This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Require binutils 2.20 or later to build glibc
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: Roland McGrath <roland at hack dot frob dot com>
- Cc: libc-alpha at sourceware dot org, "Ryan S. Arnold" <rsa at us dot ibm dot com>, Steven Munroe <sjmunroe at us dot ibm dot com>, Andreas Krebbel <Andreas dot Krebbel at de dot ibm dot com>, Martin Schwidefsky <schwidefsky at de dot ibm dot com>, Kaz Kojima <kkojima at rr dot iij4u dot or dot jp>, Thomas Schwinge <thomas at codesourcery dot com>
- Date: Sun, 26 Feb 2012 22:16:23 +0000 (UTC)
- Subject: Re: Require binutils 2.20 or later to build glibc
- References: <Pine.LNX.4.64.1202242216420.2811@digraph.polyomino.org.uk><20120224230247.D0DF92C0A1@topped-with-meat.com>
On Fri, 24 Feb 2012, Roland McGrath wrote:
> In its own right, I think 2.20 seems like a reasonable baseline.
> But it seems suspect that we're requiring a binutils years newer
> than the GCC we require. Perhaps it would be OK to require 4.4.3
> by now, which meshes better with 2.20. But I'm not at all sure.
GCC release series are maintained for longer than binutils release series;
4.3.6 was released on 27 June last year (and was the final release of
4.3).
As noted in another discussion, there is at least one reason builds with
4.3 fail right now (uchar.h, used in the build, doesn't work with 4.3). I
don't know if there might turn out to be other causes of failure as well
if that (and the duplicate typedefs issue) were fixed.
Given that there were no objections and some support for the patch
requiring at least 4.3 I propose to commit in on Monday unless any late
issues come up. We can still move to requiring 4.4 afterwards if we
decide we don't want to fix uchar.h to allow 4.3 builds and it seems
established that we aren't trying to support 3.4 through 4.2 any more.
> I vaguely recall people posting before that requiring too new
> a binutils was burdensome for some platform or other, but I don't
> recall the details. I think we should at least get each arch maintainer
> to weigh in explicitly before deciding.
We know it's required for x86_64, don't have a separate i386 maintainer,
and have SPARC maintainer support.
That makes it powerpc, s390 and sh we haven't heard from the maintainers
of. I've CC:ed those maintainers; could you comment on whether requiring
binutils 2.20 or later is OK for your architecture?
--
Joseph S. Myers
joseph@codesourcery.com