Re: [PATCH v7 00/11] error: auto propagated local_err part I

From: Markus Armbruster <armbru@redhat.com>
To: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Cc: "Kevin Wolf" <kwolf@redhat.com>,
	"Stefano Stabellini" <sstabellini@kernel.org>,
	qemu-block@nongnu.org, "Paul Durrant" <paul@xen.org>,
	"Philippe Mathieu-Daudé" <philmd@redhat.com>,
	"Michael Roth" <mdroth@linux.vnet.ibm.com>,
	qemu-devel@nongnu.org, "Greg Kurz" <groug@kaod.org>,
	"Gerd Hoffmann" <kraxel@redhat.com>,
	"Stefan Hajnoczi" <stefanha@redhat.com>,
	"Anthony Perard" <anthony.perard@citrix.com>,
	xen-devel@lists.xenproject.org, "Max Reitz" <mreitz@redhat.com>,
	"Laszlo Ersek" <lersek@redhat.com>,
	"Stefan Berger" <stefanb@linux.ibm.com>
Subject: Re: [PATCH v7 00/11] error: auto propagated local_err part I
Date: Mon, 16 Mar 2020 15:40:02 +0100	[thread overview]
Message-ID: <87r1xsfjnh.fsf@dusky.pond.sub.org> (raw)
In-Reply-To: <87abeabb-c8ee-ed6f-6b3a-b3fc24d07b89@virtuozzo.com> (Vladimir Sementsov-Ogievskiy's message of "Tue, 3 Mar 2020 11:12:46 +0300")

Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes:

> 03.03.2020 11:01, Markus Armbruster wrote:
>> Hi Vladimir,
>>
>> I've come to rather like your ERRP_AUTO_PROPAGATE() idea.  What I
>> wouldn't like is a protracted conversion.
>>
>> Once we're happy with PATCH 1-3, it's a matter of running Coccinelle and
>> reviewing its output.  I'm confident we can converge on PATCH 1-3.
>>
>> It's two weeks until soft freeze.  We need to decide whether to pursue a
>> partial conversion for 5.0 (basically this series plus the two patches
>> we identified in review of PATCH 1), or delay until 5.1.  In either
>> case, I want the conversion to be finished in 5.1.
>>
>> Please do not feel pressured to make the 5.0 deadline.
>>
>> I can queue up patches for 5.1 during the freeze.
>>
>> How would you like to proceed?
>>
>
> Hi Markus! Funny coincidence: exactly now (less than 1 hour ago), I've
> started working for the next version for these series. So, I'm going to
> resend today. Of course, I'd prefer to merge something to 5.0 if at all
> possible.

That was v8, followed by v9.  We're clearly converging.  However, the
soft freeze is tomorrow already.

You've persevered with this idea for quite a while; some impatience
would be quite excusable now.  Still, I doubt part I making 5.0 matters.
The hand-written part is likely to rebase easily, and the generated part
should be regenerated instead of rebased anyway.

What actually matters is *finishing* the job.  What does that take?

* Consensus on the hand-written part.  I think we're basically there, we
  just want to work in a few more tweaks.

* Split the generated part into reviewable batches, regenerating patches
  as necessary.  Solicit review.  First batch is part of this series,
  and v9 looks ready to me.  I assume you'll prepare the remaining
  batches.

* Queue up batches as they become ready, post pull requests.  I can do
  that.

* Update the QAPI code generator to the new Error usage.  I can do that.