[omniORB] Orbix 3.0 v.s omniORB2.8

[Robert E Gruber]

Struggled with this last night.   Orbix 3.0 apparently expects the key to be
null terminated, and adds a null octet if it is not, causing key length to
change from expected length to expected length + 1.

I assume this is incorrect behavior -- or are keys supposed to always end in
null octet?

Anyone adding the hack Sai-Lai suggested might want to make sure the last
octet is null before deciding to discard it.

-- Bob