From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:47956) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YT6Ed-0008JM-Jj for qemu-devel@nongnu.org; Wed, 04 Mar 2015 05:09:00 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1YT6Ec-0008Ok-Np for qemu-devel@nongnu.org; Wed, 04 Mar 2015 05:08:59 -0500 Date: Wed, 4 Mar 2015 11:08:43 +0100 From: Kevin Wolf Message-ID: <20150304100842.GD3465@noname.str.redhat.com> References: <1425412630-14004-1-git-send-email-mreitz@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1425412630-14004-1-git-send-email-mreitz@redhat.com> Subject: Re: [Qemu-devel] [PATCH] iotests: Fix 061's reference output List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Max Reitz Cc: Stefan Hajnoczi , qemu-devel@nongnu.org, qemu-block@nongnu.org, Markus Armbruster Am 03.03.2015 um 20:57 hat Max Reitz geschrieben: > As of commit dc523cd348c47372faa7271c9aab2030f94c290d, passing an unkown > option name for image creation no longer results in that option's name > printed. This affects iotest 061's output, fix it. > > Signed-off-by: Max Reitz > --- > tests/qemu-iotests/061.out | 1 - > 1 file changed, 1 deletion(-) > > diff --git a/tests/qemu-iotests/061.out b/tests/qemu-iotests/061.out > index e70f983..52f0a31 100644 > --- a/tests/qemu-iotests/061.out > +++ b/tests/qemu-iotests/061.out > @@ -287,7 +287,6 @@ Lazy refcounts only supported with compatibility level 1.1 and above (use compat > qemu-img: Error while amending options: Invalid argument > Unknown compatibility level 0.42. > qemu-img: Error while amending options: Invalid argument > -qemu-img: Invalid parameter 'foo' > qemu-img: Invalid options for file format 'qcow2' > Changing the cluster size is not supported. > qemu-img: Error while amending options: Operation not supported I think I would prefer a patch that primarily fixes the error message (I would consider the recent change that removed the name of the unknown option a usability regression that the test caught) and then changes the test only as a byproduct. Markus, throwing away the original error message like commit dc523cd3 does is almost always wrong. It's the callee that knows what exactly the problem is; the caller may add context, but it rarely knows the details better. Kevin