How can I detect if a .NET StreamReader found a UTF8 BOM on the underlying stream?

Question

I get a FileStream(filename,FileMode.Open,FileAccess.Read,FileShare.ReadWrite) and then a StreamReader(stream,true).

Is there a way I can check if the stream started with a UTF8 BOM? I am noticing that files without the BOM are read as UTF8 by the StreamReader.

How can I tell them apart?

Solution 1

Rather than hardcoding the bytes, it is prettier to use the API

public string ConvertFromUtf8(byte[] bytes)
{
  var enc = new UTF8Encoding(true);
  var preamble = enc.GetPreamble();
  if (preamble.Where((p, i) => p != bytes[i]).Any()) 
    throw new ArgumentException("Not utf8-BOM");
  return enc.GetString(bytes.Skip(preamble.Length).ToArray());
}

Solution 2

You can detect whether the StreamReader encountered a BOM by initializing it with a BOM-less UTF8 encoding and checking to see if CurrentEncoding changes after the first read.

var utf8NoBom = new UTF8Encoding(false);
using (var reader = new StreamReader(file, utf8NoBom))
{
    reader.Read();
    if (Equals(reader.CurrentEncoding, utf8NoBom))
    {
        Console.WriteLine("No BOM");
    }
    else
    {
        Console.WriteLine("BOM detected");
    }
}

Comments

[Cameron Taggart]

I never would have thought that this would work. Thanks! It is really too bad that the opposite isn't true. You can't pass int UTF8Encoding(true) and have it return UTF8Encoding(false).

[Martin]

@carlo-v-dango, I'd recommend adding some kind of null-check since bytes may be empty if file is empty. if (preamble.Where((p, i) => bytes.Length > i && p != bytes[i]).Any()) or whatever floats your boat.