On 05/09/2014 01:34 AM, Christopher Gregory wrote:
> Hello,
>
> This is driving me absolutely insane. I am running systemd and after many
> re-installations have got gdm and gnome working correctly to some degree.
> Thanks to following Armin's notes for tweaking BLFS.
>
> I am now getting errors when trying to start the mail program evolution,
> the bloody thing seg faults and gnome-terminal just plain refuses to
> start.
>
> gnome terminal gives this error:
>
> dbus-launch gnome-terminal
> Error creating terminal: Message did not receive a reply (timeout by
> message bus)
>
> I have installed gnome in the /usr tree completly. This was the only way
> that I could get gdm and everything working correctly. I have a FULL
> installation of gnome. I also am unable to use gedit.
>
> I have the following versions installed:
>
> dbus 1.8.2 (upgraded to that incase there was an issue with 1.8.0)
> gnome-terminal 3.12.1 (downgraded to that as 3.12.2 also did not work)
> systemd 208 (svn version of the systemd installation of lfs was a
> failure for me as it would not boot correctly.
>
>
> After I completed the LFS part of installing the base systemd system, I
> only got xorg up and running via the BLFS book. I then used jhbuild to
> pull all 281 or so packages of gnome down and then I went through and used
> their autogen arguments to install gnome into /usr only modifications
> that were made to those were to make sure that prefix=/usr was appended to
> it.
>
> I have installed all of the dependencies for gnome so this is not the issue.
>
> I am unable to get a meaningful result to post here with the output of gdb
> as I just do not know how to use it properly. All I seem to manage to get
> using the following suggested test:
>
> gdb --args /usr/libexec/gnome-terminal-server --app-id my.foo.Terminal
> (gdb) run
>
> gnome-terminal --app-id my.foo.Terminal
>
>
> New Thread 0xb634db40 (LWP 4336)]
> [New Thread 0xb58f4b40 (LWP 4337)]
> [New Thread 0xb4effb40 (LWP 4338)]
> [Thread 0xb58f4b40 (LWP 4337) exited]
> [Thread 0xb634db40 (LWP 4336) exited]
> [Thread 0xb6a81840 (LWP 4332) exited]
> [Inferior 1 (process 4332) exited normally]
>
>
> After I upgraded dbus and gtk+ I went through and recompiled EVERYTHING in
> gnome. Ending up with the same result.
>
> I want to solve this issue without having to re-format and try again.
>
> I want to use evolution for email as I have done so for a number of years,
> and do NOT want to change. I know that it is an option, but it does not
> solve these issues.
>
> I have done many hours of research to get this blasted thing up and
> running, and the conclusions I am starting to reach is that they are
> deliberatly making it work only on redhat/fedora.
>
> I do have gnome terminal compiled with debug on, but as bloody gnome wiki
> will not tell you how to use it I do not know how to. Also the times that
> I have been into the so called chat room on irc for help with gnome no one
> has said a bloody thing so all they appear to be is either bots or very
> unhelpful people.
>
> The entry from journalctl for gedit is:
>
> gedit.desktop[4241]: (gedit:4241): Gtk-CRITICAL **:
> gtk_menu_button_set_menu_model: assertion 'GTK_IS_MEN
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241): GLib-GIO-CRITICAL
> **: g_settings_bind_with_mapping: assertion 'G_IS_OB
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241): Gtk-CRITICAL **:
> gtk_statusbar_get_context_id: assertion 'GTK_IS_STATU
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241): Gtk-CRITICAL **:
> gtk_statusbar_get_context_id: assertion 'GTK_IS_STATU
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241): Gtk-CRITICAL **:
> gtk_statusbar_get_context_id: assertion 'GTK_IS_STATU
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241): Gtk-CRITICAL **:
> gtk_menu_button_set_menu_model: assertion 'GTK_IS_MEN
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241): Gtk-CRITICAL **:
> gtk_menu_button_set_menu_model: assertion 'GTK_IS_MEN
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241):
> GLib-GObject-CRITICAL **: g_signal_connect_data: assertion 'G_TYPE_CHE
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241):
> GLib-GObject-WARNING **: invalid (NULL) pointer instance
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241): GLib-GIO-CRITICAL
> **: g_settings_bind_with_mapping: assertion 'G_IS_OB
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241):
> GLib-GObject-CRITICAL **: g_object_bind_property_full: assertion 'G_IS
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241):
> GLib-GObject-CRITICAL **: g_object_bind_property_full: assertion 'G_IS
> May 09 10:58:30 lfs gedit.desktop[4241]: (gedit:4241): Gtk-CRITICAL **:
> Error building template class 'GeditWindow' for an in
>
> The info for evolution is:
>
>
> May 09 10:36:52 lfs systemd-coredump[3922]: Process 3906 (evolution)
> dumped core.
