On 10/22/2012 09:25 PM, Adam Litke wrote:
On Mon, Oct 22, 2012 at 08:56:16PM +0200, Itamar Heim wrote:
> On 10/22/2012 03:04 PM, Adam Litke wrote:
>> On Sun, Oct 21, 2012 at 09:09:52AM +0200, Itamar Heim wrote:
>>> On 07/15/2012 03:12 AM, Adam Litke wrote:
>>>> For the past few weeks I have been working on creating a schema that
fully
>>>> describes the vdsm API. I am mostly finished with that effort and I
wanted to
>>>> share the results with the team. Attached are two files: the raw schema
and an
>>>> html document with cross-linked type information.
>>>>
>>>> This should already be useful in its current form, but I have bigger
plans. I
>>>> would first like to get help to correct errors in the schema. Then, I
will
>>>> start the process of writing a code generator that will create C/gObject
code
>>>> that we can compile into a libvdsm with language bindings for python,
java, etc.
>>>>
>>>> Please take a look at the attached files and let me know what you think?
>>>>
>>>> P.S. I tried to attach these to the oVirt Wiki, but they are not
permitted file
>>>> types.
>>>>
>>> Hi Adam,
>>>
>>> that's quite a big scheme to review.
>>> have you thought about ways to solicit inputs for it?
>>> (maybe schedule per topic reviews of the new scheme/api for VM
>>> operations (virt), network (host level, vm level), storage, sla
>>> policy, etc.)?
>>
>> For the first pass, we are trying to replicate the current API as much as
>> possible. For subsequent refactoring, I'd expect the discussions to occur
in
>> the community around the patches that are implementing the proposed changes.
>>
>
> I think it would be worth highlighting these patches on mailing
> list, or even do a scheduled discussion/brainstorm around various
> parts for where it should go.
I have been trying to get some attention on it for some time. I agree it's a
good idea. I'll be presenting it at the Barcelona workshop as well. Any other
ideas on how to have a constructive and inclusive mailing list discussion would
be welcomed :) In the past, it was hard to keep the thread focused and
driving toward consensus.
I'd try to focus on one area at time, do an email thread as a prep for a
meeting to drive more consolidated discussions, then see where you reach.