Fix/flate trailer subset tags opm

Directory.Packages.props

@@ -1,36 +1,29 @@
 <Project>
-
   <PropertyGroup>
     <ManagePackageVersionsCentrally>true</ManagePackageVersionsCentrally>
     <CentralPackageTransitivePinningEnabled>true</CentralPackageTransitivePinningEnabled>
     <NuGetAuditMode>all</NuGetAuditMode>
   </PropertyGroup>
-
   <PropertyGroup>
     <RestoreSources>
       https://api.nuget.org/v3/index.json;
     </RestoreSources>
   </PropertyGroup>
-
   <ItemGroup Label="Build packages">
     <PackageVersion Include="Microsoft.SourceLink.GitHub" Version="10.0.401" />
   </ItemGroup>
-
   <ItemGroup Label="General packages">
     <PackageVersion Include="SixLabors.ImageSharp" Version="3.1.12" />
     <PackageVersion Include="Synercoding.Primitives" Version="1.0.0-rc11" />
   </ItemGroup>
-
   <ItemGroup Label="Testing packages">
-    <PackageVersion Include="coverlet.collector" Version="6.0.4" />
-    <PackageVersion Include="Microsoft.NET.Test.Sdk" Version="17.14.1" />
+    <PackageVersion Include="coverlet.collector" Version="10.1.0" />
+    <PackageVersion Include="Microsoft.NET.Test.Sdk" Version="18.10.1" />
     <PackageVersion Include="xunit" Version="2.9.3" />
-    <PackageVersion Include="xunit.runner.visualstudio" Version="3.1.1" />
+    <PackageVersion Include="xunit.runner.visualstudio" Version="4.0.0" />
     <PackageVersion Include="Pdfium.Rendering" Version="1.0.0-alpha05" />
   </ItemGroup>
-
   <ItemGroup Label="Example app packages">
-    <PackageVersion Include="Microsoft.Extensions.Logging" Version="8.0.1" />
+    <PackageVersion Include="Microsoft.Extensions.Logging" Version="10.0.12" />
   </ItemGroup>
-
-</Project>
+</Project>

src/Synercoding.FileFormats.Pdf/Content/ExtendedGraphicsState.cs

@@ -23,6 +23,11 @@ public sealed record class ExtendedGraphicsState
     /// </summary>
     public bool? OverprintNonStroking { get; init; }
 
+    /// <summary>
+    /// The overprint mode, which determines how overprinting is applied when painting in a DeviceCMYK color space.
+    /// </summary>
+    public OverprintMode? OverprintMode { get; init; }
+
     /// <summary>
     /// The current stroking alpha constant, specifying the constant shape or constant opacity value
     /// to be used for stroking operations in the transparent imaging model.
@@ -71,6 +76,8 @@ internal IPdfDictionary ToPdfDictionary()
             dictionary[PdfNames.OP] = new PdfBoolean(Overprint.Value);
         if (OverprintNonStroking.HasValue)
             dictionary[PdfNames.op] = new PdfBoolean(OverprintNonStroking.Value);
+        if (OverprintMode.HasValue)
+            dictionary[PdfNames.OPM] = new PdfNumber((int)OverprintMode.Value);
         if (CurrentAlphaConstantStroking.HasValue)
             dictionary[PdfNames.CA] = new PdfNumber(CurrentAlphaConstantStroking.Value);
         if (CurrentAlphaConstantNonStroking.HasValue)

src/Synercoding.FileFormats.Pdf/Content/OverprintMode.cs

@@ -0,0 +1,17 @@
+namespace Synercoding.FileFormats.Pdf.Content;
+
+/// <summary>
+/// Specifies how overprinting is applied when painting in a DeviceCMYK color space.
+/// </summary>
+public enum OverprintMode
+{
+    /// <summary>
+    /// Each source color component value replaces the value previously painted for the corresponding device colorant,
+    /// regardless of what the new value is.
+    /// </summary>
+    Standard = 0,
+    /// <summary>
+    /// A tint value of 0.0 for a source color component leaves the corresponding component of the previously painted color unchanged.
+    /// </summary>
+    NonZero = 1
+}

src/Synercoding.FileFormats.Pdf/Content/Text/Fonts/CID/CIDFontWriter.cs

@@ -32,12 +32,17 @@ public static void WriteType0Font(ObjectWriter writer, PdfObjectId fontId, Font
         PdfObjectId? cidToGidMapId = null;
         Dictionary<ushort, ushort>? cidToGidMap = null;
 
+        var baseFontName = ttFont.FontName;
+
         // Write CIDSystemInfo
         _writeCIDSystemInfo(writer, cidSystemInfoId);
 