> May 09 10:36:52 lfs gnome-session[3559]: Window manager warning: Log level
> 16: AT-SPI: Error in GetItems, sender=org.freedesk
> May 09 10:36:52 lfs kernel: evolution[3906]: segfault at 0 ip b3f28880 sp
> bfc00620 error 6 in libgtk-3.so.0.1301.0[b3e1d000+5
> May 09 10:36:52 lfs gnome-session[3559]: Window manager warning: Log level
> 16: AT-SPI: Error in GetItems, sender=org.freedesk
>
>
> Thanks in advance for any assistance.
>
> Regards,
>
> Christopher
>
What you are actually facing is an issue in at-spi2-core daemon, which
launches its own dbus-daemon for handing accessibility stuff. I saw
Ken's reply earlier and he did sugest rebuilding entire DBus when Xorg
Libraries are present. However, I did a quick examination of DBus code
and noticed the following
$ grep -r DBUS_BUILD_X11
tools/dbus-launch.h:#ifdef DBUS_BUILD_X11
tools/dbus-launch-x11.c:#ifdef DBUS_BUILD_X11
tools/dbus-launch.c:#ifdef DBUS_BUILD_X11
tools/dbus-launch.c:#ifdef DBUS_BUILD_X11
tools/dbus-launch.c:#endif /* DBUS_BUILD_X11 */
tools/dbus-launch.c:#ifdef DBUS_BUILD_X11
tools/dbus-launch.c:#ifdef DBUS_BUILD_X11
tools/dbus-launch.c:#ifdef DBUS_BUILD_X11
tools/dbus-launch.c:#ifndef DBUS_BUILD_X11
tools/dbus-launch.c:#else /* DBUS_BUILD_X11 */
tools/dbus-launch.c:#endif /* DBUS_BUILD_X11 */
tools/dbus-launch.c:#ifdef DBUS_BUILD_X11
config.h.in:#undef DBUS_BUILD_X11
configure.ac: AC_DEFINE([DBUS_BUILD_X11], [1], [Define to build X11
functionality])
configure:$as_echo "#define DBUS_BUILD_X11 1" >>confdefs.h
cmake/tools/CMakeLists.txt:if (DBUS_BUILD_X11)
cmake/tools/CMakeLists.txt:endif(DBUS_BUILD_X11)
cmake/tools/CMakeLists.txt:if (DBUS_BUILD_X11)
cmake/tools/CMakeLists.txt:endif (DBUS_BUILD_X11)
cmake/CMakeLists.txt: option (DBUS_BUILD_X11 "Build with X11 autolaunch
support " ON)
cmake/config.h.cmake:#cmakedefine DBUS_BUILD_X11 1
cmake/config.h.cmake:#ifdef DBUS_BUILD_X11
DBUS_BUILD_X11 is only defined when libX11, libSM and libICE are found
and as you can see, it's only specific to dbus-launch.c which is used to
build dbus-launch utility so there is no need to rebuild everything. But
you need to confirm that dbus-launch is actually linked to X11 libraries
by ie using ldd:
$ ldd $(which dbus-launch)
linux-vdso.so.1 (0x00007fffa63ff000)
libSM.so.6 => /usr/lib/libSM.so.6 (0x00007f63cbbec000)
libICE.so.6 => /usr/lib/libICE.so.6 (0x00007f63cb9d0000)
libX11.so.6 => /usr/lib/libX11.so.6 (0x00007f63cb683000)
libc.so.6 => /lib/libc.so.6 (0x00007f63cb2d6000)
You should be able to start dbus session daemon just fine then. You can
use "dbus-launch evolution" for instance to check if it works.
If that works, it means your "insert the program that starts your
session here" isn't starting the session as it should.
On my system, I have several dbus-daemon and dbus-launch processes running:
message+ 383 0.0 0.0 26104 2428 ? Ss May08 0:06
/usr/bin/dbus-daemon --system --address=systemd: --nofork --nopidfile
--systemd-activation
The first one is system-wide daemon, started on boot by systemd and
without it, no other daemon below would function.
lightdm 476 0.0 0.0 24380 776 ? S May08 0:00
/usr/bin/dbus-launch --autolaunch e726e81fd11b4b7caf861c0ee3824910
--binary-syntax --close-stderr
lightdm 539 0.0 0.0 24984 896 ? Ss May08 0:00
/usr/bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session
These two represent the actual command used to spawn dbus-daemon session
daemon for lightdm session, which is display manager I use.
armin 655 0.0 0.0 24380 776 ? S May08 0:00
dbus-launch --sh-syntax --exit-with-session
armin 656 0.0 0.0 26308 2520 ? Ss May08 0:02
/usr/bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session
These two are specific to my session
armin 998 0.0 0.0 24984 1400 ? S May08 0:00
/usr/bin/dbus-daemon --config-file=/etc/at-spi2/accessibility.conf
--nofork --print-address 3
And this one is specific for at-spi2-registryd daemon, which is launched
when any GTK+3 app is started. There is no dbus-launch command for this
one which means it isn't started that way. What you are facing is that
this one is not present, but it should. "Why" is really beyond me. Maybe
you can "strace" it by launching a GTK+3 app.
--
Note: My last name is not Krejzi.
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
