[Libvirt-cim] [PATCH] [RFC] Changes required to work with new std_indication. The trigger changes were very light, which was nice

Jay Gagnon grendel at linux.vnet.ibm.com
Fri Feb 8 15:35:03 UTC 2008


Dan Smith wrote:
> I agree with Heidi.  I think you've opened the door to take it a
> little further so that the calls in Enable/Disable and
> Activate/Deactivate are automated for the simple case.  Maybe this is
> already planned?
>
>   
Yup, those are phase two.
> JG> +static struct std_ind_state migrate = {
> JG> +        .ind_name = "Xen_ComputerSystemMigrationIndication",
> JG> +        .filter_active = false,
> JG> +        .enabled = false,
> JG> +};
>
> This is a really minor thing, but this definition has some fields that
> must always be set to false on initialization to avoid badness.  Now,
> since these are static, they are all initialized to zero, which should
> cover you in the case that someone just doesn't set filter_active and
> enabled.  However, perhaps setting the precedent with something like a
> constructor macro would be appropriate?
>
>   #define DECLARE_IND(ident, name)              \ 
>           static struct std_ind_state ident = { \
>                  .ind_name = name,              \
>                  .filter_active = false,        \
>                  .enabled = false,              \
>           };
>
> Which would make the above definition look like this:
>
>   DECLARE_IND(migrate, "Xen_ComputerSystemMigrationIndication");
>
> Thoughts?
>
>   
Ooh, I like that.  I'll definitely use it.


-- 

-Jay




More information about the Libvirt-cim mailing list