From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 83392 invoked by alias); 19 Oct 2016 03:23:08 -0000 Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org Received: (qmail 81776 invoked by uid 89); 19 Oct 2016 03:22:52 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-1.1 required=5.0 tests=AWL,BAYES_00,SPF_HELO_PASS,SPF_SOFTFAIL autolearn=no version=3.3.2 spammy=hooked, dark, 72, marks X-HELO: mail.baldwin.cx Received: from bigwig.baldwin.cx (HELO mail.baldwin.cx) (96.47.65.170) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Wed, 19 Oct 2016 03:22:41 +0000 Received: from ralph.baldwin.cx (c-73-231-226-104.hsd1.ca.comcast.net [73.231.226.104]) by mail.baldwin.cx (Postfix) with ESMTPSA id B379C10AF91; Tue, 18 Oct 2016 23:22:39 -0400 (EDT) From: John Baldwin To: gdb-patches@sourceware.org Cc: Pedro Alves Subject: Re: Buildslaves and C++11 support Date: Wed, 19 Oct 2016 03:23:00 -0000 Message-ID: <3233416.N5JnNfTiFt@ralph.baldwin.cx> User-Agent: KMail/4.14.10 (FreeBSD/11.0-PRERELEASE; KDE/4.14.10; amd64; ; ) In-Reply-To: <14083dec-ef75-e6ad-3c52-d947478ebc82@redhat.com> References: <1476448162-20203-1-git-send-email-palves@redhat.com> <14083dec-ef75-e6ad-3c52-d947478ebc82@redhat.com> MIME-Version: 1.0 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="us-ascii" X-IsSubscribed: yes X-SW-Source: 2016-10/txt/msg00556.txt.bz2 On Tuesday, October 18, 2016 12:20:00 AM Pedro Alves wrote: > On 10/17/2016 07:00 PM, Pedro Alves wrote: > > > A small update here. > > > > I plan to use the new trybot support to test a patch > > that forces the C++11 requirement, to determine which > > buildbots don't have a recent enough compiler yet. Also > > chatting with Sergio about getting this info written > > down and visible on the buildbot web frontend or extracted > > systematically out of the builds, to avoid this "in the dark" > > situation. > > I was looking at this, and noticed that > > https://sourceware.org/gdb/wiki/BuildBot#Available_Builders > > doesn't actually cover all the build slaves. The ones it does cover I > think we already know run sufficient enough gcc from the OS version. > > So I looked at the full list of configured build slaves, at: > > http://gdb-build.sergiodj.net/buildslaves > > and built the following table based on info I could find either in the > build slaves' info page, or if compiler info missing there, by doing a > web search for which gcc version a given distro ships with. > > Slave | Compiler / Info > ---------------------------------+----------------------------------------------------------- > aix-power8-1 | AIX 7.2 - it's on the gcc compile farm; I logged in, and it has GCC 6.1 > debian-jessie-mips-be-1 | Debian Jessie (comes with GCC 4.9, AFAIK); not connected anyway. > debian-jessie-s390x-1 | Debian Jessie (comes with GCC 4.9, AFAIK); connected > fedora-ppc64be-1 | Fedora 23 (comes with GCC 5.3.1) > fedora-ppc64le-1 | Fedora 23 (comes with GCC 5.3.1) > fedora-x86-64-1 | Fedora 24 (GCC 6.1) > fedora-x86-64-2 | Fedora 24 (GCC 6.1) > fedora-x86-64-3 | Fedora 24 (GCC 6.1) > fedora-x86-64-4 | Fedora 24 (GCC 6.1) > gdb-amd64-netbsd7 | NetBSD 7 (gcc version 4.8.4 (nb2 20150115)) > koobs-freebsd8 | *not connected* > koobs-freebsd9 | gcc 4.2 (fails building gdb now already) > koobs-freebsd10 | clang 3.4.1 (C++11 ready) > koobs-freebsd11 | clang 3.8.0 (C++11 ready) > marist-fedora-s390x | Fedora 23 (GCC 5.3.1) > rhel-7.1-s390x-1 | RHEL 7.1 - should be GCC 4.8 > ubuntu-aarch64-1 | Ubuntu 14.04 (GCC 4.8.2) > wildebeest-debian-jessie-i686 | Debian Jessie (GCC 4.9.2) > wildebeest-debian-jessie-x86_64 | Debian Jessie (GCC 4.9.2) > > I believe that if any is not actually building with the > system compiler, it'll probably be using a newer gcc version... > > So looks like the only one not C++11 ready is the FreeBSD 9 one. > But, that one hasn't managed to build gdb for a long while > [1] (did it ever?) and isn't even hooked to send test reports to the > testers list, AFAIK. We wouldn't really miss it, IMO. > > [1] - http://gdb-build.sergiodj.net/builders/FreeBSD-x86_64-m64 > > So looks like Mark's Debian Wheezer buildslave was really the only one > that needed updating and we're actually ready for C++11 from > the buildbot's perspective? For FreeBSD 9, I can fix the port that is used to build the package folks install to use a C++11 compiler for GDB 7.13, so that should be fine. It should also be fairly simple to fix the buildbot to install a newer compiler from a package to use instead of the system compiler as well. I can follow up with the maintainer. Personally, I am only testing GDB on FreeBSD 10.x and later generally using clang (though the clang build has a few warnings still, when I've used gcc it built cleanly). -- John Baldwin