         // Handle font subsetting if enabled and we have tracked usage
         if (writerSettings.EnableSubsetting && tracker.HasUsage)
         {
+            // Subset fonts must have their name prefixed with a tag (ISO 32000-2, 9.9.2)
+            baseFontName = _getSubsetTag(ttFont, tracker) + "+" + ttFont.FontName;
+
             var subsetter = new TrueType.FontSubsetter();
             var result = subsetter.CreateSubset(ttFont, tracker);
 
@@ -58,10 +63,10 @@ public static void WriteType0Font(ObjectWriter writer, PdfObjectId fontId, Font
         }
 
         // Write font descriptor
-        _writeFontDescriptor(writer, fontDescriptorId, ttFont, fontFileId);
+        _writeFontDescriptor(writer, fontDescriptorId, ttFont, baseFontName, fontFileId);
 
         // Write CID font (descendant font) with optional CIDToGIDMap
-        _writeCIDFont(writer, descendantFontId, ttFont, fontDescriptorId, cidSystemInfoId, tracker, cidToGidMapId);
+        _writeCIDFont(writer, descendantFontId, ttFont, baseFontName, fontDescriptorId, cidSystemInfoId, tracker, cidToGidMapId);
 
         // Write CIDToGIDMap if needed
         if (cidToGidMapId != null && cidToGidMap != null)
@@ -77,7 +82,7 @@ public static void WriteType0Font(ObjectWriter writer, PdfObjectId fontId, Font
         {
             [PdfNames.Type] = PdfNames.Font,
             [PdfNames.Subtype] = PdfNames.Type0,
-            [PdfNames.BaseFont] = PdfName.Get(ttFont.FontName),
+            [PdfNames.BaseFont] = PdfName.Get(baseFontName),
             [PdfNames.Encoding] = PdfNames.IdentityH,
             [PdfNames.DescendantFonts] = new PdfArray() { descendantFontId.GetReference() },
             [PdfNames.ToUnicode] = toUnicodeId.GetReference()
@@ -152,7 +157,20 @@ private static void _writeFontFileWithSubsetting(ObjectWriter writer, PdfObjectI
         });
     }
 
-    private static void _writeFontDescriptor(ObjectWriter writer, PdfObjectId descriptorId, TrueTypeFont font, PdfObjectId fontFileId)
+    private static string _getSubsetTag(TrueTypeFont font, FontUsageTracker tracker)
+    {
+        // Deterministic tag derived from the font name and the used characters, so identical input gives identical output.
+        var input = font.FontName + "|" + new string(tracker.UsedCharacters.Order().ToArray());
+        var hash = System.Security.Cryptography.SHA256.HashData(Encoding.UTF8.GetBytes(input));
+
+        Span<char> tag = stackalloc char[6];
+        for (int i = 0; i < tag.Length; i++)
+            tag[i] = (char)( 'A' + ( hash[i] % 26 ) );
+
+        return new string(tag);
+    }
+
+    private static void _writeFontDescriptor(ObjectWriter writer, PdfObjectId descriptorId, TrueTypeFont font, string fontName, PdfObjectId fontFileId)
     {
         var tables = font.Tables;
         var head = tables.Head!;
@@ -170,7 +188,7 @@ private static void _writeFontDescriptor(ObjectWriter writer, PdfObjectId descri
         var descriptor = new PdfDictionary()
         {
             [PdfNames.Type] = PdfNames.FontDescriptor,
-            [PdfNames.FontName] = PdfName.Get(font.FontName),
+            [PdfNames.FontName] = PdfName.Get(fontName),
             [PdfNames.Flags] = new PdfNumber(flags),
             [PdfNames.FontBBox] = new PdfArray
             {
@@ -194,14 +212,14 @@ private static void _writeFontDescriptor(ObjectWriter writer, PdfObjectId descri
         });
     }
 
-    private static void _writeCIDFont(ObjectWriter writer, PdfObjectId cidFontId, TrueTypeFont font,
+    private static void _writeCIDFont(ObjectWriter writer, PdfObjectId cidFontId, TrueTypeFont font, string baseFontName,
         PdfObjectId fontDescriptorId, PdfObjectId cidSystemInfoId, FontUsageTracker tracker, PdfObjectId? cidToGidMapId = null)
     {
         var cidFont = new PdfDictionary()
         {
             [PdfNames.Type] = PdfNames.Font,
             [PdfNames.Subtype] = PdfNames.CIDFontType2,
-            [PdfNames.BaseFont] = PdfName.Get(font.FontName),
+            [PdfNames.BaseFont] = PdfName.Get(baseFontName),
             [PdfNames.CIDSystemInfo] = cidSystemInfoId.GetReference(),
             [PdfNames.FontDescriptor] = fontDescriptorId.GetReference(),
             [PdfNames.DW] = new PdfNumber(1000) // Default width

