From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Sat, 11 Aug 2018 12:25:29 +0200 Subject: [Buildroot] Analysis of build results for 2018-08-09 In-Reply-To: References: <20180810060017.0EEFE20719@mail.bootlin.com> <20180811001555.0b515e22@windsurf> Message-ID: <20180811122529.664a5d96@windsurf> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hello, Adding Hollis, which I forgot in my initial e-mail. Hollis, please have a look at the thread, there has been some other build errors on your older build machine. On Fri, 10 Aug 2018 19:20:47 -0500, Matthew Weber wrote: > > This we've got only 10 failures on master, let's have a look at them. > > > aarch64 | host-libselinux-2.7 | NOK | http://autobuild.buildroot.net/results/66c03506e10071ec44ca509304595f4f5150e4d9 | > > > mipsel | host-libselinux-2.7 | NOK | http://autobuild.buildroot.net/results/2b41afe7841feba626fc29c9474e82d9934e6173 | > > > > Due to the use of atomic built-ins, not available in old compilers. I > > would suggest to make all packages that depend on host-libselinux to > > depend on the appropriate host gcc version. > > Hollis has sent a patch suggestion to keep libselinux compatible with > older gcc. Is there a preference on carrying the patch vs using a > host gcc version detect ? I didn't realize we had a way to depend on > host gcc version. I'll take a look at that but let me know which way > is preferred. If the issue can be worked around by a patch, it's nice because we don't have to propagate the host gcc dependency to gazillions of packages. However, Hollis patch wasn't that simple/trivial, it would be nice to have it checked by upstream. Or maybe it's just me who isn't familiar enough with the semantic of atomic intrinsics. Best regards, Thomas -- Thomas Petazzoni, CTO, Bootlin (formerly Free Electrons) Embedded Linux and Kernel engineering https://bootlin.com