Properly handle ZapfDingbats font for TrueTypeSimpleFont and add tests

UglyToad · Jan 5, 2025 · 53cf4f2 · 53cf4f2
1 parent 4430a01
commit 53cf4f2
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 7 deletions.
diff --git a/src/UglyToad.PdfPig.Tests/Integration/Documents/capas.pdf b/src/UglyToad.PdfPig.Tests/Integration/Documents/capas.pdf
diff --git a/src/UglyToad.PdfPig.Tests/Integration/ZapfDingbatsTests.cs b/src/UglyToad.PdfPig.Tests/Integration/ZapfDingbatsTests.cs
@@ -4,6 +4,17 @@
 
     public class ZapfDingbatsTests
     {
+        [Fact]
+        public void TrueTypeSimpleFont1()
+        {
+            using (var document = PdfDocument.Open(IntegrationHelpers.GetDocumentPath("capas")))
+            {
+                var page = document.GetPage(18);
+                // ZapfDingbats characters are spaces
+                Assert.Contains(" ", page.Letters.Select(l => l.Value));
+            }
+        }
+
         [Fact]
         public void Type1Standard14Font1()
         {
@@ -17,7 +28,6 @@ public void Type1Standard14Font1()
         [Fact]
         public void Type1Standard14Font2()
         {
-            // This document does not actually contain circular references
             using (var document = PdfDocument.Open(IntegrationHelpers.GetDocumentPath("MOZILLA-LINK-5251-1")))
             {
                 var page = document.GetPage(1);
@@ -33,7 +43,6 @@ public void Type1Standard14Font2()
         [Fact]
         public void Type1FontSimple1()
         {
-            // This document does not actually contain circular references
             using (var document = PdfDocument.Open(IntegrationHelpers.GetDocumentPath("MOZILLA-2775-1")))
             {
                 var page = document.GetPage(11);
@@ -44,7 +53,6 @@ public void Type1FontSimple1()
         [Fact]
         public void Type1FontSimple2()
         {
-            // This document does not actually contain circular references
             using (var document = PdfDocument.Open(IntegrationHelpers.GetDocumentPath("PDFBOX-492-4.jar-8")))
             {
                 var page = document.GetPage(1);

diff --git a/src/UglyToad.PdfPig/PdfFonts/Simple/TrueTypeSimpleFont.cs b/src/UglyToad.PdfPig/PdfFonts/Simple/TrueTypeSimpleFont.cs
@@ -31,6 +31,8 @@ internal sealed class TrueTypeSimpleFont : IFont
 
         private readonly double[] widths;
 
+        private readonly bool isZapfDingbats;
+
 #nullable disable
         public NameToken Name { get; }
 #nullable enable
@@ -63,8 +65,7 @@ public TrueTypeSimpleFont(
             Details = descriptor?.ToDetails(Name?.Data)
                       ?? FontDetails.GetDefault(Name?.Data);
 
-            // Assumption is ZapfDingbats is not possible here. We need to change the behaviour if not the case
-            System.Diagnostics.Debug.Assert(!(encoding is ZapfDingbatsEncoding || Details.Name.Contains("ZapfDingbats")));
+            isZapfDingbats = encoding is ZapfDingbatsEncoding || Details.Name.Contains("ZapfDingbats");
         }
 
         public int ReadCharacterCode(IInputBytes bytes, out int codeLength)
@@ -100,12 +101,22 @@ public bool TryGetUnicode(int characterCode, [NotNullWhen(true)] out string? val
             // If the font is a simple font that uses one of the predefined encodings MacRomanEncoding, MacExpertEncoding, or WinAnsiEncoding...
 
             //  Map the character code to a character name.
-            var encodedCharacterName = encoding.GetName(characterCode);
+            var name = encoding.GetName(characterCode);
 
             // Look up the character name in the Adobe Glyph List or additional Glyph List.
             try
             {
-                value = GlyphList.AdobeGlyphList.NameToUnicode(encodedCharacterName);
+                if (isZapfDingbats)
+                {
+                    value = GlyphList.ZapfDingbats.NameToUnicode(name);
+
+                    if (value is not null)
+                    {
+                        return true;
+                    }
+                }
+
+                value = GlyphList.AdobeGlyphList.NameToUnicode(name);
             }
             catch
             {