Support reading files with missing white space after xref in lenient mode (#906)

Support missing white space after xref

---------

Co-authored-by: Arnaud TAMAILLON <arnaud.tamaillon@younited-credit.fr>

src/UglyToad.PdfPig.Tests/Integration/CrossReferenceParserTests.cs

@@ -0,0 +1,13 @@
+namespace UglyToad.PdfPig.Tests.Integration
+{
+    public class CrossReferenceParserTests
+    {
+        [Fact]
+        public void CanReadDocumentWithMissingWhitespaceAfterXRef()
+        {
+            string path = IntegrationHelpers.GetSpecificTestDocumentPath("xref-with-no-whitespace.pdf");
+            using var document = PdfDocument.Open(path);
+            Assert.Equal(3, document.NumberOfPages);
+        }
+    }
+}

src/UglyToad.PdfPig.Tests/Parser/Parts/FileStructure/CrossReferenceTableParserTests.cs

@@ -300,18 +300,42 @@ trailer
 trailer
 << >>";
             // Strict parsing
-            var input = StringBytesTestConverter.Scanner(data);
+            var input = GetReader(data);
-            var act = () => CrossReferenceTableParser.Parse(input.scanner, 0, false);
+            var act = () => CrossReferenceTableParser.Parse(input, 0, false);
             var ex = Assert.Throws<PdfDocumentFormatException>(act);
             Assert.Equal("Found a line with 2 unexpected entries in the cross reference table: 127, 0.", ex.Message);
 
             // Lenient Parsing
-            input = StringBytesTestConverter.Scanner(data);
+            input = GetReader(data);
-            var result = CrossReferenceTableParser.Parse(input.scanner, 0, true);
+            var result = CrossReferenceTableParser.Parse(input, 0, true);
 
             Assert.Equal(6, result.ObjectOffsets.Count);
         }
 
+        [Fact]
+        public void ParsesMissingWhitespaceAfterXref()
+        {
+            var data = @"xref15 2
+0000000190 00000 n
+0000000250 00032 n
+
+trailer
+<<>>";
+            var input = GetReader(data);
+
+            // Strict parsing
+            var act = () => CrossReferenceTableParser.Parse(input, 0, false);
+            
+            var ex = Assert.Throws<PdfDocumentFormatException>(act);
+            Assert.Equal("Unexpected operator in xref position: xref15.", ex.Message);
+
+            // Lenient Parsing
+            input = GetReader(data);
+            var result = CrossReferenceTableParser.Parse(input, 0, true);
+
+            Assert.Equal(2, result.ObjectOffsets.Count);
+        }
+
         private static CoreTokenScanner GetReader(string input)
         {
             return StringBytesTestConverter.Scanner(input).scanner;

src/UglyToad.PdfPig/Parser/FileStructure/CrossReferenceParser.cs

@@ -59,7 +59,7 @@
 
                 tokenScanner.MoveNext();
 
-                if (tokenScanner.CurrentToken is OperatorToken tableToken && tableToken.Data == "xref")
+                if (CrossReferenceTableParser.IsCrossReferenceMarker(tokenScanner, isLenientParsing))
                 {
                     missedAttempts = 0;
                     log.Debug("Element was cross reference table.");

src/UglyToad.PdfPig/Parser/FileStructure/CrossReferenceTableParser.cs

@@ -13,7 +13,7 @@
     {
         private const string InUseEntry = "n";
         private const string FreeEntry = "f";
-        
+
         public static CrossReferenceTablePart Parse(ISeekableTokenScanner scanner, long offset, bool isLenientParsing)
         {
             var builder = new CrossReferenceTablePartBuilder
@@ -31,10 +31,22 @@
 
             if (scanner.CurrentToken is OperatorToken operatorToken)
             {
-                if (operatorToken.Data == "xref")
+                if (operatorToken.Data == OperatorToken.Xref.Data)
                 {
                     scanner.MoveNext();
                 }
+                else if (isLenientParsing)
+                {
+                    if (operatorToken.Data.StartsWith(OperatorToken.Xref.Data))
+                    {
+                        scanner.Seek(scanner.CurrentPosition - operatorToken.Data.Length + OperatorToken.Xref.Data.Length);
+                        scanner.MoveNext();
+                    }
+                    else
+                    {
+                        throw new PdfDocumentFormatException($"Unexpected operator in xref position: {operatorToken}.");
+                    }
+                }
                 else
                 {
                     throw new PdfDocumentFormatException($"Unexpected operator in xref position: {operatorToken}.");
@@ -106,6 +118,15 @@
             return builder.Build();
         }
 
+        public static bool IsCrossReferenceMarker(ISeekableTokenScanner scanner, bool isLenientParsing)
+        {
+            return (scanner.CurrentToken is OperatorToken operatorToken
+                    && (operatorToken.Data == OperatorToken.Xref.Data
+                        || (isLenientParsing 
+                            && operatorToken.Data.StartsWith(OperatorToken.Xref.Data)
+                            && int.TryParse(operatorToken.Data.Substring(OperatorToken.Xref.Data.Length), out _))));
+        }
+
         private static int ProcessTokens(ReadOnlySpan<IToken> tokens, CrossReferenceTablePartBuilder builder, bool isLenientParsing,
             int objectCount, ref TableSubsectionDefinition definition)
         {