When using a try/catch for charset conversion, make sure the Encoding…

… does not use a fallback char Fixes issue #88
jstedfast · Dec 12, 2014 · 5db8684 · 5db8684
1 parent 17399d8
commit 5db8684
Show file tree

Hide file tree

Showing 4 changed files with 11 additions and 6 deletions.
diff --git a/MimeKit/InternetAddress.cs b/MimeKit/InternetAddress.cs
@@ -191,8 +191,8 @@ static bool TryParseLocalPart (byte[] text, ref int index, int endIndex, bool th
 					return false;
 
 				try {
-					token.Append (Encoding.UTF8.GetString (text, start, index - start));
-				} catch (Exception ex) {
+					token.Append (CharsetUtils.UTF8.GetString (text, start, index - start));
+				} catch (DecoderFallbackException ex) {
 					if (throwOnError)
 						throw new ParseException ("Internationalized local-part tokens may only contain UTF-8 characters.", start, start, ex);
 

diff --git a/MimeKit/TextPart.cs b/MimeKit/TextPart.cs
@@ -198,8 +198,8 @@ public TextPart () : base ("text", "plain")
 
 					if (encoding == null) {
 						try {
-							return Encoding.UTF8.GetString (content, 0, (int) memory.Length);
-						} catch {
+							return CharsetUtils.UTF8.GetString (content, 0, (int) memory.Length);
+						} catch (DecoderFallbackException) {
 							// fall back to iso-8859-1
 							encoding = Encoding.GetEncoding (28591); // iso-8859-1
 						}

diff --git a/MimeKit/Utils/CharsetUtils.cs b/MimeKit/Utils/CharsetUtils.cs
@@ -31,6 +31,8 @@
 #if PORTABLE
 using EncoderReplacementFallback = Portable.Text.EncoderReplacementFallback;
 using DecoderReplacementFallback = Portable.Text.DecoderReplacementFallback;
+using EncoderExceptionFallback = Portable.Text.EncoderExceptionFallback;
+using DecoderExceptionFallback = Portable.Text.DecoderExceptionFallback;
 using DecoderFallbackBuffer = Portable.Text.DecoderFallbackBuffer;
 using DecoderFallback = Portable.Text.DecoderFallback;
 using Encoding = Portable.Text.Encoding;
@@ -41,6 +43,9 @@
 namespace MimeKit.Utils {
 	static class CharsetUtils
 	{
+		// Note: Encoding.UTF8.GetString() replaces invalid bytes with a unicode '?' character,
+		// so we use our own UTF8 instance when using GetString() if we do not want it to do that.
+		public static readonly Encoding UTF8 = Encoding.GetEncoding (65001, new EncoderExceptionFallback (), new DecoderExceptionFallback ());
 		static readonly Dictionary<string, int> aliases;
 
 		static CharsetUtils ()

diff --git a/MimeKit/Utils/ParseUtils.cs b/MimeKit/Utils/ParseUtils.cs
@@ -206,8 +206,8 @@ static bool TryParseDotAtom (byte[] text, ref int index, int endIndex, byte[] se
 					index++;
 
 				try {
-					token.Append (Encoding.UTF8.GetString (text, start, index - start));
-				} catch (ParseException ex) {
+					token.Append (CharsetUtils.UTF8.GetString (text, start, index - start));
+				} catch (DecoderFallbackException ex) {
 					if (throwOnError)
 						throw new ParseException ("Internationalized domains may only contain UTF-8 characters.", start, start, ex);