Am 25.10.14 22:35, schrieb Steffan Karger:
> Since key_state_ssl_init() returns void, it *must* succeed. By changing the
> 'if(a) { success } else ignore' into 'ASSERT(a); success', we make sure we
> don't fail (which we in practice don't do, because otherwise we would have
> had complaining users with hard to trace errors).
>
> This patch just changes the if() to an ASSERT(), but that causes a lot of
> indenting changes, making this patch look a lot bigger than it really is.
>
>
ACK. Bailing out on these kind of crypto errors seems better than to
continue in an unknown state.Arne
