Subject mime field is in ASCII. Every character excluded by the ASCII table has to be Q/encoded or base64/encoded. Content-Type field in the header has also nothing to do with the way subject is encoded. Am I correct?
However (and unfortunately) some clients (read Microsoft Outlook 6 for example) insert a string encoded in whatever (BIG5 for example) in the header, without specifying with q/base64 encoding that the string is in BIG5. How can i handle these wrongly-encoded emails? Is there a standard way, to parse these?
My goal is to have the biggest compatibility possible, even by using 3rd part paid programs; how can i do that? (sorry for my buggy english)
Subject header encoding has nothing to do with Content-Type header. There is no "perfect" way to handle Subject. I've implemented this just by a hack that tries to see if all characters of text fit in big5, if not then try next encoding in order.
Big5, utf-8, latin-1, q/base64 and finally ascii