleToArray(page.Value.PageSize);
                 }
 
+                // Adobe Acrobat errors if content streams ref'd by multiple pages, turn off
+                // dedup if on to avoid issues
+                var prev = context.AttemptDeduplication;
+                context.AttemptDeduplication = false;
+
                 var toWrite = page.Value.contentStreams.Where(x => x.HasContent).ToList();
                 if (toWrite.Count == 0)
                 {
-                    // write empty
                     pageDictionary[NameToken.Contents] = new PdfPageBuilder.DefaultContentStream().Write(context);
                 }
                 else if (toWrite.Count == 1)
@@ -529,7 +578,7 @@ namespace UglyToad.PdfPig.Writer
                     }
                     pageDictionary[NameToken.Contents] = new ArrayToken(streams);
                 }
-
+                context.AttemptDeduplication = prev;;
 
                 leafChildren[leafNum].Add(context.WriteToken(new DictionaryToken(pageDictionary)));
 

src/UglyToad.PdfPig/Writer/PdfStreamWriter.cs

@@ -30,8 +30,9 @@
             DisposeStream = disposeStream;
         }
 
-        public Stream Stream { get; protected set; }
+        public Stream Stream { get; protected set; }
-
+        public bool AttemptDeduplication { get; set; } = true;
+
         public virtual IndirectReferenceToken WriteToken(IToken token)
         {
             if (!Initialized)

src/UglyToad.PdfPig/Writer/TokenWriter.cs

@@ -80,7 +80,11 @@
         /// <param name="token">The token to write to the stream.</param>
         /// <param name="outputStream">The stream to write the token to.</param>
         public static void WriteToken(IToken token, Stream outputStream)
-        {
+        {
+            if (token == null)
+            {
+                throw new ArgumentNullException(nameof(token));
+            }
             switch (token)
             {
                 case ArrayToken array:
@@ -119,7 +123,9 @@
                     break;
                 case StringToken stringToken:
                     WriteString(stringToken, outputStream);
-                    break;
+                    break;
+                default:
+                    throw new PdfDocumentFormatException($"Attempted to write token type of {token.GetType()} but was not known.");
             }
         }
 
@@ -294,8 +300,16 @@
 
             foreach (var pair in dictionary.Data)
             {
-                WriteName(pair.Key, outputStream);
+                WriteName(pair.Key, outputStream);
-                WriteToken(pair.Value, outputStream);
+
+                // handle scenario where PdfPig has a null value under some circumstances
+                if (pair.Value == null)
+                {
+                    WriteToken(NullToken.Instance, outputStream);
+                } else
+                {
+                    WriteToken(pair.Value, outputStream);
+                }
             }
 
             outputStream.Write(DictionaryEnd, 0, DictionaryEnd.Length);