gnome-initial-setup is creating a keyring by calling
gnome-keyring-daemon --unlock. Since we don't have a password
at that time yet, it passes "" as the password. Just accept
this and create an unencrypted keyring.

https://bugzilla.gnome.org/show_bug.cgi?id=762095

Fábio Nogueira authored Feb 10, 2016 and Administrator committed Feb 10, 2016

(cherry picked from commit 1a068b97)

Use the correct enum value instead of int.

https://bugzilla.gnome.org/show_bug.cgi?id=753698

https://bugzilla.gnome.org/show_bug.cgi?id=759399

In looking for places where gnome-keyring-daemon potentially
spawns subprocesses, I found this piece of unused code. Remove it.

https://bugzilla.gnome.org/show_bug.cgi?id=756324

Let's keep the environment vars that the daemon can depend upon
to a minimum, so that we reduce the dependency to the initialization
order and we avoid inverse dependencies with gnome-shell/XWayland.
The daemon was already ignoring most of these.

https://bugzilla.gnome.org/show_bug.cgi?id=756324

Which, as the name implies, runs before the DisplayServer, ie,
before gnome-shell-wayland, and makes sure gnome-shell and
its descendants see the right environment variable.

https://bugzilla.gnome.org/show_bug.cgi?id=756324

https://bugzilla.gnome.org/show_bug.cgi?id=756865

https://bugzilla.gnome.org/show_bug.cgi?id=756812

https://bugs.freedesktop.org/show_bug.cgi?id=92383

After we are unexported, those signals will not be emitted anymore.

Previously objects were only explicitly exported on the bus when
they were ready. However now due to GDBus handler connections
they are exported earlier.

Make sure to export a prompt object before something is exported
at the same object path to take its place.

https://bugzilla.gnome.org/show_bug.cgi?id=756032

Right now gnome-keyring will keep processes around forever
in some cases.  They need to die when the session goes away,
at least.

Signed-off-by: Cosimo Cecchi <cosimoc@gnome.org>
 * Don't leak a DBus connection

Signed-off-by: Stef Walter <stefw@gnome.org>
 * Print error message if we cannot connect to bus
 * Wait for gnome-keyring-daemon when --foreground

https://bugzilla.gnome.org/show_bug.cgi?id=756059

https://bugzilla.gnome.org/show_bug.cgi?id=756059

It's not really a good idea to fork after glib has initialized,
since it has helper threads that may have taken locks etc.

This commit forks really early to prevent locks from leaking
and causing deadlock.

Signed-off-by: Cosimo Cecchi <cosimoc@gnome.org>
 * Split out separate function
 * Check return value of g_unix_open_pipe()
 * Don't fork when running foreground
 * Read login password before fork()

Signed-off-by: Stef Walter <stefw@gnome.org>
 * Since stdout is open, just print evironment directly

https://bugzilla.gnome.org/show_bug.cgi?id=756059

And use the more complex function call instead.

Fixes this warning:

/usr/include/selinux/flask.h:5:2: error: #warning "Please remove any #include's of this header in your source code." [-Werror=cpp]
 #warning "Please remove any #include's of this header in your source code."
  ^

Jan Alexander Steffens authored Oct 06, 2015 and Cosimo Cecchi committed Oct 07, 2015

Otherwise incoming calls can race with our initialization. Activating
calls are even guaranteed to arrive before we have set up the service.

https://bugzilla.gnome.org/show_bug.cgi?id=756006