add page size and more tests for cross reference table parsing

src/UglyToad.Pdf.Tests/Integration/FontSizeTestFromGoogleChromeTests.cs

@@ -3,6 +3,7 @@
     using System;
     using System.IO;
     using System.Linq;
+    using Content;
     using Xunit;
 
     public class FontSizeTestFromGoogleChromeTests
@@ -49,5 +50,16 @@
                 Assert.Equal("Hello, this is 16ptHello, this is 16px", text);
             }
         }
+
+        [Fact]
+        public void GetsCorrectPageSize()
+        {
+            using (var document = PdfDocument.Open(GetFilename()))
+            {
+                var page = document.GetPage(1);
+
+                Assert.Equal(PageSize.A4, page.Size);
+            }
+        }
     }
 }

src/UglyToad.Pdf.Tests/Integration/SinglePageSimpleTests.cs

@@ -5,6 +5,7 @@ namespace UglyToad.Pdf.Tests.Integration
     using System.Collections.Generic;
     using System.IO;
     using System.Linq;
+    using Content;
     using Xunit;
 
     public class SinglePageSimpleTests
@@ -92,6 +93,17 @@ namespace UglyToad.Pdf.Tests.Integration
             }
         }
 
+        [Fact]
+        public void GetsCorrectPageSize()
+        {
+            using (var document = PdfDocument.Open(GetFilename()))
+            {
+                var page = document.GetPage(1);
+
+                Assert.Equal(PageSize.Letter, page.Size);
+            }
+        }
+
         [Fact]
         public void LettersHavePdfBoxPositions()
         {

src/UglyToad.Pdf.Tests/Parser/Parts/CrossReference/CrossReferenceTableParserTests.cs

@@ -1,5 +1,6 @@
 namespace UglyToad.Pdf.Tests.Parser.Parts.CrossReference
 {
+    using System;
     using System.Linq;
     using IO;
     using Pdf.Cos;
@@ -228,6 +229,51 @@ trailer
             Assert.Equal(32, results[4].Generation);
         }
 
+        [Fact]
+        public void EntryPointingAtOffsetInTableThrows()
+        {
+            var input = GetReader(@"xref
+0 2
+0000000000 65535 f
+0000000010 00000 n
+trailer
+<<>>");
+
+            Action action = () => parser.TryParse(input, 0, false, objectPool, out var _);
+
+            Assert.Throws<InvalidOperationException>(action);
+        }
+
+        [Fact]
+        public void EntryWithInvalidFormatThrows()
+        {
+            var input = GetReader(@"xref
+0 22
+0000000000 65535 f
+0000aa0010 00000 n
+trailer
+<<>>");
+
+            Action action = () => parser.TryParse(input, 0, false, objectPool, out var _);
+
+            Assert.Throws<InvalidOperationException>(action);
+        }
+
+        [Fact]
+        public void ShortLineInTableReturnsFalse()
+        {
+            var input = GetReader(@"xref
+15 2
+000000019000000 n
+0000000250 00032 n
+trailer
+<<>>");
+
+            var result = parser.TryParse(input, 0, false, objectPool, out var table);
+
+            Assert.False(result);
+        }
+
         private static IRandomAccessRead GetReader(string input)
         {
             return new RandomAccessBuffer(OtherEncodings.StringAsLatin1Bytes(input));

src/UglyToad.Pdf/Content/Page.cs

@@ -6,7 +6,7 @@
     public class Page
     {
         /// <summary>
-        /// The 1 indexed page number.
+        /// The page number (starting at 1).
         /// </summary>
         public int Number { get; }
 
@@ -28,6 +28,11 @@
         /// </summary>
         public decimal Height { get; }
 
+        /// <summary>
+        /// The size of the page according to the standard page sizes or Custom if no matching standard size found.
+        /// </summary>
+        public PageSize Size { get; }
+
         internal Page(int number, MediaBox mediaBox, CropBox cropBox, PageContent content)
         {
             if (number <= 0)
@@ -42,6 +47,8 @@
 
             Width = mediaBox.Bounds.Width;
             Height = mediaBox.Bounds.Height;
+
+            Size = mediaBox.Bounds.GetPageSize();
         }
     }
 }

src/UglyToad.Pdf/Content/PageSize.cs

@@ -0,0 +1,93 @@
+namespace UglyToad.Pdf.Content
+{
+    using System.Collections.Generic;
+    using Geometry;
+
+    public enum PageSize
+    {
+        Custom = 0,
+        A0 = 3,
+        A1 = 4,
+        A2 = 5,
+        A3 = 6,
+        A4 = 7,
+        A5 = 8,
+        A6 = 9,
+        A7 = 10,
+        A8 = 11,
+        A9 = 12,
+        A10 = 13,
+        Letter = 14,
+        Legal = 15,
+        Ledger = 16,
+        Tabloid = 17,
+        Executive = 18
+    }
+
+    internal static class PageSizeExtensions
+    {
+        private static readonly Dictionary<WidthHeight, PageSize> Lookup = new Dictionary<WidthHeight, PageSize>
+        {
+            {new WidthHeight(2384, 3370), PageSize.A0},
+            {new WidthHeight(1684, 2384), PageSize.A1},
+            // Seems there is some disagreement 1190/1191
+            {new WidthHeight(1190, 1684), PageSize.A2},
+            {new WidthHeight(1191, 1684), PageSize.A2},
+            // Seems there is some disagreement 1190/1191
+            {new WidthHeight(842, 1190), PageSize.A3},
+            {new WidthHeight(842, 1191), PageSize.A3},
+            {new WidthHeight(595, 842), PageSize.A4},
+            {new WidthHeight(420, 595), PageSize.A5},
+            {new WidthHeight(298, 420), PageSize.A6},
+            {new WidthHeight(210, 298), PageSize.A7},
+            {new WidthHeight(147, 210), PageSize.A8},
+            {new WidthHeight(105, 147), PageSize.A9},
+            {new WidthHeight(74, 105), PageSize.A10},
+            {new WidthHeight(612, 792), PageSize.Letter},
+            {new WidthHeight(612, 1008), PageSize.Legal},
+            {new WidthHeight(1224, 792), PageSize.Ledger},
+            {new WidthHeight(792, 1224), PageSize.Tabloid},
+            // Again there is disagreement here
+            {new WidthHeight(540, 720), PageSize.Executive},
+            {new WidthHeight(522, 756), PageSize.Executive},
+        };
+
+        public static PageSize GetPageSize(this PdfRectangle rectangle)
+        {
+            if (!Lookup.TryGetValue(new WidthHeight(rectangle.Width, rectangle.Height), out var size))
+            {
+                return PageSize.Custom;
+            }
+
+            return size;
+        }
+
+        private struct WidthHeight
+        {
+            public decimal Width { get; }
+
+            public decimal Height { get; }
+
+            public WidthHeight(decimal width, decimal height)
+            {
+                Width = width;
+                Height = height;
+            }
+
+            public override bool Equals(object obj)
+            {
+                return obj is WidthHeight height &&
+                       Width == height.Width &&
+                       Height == height.Height;
+            }
+
+            public override int GetHashCode()
+            {
+                var hashCode = 859600377;
+                hashCode = hashCode * -1521134295 + Width.GetHashCode();
+                hashCode = hashCode * -1521134295 + Height.GetHashCode();
+                return hashCode;
+            }
+        }
+    }
+}