Change History (10)
comment:1 by , 22 years ago
Priority: | high → low |
---|
comment:2 by , 22 years ago
I have confirmed that this method does work - flawlessly. Just give me the word and I'll stick this in ASAP.
comment:4 by , 21 years ago
using ln -s /stage1/lib/ld-linux.so.2 /lib
is cleaner btw,
but whichever way we do it we still have to find a way to support other platforms linkers (ld.so.1
and the like)
comment:5 by , 21 years ago
bug_file_loc: | http://archive.linuxfromscratch.org/mail-archives/lfs-dev/2003/05/0285.html → http://archives.linuxfromscratch.org/mail-archives/lfs-dev/2003-May/034151.html |
---|
comment:6 by , 21 years ago
Resolution: | → invalid |
---|---|
Status: | new → closed |
comment:7 by , 21 years ago
Why is this marked invalid? It just needs to be tested to confirm that it works (it does according to Zack and me).
comment:8 by , 21 years ago
I'm defering to Ryan and Greg's judgements in this one. If they highly recommend leaving things the way they are, then I'm ok leaving things the way they are.
It can be reopened if Greg and/or Ryan want to look into it as well.
comment:9 by , 21 years ago
I finally got around to looking into this.
Whilst it does work as suggested (with the symlink - fails miserably without) I am still strongly against including it in the book. IMHO, it detracts from the flow, unnecessarily complicates the build method, and is likely to confuse the hell out of the very readers we are trying to educate.
Folks who need Ch 5 on a CD could certainly utilise the method, but their needs should not burden the majority of readers who will not need it. Someone who cares should write a mini-hint.
comment:10 by , 21 years ago
I fail to understand how it complicates the build instructions. Instead of patching the gcc specs file and the ldscripts after Ch 6 glibc, the method I sugggested does it before. Creating the symlink is not complicated to understand since it is done for some other stuff.
Rather than detracting from the flow, it enhances it since there is a clear demarkation and non-dependency between pre-chroot and post-chroot parts (the way it is built currently, the binutils-build directory also needs to be available after chroot, hence Ch 5 needs to be built under $LFS/).
I suggested this approach coz it was easier for creating my build scripts. The reason I closed it was the lack of enthusiasm.
Let's leave this for the near future when things have settled down a little bit.