Re: [g-a-devel]regression test fails

["benny k." <ben bagu org>]

Michael said:

> 	Can you verify that 'make check' passes in linc, ORBit2 and libbonobo
> as well ? might catch something.

make check passed in all three.


> > i just updated gail and at-spi from cvs. i still get the same regression test
> > failure when i run it by itself.  test-simple just sits at the following
> > point if registryd is running in another terminal. no output is produced 
> > by registryd.
> 
> 	Fascinating; ok it's blocking at some point. Tell me do you use the
> snapshots from Ximian ? if so then can you type:
> 
> 	export ORBIT2_DEBUG=traces:timings
> 
> 	then bonobo-slay and run in separate terminals, and post the output to
> here ? otherwise you might need to re-build ORBit2 with --enable-debug
> to get anything useful.


i'm using the vbs, not the ximian snapshots. so i guess i'll have to rebuild
ORBit2. i think i'll do a fresh check out and build of everything and
see where this gets me.

> 
> > ben booya:/gnome/head/cvs/at-spi$ test/test-simple 
> > GTK Accessibilty Module initialized
> > 
> > ...
> > 
> > Testing keyboard listeners ...
> > Application registered & listening
> > 
> > 
> > any ideas on this one?
> 
> 	Sadly not - dead hard to debug like this; possibly a stack trace from
> both ends might be helpful from gdb - just press C-z and type 'bt' at
> each end, and post that.


from registryd:

(gdb) bt
#0  0x407df2b0 in poll () from /lib/libc.so.6
#1  0x406c8bdf in g_main_context_poll (context=0x8062860, timeout=-1,
	priority=2147483647, fds=0x80643b8, n_fds=6) at gmain.c:2547
#2  0x406c819a in g_main_context_iterate (context=0x8062860, block=1,
	dispatch=1, self=0x8062928) at gmain.c:2224
#3  0x406c889f in g_main_loop_run (loop=0x8063080) at gmain.c:2449
#4  0x400c9d7f in bonobo_main () at bonobo-main.c:290
#5  0x0804c1f1 in main (argc=1, argv=0xbffffc34) at registry-main.c:63
#6  0x4072f65f in __libc_start_main () from /lib/libc.so.6
(gdb) 

did you want a bt from test-simple as well? test-simple appears to be a
wrapper for stuff in the .libs directory


all of these results have come from a debian testing box. i have another
machine running redhat 7.1. the redhat box was updated aprox 1.5 weeks ago.

anyway, the regression tests passes on this machine. so i guess that there is
something wrong with my debian setup.

thanks for you help.

ben