Why does my system still want gcc 3.4.9?
 

I have been battling this weirdness for several months, and it has
been getting worse and worse. Now I can't even unpack half the man
pages.

The short of it is that a lot of binaries on my system are linked
against gcc 3.4.9, even if I remerge them from scratch. It happened
with gcc 4.2.2, I emerged 4.2.3 and it still happens. The most common
broken binary is /usr/bin/lzma, which I have remerged several times
since the 4.2.3 emerge to no avail.

# ldd /usr/bin/lzma
/usr/bin/lzma: /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/libstdc++.so.6: version `GLIBCXX_3.4.9' not found (required by /usr/bin/lzma)
linux-gate.so.1 => (0xffffe000)
libstdc++.so.6 => /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/libstdc++.so.6 (0xb7e3a000)
libm.so.6 => /lib/libm.so.6 (0xb7e15000)
libgcc_s.so.1 => /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/libgcc_s.so.1 (0xb7e09000)
libc.so.6 => /lib/libc.so.6 (0xb7cd4000)
/lib/ld-linux.so.2 (0xb7f63000)

gcc-config is happy:

# gcc-config -l
[1] i686-pc-linux-gnu-4.1.2
[2] i686-pc-linux-gnu-4.2.3 *

The bad symlinks for gcj components (reported by revdep-rebuild) have
been fixed.

There is an old gcc hanging around, which I have been tempted to move
out of the way and see what happens, but I don't like broken
unbootable systems.

$ ll /usr/lib/gcc/i386-pc-linux-gnu/
total 0
drwxr-xr-x 2 root root 80 Jun 12 2006 3.4.4

with the other two versions elsewhere:

$ ll /usr/lib/gcc/i686-pc-linux-gnu/
total 4
drwxr-xr-x 4 root root 2464 Dec 27 15:07 4.1.2
drwxr-xr-x 5 root root 2504 May 12 11:58 4.2.3

This is a ~x86 system.

--
... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
GPG = E987 4493 C860 246C 3B1E 6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
--
gentoo-user@lists.gentoo.org mailing list

Why does my system still want gcc 3.4.9?
 

Am Mittwoch, 14. Mai 2008 schrieb ext felix@crowfix.com:

> The short of it is that a lot of binaries on my system are linked
> against gcc 3.4.9, even if I remerge them from scratch.

There is no such thing as gcc 3.4.9.

> It happened
> with gcc 4.2.2, I emerged 4.2.3 and it still happens. The most common
> broken binary is /usr/bin/lzma, which I have remerged several times
> since the 4.2.3 emerge to no avail.
>
> # ldd /usr/bin/lzma
> /usr/bin/lzma: /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/libstdc++.so.6:
> version `GLIBCXX_3.4.9' not found (required by /usr/bin/lzma)

You emerged it with 4.2.3, but _at_runtime_ it's trying to get libstdc++
from 4.1.2. Check your environment, especially all variables with PATH in
their name (env|grep PATH), as well as /etc/ld.so.conf.

Try: LD_LIBRARY_PATH=/usr/lib/gcc/i686-pc-linux-gnu/4.2.3 ldd /usr/bin/lzma

If this works, try running "env-update && source /etc/profile", then
just "ldd /usr/bin/lzma" again.

> gcc-config is happy:

It doesn't matter how happ gcc-config is. Your runtime linker is not.

> There is an old gcc hanging around, which I have been tempted to move
> out of the way and see what happens, but I don't like broken
> unbootable systems.

Rename it and run revdep-rebuild. If everything is fine you can remove it.

HTH...

Dirk
--
Dirk Heinrichs | Tel: +49 (0)162 234 3408
Configuration Manager | Fax: +49 (0)211 47068 111
Capgemini Deutschland | Mail: dirk.heinrichs@capgemini.com
Wanheimerstraße 68 | Web: http://www.capgemini.com
D-40468 Düsseldorf | ICQ#: 110037733
GPG Public Key C2E467BB | Keyserver: wwwkeys.pgp.net

Why does my system still want gcc 3.4.9?
 

