Re: [lfs-support] Force i386

[Rob]

William Harrington  wrote:
> If we get a break in real world prioritized responsibilities, then
> the book may get some excellent updates. Until then, we
> are working at a frivolous pace.

Absolutely. I did not mean that as a criticism at all.
In fact, I intend looking at the CLFS roadmap and seeing if I can
apply it on newer versions. Not on a production box, of course,
but I'll build from the book first and then see if I can update it.
Thank you all for the excellent work.
-- 
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

Re: [lfs-support] Force i386

[William Harrington]

On Mon, 05 Sep 2016 06:02:20 -0500
"Rob"  wrote:

> Well it's using kernel 3.14. We're at kernel 4.7.2. I am not sure about the 
> security risks of running such old stuff. The copyright date is 2014. That's 
> what I meant by a few years behind.

Both LFS and CLFS are a guide. Since we aren't bleeding edge with distros 
doesn't mean the instructions won't result the required position. We have 
plenty of tickets to keep us busy at http://trac.clfs.org. The fact that the 
book is at a 3.14 kernel means nothing. The stable book is the latest release. 
The great new features are in the dev book. Follow that and review the tickets 
for the upcoming updates to the dev book for the future release. If we get a 
break in real world prioritized responsibilities, then the book may get some 
excellent updates. Until then, we are working at a frivolous pace.

Sincerely,
William Harrington
-- 
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

Re: [lfs-support] Force i386

[Pierre Labastie]

On 05/09/2016 13:02, Rob wrote:

Ken Moffat  wrote:

Building LFS for the first time is hard. Many people have trouble.
And multilib is harder still (or perhaps just "more tedious", but
"harder" is probably a better starting assumption).

I wouldn't say harder but tedious definitely. There aren't a lot of 
instructions out there, either. The stuff I found says things like, enable 
multilib repository in xx distribution; or edit your apt sources and install 
these packages. Nothing about enabling a multilib toolchain from source. I'm 
beginning to think I'll have to read C programming books to figure everything 
out.


Cross-lfs may well be using older packages than LFS - they have even
fewer developers and builders than we do - but to suggest they are
"a few years behind" is an overstatement.

Well it's using kernel 3.14. We're at kernel 4.7.2. I am not sure about the 
security risks of running such old stuff. The copyright date is 2014. That's 
what I meant by a few years behind.



You should use development:
http://clfs.org/files/BOOK/CLFS-SYSTEMD-TRUNK-GIT-20160729-CHUNKS.tar.bz2 
(for systemd)

or
http://clfs.org/files/BOOK/CLFS-SYSVINIT-TRUNK-SYSVINIT-20160729-CHUNKS.tar.bz2 
(for SYS V)


or you can render the books yourself from the git repo:
-
cd
git clone git://git.clfs.org/cross-lfs.git
cd cross-lfs/BOOK
# Next line only for SYS V
git checkout sysvinit
make
--
Point your browser to file://path>/cross-lfs/render/index.html



Pierre


--
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

Re: [lfs-support] Force i386

[akhiezer]

> From: "Rob" <captinlo...@gmail.com>
> Date: Mon, 05 Sep 2016 06:02:20 -0500
> Subject: Re: [lfs-support] Force i386
>
> Ken Moffat <zarniwh...@ntlworld.com> wrote:
> > Building LFS for the first time is hard. Many people have trouble.
> > And multilib is harder still (or perhaps just "more tedious", but
> > "harder" is probably a better starting assumption).
>
> I wouldn't say harder but tedious definitely. There aren't a lot of 
> instructions out there, either. The stuff I found says things like, enable 
> multilib repository in xx distribution; or edit your apt sources and install 
> these packages. Nothing about enabling a multilib toolchain from source. I'm 
> beginning to think I'll have to read C programming books to figure everything 
> out.
>


Slackware's multilib acknowledges 'Slamd64' and 'CLFS' as the two
principal references used:
http://alien.slackbook.org/dokuwiki/doku.php?id=slackware:multilib  .


