Page MenuHome GnuPG

German special characters will not be decrypted in a proper way
Closed, WontfixPublic


Current version was installed and a PGP encrypted and signed message was written. Encyption of the German pristin test was done in the Kleopatra notepad and then copied into Outlook 2016 to send it.

I opened the already sent message in Outlook and I decrypted it there and recognized that all special German characters (Umlaute) could not be decrypted in a proper way. A minor problem, but I want to address it in the forum.


Kleopatra 3.1.11 - gpg4win 3.1.11

Event Timeline

aheinecke claimed this task.
aheinecke added subscribers: werner, aheinecke.


there is no way to handle this correctly with such messages. The PGP Standard says that PGP Messages may have every encoding. This is a reason why we always talk about "you should use PGP/MIME" as this is basically PGP with added handling for content meta information like the text encoding.

Usually it is UTF-8 though, but in the case of GpgOL we take the encoding preferably from the Mail header. So we use the encoding which the Mail says it is in.

If you copy & paste a PGP Message into Outlook it does not know that there are special characters in there and Outlook then says. "Hey all that base64 code can be encoded as 7-Bit ASCII" and the sent mail will use the local 8 bit encoding. So your mail Probably will claim "I'm a Latin-1 Mail" while Kleoptra encoded the special characters as UTF-8.

When decrypting GpgOL then interprets the content of the PGP Message with Latin-1.

@werner always says that I should interpret PGP Messages always as UTF-8 even though the standard says differently. And e.g. in T4130 the PGP Message is not UTF-8 and I have seen others out there that send mails not in UTF-8.

I am closing this as "Wontfix" although it rather is a "Cantfix". The solution is not to use PGP Inline and rather to use PGP MIME. T4130 is still open regarding PGP Inline handling. We might change it in the future to use UTF-8 by default, then at least we will be compatible with Kleopatra as Kleopatra always uses UTF-8 encoding.

Thanks for the report anyway. It is always interesting to see what the problems are in the wild.