Port IPTC encoding fix from Java library

Ports drewnoakes/metadata-extractor#615
drewnoakes · May 22, 2023 · c874c0e · c874c0e
1 parent 6e92951
commit c874c0e
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 1 deletion.
diff --git a/MetadataExtractor.Tests/Formats/Iptc/Iso2022ConverterTest.cs b/MetadataExtractor.Tests/Formats/Iptc/Iso2022ConverterTest.cs
@@ -11,7 +11,9 @@ public sealed class Iso2022ConverterTest
         public void ConvertEscapeSequenceToEncodingName()
         {
             Assert.Equal("UTF-8", Iso2022Converter.ConvertEscapeSequenceToEncodingName(new byte[] { 0x1B, 0x25, 0x47 }));
+            //Assert.Equal("ISO-8859-1", Iso2022Converter.ConvertEscapeSequenceToEncodingName(new byte[] { 0x1B, 0x2E, 0x41 }));
             Assert.Equal("ISO-8859-1", Iso2022Converter.ConvertEscapeSequenceToEncodingName(new byte[] { 0x1B, 0xE2, 0x80, 0xA2, 0x41 }));
+            Assert.Equal("ISO-8859-1", Iso2022Converter.ConvertEscapeSequenceToEncodingName(new byte[] { 0x1B, 0x2D, 0x41 }));
             Assert.Null(Iso2022Converter.ConvertEscapeSequenceToEncodingName(new byte[] { 1, 2, 3, 4 }));
         }
     }

diff --git a/MetadataExtractor/Formats/Iptc/Iso2022Converter.cs b/MetadataExtractor/Formats/Iptc/Iso2022Converter.cs
@@ -7,6 +7,7 @@ public static class Iso2022Converter
         private const int Dot = 0xe280a2;
         private const byte LatinCapitalA = (byte)'A';
         private const byte LatinCapitalG = (byte)'G';
+        private const byte MinusSign = (byte)'-';
         private const byte PercentSign = (byte)'%';
         private const byte Esc = 0x1B;
 
@@ -19,12 +20,15 @@ public static class Iso2022Converter
             if (bytes.Length > 3 && bytes[0] == Esc && (bytes[3] | (bytes[2] << 8) | (bytes[1] << 16)) == Dot && bytes[4] == LatinCapitalA)
                 return "ISO-8859-1";
 
+            if (bytes.Length > 2 && bytes[0] == Esc && bytes[1] == MinusSign && bytes[2] == LatinCapitalA)
+                return "ISO-8859-1";
+
             return null;
         }
 
         /// <summary>Attempts to guess the encoding of a string provided as a byte array.</summary>
         /// <remarks>
-        /// Encodings trialled are, in order:
+        /// Encodings trialed are, in order:
         /// <list type="bullet">
         ///   <item>UTF-8</item>
         ///   <item>ISO-8859-1</item>