Re: [Qemu-devel] [PATCH v8 15/18] qapi: Move duplicate member checks to schema check()

[Eric Blake <eblake@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1752 bytes --]

On 10/13/2015 11:13 AM, Markus Armbruster wrote:

>>>
>>> I've come to the conclusion that we should get rid of the self-inflicted
>>> pain before we attempt to detect all collisions.
>>
>> Then that sounds like I should try harder to get the kind/type naming,
>> the boxed base naming, and even the anonymous union naming all hoisted
>> into this subset, and spin a v9?
> 
> I can take PATCH 01-09,12 into my tree right away, with PATCH 07's two
> redundant is_implicit() methods dropped, and PATCH 12's comment touched
> up.

Okay.

> 
> I could take PATCH 10, but let's at least try to make a plan for
> c_name() first.  If we fail, I'll take the patch, perhaps less the % to
> + change, and we'll revisit c_name() later when we see more clearly.

At this point, I'm not sure whether 10 disappears completely after the
type/kind fix, so that alone is a good enough reason to leave 10 out of
your tree for another round.

> 
> You want to move PATCH 11 to later in the queue, and I like that.
> 
> PATCH 13 needs a fix squashed in, and a few nits touched up.  If you
> want me to do that on commit, please propose a patch for me to squash
> in.  But a respin is probably easier for all.
> 
> PATCH 14 is fine, but it depends on 13.
> 
> I haven't finished review of PATCH 15-18.
> 
> Taken together, I think the easiest way forward is I take 01-09,12, and
> you respin the rest after we finish its review.  Makes sense?
> 

Sounds like we're agreed then: take the obvious patches into your tree,
and let me rework the tail of this subset on top of cleanups that reduce
self-inflicted collisions.

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


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 604 bytes --]