> On 24 Jun 2015, at 16:11, Ian Jackson <ian.jack...@eu.citrix.com> wrote:
>
> Rob Hoes writes ("[PATCH RFC 9/9] libxl: introduce specific error codes in
> libxl_device_nic_add"):
>> Signed-off-by: Rob Hoes <rob.h...@citrix.com>
> ...
>> + # NIC parameters could not be determined
>> + (ENUM_PREV, "NIC_SCRIPT_UNDETERMINED"),
>> + (ENUM_PREV, "NIC_DEVID_UNDETERMINED"),
>
> Perhaps we could have a coherent naming scheme ? These are invalid
> parameter errors, aren't they ? That is, libxl's caller specified
> something wrong.
>
> So maybe
> INVALID_NIC_SCRIPT_UNDETERMINED
> ?
>
Also see my other reply on PATCH 8/9. I think that the distinction between
UNDETERMINED and INVALID is useful.
I agree that a naming scheme would be good to have. I’ve tried to do that at
least for the codes related to libxl_device_<type> structures, where I’ve
included <device_type>_<field> in the error code, plus a condition (description
of the problem) such as INVAL or UNDETERMINED.
Rob
> What do others think ?
>
> Ian.
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
