Re: ORB problem, was Re: Problem moving applets between panels

[Raja R Harinath <harinath cs umn edu>]

Manish Vachharajani <mvachhar@vger.rutgers.edu> writes:
> Doh, maybe it is an ORB problem, here is the backtrace.  The printed error
> message follows.
> 
> #0 0x4036a6a1 in __kill ()  
> 
> #1 0x4036a4cf in raise (sig=6) at ../sysdeps/posix/raise.c:27
> 
> #2 0x4036b6df in abort () at ../sysdeps/generic/abort.c:83
> 
> #3 0x4004bbc6 in g_error (format=0x40040b6e "file %s: line %d (%s): 
> \"%s\"")  at gmessages.c:64
> 
> #4 0x4003ff28 in giop_recv_reply_decode_message (buf=0x80fe968)  at
> giop-msg-buffer.c:673
[snip] 
> The output error message:
> 
> ** ERROR **: file giop-msg-buffer.c: line 673
> (giop_recv_reply_decode_message): "GIOP_RECV_BUFFER(buf)->cur <
> (GIOP_RECV_BUFFER(buf)->message_body +
> GIOP_MESSAGE_BUFFER(buf)->message_header.message_size)" 

This is only half the story.  This message is spewed by the `panel'.
What error did the applet print (unfortunately, this is also prefixed
just by ** ERROR **, and since it is fork/execed, gdb cannot debug the
applet). 

Anyway, I've seen this happen when the ORBit client abruptly disconnects
from the ORBit server (the applet being the client, and panel being the
server in this case).  The real problem is probably in the applet part
of the code.  For me, the applet usually crashes saying something like
Assertion "sum == res" failed.  A bit of sleuthing shows that a writev
call EINVALs.

- Hari
-- 
Raja R Harinath ------------------------------ harinath@cs.umn.edu
"When all else fails, read the instructions."      -- Cahn's Axiom
"Our policy is, when in doubt, do the right thing."   -- Roy L Ash