message/rfc822 attachment with blank Content-Transfer-Encoding header causes attachment to become corrupted when loaded and prepared by mimekit

[The-Nutty]

Describe the bug
A clear and concise description of what the bug is.

Platform (please complete the following information):

• OS: Windows, Linux
• .NET Runtime: NET8
• .NET Framework: NET8
• MimeKit Version: 4.1.0 and 4.6.0 tested

To Reproduce

1. In a new .NET console app with Mimekit installed, paste the following

var msg = MimeMessage.Load("demo.eml");
msg.Prepare(EncodingConstraint.None);

msg.WriteTo("output.eml");

With demo.eml set to copy to output directory.

2. Run the script and it will generate output.eml
3. Open both in a message viewer (IE outlook): demo.eml will work fine with the attached email text being readable "A medium-severity alert has been triggered....." output.eml the attachment will instead show "--=3D-th9mpsHWlvyU5398Ajg2gQ=3D=3D
   Content-Type: multipart/alternative; boundary=3D"=3D-opcn9iLL5vM98wIjSIhg+Q=
   =3D=3D"

--=3D-opcn9iLL5vM98wIjSIhg+Q=3D=3D
......."

Expected behaviour
The attachment should not be corrupted when loaded, prepared and saved by mimekit.

Additional context
This seems to only occur when time message/rfc822 header has a Content-Transfer-Encoding: header, this does not happen with either Content-Transfer-Encoding: quoted-printable or a no Content-Transfer-Encoding at all.

[jstedfast]

My gut instinct is that this is where things are going wrong:

https://github.com/jstedfast/MimeKit/blob/master/MimeKit/ParserOptions.cs#L358-L383

Technically, what is likely going wrong is the IsEncoded() logic here: https://github.com/jstedfast/MimeKit/blob/master/MimeKit/ParserOptions.cs#L282-L294

but the first link I pasted contains the "justification" for said logic and why it's being used at all.

2 possible fixes that might work:

1. Modify IsEncoded(ContentEncoding encoding) to add ContentEncoding.Default to the list of values that return false
2. Modify IsEncoded(IList<Header> headers) to return false if the TryParse() fails

[jstedfast]

I'll want to add unit tests for this (for both MimeParser and ExperimentalMimeParser) before I call this "complete", but in the meantime, I figured I'd push those potential fixes and you'll be able to test the nuget package (that will get pushed to myget.org) when the CI/CD pipeline finishes pushing out a new build/package.

[jstedfast]

FWIW, I'm also planning to push a new release probably this weekend or maybe before then if I get the free time.

[The-Nutty]

Brilliant, thanks for your help really appreciate the quick response.

[jstedfast]

Just added some unit tests and it seems like it works with my fixes.