Re: ORBit-0.5.17 not working with MICO 2.3.7?

[<dahaverk rockwellcollins com>]

Hi Michael,

>On Mon, 2002-07-22 at 13:22, dahaverk@rockwellcollins.com wrote:
>> Another thought I've had after reading through the spec is that "ORBit"
and
>> "MICO" may not be agreeing on the semantics of the "oneway"
operation....
>> Thus, ORBit is expecting an "ACK" or something back from MICO and that
>> isn't happening since MICO recognized that it was a "oneway".

 >ORBit shouldn't expect an ack, so that shouldn't be an issue. It's
rather hard to debug anything with ORBit1 though sadly.

 That's good to know.  I haven't quite figured out what MICO thinks it
wants on it's side.  Just a matter of slogging through the code until I hit
the problem.

 I'm not excluding the idea that I've disabled something important in my
hacked version of ORBit1 on my embedded target.

 I'm also hampered by the lack of complete support for the MICO command
line options.  E.g. Hooking up to a Name server via  "--ORBNamingAddr
inet:192.168.50.1:8888" for example.   I've been looking at the code but I
haven't quite figured out how I need to convert the inet:<ip_addr>:port#
into something ORBit can use.

 It also looks to me like the "alloca" in CORBA_init needs to have a "+ 1"
in the string calculation so the null character doesn't stomp on somthing
important. I haven't looked at the ORBit2 init function to see if it has
the same problem.  It seem every place else I've looked that is doing an
alloc for a string properly adds "+1" to the length to allow for the null
character.

>> Is there some kind of GIOP/IIOP debug option that could be turned on in
>> ORBit that would help us diagnose this problem?

 >Only in ORBit2 - which is best enabled for this by:

 >$ export ORBIT2_DEBUG=traces:errors

 I found several compile time "#define"s to enable.  That may help also.

 >Sorry, but ORBit-0.5.17 is really hard to support,

 Alas...  Catch 22... I wanted the C++ support, but ORBit2 was lacking
 there...  Also I wanted Ada support.  I haven't actually gotten to the
 point of trying the ORBit1 Ada support yet.  But, will have to eventually.

 It would help also if the C++ support was more compatible with MICO.  e.g.
 the namespace usage.


Cheers,

David Haverkamp