From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:48902)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1Vm1rG-00037C-CD
	for qemu-devel@nongnu.org; Thu, 28 Nov 2013 08:42:24 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1Vm1r6-0007wS-HG
	for qemu-devel@nongnu.org; Thu, 28 Nov 2013 08:42:18 -0500
Received: from mx1.redhat.com ([209.132.183.28]:23047)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1Vm1r6-0007v5-8c
	for qemu-devel@nongnu.org; Thu, 28 Nov 2013 08:42:08 -0500
Message-ID: <5297480B.9050501@redhat.com>
Date: Thu, 28 Nov 2013 14:41:31 +0100
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <1385601858-8065-1-git-send-email-imammedo@redhat.com>
	<CAEgOgz6-=w7ogzBqAeOjHQuJhR3P_ovGv5iTmQEfdGBJBt5-tg@mail.gmail.com>
	<20131128142323.01a88c7e@nial.usersys.redhat.com>
In-Reply-To: <20131128142323.01a88c7e@nial.usersys.redhat.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH] qom: abort on error in property setter if
 caller passed errp == NULL
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Igor Mammedov <imammedo@redhat.com>
Cc: Peter Crosthwaite <peter.crosthwaite@xilinx.com>, "Michael S. Tsirkin" <mst@redhat.com>, "qemu-devel@nongnu.org Developers" <qemu-devel@nongnu.org>, Anthony Liguori <anthony@codemonkey.ws>, =?ISO-8859-1?Q?Andreas_F=E4rber?= <afaerber@suse.de>

Il 28/11/2013 14:23, Igor Mammedov ha scritto:
> > object_property_set(Foo, bar, "baz", &abort_on_err);
> 
> that is just another way to put burden on caller, instead of doing it
> in one place.

It's also much more self-documenting.

The problem is that there is one very good case where you want the
silent-don't-care behavior: when you don't care about the exact error,
and you can figure out whether there was an error from the returned
value of the function.  This doesn't apply to object_property_set of
course, but it is the reason why NULL Error* has silent-don't-care behavior.

Now, let's look at the alternatives:

* keep silent don't care
  + consistent
  + predictable
  - not always handy

* only modify object_property_set
  + mostly does the right thing
  - inconsistent with other Error* functions
  - inconsistent with _nofail functions

* Peter's alternative
  + self-documenting
  + consistent
  + predictable

* make Error* mandatory for all void functions
  + consistent
  + almost predictable (because in C you can ignore return values)
  - not necessarily does the right thing (e.g. cleanup functions)
  - requires manual effort to abide to the policy

I vote for Peter's proposal, or for adding object_property_set_nofail.
No particular preference.

Another variant: modify object_property_set to add a g_warning.  I think
it's fine.  It reduces the inconsistency, and still simplifies debugging.

Paolo