Date: Tue, 03 Sep 2019 14:06:42 -0000 From: John Baldwin <jhb@FreeBSD.org> To: Enji Cooper <ngie@FreeBSD.org>, src-committers@freebsd.org, svn-src-user@freebsd.org Subject: Re: svn commit: r346108 - user/ngie/fix-up-cxxstd/share/mk Message-ID: <c293316b-c680-5200-0aa1-7fa18a358565@FreeBSD.org> In-Reply-To: <201904102219.x3AMJlRu006058@repo.freebsd.org> References: <201904102219.x3AMJlRu006058@repo.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 4/10/19 3:19 PM, Enji Cooper wrote: > Author: ngie > Date: Wed Apr 10 22:19:47 2019 > New Revision: 346108 > URL: https://svnweb.freebsd.org/changeset/base/346108 > > Log: > Rework r345708 > > * Default to gnu++14 with clang++ and g++ 6.x+ according to the defaults of > each compile toolchain. > * Do not override user-specified values for `-std=*`. Some ports rely on this > functionality. > > This patch is similar to the one submitted by jbeich, however, it aims to > optimize the latter case by not testing the compiler version, etc, at all. > > Reported by: jbeich One thing I don't understand is why we'd set the value to the default value. If we want to use the compiler's default it seems like we should just not set the flag at all. Trying to keep the list in sync across compiler versions seems fragile. If, OTOH, we want to ensure that FreeBSD only uses C++11 so that in theory any C++11 compiler can compile it, then forcing C++11 even on newer compilers as the original change did would seem to make sense (and seems to be a legitimate reason to force -std). However, trying to just match the compiler's default seems to be a giant NOP, but one that requires ongoing maintenance. Can you describe the goal of what the -std changes are trying to do? -- John Baldwin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?c293316b-c680-5200-0aa1-7fa18a358565>