headers
John Wyatt | 1 Nov 2003 08:01
Picon
Favicon

Glibc configure/build problems (stage 1, very vague)

Basically in the first build of glibc in chap 5
configure  or build headers(or what Glibc uses) does
not seem to copy(or declare) the proper files so they
can be compiled, this is in the make or the make check
stage.

I list my build script so if there is any errors.
(motified one to fit the email of Yahoo)

# Glibc 2.3.2
# Needs linuxthreads to be compatible with most GNU
Linux distros
# Based on LFS, written by John Wyatt

echo "Glibc"
cp --copy-contents /opt/packages/glibc-2.3.2.tar.bz2
$LFS/sources/packages/glibc-2.3.2.tar.bz2
cp --copy-contents
/opt/packages/glibc-linuxthreads-2.3.2.tar.bz2
$LFS/sources/packages/glibc-linuxthreads-2.3.2.tar.bz2
cd $LFS/tools
tar --bzip2 --extract -f
$LFS/sources/packages/glibc-2.3.2.tar.bz2
mkdir /tools/etc
touch /tools/etc/ld.so.conf
mkdir glibc-build
cd $LFS/tools/glibc-2.3.2
patch -Np1 -i /opt/packages/glibc-2.3.2-sscanf-1.patch
cd $LFS/tools/glibc-build
tar --bzip2 --extract -f /
(Continue reading)

Permalink | Reply | 0 comments |
headers
John Wyatt | 1 Nov 2003 08:19
Picon
Favicon

Glibc configure/build problems (stage 1, very vague)

(forgot to include this)
te_IN.UTF-8.../bin/sh: line 1:
/tools/glibc-build/elf/ld.so: No such file or
directory
 done
tg_TJ.KOI8-T.../bin/sh: line 1:
/tools/glibc-build/elf/ld.so: No such file or
directory
 done
th_TH.TIS-620.../bin/sh: line 1:
/tools/glibc-build/elf/ld.so: No such file or
directory
 done
tl_PH.ISO-8859-1.../bin/sh: line 1:
/tools/glibc-build/elf/ld.so: No such file or
directory
 done
tr_TR.ISO-8859-9.../bin/sh: line 1:
/tools/glibc-build/elf/ld.so: No such file or
directory
 done
uk_UA.KOI8-U.../bin/sh: line 1:
/tools/glibc-build/elf/ld.so: No such file or
directory
 done
ur_PK.UTF-8.../bin/sh: line 1:
/tools/glibc-build/elf/ld.so: No such file or
directory
 done
uz_UZ.ISO-8859-1.../bin/sh: line 1:
(Continue reading)

Permalink | Reply | 0 comments |
headers
linux | 1 Nov 2003 08:20

depmod

hey all:

why am i getting this error when i run "depmod" or "depmod -a" or
depmod -a -F /usr/src/linux/System.map ?

----
bash-2.05a# depmod
depmod: *** Unresolved symbols in
/lib/modules/2.4.19/kernel/drivers/char/drm/gamma.o
depmod: *** Unresolved symbols in
/lib/modules/2.4.19/kernel/drivers/char/drm/sis.o
depmod: *** Unresolved symbols in
/lib/modules/2.4.19/kernel/drivers/usb/usb-uhci.o
bash-2.05a#
---

Any Solution for that?

Thanks in Adv
Paras.

--

-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
Permalink | Reply | 0 comments |
headers
James Richard Tyrer | 1 Nov 2003 09:02
Picon
Favicon

Re: Won't boot -- LILO appears to work OK.

Thanks for your ideas.

I checked everything again and it still wouldn't work so I built LILO 
again, and it boots fine.

--
JRT

--

-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
Permalink | Reply | 0 comments |
headers
Stephane Charette | 1 Nov 2003 09:09

Re: checking size of long double... configure: error: cannot compute sizeof (long double), 77

Greg Schafer suggested:
>Dunno for sure, but this is possibly caused by the binutils-pass1
>not being linked statically. Please check this by running:
>
>file /tools/bin/ld
>
>If it says dynamic, redo the binutilspass1 build properly then try
>glibc again.

No, it is static.  Here is the output:

-> /tools/bin/ld: ELF 32-bit LSB executable, Intel 80386, version 1
-> (SYSV), for GNU/Linux 2.2.5, statically linked, not stripped

Any other ideas on the cause of this problem:

On Thu, 30 Oct 2003 01:05:46 -0800, Stephane Charette wrote:

>>Back on August 18 2003, Ashley Rolleston reported the following:
>>
>>- chapter 6, LFS 4.1:
>>
>>	checking for long double... yes
>>	checking size of long double... configure: error: cannot compute sizeof (long double), 77
>>	See `config.log' for more details.
>>
>>Now I'm following the steps for LFS 5.0-pre3 (host is RH 9) and I've encountered the exact same error in
chapter 5 while trying to install glibc.  This is what I get:
>>
>>	../glibc-2.3.2/configure --prefix=/tools --disable-profile --enable-add-ons
(Continue reading)

Permalink | Reply | 0 comments |
headers
Kevin Boergens | 1 Nov 2003 13:32
Picon

Symlink on /mnt did the trick!

