On Mon, May 02, 2011 at 04:45:59PM -0500, Bruce Dubbs wrote: > Manuel Gonzalez Montoya wrote: > > On Mon, May 2, 2011 at 2:45 AM, Brett <[email protected]> wrote: > >> Hi > >> > >> I am trying to build LFS SVN-20110427 on an AMD64 system > >> > >> I get a consistent failure when running the checks for Binutils in > >> chapter 6. This is the output where the error occurs: > >> > > > > For the error in new.cc you need to remove the include > > <exception_defines.h> > > from the file ld/testsuite/ld-elf/new.cc > > or see http://comments.gmane.org/gmane.linux.lfs.support/32431 > > sed -i "/exception_defines.h/d" ld/testsuite/ld-elf/new.cc > > > For the errors in the selective test find the following line in the > > file ld/testsuite/ld-selective/selective.exp > > set cxxflags "-fvtable-gc -fno-exceptions -fno-rtti" > > and remove -fvtable-gc parameter > > or see http://sourceware.org/ml/binutils/2010-12/msg00189.html > > sed -i "s/-fvtable-gc //" ld/testsuite/ld-selective/selective.exp > > I created a ticket to make these fixes. Thanks. > > -- Bruce > Glad Manuel has provided a fix to disable that test (although it comes down to "what I don't know about cannot harm me ;)". I haven't built recently, but I'd just like to note that (at least on x86_64) during the past 3 years I've *often* seen failures in 'ld' tests, but everything has still worked ok. I think the ld tests are *good* at testing corner cases, and that seeing some failures there is not necessarily a reason to worry.
The fix is nice and tidy, but I do wonder if it might be better to accept [ at least for people using the development book ] that tests fail when the infrastructure changes (in this case, newer gcc) ? But, I've no strong feelings either way. ĸen -- das eine Mal als Tragödie, das andere Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
