[omniORB] Abort in omniORB2.8.0

[Duncan Grisby]

On Tuesday 15 January, "Hinkofer, Fa. pds, E230, D" wrote:

> We are using the recommended snapshot of omniORB2.8.0. But now we
> have a problem with our application and omniORB2.8.0. Our
> application requests and receives tiff-images via
> omniORB2.8.0. After processing a tiff-image our application
> immediatly sends back the results of processing via omniORB2.8.0. In
> a range of 100.000 to 600.000 processed images we get an abort in
> omniORB2.8.0. Look to the following stack-dump (created by dbx):

[...]
>   [9] __Crun::ex_throw(0x0, 0x0, 0xfef6a880, 0xfef6a054, 0x0,
> 0xfef6a8c8),
> at 0xfef54eac
>   [10] CORBA::Object::_duplicate(obj3D ???) (optimized), at
> 0xff2af304 (line
> ~127) in "corbaObject.cc"

The abort is due to an uncaught exception. The only exception thrown
by CORBA::Object::_duplicate() is BAD_PARAM. It does that when it
thinks the object reference being duplicated is invalid -- either
because it has been deallocated, or because it has been trampled on by
other code.

Most likely there is a bug in your code, either releasing an object
reference too many times, or trashing the heap. If possible, try
running your application with Purify or something similar. It's not
impossible that there's still a problem with the compiler, or with
omniORB itself, but I think that's less likely.

Cheers,

Duncan.

-- 
 -- Duncan Grisby  \  Research Engineer  --
  -- AT&T Laboratories Cambridge          --
   -- http://www.uk.research.att.com/~dpg1 --