> > Cross-lfs may well be using older packages than LFS - they have even
> > fewer developers and builders than we do - but to suggest they are
> > "a few years behind" is an overstatement.  
>
> Well it's using kernel 3.14. We're at kernel 4.7.2. I am not sure about the 
> security risks of running such old stuff. The copyright date is 2014. That's 
> what I meant by a few years behind.
>


You can usually drop-in a newer kernel no-probs.


>
> > My suggestion to Rob is that he should build LFS, plus whichever
> > parts of BLFS are useful to him.  
>
> This is my third build of LFS/BLFS. The first two times, I built for i686. 
> This time, I wanted to build for x86_64, but still be able to run a couple of 
> i686 binaries and libes on there. Here is where I ran into problems, and 
> which is what prompted my questions.
> I eventually gave up trying to figure out the multilib/arch stuff, and 
> yesterday I just built LFS 7.10 on an i686 machine.


Another way is to run a 32-bit os as a vm.


> As I said earlier, I think I'll have to find a good C reference and try to 
> figure everything out. The problem I've found with that is a lot of C books 
> figure you already have everything installed already--all your toolchains and 
> stuff. So they don't talk about it.


Or just read gcc docs first.



rgds,
akh


p.s. pls wrap your lines to ca 74-80 chars' width.





--
-- 
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

Re: [lfs-support] Force i386

[Rob]

Ken Moffat  wrote:
> Building LFS for the first time is hard. Many people have trouble.
> And multilib is harder still (or perhaps just "more tedious", but
> "harder" is probably a better starting assumption).

I wouldn't say harder but tedious definitely. There aren't a lot of 
instructions out there, either. The stuff I found says things like, enable 
multilib repository in xx distribution; or edit your apt sources and install 
these packages. Nothing about enabling a multilib toolchain from source. I'm 
beginning to think I'll have to read C programming books to figure everything 
out.

> Cross-lfs may well be using older packages than LFS - they have even
> fewer developers and builders than we do - but to suggest they are
> "a few years behind" is an overstatement.  

Well it's using kernel 3.14. We're at kernel 4.7.2. I am not sure about the 
security risks of running such old stuff. The copyright date is 2014. That's 
what I meant by a few years behind.


> My suggestion to Rob is that he should build LFS, plus whichever
> parts of BLFS are useful to him.  

This is my third build of LFS/BLFS. The first two times, I built for i686. This 
time, I wanted to build for x86_64, but still be able to run a couple of i686 
binaries and libes on there. Here is where I ran into problems, and which is 
what prompted my questions.
I eventually gave up trying to figure out the multilib/arch stuff, and 
yesterday I just built LFS 7.10 on an i686 machine. As I said earlier, I think 
I'll have to find a good C reference and try to figure everything out. The 
problem I've found with that is a lot of C books figure you already have 
everything installed already--all your toolchains and stuff. So they don't talk 
about it.
-- 
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

Re: [lfs-support] Force i386

[Ken Moffat]

On Mon, Sep 05, 2016 at 01:45:45AM +0200, Tim Tassonis wrote:
> On 09/04/2016 10:12 AM, Rob wrote:
> > Thanos Baloukas  wrote:
> > http://lists.linuxfromscratch.org/pipermail/lfs-support/2016-September/050333.html
> > 
> > In that thread, you linked to a script collection, on github. The 
> > instructions in there did not work (not surprising), and asking google just 
> > linked to a bunch of pages about installing multiarch packages using 
> > various distro package managers. Not useful here.
> > So, I gave up on trying to build a multilib toolchain.
> > 
> 
> There is a working instruction from Armin to quite easily get a 32-bit glibc
> on a 64-bit lfs system.
> 
> As I only needed a glibc to get my binary 32-bit brother printer drivers
> working, that was enough for me.
> 
> If that is all you need, I can send you the instructions. If you however
> need more, I'm clueless as well.
> 
Unlike 10 years ago, people building from source (which is what LFS
is all about) only need multilib if they have to run 32-bit binary
packages.

