orbit memleak

[bob thestuff net]

I have done some more testing on that memleak problem I sent in last night
and have found out why it is leaking. In the generated code, it generates:
environment__free(&env, NULL, CORBA_FALSE);

it really needs to be
environment__free(&env, NULL, CORBA_TRUE);

That is why it is leaking the strings from the hash.

Am I going to have to modify all the generated code? Or would it be better
if I freed the strings myself? Or better yet, could that last option be
put into the servant struct to allow the user defined code to specify
wether or not the strings will be freed?

Thanks,
Bob