On Tue, 2017-10-10 at 16:10 +0100, Daniel P. Berrange wrote:
> Wouldn't it be much safer to leave the newly-added NULL
checks
> in place so that the qemuDomainMachineIs*() functions are locally
> correct instead of relying on external guarantees in order not
> to crash on the user?
Are these qemuDomainMachineIs* functions the only places that makes
assumptions about machine being non-NULL though ?
I guess we don't know.
And that's why assumptions should be verified locally :)
I wouldn't want to
encourage the idea that it is OK to have a NULL machine value, because
that would just push crashes to elsewhere in the code if people rely
on it.
Okay, I'll drop those hunks from the respin then.
--
Andrea Bolognani / Red Hat / Virtualization