This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: canadian cross trouble with libiberty location


DJ Delorie wrote:
   --build=i686-pc-linux-gnu \
   --host=x86_64-unknown-linux-gnu \
   --target=i686-unknown-linux-gnu \


This isn't a true canadian, since build==target.

But build *isn't* equal to target; the manufacturer field differs. Or is that field ignored here? That'd be sad; that trick has been really handy for building cross-compilers.

> I refer to these as
"crossback" builds and they are even more rare than canadians.  The
documentation section "Configure Terms and History" talks about the
different types of builds.

You make it sound like I'm after the Crumple-Horned Snorkack!


Most likely, there's somewhere testing "$build != $target" where it
should test "$build != $host" (or something similar, like
$build_alias), or getting the build vs target subdirectories confused.
But one drawback to even getting this right is that the build
libiberty and the target libiberty reside in the same directory yet
are built with different compilers, which may cause problems.  More
recent sources use build-<foo> to avoid this conflict.

OK, I'll try gcc-3.4.0 then, and see if that behaves any better.


One workaround is to use --build=i586-pc-linux-gnu instead of i686,
just so that build and target are different.

Groan. Then what do I do for --build=x86_64-unknown-linux-gnu \ --host=i686-unknown-linux-gnu \ --target=x86_64-unknown-linux-gnu ? I hope I don't have to do the 586 trick.

- Dan

--
My technical stuff: http://kegel.com
My politics: see http://www.misleader.org for examples of why I'm for regime change


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]