LinuxQuestions.org - [SOLVED] Odd behaviour from 1st compile.

- Linux From Scratch (https://www.linuxquestions.org/questions/linux-from-scratch-13/)

- - Odd behaviour from 1st compile. (https://www.linuxquestions.org/questions/linux-from-scratch-13/odd-behaviour-from-1st-compile-4175638529/)

Odd behaviour from 1st compile.

Starting Chapter 5 - binutils-pass 1 onwards I am getting weird behaviour. In the make compiles halt on

Code:

checking how to print strings...

If I interrupt with Ctrl_C, it throws me an error with zlib; If I hit returm, It prompts for a password :-/. I have tried the lfs user's password, and it appears to loop. If I enter no password, I get a 2nd password prompt; Hit return again, and it proceeds as if nothing had happened:-//. No warning, no error :-o.

Code:

lfs:~$ cat .bash_profile

exec env -i HOME=$HOME TERM=$TERM PS1='\u:\w\$ ' /bin/bash

lfs:~$ cat .bashrc

set +h

umask 022

LFS=/media/hd0

LC_ALL=POSIX

LFS_TGT=$(uname -m)-lfs-linux-gnu

PATH=/tools/bin:/bin:/usr/bin

export LFS LC_ALL LFS_TGT PATH

I'm building on Slackware64-current, and my version throws the odd wartning about headers on a compile (usually sysdeps.h), but they never interfered with anything.

EDIT: These errors come from configure scripts. Binutils and gcc run a few each.

I would run the version-ck.sh and run echo $LFS Then remount the target partition mount -v -t ext4 /dev/<xxx> $LFS Using this LFS=/media/hd0 is not standard and can introduce issues.

1st compile reply

I'd start over with Sections 2 and 4. Not using export LFS=/mnt/lfs may have tripped something up in those.

Thanks for the reply.

I think I can call this solved. It's a gremlin from my base system (too many updates, no recent clean install) but it's harmless. I went on, & things are fine. It was big in binutils & gcc, but they have several configure scripts in both of them. The 'make test' stage in things is coming out as clean as you would expect for chapter 5. I'm on bison now in Ch. 5-17.1 and all is smooth. The definitive test is the make test on gcc in chapter 6. I'll scrub it all if that looks poorly.

On this box, building a kernel throws a bucketload of warnings, but the kernel comes out just fine.

Yep, it's solved all right.

On chapter 6 now, passing with flying colors on Binutils, Glibc & Gcc, in fact on everything. The problem vanished in the chroot. It was satisfying to have binutils not stop with stupid nonsense; it has a lot of small compiles with individual configure scripts, and was the first thing in chapter 5 and the major offender, prompting the query.