This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] avoid ABI warnings on arm when running the libstdc++v3 testsuite
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: Paolo Carlini <paolo dot carlini at oracle dot com>
- Cc: Matthias Klose <doko at ubuntu dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>, "libstdc++ at gcc dot gnu dot org" <libstdc++ at gcc dot gnu dot org>
- Date: Fri, 15 Jan 2010 00:22:00 +0000 (UTC)
- Subject: Re: [patch] avoid ABI warnings on arm when running the libstdc++v3 testsuite
- References: <4B4F5A66.6020009@ubuntu.com> <4B4F5C6A.9050808@oracle.com>
On Thu, 14 Jan 2010, Paolo Carlini wrote:
> On 01/14/2010 06:54 PM, Matthias Klose wrote:
> > The attached patch runs the testsuite with -Wno-abi on arm targets. Ok
> > for the trunk and the 4.4 branch? Should the target be more restricted?
> Does not seem the right fix to me, the real issue is that *user code*
> will warn, unless the user passes -Wno-abi! How happens that a system
> header spills a warning, again? We have been working on those issues for
(a) It's a note (inform ()), not a warning; the general system header
warning suppression may not work for those.
(b) I don't see a report of the compilation options. Are the header
directories in the source and build trees being properly included with
-isystem rather than -I so they are marked as system headers? If not,
that's a testsuite bug. When testing the original patch I would have been
using installed testing, in which the headers would properly have been
marked as system headers. (Though I see the relevant header does indeed
use #pragma GCC system_header, so this particular issue shouldn't have
caused a problem.)
I definitely made sure when testing the patch that the mangling of va_list
did not affect the libstdc++.so ABI.
--
Joseph S. Myers
joseph@codesourcery.com