This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: Will the gold linker ever be NON Beta ?
- From: Markus Trippelsdorf <markus at trippelsdorf dot de>
- To: Cary Coutant <ccoutant at gmail dot com>
- Cc: Ian Lance Taylor <iant at google dot com>, Gordon McConnell <Gordon dot McConnell at synopsys dot com>, "binutils at sourceware dot org" <binutils at sourceware dot org>
- Date: Thu, 14 Sep 2017 11:43:29 +0200
- Subject: Re: Will the gold linker ever be NON Beta ?
- Authentication-results: sourceware.org; auth=none
- References: <9328C5B713035A448835151B68585C3001845DC86F@us01wembx1.internal.synopsys.com> <CAKOQZ8yMB8GdP49tCx0yug-8kYW1ign8Ag6jdNrimK+v0kxFuw@mail.gmail.com> <CAJimCsE9Zj5-fqSquJhui84ZJiNC-J4FZtgRybLpsyR9LtxGxA@mail.gmail.com>
On 2017.06.22 at 11:04 -0700, Cary Coutant wrote:
> On Wed, Jun 21, 2017 at 5:30 PM, Ian Lance Taylor via binutils
> <binutils@sourceware.org> wrote:
> > On Wed, Jun 21, 2017 at 7:17 AM, Gordon McConnell
> > <Gordon.McConnell@synopsys.com> wrote:
> >>
> >> The main binutils page at https://www.gnu.org/software/binutils/ states:
> >> gold - A new, faster, ELF only linker, still in beta test.
> >>
> >> I've found some old threads about gold's Beta status - the newest appears to be this one from April 2015: https://sourceware.org/ml/binutils/2015-04/msg00051.html
> >>
> >> Has anything changed since then, such that the Beta status might be altered any time soon ?
> >
> > As far as I know nothing has changed, which is to say, no GNU/Linux
> > distro has yet chosen to make gold the standard linker.
>
> I'd argue that a better criterion for when to emerge from a beta state
> is feature completeness. The only area that I know of where gold could
> be considered incomplete is in the scripting language support, but
> full support for the Gnu ld scripting language was never a stated goal
> of the project -- the goal was simply to support enough scripting to
> enable the glibc build with gold. I think we're there, but if anyone
> knows anything that's still preventing gold from building glibc,
> please point me to the PR(s) and I'll prioritize those.
Just an update. Glibc can now be build with gold. The only issue left is
an __asm__(.symver ...) related testsuite error, where ld.bfd and
ld.gold behave differently See
https://sourceware.org/bugzilla/show_bug.cgi?id=22011 for a testcase.
--
Markus