This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: A patch for the toplevel Makefile
On Thu, May 18, 2000 at 07:52:41PM -0300, Alexandre Oliva wrote:
> On May 18, 2000, "H . J . Lu" <hjl@lucon.org> wrote:
>
> > +if [ x${target} != x${build} -a x${build} = x${host} ]; then
>
> The use of `-a' is not fully portable. Instead, use multiple
> invocations of `[' separated by `&&'. In fact, instead of `[', it is
> recommended to use `test', because it can be safely moved into
> configure.in and processed by autoconf.
Ok. I will update it.
>
> I'm not sure I understand the role of $build in this test. Why does
> it make any difference if its equal to host or not?
>
> > + echo "Please create the directory:" 1>&2
> > + echo 1>&2
> > + echo " ${prefix}/lib/gcc-lib/${target_alias}/${gcc_version}" 1>&2
> > + echo 1>&2
> > + echo "first for successful cross-compilation." 1>&2
>
> I'm not sure I understand why this is necessary. I've never had to
> create any directory in the installation tree before I could get a
> fully functional cross toolchain. But then, I build all tools in a
> single build tree. Are you talking about the scenario in which, say,
> the assembler and the linker are already installed, and then you're
> starting to build the compiler? How about testing whether `ld' and
No. It is because I don't use newlib. My glibc has been installed
under ${prefix}/${target_alias}.
BTW, there is another bug in my old patch. I am waiting for feedback
from Ian.
H.J.