[Glade-users] xlib error corba/glademm app

[dwalter syr edu (David Walter)]

I have started learning to use glade for a user
interface for a corba application.

The program seems to run fine. Connections are
established and the call back to the client and
server are okay. Depending one or many sequential
calls can be made successfully on between the
client gui and server application.


Then after  a some (indeterminate) amount  of time
the application gives a message like:

      Xlib: unexpected async reply (sequence 0x1dd)!

The sequence number given is inconsistent, that
is, is is different every time.

On a multiprocessor system this happens sooner
than on a single processor system -- if that
suggest anything.


I am using omniORB3 which uses pthreads and
generates several threads of control for the
communication side of the program.

Linux eaglepeak 2.2.14 #22 Tue Oct 31 08:30:01 EST 2000 i686 unknown

glademm V0.5_10 (glade to Gtk-- converter)

Gnome Glade 0.5.11

This is using the c++ generated source from
glade--.

This is basically a two way conversation with data
send initiated in a button click call back, and
update of the screen is asynchronously handled by
the corba client thread.

Questions:
        Is this a thread problem, and if so can
the thread be fixed in the call back function?

I read in the archive that you can't use

  gdk_threads_enter / gdk_threads_leave inside a callback.

Does this include any callback initiated by the
interface.

If so what about the threads under the omni code? 

One   thread  in the  omni   client updates a text
widget, there I tried guarding  the update of  the
text          widget    by gdk_threads_enter     /
gdk_threads_leave pair, but  this seems to have no
effect.

        Is gdk_threads_enter / gdk_threads_leave
== to the macro's GDK_THREADS_ENTER /
GDK_THREADS_LEAVE?


Thanks so much for any help.

David

please cc me as I am not subscribed.

Thanks again.