Re: help needed with array out arguments



Hi Gergo,

On Tue, 2002-04-02 at 09:59, ERDI Gergo wrote:
> On 2 Apr 2002, Michael Meeks wrote:
> > 	With that table you can construct all the rest of it trivially.
> 
> I tried that, and came up with what you describe below. However, when that
> didn't work, I tried the dynamic one I've sent in my initial mail. That
> didn't work either.

	Extraordinary.

> So my actual client code is:
> 
>         OuterModule_Testmodule_MyArrAtomic_slice numbers[6];
> 
>         OuterModule_Testmodule_IFace1_method4 (_orbitcpp_get_c_object (),
> 					       numbers,
> 					       _ev._orbitcpp_get_c_object ());
> 
> and the actual server code is empty (it doesn't even touch the out
> argument):

	Are you sure you are using an array of things of fixed size ? it all
depends; the best place to look is ORBit2/test/everything/client.c
(testFixedLengthArray) and the corresponding (testVariableLengthArray) -
also the server impls. [ test/everything was originally snarfed from the
ORBit C++ regression suite ].

> #4  0x40045289 in ORBit_marshal_arg (buf=0x80612c8, val=0x4, tc=0x8055a20)
>     at corba-any.c:367

	That's clearly broken.

> #5  0x4003faf9 in ORBit_small_invoke_adaptor (adaptor_obj=0x8061088, 
>     recv_buffer=0x8066b40, m_data=0x80576a0, data=0xbffff300,
> ev=0xbffff350)
>     at orbit-small.c:865

	Possibly a bug in orbit-small, but why does it work for our regression
tests ?

	Regards,

		Michael.

-- 
 mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot




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