Greetings from a long time lurker. Jeremy's post prompted me to ask a question I've had for some time now. Why have a crooslfs if you end building lfs natively anyway?
I've been playing with lfs for about a year now. I turned to the crooslfs project when I wanted to turn an old pentium machine into dedicated firewall for my home network. Why crosslfs? I started building the new lfs system from the livecd. This was painfully slow. So I though I could build it using cross techniques and build the image on my main box and then move the drive I used back to the old machine. Needless to say using crosslfs I still would have ended up building a substantial portion on the old machine. I realize that a pentium is not really a cross build since my main machine is an Athlon but I'm sure there are cases (embedded systems) where you can't really just boot and continue the process. What I'm playing with now is a hybrid method. I basically build the temptools phase of lfs. Then using techniques from croosgcc, build a croos tool chain for my new target. I then use these tools to build the target image. The cross tools link against the libs built in the croostools phase but they should end up being functionally equivalent to the glibc we build at the start of what would be the chroot phase (chap 6). To me this seems like a worthy goal for crosslfs. To actually build the entire lfs system on one machine and the be able to move it to the machine being targeted. Am I missing something? -- Regards, Jeff Cousino -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
