Message ID | 871v3kchxx.fsf@kepler.schwinge.homeip.net |
---|---|
State | Accepted, archived |
Headers | show |
On Mon, 7 Feb 2011, Thomas Schwinge wrote: > Hallo! > > On Sun, 6 Feb 2011 21:55:53 -0800 (PST), Roland McGrath <roland@frob.com> wrote: > > I don't think we need to support a -bsd command-line option. > > Here is a patch. OK to commit? As the Hurd maintainer for GCC, you don't need approval to commit this patch....
Hallo! On Mon, 7 Feb 2011 18:14:42 +0000 (UTC), "Joseph S. Myers" <joseph@codesourcery.com> wrote: > On Mon, 7 Feb 2011, Thomas Schwinge wrote: > > On Sun, 6 Feb 2011 21:55:53 -0800 (PST), Roland McGrath <roland@frob.com> wrote: > > > I don't think we need to support a -bsd command-line option. > > > > Here is a patch. OK to commit? > > As the Hurd maintainer for GCC, you don't need approval to commit this > patch.... I see -- then let's say I posted it for review, but nobody had any comments, so I now committed it. :-) On Mon, 07 Feb 2011 10:28:10 +0100, I wrote: > I will continue to see whether the GNU/Hurd spec files can be simplified > / streamlined. Joseph, is that actually a sensible use of my time, given that you're generally working on untangling these spec definitions? (I'm of course not asking you to specifically work on GNU/Hurd bits, so if I'm misinterpreting your current work, please tell so, and I will continue.) Grüße, Thomas
On Tue, 8 Feb 2011, Thomas Schwinge wrote: > On Mon, 07 Feb 2011 10:28:10 +0100, I wrote: > > I will continue to see whether the GNU/Hurd spec files can be simplified > > / streamlined. > > Joseph, is that actually a sensible use of my time, given that you're > generally working on untangling these spec definitions? (I'm of course > not asking you to specifically work on GNU/Hurd bits, so if I'm > misinterpreting your current work, please tell so, and I will continue.) I think further disentangling (so that linux.h headers are not used for non-Linux targets) will wait until 4.7 (once the obsoleted targets have been reviewed, only on x86 will there be non-Linux targets using linux.h and so only on x86 will the architecture-specific linux*.h need cleaning up).
diff --git a/gcc/config/gnu.h b/gcc/config/gnu.h index 1299376..db9d63c 100644 --- a/gcc/config/gnu.h +++ b/gcc/config/gnu.h @@ -22,11 +22,11 @@ along with GCC. If not, see <http://www.gnu.org/licenses/>. /* Provide GCC options for standard feature-test macros. */ #undef CPP_SPEC -#define CPP_SPEC "%{posix:-D_POSIX_SOURCE} %{bsd:-D_BSD_SOURCE}" +#define CPP_SPEC "%{posix:-D_POSIX_SOURCE}" -/* Default C library spec. Use -lbsd-compat for gcc -bsd. */ +/* Default C library spec. */ #undef LIB_SPEC -#define LIB_SPEC "%{pthread:-lpthread} %{bsd:-lbsd-compat} %{pg|p|profile:-lc_p;:-lc}" +#define LIB_SPEC "%{pthread:-lpthread} %{pg|p|profile:-lc_p;:-lc}" /* Standard include directory. In GNU, "/usr" is a four-letter word. */ #undef STANDARD_INCLUDE_DIR diff --git a/gcc/config/i386/gnu.h b/gcc/config/i386/gnu.h index 077933b..2ce6aea 100644 --- a/gcc/config/i386/gnu.h +++ b/gcc/config/i386/gnu.h @@ -27,7 +27,7 @@ along with GCC. If not, see <http://www.gnu.org/licenses/>. #define TARGET_VERSION fprintf (stderr, " (i386 GNU)"); #undef CPP_SPEC -#define CPP_SPEC "%{pthread:-D_REENTRANT} %{posix:-D_POSIX_SOURCE} %{bsd:-D_BSD_SOURCE}" +#define CPP_SPEC "%{pthread:-D_REENTRANT} %{posix:-D_POSIX_SOURCE}" #undef CC1_SPEC #define CC1_SPEC "%(cc1_cpu)"
Hallo! On Sun, 6 Feb 2011 21:55:53 -0800 (PST), Roland McGrath <roland@frob.com> wrote: > I don't think we need to support a -bsd command-line option. Here is a patch. OK to commit? I will continue to see whether the GNU/Hurd spec files can be simplified / streamlined. gcc/ 2011-02-07 Thomas Schwinge <thomas@schwinge.name> * config/gnu.h (CPP_SPEC, LIB_SPEC): Remove handling of -bsd option. * i386/config/gnu.h (CPP_SPEC): Likewise. --- gcc/config/gnu.h | 6 +++--- gcc/config/i386/gnu.h | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-)