Date: Sat, 21 Feb 2015 21:00:23 +0100 From: John Marino <freebsd.contact@marino.st> To: Jan Beich <jbeich@FreeBSD.org> Cc: svn-ports-head@freebsd.org, svn-ports-all@freebsd.org, marino@freebsd.org, ports-committers@freebsd.org Subject: Re: svn commit: r379485 - in head: mail/thunderbird/files www/firefox-esr/files www/libxul/files Message-ID: <54E8E3D7.7060102@marino.st> In-Reply-To: <4mqf-78zc-wny@FreeBSD.org> References: <201502210822.t1L8MSNt008911@svn.freebsd.org> <54E8444B.4090707@marino.st> <4mqf-78zc-wny@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2/21/2015 13:19, Jan Beich wrote: > John Marino <freebsd.contact@marino.st> writes: > >>> Author: jbeich >>> Date: Sat Feb 21 08:22:28 2015 >>> New Revision: 379485 >>> URL: https://svnweb.freebsd.org/changeset/ports/379485 >>> QAT: https://qat.redports.org/buildarchive/r379485/ >>> >>> Log: >>> Fix invalid conversions of nullptr to bool to unbreak clang 3.6 build >>> >>> PR: 197857 >>> Submitted by: dim (initial version) >>> Obtained from: https://github.com/mozilla/gecko-dev/commit/dd19f59 >>> >> >> All the mozilla stuff is busted on gcc5 for this reason: firefox, >> seamonkey, libxul, thunderbird. I have high hopes this patch fixes it. >> Are you planning to patch them all? > > Isn't USES=compiler:c++11-lang supposed to pick Clang over GCC? Not on DragonFly. > This commit only backports the fix from Firefox 32, SeaMonkey 2.29. > > If you haven't filed a bug don't expect me to know about the issue. > My testing is limited due to not having a buildfarm/fast hardware > while (any) GCC requires more memory/buildtime than Clang + extra > time spent building the compiler itself. So, just knowing which > file/which line fails to compile would have saved a lot of time. GCC 5 isn't even in pre-release yet, but this issue with all of Mozilla is pretty well documented because Fedora 22 will release with GCC-5 and all the mozilla stuff failed to build. I've got my own list of around ~180 ports that failed to build although some of them were due to not passing flags and not setting the environment. I only found out this week about Mozilla and I'm still working the big package failures (e.g. I just fixed boost-libs which took out over 800 ports). I only expected a problem with -esr ports because I have to assume these errors will be fixed before next year. Anyway, I was sincere with my thanks. It was a welcome sight to see these fixed before I had to do something with them. BTW, I've already checked and they now build on gcc-5. John
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?54E8E3D7.7060102>