On 02/23/2016 06:24 AM, Alberto Garcia wrote:
> On Tue 23 Feb 2016 02:17:23 PM CET, Eric Blake wrote:
> 
>> Commit message should say why we need a third event, rather than
>> reusing either of the other two (my guess: because you don't have a
>> location, and don't want to modify the existing two to report a
>> location - but why not just use 'sector-num':0, 'sectors-count':<size
>> of file> to report the entire file as the location?)
> 
> I would also be fine with that solution.

I would also be fine if we added an optional enum member to the existing
event that said which operation failed ('read', 'write', 'flush') -
adding optional output members is safe, while converting existing
mandatory output members to optional may confuse existing clients.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to