Comment 33 for bug 867424

Created attachment 8605868
Stop using libgnome and libgnomeui on Linux v2

I've updated the patch above, and have had some success with it, but it still needs some work in order to disconnect from the SM at the right time. It's been compiled and tested a little on top of 37.0.1, it just needed a couple of trivial conflicts resolved to rebase it to master, but hasn't been tested there.

Changes from v1
---------------

* Rebase to 37.0.1 and resolve conflicts
  - Drop MOZ_WIDGET_GTK == 2 checks
  - Unbitrot to match "Bug 777292 part 2 - Change all nsnull to
nullptr" (Note that this was actually Bug 626472)
  - Unbitrot to match "Bug 773151: Convert nsCAutoString->nsAutoCString
CLOSED TREE r=bsmedberg"
  - Unbitrot to match "Bug 784739 - Switch from NULL to nullptr in
toolkit/;r=ehsan"
  - Avoid "Wwrite-strings" warning for static string conversion
* Further changes to the functionality of the patch
  - Don't DisconnectFromSM() until destructor.
  - Add comments explaining the rationale for disconnect timing wrt
shutdown phases
  - Expand some comments with more detail on the standard or
implementation issues
  - Don't IceSetErrorHandler, and add comment about the motivation for
IceSetIOErrorHandler
  - Rename client_id to prev_client_id; Rename new_client_id to
client_id
  - Set SM_CLIENT_ID on client leader window as per ICCM specified

Questions
---------

* Is the #define gtk_function gtk_function_, #undef gtk_function dance
in the gtk compat defines the right approach?
* are gio watch callbacks always guaranteed to be on the main thread?
* what should I use to log instead of printf?

Remaining issues
----------------
* Doesn't solve bug #557601 - DisconnectFromSM is called at the right
time when quitting firefox, but is too early on shutdown. I don't
understand why yet.
* The interaction call/response with the session manager needs some work
* Doesn't set a couple of SM properties that gnome-client did
* Doesn't use grabs to prevent interaction at the wrong time like
gnome-client did