Re: Whose end is this broken on. . .




On Sat, 5 Dec 1998, Michael Bruce wrote:
[snip]
> Oh, as another data point that I forgot to mention in my original
> message: when I try to add an applet, the panel hangs completely, and
> swallows all keyboard input (except for the X-server keys), so I have
> to alt-tab to a VT and kill it.

I've got two machines with similar GNOME installations on it.  One works
fine the other has had the exact panel/appplet problems Michael Bruce is
describing for over a week now.  The difference is my machine with the
panel problems doesn't even have esound built (no sound card), so the
problem certainly isn't with esound.  The problems hit me at the same time
as gettext became a requirement, but I have no evidence that that was the
problem.

I keep doing complete wipes, fresh anoncvs checkouts/recompile/reinstalls
of (in order) glib, gtk+, ORBit, gnome-libs and gnome-core, in the hopes
that something got fixed somewhere.  If there is something more I should
be updating, please let me know.  Here is some more information on the
problem, note that the applet I am using in all the tests below is the
fish_applet.

* If I run the applet from panel, panel freezes as he described.  If I
  kill the appplet's process panel starts working again.

* If I am running panel, and run the applet from the command line, the
  panel does not freeze, and the applet just hangs (no error messages,
  no segfault, nothing).  Running ps shows the applet as sleeping with
  no CPU time used.  The applet responds to control-C normally.  This
  happens both if I run the applet with no arguments and also if I run it
  with the same "--activate-goad-server fish_applet" argument that
  panel uses.

* If I run gdb on fish_applet, and interrupt it after a while, I get
  the following gdb (v4.17) sessions:

  (gdb) set args --activate-goad-server fish_applet
  (gdb) run
  Starting program: /opt/gnome/bin/fish_applet --activate-goad-server
  fish_applet

  Program received signal SIGINT, Interrupt.
  0x404d9b3e in __select ()
  (gdb) bt
  #0  0x404d9b3e in __select ()
  #1  0x4009fed4 in __DTOR_END__ ()
  #2  0x4009aff9 in giop_main_next_message ()
  #3  0x4009c20c in giop_recv_reply_buffer_use_multiple ()
  #4  0x4009d260 in giop_recv_reply_buffer_use ()
  #5  0x40026fdc in CosNaming_NamingContext_resolve ()
  #6  0x4001d8e2 in gnome_name_service_get ()
  #7  0x4001e859 in goad_server_activate ()
  #8  0x4001e68f in goad_server_activate_with_repo_id ()
  #9  0x4000fa01 in gnome_panel_applet_init_corba ()
  #10 0x4001243f in applet_widget_init ()
  #11 0x804af8f in main ()

  ----------------------------------------------------
  (gdb) run
  Starting program: /opt/gnome/bin/fish_applet 

  Program received signal SIGINT, Interrupt.
  0x404d4f74 in __read ()
  (gdb) bt
  #0  0x404d4f74 in __read ()
  #1  0x40502e7c in __DTOR_END__ ()
  #2  0x404a740c in _IO_file_underflow (fp=0x8084c28) at fileops.c:303
  #3  0x404a7fe8 in __underflow (fp=0x8084c28) at genops.c:249
  #4  0x404a55d3 in _IO_getline (fp=0x8084c28, buf=0xbfffe8fc "", n=2047, 
      delim=10, extract_delim=1) at iogetline.c:47
  #5  0x404a4dc1 in _IO_fgets (buf=0xbfffe8fc "", n=2048, fp=0x8084c28)
      at iofgets.c:47
  #6  0x4001f028 in goad_server_activate_exe ()
  #7  0x4001ea87 in goad_server_activate ()
  #8  0x4001e68f in goad_server_activate_with_repo_id ()
  #9  0x4000fa01 in gnome_panel_applet_init_corba ()
  #10 0x4001243f in applet_widget_init ()
  #11 0x804af8f in main ()

  Both of these traces are consistant, that is if I set the arguments
  and wait, I always get the first one.  If I don't, I always get the
  second one.  I have no idea on how to do a trace on the situation
  where panel runs the applet, if someone can tell me, I'll gladly
  send in that trace.

Hopefully this might make it easier to isolate the problem.  Thank you
very much.

-Gleef



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]