ByteString.ToString is sometimes broken for Unicode encoding #3147
Description
Hi.
Currently, ByteString.ToString(Encoding) calls Encoding.GetString for each underlying segment that comprises the ByteString. Unfortunately, this breaks when you use an encoding such as System.Text.Encoding.Unicode and the underlying segments do not represent entire characters.
Here's a quick repro that works on Akka.NET 1.3.1 (tested on Windows, but it's probably the same everywhere):
[Fact]
public void UnicodeByteString_Failure()
{
const string expected = "ABC";
Encoding encoding = Encoding.Unicode;
byte[] rawData = encoding.GetBytes(expected);
ByteString data = ByteString.Empty;
data += ByteString.CopyFrom(rawData, 0, 3); // 1 and a half characters
data += ByteString.CopyFrom(rawData, 3, 3); // 1 and a half characters
Assert.Equal(rawData.Length, data.Count);
string actual = data.ToString(encoding);
Assert.Equal(expected, actual);
}Note that if you compact the ByteString before calling ToString then it works fine.
[Fact]
public void UnicodeByteString_Success()
{
const string expected = "ABC";
Encoding encoding = Encoding.Unicode;
byte[] rawData = encoding.GetBytes(expected);
ByteString data = ByteString.Empty;
data += ByteString.CopyFrom(rawData, 0, 3);
data += ByteString.CopyFrom(rawData, 3, 3);
Assert.Equal(rawData.Length, data.Count);
data = data.Compact();
string actual = data.ToString(encoding);
Assert.Equal(expected, actual);
}I'd suggest that perhaps ToString should either call Compact or manually assemble a byte array before calling Encoding.GetString.
I'm happy to open a PR for this but wanted to discuss it with you, first, since it technically represents a change in behaviour.