How does an email client read the content-type headers for encoding?

Question

It is possible to send an email with different content types: text/html, text/plain, mime, etc. It also is possible to use different encodings, including (according to the RFCs) for header fields: us-ascii, utf8, etc.

How do you solve the chicken and egg problem? The content-type header is just one of several headers. If the headers can be any encoding, how does a mail server or client know how to read the content-type header if it does not know what encoding the headers themselves are in?

I can see it if the first line, eg had to be the content-type and it had to be in a pre-agreed encoding, (eg ascii), but that is not the case.

How do you parse a stream of bytes whose encoding is embedded as a string inside that very same stream?

Answer 1

Headers are defined to be in ascii. They can be in utf-8 if agreed to out of band, such as via the smtp or imap utf-8 capability extensions.

Internationalization in headers is performed via "encoded words", where the encoding is part of the header data. (This looks like a string such as =?iso8859-1?q?sample_header_data?=). See rfc2047.

Content Type headers do not apply to headers themselves, only the body content.