From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:32802) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1dqFex-00033t-U2 for qemu-devel@nongnu.org; Fri, 08 Sep 2017 05:33:20 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1dqFeq-0006BS-KT for qemu-devel@nongnu.org; Fri, 08 Sep 2017 05:33:11 -0400 Received: from mx1.redhat.com ([209.132.183.28]:49712) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1dqFeq-0006AV-EO for qemu-devel@nongnu.org; Fri, 08 Sep 2017 05:33:04 -0400 From: Markus Armbruster References: <20170831142430.16665-1-berrange@redhat.com> Date: Fri, 08 Sep 2017 11:33:01 +0200 In-Reply-To: <20170831142430.16665-1-berrange@redhat.com> (Daniel P. Berrange's message of "Thu, 31 Aug 2017 15:24:25 +0100") Message-ID: <87r2vhfsc2.fsf@dusky.pond.sub.org> MIME-Version: 1.0 Content-Type: text/plain Subject: Re: [Qemu-devel] [PATCH v2 0/5] qapi: support py2 & py3 in parallel List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Daniel P. Berrange" Cc: qemu-devel@nongnu.org, Michael Roth , David Michael "Daniel P. Berrange" writes: > Since I claimed that supporting py2 & py3 in parallel would be easy > for QEMU, I figured I ought to actually give it a try to backup that > assertion. > > This small patch series is the result of that effort. I tested this > series on Fedora 26 using 2.7.13 and Python 3.6.2. > > To test with py3, I hacked config-host.mak to change the PYTHON > variable to point to 'python3' binary, then compared the following > generated content for the files: > > qmp-commands.h qapi-types.h qapi-visit.h qapi-event.h > qmp-marshal.c qapi-types.c qapi-visit.c qapi-event.c > qmp-introspect.c qmp-introspect.h > > with that generated under py2 to see they are identical. > > It is possible there's still more bugs hiding that could impact > on 2.6 or earlier versions of 3.x or 2.7.x, so this probably > needs a bit wider testing, but I think the series illustrates > the broad scope of the changes we can expect. Only the need > to adapt to different module import locations adds to the > line count, and that's fairly minimal. This hasn't made it to the front of my review queue, but I got a quick question meanwhile. I guess this was triggered by the discussion of David's "[PATCH] scripts: Support building with Python 3". How does it differ from David's patch? Is it derived from it? Or is it an independently developed replacement?