Re: IOR reference of ORbit2 not recognized by mico

[Mark McLoughlin <mark skynet ie>]

Hey Duncan,

On Wed, 20 Mar 2002, Duncan Grisby wrote:

> On Wednesday 20 March, Mark McLoughlin wrote:
>
> > 	If mico doesn't recognise a component, it should ignore it.
> > There is enough info in the other components for mico to connect to
> > the object.
>
> The IOR is invalid. Try giving it to
>
>   http://www.parc.xerox.com/istl/projects/ILU/parseIOR/
>
> and you get
>
> ...
> (iiop.c:parse_MultiComponent_Profile):  profile contains 2 components
> (iiop.c:parse_MultiComponent_Profile):  component 1 of type 5, 24 bytes
> Attempt to read byte sequence of length 83951617, which exceeds
> IIOPMaxStringSize value of 4294967295 or call limit of 65535.
> (iiop.c:parse_MultiComponent_Profile): Invalid encapsulated profile
> detected:
>
> omniORB 4's catior agrees that it's invalid, but doesn't give such a
> helpful error message. (omniORB 3 and earlier don't try to look into
> multiple component profiles, so they don't notice that it's invalid.)
>
> The second component of the multiple component profile claims that it
> has a marshalled length of 83951617 octets, which it clearly doesn't,
> so MICO et al are quite right to complain that the IOR is invalid.

	Super ! That's the the kind of info which makes this easier to
debug ...

	Believe me, I was under no false sense of security that this
wasn't an ORBit2 bug, since it wasn't too long ago that I found some
similar ones getting it to work with Java ..

	Anyway, I'll try to get some time to fix this soon unless
somebody else finds the fix.

Cheers,
Mark.