On Mittwoch, 14. Mai 2008, felix@crowfix.com wrote:
> I have been battling this weirdness for several months, and it has
> been getting worse and worse. Now I can't even unpack half the man
> pages.
>
> The short of it is that a lot of binaries on my system are linked
> against gcc 3.4.9, even if I remerge them from scratch. It happened
> with gcc 4.2.2, I emerged 4.2.3 and it still happens. The most common
> broken binary is /usr/bin/lzma, which I have remerged several times
> since the 4.2.3 emerge to no avail.
>
> # ldd /usr/bin/lzma
> /usr/bin/lzma: /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/libstdc++.so.6: version
> `GLIBCXX_3.4.9' not found (required by /usr/bin/lzma) linux-gate.so.1 =>
> (0xffffe000)
> libstdc++.so.6 =>
> /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/libstdc++.so.6 (0xb7e3a000) libm.so.6
> => /lib/libm.so.6 (0xb7e15000)
> libgcc_s.so.1 => /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/libgcc_s.so.1
> (0xb7e09000) libc.so.6 => /lib/libc.so.6 (0xb7cd4000)
> /lib/ld-linux.so.2 (0xb7f63000)
>
> gcc-config is happy:
>
> # gcc-config -l
> [1] i686-pc-linux-gnu-4.1.2
> [2] i686-pc-linux-gnu-4.2.3 *
>
> The bad symlinks for gcj components (reported by revdep-rebuild) have
> been fixed.
>
> There is an old gcc hanging around, which I have been tempted to move
> out of the way and see what happens, but I don't like broken
> unbootable systems.
>
> $ ll /usr/lib/gcc/i386-pc-linux-gnu/
> total 0
> drwxr-xr-x 2 root root 80 Jun 12 2006 3.4.4
>
> with the other two versions elsewhere:
>
> $ ll /usr/lib/gcc/i686-pc-linux-gnu/
> total 4
> drwxr-xr-x 4 root root 2464 Dec 27 15:07 4.1.2
> drwxr-xr-x 5 root root 2504 May 12 11:58 4.2.3
>
> This is a ~x86 system.

there are/were probably a bunch of *la files left which have that stale crap
in them. Grep for it. Then remove them.

--
gentoo-user@lists.gentoo.org mailing list

Why does my system still want gcc 3.4.9?
 

On Wed, May 14, 2008 at 08:19:38AM +0200, Dirk Heinrichs wrote:

> There is no such thing as gcc 3.4.9.
>
> You emerged it with 4.2.3, but _at_runtime_ it's trying to get libstdc++
> from 4.1.2. Check your environment, especially all variables with PATH in
> their name (env|grep PATH), as well as /etc/ld.so.conf.

I see the difference. I had thought if it was looking for 3.4.9, it
must have been compiled with 3.4.9. But I just edited (read only!)
libstdc++.so... and it has the string 3.4.9 in it, presumably to
satisfy multiple versions.

> Try: LD_LIBRARY_PATH=/usr/lib/gcc/i686-pc-linux-gnu/4.2.3 ldd /usr/bin/lzma

This works.

> If this works, try running "env-update && source /etc/profile", then
> just "ldd /usr/bin/lzma" again.

Unfortunately, this doesn't work. But at least I am now pointed in
the right direction of being a runtime problem solved (temporarily) by
LD_LIBRARY_PATH. Now I need to get me some sleep and get a fresh start.

--
... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
GPG = E987 4493 C860 246C 3B1E 6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
--
gentoo-user@lists.gentoo.org mailing list

Why does my system still want gcc 3.4.9?
 

Am Mittwoch, 14. Mai 2008 schrieb felix@crowfix.com:

> > Try: LD_LIBRARY_PATH=/usr/lib/gcc/i686-pc-linux-gnu/4.2.3 ldd
> > /usr/bin/lzma
>
> This works.

Good.

> > If this works, try running "env-update && source /etc/profile", then
> > just "ldd /usr/bin/lzma" again.
>
> Unfortunately, this doesn't work.

Did you check /etc/ld.so.conf, maybe 4.1.2 is still listed before 4.2.3?

You may also need to clean up /etc/env.d a bit and running gcc-config again
afterwards also seems to be a good idea.

HTH...

Dirk

Why does my system still want gcc 3.4.9?
 

On Wed, May 14, 2008 at 06:40:24PM +0200, Dirk Heinrichs wrote:

> Did you check /etc/ld.so.conf, maybe 4.1.2 is still listed before 4.2.3?
>
> You may also need to clean up /etc/env.d a bit and running gcc-config again
> afterwards also seems to be a good idea.

Thanks. I started looking at it last night, but it was a warm night
after a long day and I left it for this morning. 4.1.2 is in
ld.so.conf, and in two env.d files:

# ls -l `grep -l 4.1.2 05*`
-rw-r--r-- 1 root root 243 Feb 16 2007 05compiler
-rw-r--r-- 1 root root 293 May 12 01:58 05gcc-i686-pc-linux-gnu

The 05compiler file is so old that I suspect it is some kind of
flotsam and needs to be deleted. The 05gcc file only includes 4.1.2
on the LDPATH line, but after 4.2.3, and its MANPATH, INFOPATH, PATH,
and ROOTPATH entries are all 4.2.3 only. "equery belongs" doesn't
know about either one.

So I moved 05compiler out of the way (I want to preserve that time
stamp just in case I do need to restore it), ran env-update again, and
now lzma is happy. I can run man again! Things which failed emerge
now build -- 323 to go.

I wonder what lessons I have learned? I misled myself into thinking
it was a compile problem because I didn't realize one lib could handle
multiple versions. I probably didn't follow post merge instructions
somewhen and that started the bitrot. I didn't take the hint when
remerging lzma several times made no difference.

