sys-libs/pam-1.0.1, 'undefined reference to `yywrap'

[Viridior]

[environment]
$ROOT: /usr/armv7a-softfloat-linux-gnueabi/
$SYSROOT: /usr/armv7a-softfloat-linux-gnueabi/
$PORTAGE_CONFIGROOT: /usr/armv7a-softfloat-linux-gnueabi/
$CBUILD: i686-pc-linux-gnu

[software]
sys-libs/pam-1.0.1

[command]
$xmerge pam

[error]
pam_conv_l.o: In function `yylex':
pam_conv_l.c:(.text+0x1654): undefined reference to `yywrap'

[discussion]
Discussed here... http://bugs.gentoo.org/show_bug.cgi?id=230231, fix should be to reinstall sys-devel/flex.

[Tscheggsches (tscheggsches)]

x86_64-pc-linux-gnu-gcc -march=nocona -O2 -pipe -D_GNU_SOURCE -W -Wall -Wbad-function-cast -Wcast-align -Wcast-qual -Wmissing-declarations -Wmissing-prototypes -Wpointer-arith -Wreturn-type -Wstrict-prototypes -Wwrite-strings -Winline -Wshadow -Wl,-O1 -Wl,--as-needed -Wl,--enable-new-dtags -Wl,--sort-common -Wl,--as-needed -Wl,-O1 -o pam_conv1 pam_conv_l.o pam_conv_y.o
pam_conv_l.o: In function `yylex':
pam_conv_l.c:(.text+0xb45): undefined reference to `yywrap'
collect2: ld returned 1 exit status
make[4]: *** [pam_conv1] Error 1
make[4]: Leaving directory `/var/tmp/portage/sys-libs/pam-1.0.1/work/Linux-PAM-1.0.1/conf/pam_conv1'
make[3]: *** [all] Error 2
make[3]: Leaving directory `/var/tmp/portage/sys-libs/pam-1.0.1/work/Linux-PAM-1.0.1/conf/pam_conv1'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory `/var/tmp/portage/sys-libs/pam-1.0.1/work/Linux-PAM-1.0.1/conf'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/var/tmp/portage/sys-libs/pam-1.0.1/work/Linux-PAM-1.0.1'
make: *** [all] Error 2
 *
 * ERROR: sys-libs/pam-1.0.1 failed.
 * Call stack:
 * ebuild.sh, line 49: Called src_compile
 * environment, line 3086: Called die
 * The specific snippet of code:
 * emake sepermitlockdir="/var/run/sepermit" || die "emake failed"
 * The die message:
 * emake failed

Reproducible: Always

Steps to Reproduce:
1.emerge -av --oneshot pam. Compile also failed on older versions

Expected Results:
Clean compile.

emerge --info
Portage 2.1.4.4 (default-linux/amd64/2007.0/desktop, gcc-4.2.4, glibc-2.6.1-r0, 2.6.25-gentoo-r5 x86_64)
=================================================================
System uname: 2.6.25-gentoo-r5 x86_64 Intel(R) Core(TM)2 Duo CPU T7500 @ 2.20GHz
Timestamp of tree: Mon, 30 Jun 2008 17:45:03 +0000
ccache version 2.4 [enabled]
app-shells/bash: 3.2_p33
dev-java/java-config: 1.3.7, 2.1.6
dev-lang/python: 2.4.4-r13
dev-python/pycrypto: 2.0.1-r6
dev-util/ccache: 2.4-r7
sys-apps/baselayout: 2.0.0
sys-apps/openrc: 0.2.5
sys-apps/sandbox: 1.2.18.1-r2
sys-devel/autoconf: 2.13, 2.61-r2
sys-devel/automake: 1.5, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10.1
sys-devel/binutils: 2.18-r1
sys-devel/gcc-config: 1.4.0-r4
sys-devel/libtool: 1.5.26
virtual/os-headers: 2.6.23-r3
ACCEPT_KEYWORDS="amd64"
CBUILD="x86_64-pc-linux-gnu"
CFLAGS="-march=nocona -O2 -pipe"
CHOST="x86_64-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/kde/4.0/env /usr/kde/4.0/share/config /usr/kde/4.0/shutdown /usr/share/config"
CONFIG_PROTECT_MASK="/etc/env.d /etc/env.d/java/ /etc/fonts/fonts.conf /etc/gconf /etc/gentoo-release /etc/revdep-rebuild /etc/terminfo /etc/texmf/web2c /etc/udev/rules.d"
CXXFLAGS="-march=nocona -O2 -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="ccache distlocks metadata-transfer parallel-fetch sandbox sfperms strict unmerge-orphans userfetch userpriv usersandbox"
GENTOO_MIRRORS="http://distfiles.gentoo.org http://distro.ibiblio.org/pub/linux/distributions/gentoo"
LDFLAGS="-Wl,-O1 -Wl,--as-needed -Wl,--enable-new-dtags -Wl,--sort-common"
LINGUAS="de"
MAKEOPTS="-j4"
PKGDIR="/usr/portage/packages"
PORTAGE_R...

[Tscheggsches (tscheggsches)]

Created attachment 158991
/var/tmp/portage/sys-libs/pam-1.0.1/temp/environment

[Kylesch (kylesch)]

I ran into this as well while testing 2008 beta 2.
re-emerging flex as described in Bug 229639 https://bugs.gentoo.org/show_bug.cgi?id=229639 fixed it for me.

[Diego Elio Pettenò (flameeyes)]

*** Bug 230545 has been marked as a duplicate of this bug. ***

[Tscheggsches (tscheggsches)]

(In reply to comment #2)
> I ran into this as well while testing 2008 beta 2.
> re-emerging flex as described in Bug 229639
> https://bugs.gentoo.org/show_bug.cgi?id=229639 fixed it for me.