Hello!

This is not the completely right place to ask, but I would be happy if 
you helped me nevertheless. I don't have any installed system, so I 
started to build my own system with a knoppix live system from CD.

I partly used the instructions form lfs and partly the help from 
"building a minimal system from source code" from Greg o Keefe.
I compiled grup, linux, init, glibc, ncurses, util-linux and bash 
(nothing static), but when I tried to chroot to /mnt (There I mounted 
the new system), bash didn't find the needed libaries.

I got nearly crazy, until I made a symlink in the target system from mnt 
  to . and now everything works fine!

Now my question: Obviously the libary paths are messed up. Is there any 
possibility to fix this or have I to start again from scratch?

TIA,
	Kevin
-- 
http://www.boergens.de

--

-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
Permalink | Reply | 2 comments |
headers
Ken Moffat | 1 Nov 2003 14:51

stderr and scripting

I'm writing a new set of scripts to build my lfs's (looking much more
like the book, flagging each package to enable restarts, calculating
space and time, ...) and I'm having problems with stderr messages in the
logging.

 A typical line will be something like

make 2>&1 >>$LOG

 What I'm noticing, particularly within `make check' (gcc and glibc,
possibly also on others) is that warnings and errors are appearing on
the console.  I've just looked at a glibc failure (seems to be in the
math tests - this is a K6-2, so perhaps not unexpected) and realised
that the relevant errors and warnings are definitely _not_ in the log.

 I'm assuming I've just exposed my lack of scripting knowledge, so would
anybody care to apply a clue bat (gently, please) ?

Ken
-- 
Will code payroll for food.

--

-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
Permalink | Reply | 3 comments |
headers
Bill's LFS Login | 1 Nov 2003 15:20

Re: stderr and scripting

On Sat, 1 Nov 2003, Ken Moffat wrote:

> I'm writing a new set of scripts to build my lfs's (looking much more
> like the book, flagging each package to enable restarts, calculating
> space and time, ...) and I'm having problems with stderr messages in the
> logging.
>
>  A typical line will be something like
>
> make 2>&1 >>$LOG
>
>  What I'm noticing, particularly within `make check' (gcc and glibc,
> possibly also on others) is that warnings and errors are appearing on
> the console.  I've just looked at a glibc failure (seems to be in the
> math tests - this is a K6-2, so perhaps not unexpected) and realised
> that the relevant errors and warnings are definitely _not_ in the log.
>
>  I'm assuming I've just exposed my lack of scripting knowledge, so would
> anybody care to apply a clue bat (gently, please) ?

Just reverse the order oof the two redirections and you're good to go.
The 'man bash' states that order is significant, and it is, but fails to
also mention the exception, which is not involved woth your problem
here, but I thought I would mention it. If you are feeding a pipe, &2
has to be redirectoed to &1 *before*. Otherwise no-go. So

   cat xyz 2>&1 | less  # Bad example, but it demonstrates the point

is needed even though the man page indicates that this should not work
as it does.
(Continue reading)

Permalink | Reply | 2 comments |
headers
Ken Moffat | 1 Nov 2003 16:28

Re: stderr and scripting

On Sat, 1 Nov 2003, Bill's LFS Login wrote:

>
> Just reverse the order oof the two redirections and you're good to go.
> The 'man bash' states that order is significant, and it is, but fails to
> also mention the exception, which is not involved woth your problem
> here, but I thought I would mention it. If you are feeding a pipe, &2
> has to be redirectoed to &1 *before*. Otherwise no-go. So
>
>    cat xyz 2>&1 | less  # Bad example, but it demonstrates the point
>
> is needed even though the man page indicates that this should not work
> as it does.
>

Thanks, Bill.  I'll give it a go like that as soon as it gets to the
end of installing glibc.  Funnily enough, I did have it like that at one
time, but "corrected" it.  I guess my former way of doing things ( make
... 2>&1 | tee -a $LOG ) led me to overlook what the man page might say
on this.

Ken
-- 
Will code payroll for food.

--

-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
(Continue reading)

Permalink | Reply | 1 comment |
headers
Stephane Charette | 1 Nov 2003 17:59

Solution found! (was: Re: checking size of long double... configure: error: cannot compute sizeof (long double), 77)

With regards to this error:

>>>	checking size of long double... configure: error: cannot compute sizeof (long double), 77
>>>	See `config.log' for more details.
[...cut...]
>>>	/tools/bin/ld: warning: ld-linux.so.2, needed by /lib/libc.so.6, not found (try using -rpath or -rpath-link)

I started over again from the start of chapter 5.  In doing so, I found that I was using
gcc-core-3.3.2.tar.bz2 while LFS uses gcc-core-3.3.1.tar.bz2.

Everything now seems fine.  Without going back and trying it all over again, I'd guess that the problem was
caused by the use of GCC v3.3.2 instead of GCC v3.3.1.

Thanks,

Stéphane Charette

--

-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
Permalink | Reply | 501 comments |

Gmane