Paul Rogers wrote:
Works for me:
$ tar --version | head -n1
tar (GNU tar) 1.29
$ unset LANG
$ ls -ld patch*
-rw-rw-r-- 1 bdubbs bdubbs 727704 Feb 25 14:29 patch-2.7.5.tar.xz
$ tar -xf patch-2.7.5.tar.xz
drwxrwxr-x 8 bdubbs bdubbs 4096 Mar 6 2015 patch-2.7.5
-rw-rw-r-- 1 bdubbs bdubbs 727704 Feb 25 14:29 patch-2.7.5.tar.xz
That's not in Chapter 5, but a full LFS-8.0 system, but that shouldn't
make any difference.
Given that I have not identified exactly what is causing my results...
Firstly, you're doing an extract, I'm doing a create (to preserve the
as-built bare, pristine FHS). Secondly, with your extract you're
letting tar discover the type of compression used and deal with it, not
specifying a compression option on the command line. Thirdly, as you
say, you're doing that in a full-up complete system, whereas I'm doing
it in the very constrained initial stage in Ch6, in the chroot jail,
with nothing but Ch5 to work with. Pardon my asking, but is this a fair
comparison?
The only thing we support is extracting in the early stages of Chapter 6.
The first place you need tar in Chapter 6 is 6.7. Linux-4.10.9 API
Headers. That works fine.
You do not have to specify the type of compression for tar when
extracting. It's been that way for several years.
Paul Rogers wrote:
I tried testing that "LANG" bug by exporting "LANG=C" before rerunning
the compressed tar, but it failed the same way. Perhaps it had to be
defined when tar was made? I didn't think that's what was implied. But
if I was going to have to rebuild tar, well, I decided to blow it all
away and start over from scratch, using tar-1.28, which caused no
problems with the LFS-7.7 host build. Did that and came back to my
first script in Ch6 to make and tarball (compressed) the bare, clean
FHS. No problems! I'm blaming tar-1.29. I stopped here for the time
being, I've a doctor's appointment to go off to. I'll try the rest of
CH6 this evening--should be good for a couple hours of grinding.
If the problem is tar-1.29, why does it work perfectly for all the
editors? There is nothing special in a disk utility like this that is
processor or system dependent.
Let me ask if any of the editors has tried to do as I have done? Do you
try to make a compressed tarball of the pristine directory structure at
the end of Ch6.5, before proceeding to the next step? Or do you just,
to borrow Ken's words, rush ahead with Ch6.6? What do you ever use tar
for in the initial stage of the chroot environment? Has anybody
actually tested it under those conditions?
Since you are doing something beyond the book, why don't you create the
tarball with the host system? There is really no need to do that inside
chroot.
My build scripts, dictated "by the book", make a reproducible
system--with intent. There was no difference between the first Ch5
build and the second but swapping out tar-1.29 for tar-1.28. The first
one failed, the second one ran successfully. Is it unreasonable to
sugggest there was some glitch introduced by that update?
As to a system dependency, at the point I had the failure I wasn't
running a functioning system! I was trying to build one in a
(deliberately) highly constrained chroot environment. Could the
tar-1.29 update have made some assumption that it has such things as a
functional locale with things like LANG defined? I don't know, but it
seems possible.
To mention the definition of "Update" from Boodle's "Devil's DP
Dictionary": to take out old bugs and install new ones.
LFS-8.0 has been out for about two months. If there was a problem,
certainly we would have heard about it before now.
So it would seem, but I am not questioning it works in a fully installed
and configured system. Apples to oranges.
OK. As Ken says, if you don't follow the book and something breaks, you
get to keep the pieces.
-- Bruce
--
http://lists.linuxfromscratch.org/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
Do not top post on this list.
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
http://en.wikipedia.org/wiki/Posting_style
