Werner Koch wk at gnupg.org
Tue Mar 29 16:19:11 CEST 2016

On Fri, 25 Mar 2016 15:32, bre at pagekite.net said:

> This is a chicken-and-egg problem. Until the tool is made widely
> available, people will not use it - most people don't even know

It is actually a tool to help with gpgme development.  However, Ben
Kibbey seems to be using it for some of his software which is the reason
for the XML output stuff.

I have mentioned it only because we are evaluating ways to interact with
web browsers and gpgme-tool might be useful to prototype a Native
Messaging based browser extension.

> This is one of the complaints/wishes us Mailpile folks had, for
> some sort of stable socket/stdio-based programmatic API for
> talking to GnuPG. This sort of interface would make it much more

A socket based interface exists for years if you are going to use the
UI-server approach we are using in Kleopatra, GpgOL and GpgEX.  gpgme
even provides high level access functions.  If you want to try this you
need an UI server (Kleopatra, GPA, or whatever you want to write) and
then test it with

  gpg-connect-agent --uiserver

(try the "HELP" command)

A stdio based interfaces exists for more than 20 years.  For example gpg
can be used as a drop-in replacement for mutt's pgp support.  But be
aware that a stdio based interface has several problems which you can
only solve with several channels or at least descriptor passing.

> Requiring that a ruby, python or node.js dev know to install
> GnuPG from the C sources and build this tool is a non-starter,

They only need to install their language binding for GPGME.

> So pretty please, ship it! :-)

It is not useful right now.



Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.

More information about the Gnupg-users mailing list