Return-Path: X-Original-To: pgbuildfarm-members@lists.pgfoundry.org Delivered-To: pgbuildfarm-members@lists.pgfoundry.org Received: from localhost (unknown [188.227.186.44]) by pgfoundry.org (Postfix) with ESMTP id 042A4A0078B for ; Thu, 8 Sep 2016 03:12:11 +0000 (UTC) Received: from pgfoundry.org ([188.227.186.71]) by localhost (maia-lon.uk.hub.org [188.227.186.44]) (maiad, port 10024) with ESMTP id 15963-09 for ; Thu, 8 Sep 2016 03:12:10 +0000 (UTC) X-Greylist: delayed 00:59:58 by SQLgrey-1.8.0 Received: from sss.pgh.pa.us (sss.pgh.pa.us [66.207.139.130]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by pgfoundry.org (Postfix) with ESMTPS id 8C72AA00773 for ; Thu, 8 Sep 2016 03:12:10 +0000 (UTC) Received: from sss1.sss.pgh.pa.us (localhost [127.0.0.1]) by sss.pgh.pa.us (8.14.4/8.14.4) with ESMTP id u882C91Q015599; Wed, 7 Sep 2016 22:12:09 -0400 From: Tom Lane To: Daniel Black In-reply-to: <773008e7-db08-a525-c887-de9aee60fdce@au.ibm.com> References: <773008e7-db08-a525-c887-de9aee60fdce@au.ibm.com> Comments: In-reply-to Daniel Black message dated "Mon, 05 Sep 2016 16:08:09 +1000" Date: Wed, 07 Sep 2016 22:12:09 -0400 Message-ID: <15598.1473300729@sss.pgh.pa.us> Cc: pgbuildfarm-members@lists.pgfoundry.org Subject: Re: [Pgbuildfarm-members] new build members X-BeenThere: pgbuildfarm-members@lists.pgfoundry.org X-Mailman-Version: 2.1.18-1 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 08 Sep 2016 03:12:11 -0000 Content-Length: 1013 Lines: 24 Daniel Black writes: > I've configured a bunch of builds to add to the build farm. The list is: Hi Daniel! Thanks for bringing all those new critters online. However, may I suggest that the ccache+clang ones would be a lot more useful if you added this to their configuration: build_env =>{ + # needed to make ccache play nice with clang + CCACHE_CPP2 => '1', Lack of this option is what's causing a lot of those useless warnings (it has to do with clang being fed preprocessor output, which triggers some of its more dubious warning heuristics). I think this has become the default behavior as of ccache 3.3, but in older versions you gotta ask for it. As an example of the importance of this, http://buildfarm.postgresql.org/cgi-bin/show_stage_log.pl?nm=devario&dt=2016-09-08%2000%3A44%3A53&stg=make contains something north of 1000 warnings, most of which are entirely bogus. Makes it hard to see if there are any that are meaningful. regards, tom lane