RE: "any within struct" ORBit decoding problem

[Mark McLoughlin <mark skynet ie>]

Hey Huw,

On Tue, 1 Jan 2002, Huw Rogers wrote:

> The alignment fix that was posted only fixed the
> receive side. I've been testing with sending
> and receiving sequence<any> types between ORBit
> and OmniORB on Solaris. The enclosed patch fixes
> both send and receive (patch is against
> released 0.5.12).
>
> The first couple of chunks of the patch
> fix unrelated compilation problems with ORBit on
> non-gcc compilers (I'm testing with Sun's compilers
> on Solaris) (invalid pointer arithmetic on void*
> types, and using dynamic initializers where static
> are required).
>
> The last two chunks of the patch fix the
> specific interoperability issue with
> typecode marshalling/demarshalling alignment.

	Great catch! And the fix is sound ... I've committed to the
orbit-stable-0-5 branch.

Cheers,
Mark.