Thanks again. I hope I don't need to ask for more help :-)

--
... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
GPG = E987 4493 C860 246C 3B1E 6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
--
gentoo-user@lists.gentoo.org mailing list

Why does my system still want gcc 3.4.9?
 

Am Mittwoch, 14. Mai 2008 schrieb ext felix@crowfix.com:

> So I moved 05compiler out of the way (I want to preserve that time
> stamp just in case I do need to restore it), ran env-update again, and
> now lzma is happy. I can run man again! Things which failed emerge
> now build -- 323 to go.

Fine.

Bye...

Dirk
--
Dirk Heinrichs | Tel: +49 (0)162 234 3408
Configuration Manager | Fax: +49 (0)211 47068 111
Capgemini Deutschland | Mail: dirk.heinrichs@capgemini.com
Wanheimerstraße 68 | Web: http://www.capgemini.com
D-40468 Düsseldorf | ICQ#: 110037733
GPG Public Key C2E467BB | Keyserver: wwwkeys.pgp.net

Why does my system still want gcc 3.4.9?
 

Another way (simple) to resolve the issue is to re-emerge lzma-utils with "nocxx" flag

Sylvain Chouleur

> Date: Wed, 14 May 2008 11:28:10 -0700
> From: felix@crowfix.com
> To: gentoo-user@lists.gentoo.org
> Subject: Re: [gentoo-user] Why does my system still want gcc 3.4.9?
>
> On Wed, May 14, 2008 at 06:40:24PM +0200, Dirk Heinrichs wrote:
>
> > Did you check /etc/ld.so.conf, maybe 4.1.2 is still listed before 4.2.3?
> >
> > You may also need to clean up /etc/env.d a bit and running gcc-config again
> > afterwards also seems to be a good idea.
>
> Thanks. I started looking at it last night, but it was a warm night
> after a long day and I left it for this morning. 4.1.2 is in
> ld.so.conf, and in two env.d files:
>
> # ls -l `grep -l 4.1.2 05*`
> -rw-r--r-- 1 root root 243 Feb 16 2007 05compiler
> -rw-r--r-- 1 root root 293 May 12 01:58 05gcc-i686-pc-linux-gnu
>
> The 05compiler file is so old that I suspect it is some kind of
> flotsam and needs to be deleted. The 05gcc file only includes 4.1.2
> on the LDPATH line, but after 4.2.3, and its MANPATH, INFOPATH, PATH,
> and ROOTPATH entries are all 4.2.3 only. "equery belongs" doesn't
> know about either one.
>
> So I moved 05compiler out of the way (I want to preserve that time
> stamp just in case I do need to restore it), ran env-update again, and
> now lzma is happy. I can run man again! Things which failed emerge
> now build -- 323 to go.
>
> I wonder what lessons I have learned? I misled myself into thinking
> it was a compile problem because I didn't realize one lib could handle
> multiple versions. I probably didn't follow post merge instructions
> somewhen and that started the bitrot. I didn't take the hint when
> remerging lzma several times made no difference.
>
> Thanks again. I hope I don't need to ask for more help :-)
>
> --
> ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
> Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
> GPG = E987 4493 C860 246C 3B1E 6477 7838 76E9 182E 8151 ITAR license #4933
> I've found a solution to Fermat's Last Theorem but I see I've run out of room o
> --
> gentoo-user@lists.gentoo.org mailing list
>

Tous vos amis discutent sur Messenger, et vous ? Téléchargez Messenger, c'est gratuit !

Why does my system still want gcc 3.4.9?
 

Am Donnerstag, 15. Mai 2008 schrieb ext Sylvain Chouleur:
> Another way (simple) to resolve the issue is to re-emerge lzma-utils with
> "nocxx" flag

No. It's just a workaround that works with certain packages, but fails for
others (where C++ is not optional). The solution is to fix the broken
environment.

Bye...

Dirk
--
Dirk Heinrichs | Tel: +49 (0)162 234 3408
Configuration Manager | Fax: +49 (0)211 47068 111
Capgemini Deutschland | Mail: dirk.heinrichs@capgemini.com
Wanheimerstraße 68 | Web: http://www.capgemini.com
D-40468 Düsseldorf | ICQ#: 110037733
GPG Public Key C2E467BB | Keyserver: wwwkeys.pgp.net

Why does my system still want gcc 3.4.9?
 

On Thu, May 15, 2008 at 10:31:09AM +0200, Dirk Heinrichs wrote:
> Am Donnerstag, 15. Mai 2008 schrieb ext Sylvain Chouleur:
> > Another way (simple) to resolve the issue is to re-emerge lzma-utils with
> > "nocxx" flag
>
> No. It's just a workaround that works with certain packages, but fails for
> others (where C++ is not optional). The solution is to fix the broken
> environment.

Absolutely. Many other packages were failing to merge too. This is
the right way to fix it.

--
... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
GPG = E987 4493 C860 246C 3B1E 6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o
--
gentoo-user@lists.gentoo.org mailing list