GnuPG in 2016

This is an overview of what happened in the first half of 2016 in the GnuPG project and community.


There has been one release of the current stable branch of GnuPG, version 2.0.30. Our development branch GnuPG modern saw two releases, 2.1.11, and 2.1.12.

Among many bugfixes and other improvements GnuPG 2.1.11 added a new command --export-ssh-key that replaces the gpgkey2ssh tool, and we now use the CA certificate of to secure communications with the servers in the pool without further configuration. Furthermore, we now print a warning if a GnuPG component is using an older version of our backend daemons.

2.1.12 brought support for a new experimental "Web Key Directory" key location service, read-support for a new private key protection format and the new extended private key format, improved Tofu support, use of the new libusb 1.0 API, and countless small features and fixes.

Daniel Genkin, Lev Pachmanov, Itamar Pipman, and Eran Tromer from the Tel Aviv University demonstrated a practical way to extract ECDH keys via low-bandwidth electromagnetic attacks on PCs1. Libgcrypt version 1.6.5 and an updated Windows installer for GnuPG 2.1.11 has been released to mitigate this attack.

In April we released a new stable version of Libgcrypt, version 1.7. It retains full API and ABI compatibiliy to the 1.6 series. Its main features are new algorithms, curves, and performance improvements.

Andre worked on PGP/MIME support in GpgOL, and the current stable version 2.3.1 of Gpg4win includes an option to enable experimental support for sending GPG/MIME and S/MIME. He also worked on Kleopatra, removing some dependencies on KDE components and DBUS, and porting it to newer KDE frameworks, making it easier to build, maintain and hack. A beta version of Gpg4win 3.0.0 has been released featuring the leaner Kleopatra with newer KDE libraries. Andre also worked on registering file extensions in Windows integrating encryption and decryption of files into the desktop environment, and implementing a 'Show Password' feature for the pinentries.

Jussi worked on libgcrypt, most notably adding an ARM assembly implementation of SHA-512, and Intel PCMUL implementations for CRC algorithms.

Justus worked on a new test framework for GnuPG and related projects, a new extensible storage format for private keys, implemented elliptic-curve cryptography in libssh using libgcrypt as backend, and triaged and fixed GnuPG bugs.

Kai evaluated Mailpile. As the future of Mozilla Thunderbird - and with it Enigmails - is uncertain, we decided to explore alternative end-user-friendly mail clients with support for GnuPG.

Gniibe worked on GnuPG and libgcrypt. He ported GnuPG's scdaemon to libusb 1.0, and worked with the team from Tel Aviv University to protect libgcrypt against the sidechannel attack. Gniibe also maintains GnuPG-related packages in Debian. He updated both Poldi and Scute. Gniibe designs and sells the FST-1, a small microcontroller that depending on the firmware either acts as an OpenPGP smart card or a hardware random number generator. As the stock is running low and some components cannot be sourced any longer, he is working on the successor featuring a simpler design, a button, and support for elliptic curve cryptography.


Werner received the Award for the Advancement of Free Software for his work on GnuPG.


Andre asked how to obtain a key in a format suitable for SSH's authorized_keys file from an OpenPGP public certificate now that gpgkey2ssh has been deprecated. The discussion was taken to the bug tracker, and a solution released with GnuPG 2.1.11.

Lachlan J. Gunn, Andrew Allison, and Derek Abbott wrote a paper about how Tor can be used to detect malicious key servers2. The announcement contains more information including a link to their implementation.

Daniel reported that there have been discussions on building a Live-CD for securely managing OpenPGP master keys and smart cards over at debian-devel, and he asked if anyone knows about such an effort, or has comments on his proposal. The discussion went into all kinds of directions, but the consensus was that this is a useful idea.

Bernhard wrote about work being done for the EasyGpg2016 project. They produced user archetypes and stories that should guide the refinement of GnuPG and related tools, and a key distribution concept.


We received this year until now about 90 donations summing up to 3000 Euro. Thanks for that help. Sure, that is not enough to pay our costs, but fortunately we still have enough money in our accounts to keep bread, water, and beer on our tables for this year and somewhat longer. We will eventually run a new donation campaign, though.

It is probably less known that that we offer SEPA payments which can be used for recurring donations. We have not received many payments through SEPA yet and thus this method is only semi-automated and the cause for delays between a donation and a confirmation mail. There is an obvious way to make us automate this ;-)

Lastly, let us confirm that we were meanwhile able to clarify our perceived problem with the Facebook donation promise. This was all due to an unfortunate misunderstanding between us. Facebook will keep on supporting GnuPG in 2016 with a donation of 50000 USD.

About this news posting

We try to write a news posting each month (though we must admit that we slipped a little in 2016). However, other work may have a higher priority (e.g. security fixes) and thus there is no promise for a fixed publication date. If you have an interesting topic for a news posting, please send it to us. A regular summary of the mailing list discussions would make a nice column on this news.



Daniel Genkin, Lev Pachmanov, Itamar Pipman, Eran Tromer, ECDH key-extraction via low-bandwidth electromagnetic attacks on PCs, proc. RSA Conference Cryptographers' Track (CT-RSA) 2016, LNCS 9610, 219-235, Springer, 2016,


Lachlan J. Gunn, Andrew Allison, Derek Abbott, Verifying public keys without trust: How anonymity can guarantee data integrity, arXiv preprint arXiv:1602.03316, 2016,