I just upgraded my desktop to Rawhide. Having problems starting the packaged Firefox, which is now GTK+ 3 on Rawhide. Here's the data I have so far: 1) The GTK+ 3 package does not work with my existing user account. 2) Moving ~/.mozilla and ~/.cache/mozilla does not help. 3) Starting with firefox -safe-mode does not help. (neither does 2) and 3) combined). 4) It doesn't work in GNOME or Xfce (so, not a desktop issue). 5) It *does* work with a new user account. 6) Upstream's static build - firefox-34.0.5.tar.bz2 - *does* work. 7) A rebuild of the Fedora package against GTK+ 2 instead of GTK+ 3 - all I did was change both %define toolkit_gtk3 lines to 0, bump the rev, and rebuild - *does* work. 8) Kevin Fenzi says the GTK+ 3 build works OK for him. When I try to run it from the console, I see: ======== (process:19332): GLib-CRITICAL **: g_slice_set_config: assertion 'sys_page_size == 0' failed (process:19332): GLib-CRITICAL **: g_slice_set_config: assertion 'sys_page_size == 0' failed (firefox:19332): GLib-GObject-WARNING **: The property GtkSettings:gtk-menu-images is deprecated and shouldn't be used anymore. It will be removed in a future version. (firefox:19332): GLib-GObject-WARNING **: The property GtkSettings:gtk-button-images is deprecated and shouldn't be used anymore. It will be removed in a future version. (firefox:19332): GLib-GObject-WARNING **: cannot register existing type 'GtkWidget' (firefox:19332): GLib-GObject-CRITICAL **: g_type_add_interface_static: assertion 'G_TYPE_IS_INSTANTIATABLE (instance_type)' failed (firefox:19332): GLib-GObject-WARNING **: cannot register existing type 'GtkBuildable' (firefox:19332): GLib-GObject-CRITICAL **: g_type_interface_add_prerequisite: assertion 'G_TYPE_IS_INTERFACE (interface_type)' failed (firefox:19332): GLib-CRITICAL **: g_once_init_leave: assertion 'result != 0' failed (firefox:19332): GLib-GObject-CRITICAL **: g_type_add_interface_static: assertion 'G_TYPE_IS_INSTANTIATABLE (instance_type)' failed (firefox:19332): GLib-GObject-CRITICAL **: g_type_register_static: assertion 'parent_type > 0' failed ======= at which point it sticks. If I strace it, the main process is sitting at a FUTEX_WAIT. If I strace the threads, one of them seems to keep repeating this: futex(0x7f9dbafc774c, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 223, {1418274200, 855170000}, ffffffff) = -1 ETIMEDOUT (Connection timed out) futex(0x7f9dbafe63a0, FUTEX_WAKE_PRIVATE, 1) = 0 futex(0x7f9dbafc774c, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 225, {1418274201, 855365000}, ffffffff) = -1 ETIMEDOUT (Connection timed out) futex(0x7f9dbafe63a0, FUTEX_WAKE_PRIVATE, 1) = 0 futex(0x7f9dbafc774c, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 227, {1418274202, 855562000}, ffffffff) = -1 ETIMEDOUT (Connection timed out) futex(0x7f9dbafe63a0, FUTEX_WAKE_PRIVATE, 1) = 0 but I couldn't get much further than that. Kevin's console output in his 'working' case does not contain all the same glib errors as mine - only the sys_page_size failures and the GtkSettings deprecation warnings, I believe.
Maybe bug 1161246 ?
Neither the OSK nor the screen reader are enabled, but I do think some of the searches I did on the glib errors indicated they were somehow to do with a11y stuff. I just couldn't figure out any way to proceed from there - there doesn't seem to be some handy switch to flip to turn ATK or whatever 'on' or 'off', or even an obvious state indicator anywhere... I *do* have a Japanese input method configured (but not usually selected). I wonder if that could have anything to do with it/
Can you please investigate output of ld.so? If there's a library mismatch it should be there. LD_DEBUG=all firefox may do the trick.
Please try to run $GNOME_ACCESSIBILITY=0 firefox
yeah, I saw the other bug and had it on my list to test yesterday but wound up distracted by an anaconda bug...I'll get to it today.
Yep, GNOME_ACCESSIBILITY=0 firefox works.
*** This bug has been marked as a duplicate of bug 1161246 ***