src/Synercoding.FileFormats.Pdf/IO/Filters/FlateDecode.cs

@@ -33,25 +33,10 @@ public byte[] Encode(byte[] input, IPdfDictionary? parameters)
 
         using (var outputStream = new MemoryStream())
         {
-            const CompressionLevel LEVEL = CompressionLevel.SmallestSize;
-
-            var (method, flags) = _getHeader(LEVEL);
-            outputStream.WriteByte(method);
-            outputStream.WriteByte(flags);
-
-            using (var flateStream = new DeflateStream(outputStream, LEVEL, leaveOpen: true))
-                flateStream.Write(input);
+            // ZLibStream writes both the zlib header and the Adler-32 trailer, which FlateDecode requires.
+            using (var zlibStream = new ZLibStream(outputStream, CompressionLevel.SmallestSize, leaveOpen: true))
+                zlibStream.Write(input);
             return outputStream.ToArray();
         }
     }
-
-    private (byte CompressionMethod, byte Flags) _getHeader(CompressionLevel compressionLevel)
-        => compressionLevel switch
-        {
-            CompressionLevel.Optimal => (0x78, 0x9C),
-            CompressionLevel.Fastest => (0x78, 0x5E),
-            CompressionLevel.SmallestSize => (0x78, 0xDA),
-            CompressionLevel.NoCompression => (0x78, 0x01),
-            var level => throw new NotImplementedException("Unsupported compression level: {level}")
-        };
 }

src/Synercoding.FileFormats.Pdf/PackageDetails.props

@@ -10,7 +10,7 @@
     <Product>Synercoding.FileFormats.Pdf</Product>
     <Title>Synercoding.FileFormats.Pdf</Title>
     <Description>Contains classes which makes it easy to quickly create a pdf file.</Description>
-    <PackageReleaseNotes>Fixed a broken constructor.</PackageReleaseNotes>
+    <PackageReleaseNotes>Fixed a broken constructor. Fixed FlateDecode streams missing the zlib Adler-32 trailer, which caused "incorrect data check" warnings in Ghostscript. Subset fonts now get the required six-letter tag prefix (e.g. ABCDEF+FontName). Added OverprintMode (OPM) support to the extended graphics state.</PackageReleaseNotes>
     <PackageReadmeFile>README.md</PackageReadmeFile>
     <PackageLicenseFile>LICENSE</PackageLicenseFile>
   </PropertyGroup>

src/Synercoding.FileFormats.Pdf/Primitives/PdfName.cs

@@ -108,6 +108,7 @@ public sealed class PdfName : IPdfPrimitive, IEquatable<PdfName>
         { "OneColumn",        new PdfName("OneColumn") },
         { "OP",               new PdfName("OP") },
         { "op",               new PdfName("op") },
+        { "OPM",              new PdfName("OPM") },
         { "Ordering",         new PdfName("Ordering") },
         { "Overlay",          new PdfName("Overlay") },
         { "P",                new PdfName("P") },

src/Synercoding.FileFormats.Pdf/Primitives/PdfNames.cs

@@ -212,6 +212,8 @@ public static PdfName OP
         => PdfName.Get(nameof(OP));
     public static PdfName op
         => PdfName.Get(nameof(op));
+    public static PdfName OPM
+        => PdfName.Get(nameof(OPM));
     public static PdfName Ordering
         => PdfName.Get(nameof(Ordering));
     public static PdfName Overlay

tests/Synercoding.FileFormats.Pdf.Tests/Content/Text/Fonts/TrueType/FontSubsettingVisualTests.cs

@@ -112,6 +112,30 @@ public void Test_SubsetFont_RepeatedCharacters_VisuallyEqual()
             $"Subset PDF should be smaller. Full: {fullFontPdf.Length}, Subset: {subsetFontPdf.Length}");
     }
 
+    [Fact]
+    public void Test_SubsetFont_FontNames_HaveSubsetTag()
+    {
+        var fontName = Font.Load(_testFontPath).FontName;
+        var pdf = System.Text.Encoding.ASCII.GetString(_createPdfWithFont("Tagged", enableSubsetting: true));
+
+        var matches = System.Text.RegularExpressions.Regex.Matches(pdf, @"/(?:BaseFont|FontName) */([^\s/<>\[\]]+)");
+
+        // Type0 BaseFont, CIDFont BaseFont and FontDescriptor FontName
+        Assert.Equal(3, matches.Count);
+        var names = matches.Select(m => m.Groups[1].Value).Distinct().ToArray();
+        var name = Assert.Single(names);
+        Assert.Matches(@"^[A-Z]{6}\+", name);
+        Assert.EndsWith(fontName, name);
+    }
+
+    [Fact]
+    public void Test_FullFont_FontNames_HaveNoSubsetTag()
+    {
+        var pdf = System.Text.Encoding.ASCII.GetString(_createPdfWithFont("Untagged", enableSubsetting: false));
+
+        Assert.DoesNotMatch(@"/(?:BaseFont|FontName) */[A-Z]{6}\+", pdf);
+    }
+
     /// <summary>
     /// Creates a PDF with the specified text and subsetting setting.
     /// Based on _writeTextOnlyPdf from ConsoleTester.

