> I suppose we could use memcmp on the as variable itself, to inspect
> the actual stored padding rather than the returned copy of it.

Yes, that's probably the only safe stance when optimization is enabled.

-- 
Eric Botcazou


Reply via email to