From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:52857) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1b7iNO-0004Uw-2U for qemu-devel@nongnu.org; Tue, 31 May 2016 08:02:31 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1b7iNI-0006CK-W3 for qemu-devel@nongnu.org; Tue, 31 May 2016 08:02:25 -0400 Received: from mx1.redhat.com ([209.132.183.28]:46161) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1b7iNI-0006CF-Pg for qemu-devel@nongnu.org; Tue, 31 May 2016 08:02:20 -0400 References: <20160527135437.20474-1-famz@redhat.com> <20160527135437.20474-5-famz@redhat.com> <20160531110058.GA24019@ad.usersys.redhat.com> From: Paolo Bonzini Message-ID: <56a9a421-be5a-2e81-c4d2-aa1d41219ff4@redhat.com> Date: Tue, 31 May 2016 14:02:16 +0200 MIME-Version: 1.0 In-Reply-To: <20160531110058.GA24019@ad.usersys.redhat.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH v6 04/15] Makefile: Rules for docker testing List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Fam Zheng Cc: qemu-devel@nongnu.org, =?UTF-8?Q?Alex_Benn=c3=a9e?= On 31/05/2016 13:00, Fam Zheng wrote: > On Tue, 05/31 10:51, Paolo Bonzini wrote: >>> diff --git a/tests/docker/Makefile.include b/tests/docker/Makefile.include >>> new file mode 100644 >>> index 0000000..372733d >>> --- /dev/null >>> +++ b/tests/docker/Makefile.include >>> @@ -0,0 +1,121 @@ >>> +# Makefile for Docker tests >>> + >>> +include $(SRC_PATH)/rules.mak >> >> Why include this _and_ include tests/docker/Makefile.include from the >> top Makefile? > > This is for quiet-command, which is only conditionally included by top > Makefile. Ah, it's for the case when configure has not been executed yet and the toplevel Makefile "assumes we are in the search tree". >> >> I think you should do one of this: >> >> a) drop this inclusion; nice, but it pollutes the toplevel makefile a bit >> >> b) do the following: >> >> - link this file into the build tree in configure >> >> - include ../../config-host.mak > > I prefer we support running from the src tree without running configure, but > $(MAKE) invocations doesn't propagate make variables such as SRC_PATH... > >> >> - add to the toplevel Makefile a rule like >> >> docker docker-%: >> $(MAKE) -C tests/docker $@ > > ... and explicitly passing it (and $(V), etc.) here seems very ad-hocery. V and others would be passed down to the recursive make. Only SRC_PATH would not be passed down. >> >> I prefer the latter. Either would make patch 3 unnecessary. > > Maybe I should make patch 3 a patch to make top Makefile include rules.mak > unconditionally? Yeah, that would be good. I'm still a bit undecided about the pollution introduced by tests/docker/Makefile.include, but I guess that's okay. By the way, could you prepare a patch to rename tests/Makefile to tests/Makefile.include? It's a good convention. Thanks, Paolo