I'm using -Os since I'll be installing it on a USB drive.

I've compiled the package from scratch now, and everything works OK now.
Thanks for the tips.


On Sun, Sep 1, 2013 at 7:05 PM, Ken Moffat <zarniwh...@ntlworld.com> wrote:

> On Sun, Sep 01, 2013 at 06:23:47PM +0200, Klemen Košir wrote:
> > I followed the instructions, but I set several environment variables:
> >
> > MAKEFLAGS="-j 4"
> > CFLAGS="-march=corei7 -Os -pipe"
> > CXXFLAGS="${CFLAGS}"
>
>  For some time, -Os has been deprecated in kernel builds.  It got
> introduced there with the hope it would improve cache handling, but
> in too many cases it produced bad code.  You have a recent machine,
> so your memory is probably bigger (for DDR3 the minimum chip size
> seems to be 2GB), so -Os is probably not going to be worthwhile.
> >
> > The error:
> >
> > ../../binutils-2.23.2/binutils/stabs.c: In function 'parse_stab_type':
> > ../../binutils-2.23.2/binutils/stabs.c:2797:57: error: 'physname' may be
> > used uninitialized in this function [-Werror=maybe-uninitialized]
> >       variants[cvars] = debug_make_static_method_variant (dhandle,
> >                                                          ^
> > ../../binutils-2.23.2/binutils/stabs.c:2596:16: note: 'physname' was
> > declared here
> >     const char *physname;
> >                 ^
> > cc1: all warnings being treated as errors
> > make[4]: *** [stabs.o] Error 1
> >
>  And the problem is -Werror.  Specifically, warning that a variable
> may be uninitialized.  That just means that gcc can't tell if the
> variable is initialized or not.
>
>  Unfortunately, Werror gets added to WARN_CFLAGS by some or all of
> the configure scripts in the subdirectories when you are building
> with gcc.
>
> ĸen
> --
> das eine Mal als Tragödie, dieses Mal als Farce
> --
> http://linuxfromscratch.org/mailman/listinfo/lfs-dev
> FAQ: http://www.linuxfromscratch.org/faq/
> Unsubscribe: See the above information page
>
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page

Reply via email to