AvalonEdit encoding detection: add support for UTF-8 without BOM.

We now don't add a BOM when saving an UTF-8 that didn't have one when it was loaded.

src/AddIns/DisplayBindings/AvalonEdit.AddIn/Src/ChooseEncodingDialog.xaml

@@ -13,13 +13,8 @@
 			Content="{core:Localize Dialog.Options.IDEOptions.TextEditor.General.FontGroupBox.FileEncodingLabel}" />
 		<ComboBox
 			Name="encodingComboBox"
-			Margin="16,0,8,8">
-			<ComboBox.ItemTemplate>
-				<DataTemplate>
-					<TextBlock Text="{Binding DisplayName}"/>
-				</DataTemplate>
-			</ComboBox.ItemTemplate>
-		</ComboBox>
+			DisplayMemberPath="DisplayName"
+			Margin="16,0,8,8" />
 		<widgets:UniformGridWithSpacing
 			Columns="2"
 			Margin="4,4,10,10"

src/Libraries/AvalonEdit/ICSharpCode.AvalonEdit/TextEditor.cs

@@ -946,11 +946,24 @@ namespace ICSharpCode.AvalonEdit
 			}
 		}
 		
+		/// <summary>
+		/// Encoding dependency property.
+		/// </summary>
+		public static readonly DependencyProperty EncodingProperty =
+			DependencyProperty.Register("Encoding", typeof(Encoding), typeof(TextEditor));
+		
 		/// <summary>
 		/// Gets/sets the encoding used when the file is saved.
 		/// </summary>
+		/// <remarks>
+		/// The <see cref="Load(Stream)"/> method autodetects the encoding of the file and sets this property accordingly.
+		/// The <see cref="Save(Stream)"/> method uses the encoding specified in this property.
+		/// </remarks>
 		[DesignerSerializationVisibility(DesignerSerializationVisibility.Hidden)]
-		public Encoding Encoding { get; set; }
+		public Encoding Encoding {
+			get { return (Encoding)GetValue(EncodingProperty); }
+			set { SetValue(EncodingProperty, value); }
+		}
 		
 		/// <summary>
 		/// Saves the text to the stream.
@@ -962,8 +975,11 @@ namespace ICSharpCode.AvalonEdit
 		{
 			if (stream == null)
 				throw new ArgumentNullException("stream");
-			StreamWriter writer = new StreamWriter(stream, this.Encoding ?? Encoding.UTF8);
-			writer.Write(this.Text);
+			var encoding = this.Encoding;
+			var document = this.Document;
+			StreamWriter writer = encoding != null ? new StreamWriter(stream, encoding) : new StreamWriter(stream);
+			if (document != null)
+				document.WriteTextTo(writer);
 			writer.Flush();
 			// do not close the stream
 			this.IsModified = false;

src/Libraries/AvalonEdit/ICSharpCode.AvalonEdit/Utils/FileReader.cs

@@ -36,6 +36,22 @@ namespace ICSharpCode.AvalonEdit.Utils
 			}
 		}
 		
+		static bool IsASCIICompatible(Encoding encoding)
+		{
+			byte[] bytes = encoding.GetBytes("Az");
+			return bytes.Length == 2 && bytes[0] == 'A' && bytes[1] == 'z';
+		}
+		
+		static Encoding RemoveBOM(Encoding encoding)
+		{
+			switch (encoding.CodePage) {
+				case 65001: // UTF-8
+					return UTF8NoBOM;
+				default:
+					return encoding;
+			}
+		}
+		
 		/// <summary>
 		/// Reads the content of the given stream.
 		/// </summary>
@@ -125,6 +141,8 @@ namespace ICSharpCode.AvalonEdit.Utils
 			}
 		}
 		
+		static readonly Encoding UTF8NoBOM = new UTF8Encoding(false);
+		
 		static StreamReader AutoDetect(Stream fs, byte firstByte, byte secondByte, Encoding defaultEncoding)
 		{
 			int max = (int)Math.Min(fs.Length, 500000); // look at max. 500 KB
@@ -187,21 +205,20 @@ namespace ICSharpCode.AvalonEdit.Utils
 			fs.Position = 0;
 			switch (state) {
 				case ASCII:
+					return new StreamReader(fs, IsASCIICompatible(defaultEncoding) ? RemoveBOM(defaultEncoding) : Encoding.ASCII);
 				case Error:
-					// when the file seems to be ASCII or non-UTF8,
+					// When the file seems to be non-UTF8,
 					// we read it using the user-specified encoding so it is saved again
 					// using that encoding.
 					if (IsUnicode(defaultEncoding)) {
 						// the file is not Unicode, so don't read it using Unicode even if the
 						// user has choosen Unicode as the default encoding.
 						
-						// If we don't do this, SD will end up always adding a Byte Order Mark
-						// to ASCII files.
 						defaultEncoding = Encoding.Default; // use system encoding instead
 					}
-					return new StreamReader(fs, defaultEncoding);
+					return new StreamReader(fs, RemoveBOM(defaultEncoding));
 				default:
-					return new StreamReader(fs);
+					return new StreamReader(fs, UTF8NoBOM);
 			}
 		}
 	}