On 6/5/06, Joe Ciccone <[EMAIL PROTECTED]> wrote:
I've probably done about a dozen builds by now and I haven't seen libtool puke yet.
I've seen it mentioned in mailing lists, but maybe it's not as bad as it seems. http://gcc.gnu.org/ml/libstdc++/2005-07/msg00067.html P.S. Remember the name Enrico Weigelt. :)
Currently, there isn't a toolchain adjustment. First you install your linux headers. Then you build binutils. Then you install glibc headers (tricking glibc into thinking nptl for that arch exists). Then you build a cross-compiler for your arch and install glibc into the sysroot.
Very interesting. No more temporary system, huh? The whole system is cross-compiled straight away? This looks really cool.
In the end you have a toolchain where you can pass -L/usr/lib -lssl and ld will look for libssl.so in [sysroot]/usr/lib. It's a very interesting build process. I'm uploading the current book to http://www.linuxfromscratch.org/~jciccone/cross-lfs-2.0/ right now. Comments are always welcome.
So, --with-sysroot is only needed while building the toolchain? It looks like the only other trick is using DESTDIR to get the install to go to the right spot. I imagine perl is causing all kind of nightmares. This looks awesome. Not to bash the work being done for CLFS-1.0, but this looks like the real deal. I'd love to contribute some time down the road when I've got some time to get my hands dirty with a cross-compiler. Keep up the good work! -- Dan -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