tests/Synercoding.FileFormats.Pdf.Tests/IO/Filters/FlateDecodeTests.cs

@@ -0,0 +1,38 @@
+using Synercoding.FileFormats.Pdf.IO.Filters;
+using System.Buffers.Binary;
+using System.IO.Compression;
+
+namespace Synercoding.FileFormats.Pdf.Tests.IO.Filters;
+
+public class FlateDecodeTests
+{
+    [Fact]
+    public void Test_Encode_WritesZlibHeaderAndAdler32Trailer()
+    {
+        var input = System.Text.Encoding.ASCII.GetBytes("Wikipedia");
+
+        var encoded = new FlateDecode().Encode(input, null);
+
+        // zlib header: deflate method with 32K window, and a header checksum that is a multiple of 31
+        Assert.Equal(0x78, encoded[0]);
+        Assert.Equal(0, ( ( encoded[0] << 8 ) | encoded[1] ) % 31);
+
+        // Adler-32 of "Wikipedia" is 0x11E60398, stored big-endian at the end of the stream
+        Assert.Equal(0x11E60398u, BinaryPrimitives.ReadUInt32BigEndian(encoded.AsSpan(encoded.Length - 4)));
+    }
+
+    [Fact]
+    public void Test_Encode_RoundTripsThroughZlib()
+    {
+        var input = new byte[10_000];
+        new Random(42).NextBytes(input);
+
+        var encoded = new FlateDecode().Encode(input, null);
+
+        using var decompressed = new MemoryStream();
+        using (var zlib = new ZLibStream(new MemoryStream(encoded), CompressionMode.Decompress))
+            zlib.CopyTo(decompressed);
+
+        Assert.Equal(input, decompressed.ToArray());
+    }
+}

tests/Synercoding.FileFormats.Pdf.Tests/Synercoding.FileFormats.Pdf.Tests.csproj

@@ -22,13 +22,13 @@
     <Content Include="..\Font-Files\juliett-font\JuliettRegular-7OXnA.ttf" Link="TestFiles\Fonts\JuliettRegular-7OXnA.ttf">
       <CopyToOutputDirectory>PreserveNewest</CopyToOutputDirectory>
     </Content>
-    <Content Include="..\Pdf-Files\encrypted-aes-128bits.pdf" Link="TestFiles\Pdfs\encrypted-aes-128bits.pdf" >
+    <Content Include="..\Pdf-Files\encrypted-aes-128bits.pdf" Link="TestFiles\Pdfs\encrypted-aes-128bits.pdf">
       <CopyToOutputDirectory>PreserveNewest</CopyToOutputDirectory>
     </Content>
     <Content Include="..\Pdf-Files\encrypted-aes-256bits.pdf" Link="TestFiles\Pdfs\encrypted-aes-256bits.pdf">
       <CopyToOutputDirectory>PreserveNewest</CopyToOutputDirectory>
     </Content>
-    <Content Include="..\Pdf-Files\encrypted-rc4-128bits.pdf" Link="TestFiles\Pdfs\encrypted-rc4-128bits.pdf" >
+    <Content Include="..\Pdf-Files\encrypted-rc4-128bits.pdf" Link="TestFiles\Pdfs\encrypted-rc4-128bits.pdf">
       <CopyToOutputDirectory>PreserveNewest</CopyToOutputDirectory>
     </Content>
   </ItemGroup>
@@ -50,5 +50,19 @@
     <PackageReference Include="bblanchon.PDFium.Linux" VersionOverride="$(PdfiumVersion)" Condition="'$(RuntimeIdentifier)'=='' Or $(RuntimeIdentifier.StartsWith('linux-'))" />
     <PackageReference Include="bblanchon.PDFium.macOS" VersionOverride="$(PdfiumVersion)" Condition="'$(RuntimeIdentifier)'=='' Or '$(RuntimeIdentifier)'=='osx-x64' Or '$(RuntimeIdentifier)'=='osx-arm64'" />
   </ItemGroup>
+
+  <ItemGroup>
+    <PackageReference Update="coverlet.collector">
+      <PrivateAssets>all</PrivateAssets>
+      <IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
+    </PackageReference>
+  </ItemGroup>
+
+  <ItemGroup>
+    <PackageReference Update="xunit.runner.visualstudio">
+      <PrivateAssets>all</PrivateAssets>
+      <IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
+    </PackageReference>
+  </ItemGroup>
 
 </Project>