Content-Disposition: filename* vs. Content-Type: name

[cyberangel]

Hello.

First of all thank you for your excellent library for handling MIME. It saves us a lot of headache parsing e-mails.

I'm having an issue with file attachments with names encoded in UTF-8, as in this example:

Content-Type: image/jpeg; name="=?utf-8?q?SBD_=C5=A0kodov=C3=A1k=2Ejpg?="
Content-Transfer-Encoding: base64
Content-Disposition: attachment; size=1049971;
    filename*="utf-8''SBD%20%C5%A0kodov%C3%A1k%2Ejpg"
Content-ID: <m1548006d0c3m73d>

The filename is supposed to be "SBD Škodovák.jpg" and the Content-Type.Name properly decodes the quoted string.
But the ContentDisposition.FileName decodes as "SBD Škodovák.jpg"". I see it is URL encoded ->
"utf-8''SBD%20%C5%A0kodov%C3%A1k%2Ejpg" -> HttpUtility.UrlDecode -> "utf-8''SBD Škodovák.jpg".

Clearly there is something amiss and I've modified my program to use The Content-Type.Name property instead.
But I would like to hear your opinion - is it a bug in MimeKit.ContentDisposition or is the client composing the message at fault?

Thanks.
Honza

[jstedfast]

The problem is that the filename parameter is inside of a quoted string (it's not supposed to be), which means that the charset is parsed as "utf-8 (note the leading dquote) which of course is a bogus charset and so it falls back to iso-8859-1.

[jstedfast]

I've fixed the code to work around this issue.

The plan is to release a new version to nuget.org this weekend.

[cyberangel]

Thank you.

[michael-o]

FTR, @cyberangel Your content id is wrong.

[cyberangel]

FTR, @cyberangel Your content id is wrong.

This was copied without change, so whatever the Content-ID was there was copied with it.