Bastien,

 What's status the bug now? I'm interesting in this issue.

 Any ideas to do it?

You'll need to list the drives connected to the computer using g_volume_monitor_get_connected_drives() and for each one, check whether they can be ejected using g_volume_can_eject() (and are an optical drive, though you can leave that one out for now, as it would be pretty hard to do right now).

Then call g_volume_eject() on the drive, and dismiss the popup after calling g_volume_eject_finish().

To test, disable the media-keys plugin in GConf, and launch test-media-keys in the plugins/media-keys directory.

Created attachment 155649 [details] [review]
Replace "eject" spawn with GIO code

Review of attachment 155649 [details] [review]:

::: plugins/media-keys/gsd-media-keys-manager.c
@@ -587,3 @@
 
-        command = gconf_client_get_string (manager->priv->conf_client,
-                                           GCONF_MISC_DIR "/eject_command",

Is the key in the schema? If so, we should remove it.

@@ -592,3 @@
-                execute (manager, command, FALSE, FALSE);
-        } else {
-                execute (manager, EJECT_COMMAND, FALSE, FALSE);

EJECT_COMMAND should go, too.

(In reply to comment #4)
> Review of attachment 155649 [details] [review]:
> 
> ::: plugins/media-keys/gsd-media-keys-manager.c
> @@ -587,3 @@
> 
> -        command = gconf_client_get_string (manager->priv->conf_client,
> -                                           GCONF_MISC_DIR "/eject_command",
> 
> Is the key in the schema? If so, we should remove it.

It's not.

> @@ -592,3 @@
> -                execute (manager, command, FALSE, FALSE);
> -        } else {
> -                execute (manager, EJECT_COMMAND, FALSE, FALSE);
> 
> EJECT_COMMAND should go, too.

Good point, will do.

Created attachment 155716 [details] [review]
Replace "eject" spawn with GIO code

Attachment 155716 [details] pushed as 325f4f4 - Replace "eject" spawn with GIO code