I'm sure Armin's instructions work, but whether somebody who has
not yet successfully built LFS (plus whatever parts of BLFS appeal
to him) has the experience to follow them seems unlikely.

Building LFS for the first time is hard. Many people have trouble.
And multilib is harder still (or perhaps just "more tedious", but
"harder" is probably a better starting assumption).

Cross-lfs may well be using older packages than LFS - they have even
fewer developers and builders than we do - but to suggest they are
"a few years behind" is an overstatement.  In a cross-build it is
not safe to assume that new versions of packages will just drop in
without problems.  Even in LFS we sometimes hit problems with newer
versions.

My suggestion to Rob is that he should build LFS, plus whichever
parts of BLFS are useful to him.  Perhaps more than once.  the first
time, do it manually, typing in the commands. After that, he should
 work out what went wrong, and *how* he can automate (script) his
builds in a way that suits him.

Most people will then need to learn what happens when their own take
on scripting fails, but I will acknowledge that maybe he is more
clever than I am, and can grasp that immediately, so that only one
scripted LFS build is needed (I just wish, after using LFS for
perhaps 12 or 14 years, that I had developed the ability to spot my
own scripting errors before they break my builds).

And after that, he should be in a position to either follow Armin's
instructions for whatever 32-bit binary software he wants to use, or
to follow cross-lfs for a full bi-arch system.

ĸen
-- 
`I shall take my mountains', said Lu-Tze. `The climate will be good
for them.' -- Small Gods
-- 
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

Re: [lfs-support] Force i386

[Tim Tassonis]

On 09/04/2016 10:12 AM, Rob wrote:

Thanos Baloukas  wrote:
http://lists.linuxfromscratch.org/pipermail/lfs-support/2016-September/050333.html

In that thread, you linked to a script collection, on github. The instructions 
in there did not work (not surprising), and asking google just linked to a 
bunch of pages about installing multiarch packages using various distro package 
managers. Not useful here.
So, I gave up on trying to build a multilib toolchain.



There is a working instruction from Armin to quite easily get a 32-bit 
glibc on a 64-bit lfs system.


As I only needed a glibc to get my binary 32-bit brother printer drivers 
working, that was enough for me.


If that is all you need, I can send you the instructions. If you however 
need more, I'm clueless as well.



Cheers
Tim

--
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

Re: [lfs-support] Force i386

[Thanos Baloukas]

On 04/09/2016 08:14 πμ, Rob wrote:

I can't seem to get multilib GCC to compile properly. It stops haflway through 
compiling with
"checking dynamic linker characteristics... configure: error: Link tests are not 
allowed after GCC_NO_EXECUTABLES."
and won't go any further. Sounds like something went wrong with binutils, but 
I'm not sure how to fix.
So, is there any way to force i686 on a x86_64 system? Perhaps change $LFS_TGT 
to i386-linux-gnu?


You don't mention what changes to the instructions led to that error.
Did you read the replies to

http://lists.linuxfromscratch.org/pipermail/lfs-support/2016-September/050333.html

Multilib is not simple. It needs changes to the whole toolchain -
binutils, gcc, glibc. The changes may be configure options,
variables, patches etc.

--
Thanos
--
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

[lfs-support] Force i386

[Rob]

I can't seem to get multilib GCC to compile properly. It stops haflway through 
compiling with
"checking dynamic linker characteristics... configure: error: Link tests are 
not allowed after GCC_NO_EXECUTABLES."
and won't go any further. Sounds like something went wrong with binutils, but 
I'm not sure how to fix.
So, is there any way to force i686 on a x86_64 system? Perhaps change $LFS_TGT 
to i386-linux-gnu?
-- 
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