nbt bug2?

[Andrew Tridgell]

Metze,

 > I think the goto failed; is a bug, if talloc_zero() fails we should
 > not free a NULL pointer there are other places with bugs like this
 > in your latest patches...

no, that was deliberate. It makes the code more readable to have a
single failure path, and talloc_free(NULL) is a noop. Otherwise we
need to have two separate failure paths.

This is way talloc_free(NULL) is allowed (abartlet initially suggested
this a while back I think).

Cheers, Tridge