This patch allows sparc32 to work with more recent than gcc-3.3. The
problem is that a symbol needed by libgcc is resolved by libc.
Signed-off-by: Jeff Bailey <jbailey@ubuntu.com>
--- klibc-1.1.16.orig/scripts/Kbuild.klibc
+++ klibc-1.1.16/scripts/Kbuild.klibc
@@ -253,7 +253,8 @@
$(EXTRA_KLIBCLDFLAGS) \
$(KLIBCCRT0) \
$(link-deps) \
- $(KLIBCLIBC) $(KLIBCLIBGCC); \
+ $(KLIBCLIBC) $(KLIBCLIBGCC) \
+ $(KLIBCLIBC); \
cp -f $@ $@.g; \
$(KLIBCSTRIP) $(KLIBCSTRIPFLAGS) $@
On Tue, Jan 31, 2006 at 12:08:55PM +0000, Jeff Bailey wrote:> This patch allows sparc32 to work with more recent than gcc-3.3. The > problem is that a symbol needed by libgcc is resolved by libc. > > Signed-off-by: Jeff Bailey <jbailey@ubuntu.com> > > --- klibc-1.1.16.orig/scripts/Kbuild.klibc > +++ klibc-1.1.16/scripts/Kbuild.klibc > @@ -253,7 +253,8 @@ > $(EXTRA_KLIBCLDFLAGS) \ > $(KLIBCCRT0) \ > $(link-deps) \ > - $(KLIBCLIBC) $(KLIBCLIBGCC); \ > + $(KLIBCLIBC) $(KLIBCLIBGCC) \ > + $(KLIBCLIBC); \It looks like a papering over something repeating $(KLIBCLIBC) twice. Have you considered using: --start-group $(KLIBCLIBC) $(KLIBCLIBGCC) --end-group so we get all symbols resolved that can be resolved, independent on order they are defined? Sam