Yes, re-emerging flex did it! And yes, I also used 2008 beta 1 or 2 for the initial install.

[Kevstar-31 (kevstar-31)]

(In reply to comment #4)
> (In reply to comment #2)
> > I ran into this as well while testing 2008 beta 2.
> > re-emerging flex as described in Bug 229639
> > https://bugs.gentoo.org/show_bug.cgi?id=229639 fixed it for me.
>
> Yes, re-emerging flex did it! And yes, I also used 2008 beta 1 or 2 for the
> initial install.
>

Is there a way to use this workaround in cataylst?

[Jeroen Roovers (jer-gentoo)]

*** Bug 229639 has been marked as a duplicate of this bug. ***

[Kevstar-31 (kevstar-31)]

is bug #222415 related?

[Anonymous bin Ich (ichbinanon)]

The bug still exists in 2008.0-r1

[Sgreadly (sgreadly)]

I can confirm too that this bug still exists on 2008.0-r1 amd64

[Jeroen Roovers (jer-gentoo)]

<email address hidden> : Don't CC unrelated teams, please.

[Kevstar-31 (kevstar-31)]

What if the latest version of flex is masked during bootstraping and pam depends on the latest version of flex.

[Kevstar-31 (kevstar-31)]

Could flex be divided into two separate packages (ie flex-dynamic flex-static)?

[kyron (kyron-neuralbs)]

I hit this bug attempting to build a stage3 (catalyst). 1.0.2 has the same issue.

[Jeff Huffman (tejing)]

I also hit this bug trying to build a i686 stage3 with catalyst. does anyone have a workaround that can be used with catalyst?

[Jacob Godserv (fun2program8)]

Let's keep the "Confirmed" status to when at least one other person, developer or user, has said "me too!"

Also, the assigned person is the one who should fix the bug. If you're sure you want to tackle this one, assign yourself again.

[Jacob Godserv (fun2program8)]

I'm going to mark this bug as Won't Fix until we decide whether or not we need to patch this.

[Gengor (gengor)]

*** Bug 262091 has been marked as a duplicate of this bug. ***

[Marat Radchenko (slonopotamus)]

Great. This means that catalyst is effectively unusable.

[Maksim Melnikau (maxposedon)]

I think I catch problem:

1. pam during compilation compiles statically conf/pam_conv1/pam_conv1 binary
(untility to convert pam.conf files to a pam.d/)

2. and this pam_conv1 don't goes to image and don't install (it is kind of useless?) so flex as rdep fixes compilation pam_conv1 but pam_conv1 won't be installed...

3. possible solutions:
a) put! flex as RDEP
b) create patch that disables compilation pam_conv1

[Gentoo-bugs-t (gentoo-bugs-t)]

(In reply to comment #17)
> Great. This means that catalyst is effectively unusable.

I've been using catalyst since last year to create current tarballs with no problem. I run it each and every night after I sync.

Perhaps you need to use a more recent stage3 seed stage than 2008 one?

You can download a newer stage3 tarball from your favorite mirror, for example for x86:
http://gentoo.mirror/experimental/x86/autobuilds/

[Marat Radchenko (slonopotamus)]

(In reply to comment #19)
> (In reply to comment #17)
> > Great. This means that catalyst is effectively unusable.
>
> I've been using catalyst since last year to create current tarballs with no
> problem. I run it each and every night after I sync.
>
> Perhaps you need to use a more recent stage3 seed stage than 2008 one?
>
> You can download a newer stage3 tarball from your favorite mirror, for example
> for x86:
> http://gentoo.mirror/experimental/x86/autobuilds/
>

I can't take any seed because I created my own for new arch (armv6j). Additional info: broken flex appeared during stage1 build and flex wasn't rebuilt for stage2.

[Gentoo-bugs-t (gentoo-bugs-t)]

(In reply to comment #20)
> (In reply to comment #19)
> > (In reply to comment #17)
> > > Great. This means that catalyst is effectively unusable.
> >
> > I've been using catalyst since last year to create current tarballs with no
> > problem. I run it each and every night after I sync.
> >
> > Perhaps you need to use a more recent stage3 seed stage than 2008 one?
> >
> > You can download a newer stage3 tarball from your favorite mirror, for example
> > for x86:
> > http://gentoo.mirror/experimental/x86/autobuilds/
> >
>
> I can't take any seed because I created my own for new arch (armv6j).
> Additional info: broken flex appeared during stage1 build and flex wasn't
> rebuilt for stage2.
>

why not chroot into the stage2 and update things by hand and then re-run catalyst as normal for the stage2?

[Diego Elio Pettenò (flameeyes)]

Really thsi is not something I'm prone to fixing pam-side..

[Maksim Melnikau (maxposedon)]

(In reply to comment #22)
> Really thsi is not something I'm prone to fixing pam-side..
>
Wasn't my comment #18 was clear enough problem description? If static palm_conv will be removed from build - problem fixed.

[Diego Elio Pettenò (flameeyes)]

I try to hack on pam's buildsystem on a strictly-needed basis: upstream is quite unfriendly, I'm not paid to do this job, I'm not even thanked (most time cursed instead) and I dislike the idea of keeping a branch of patches for Gentoo just because upstream “can't see what the problem is”.

[Denissia (denissia)]

I don't know why, but after I have made "emerge flex" error with installing pam has gone!
Thanks god, now I can use openssh and sudo.

[Maksim Melnikau (maxposedon)]

(In reply to comment #25)
> I don't know why, but after I have made "emerge flex" error with installing pam
If you interested why - read comment #18

and I still don't understand why bug `can'f fix...`

[Dennis Schridde (devurandom)]

I hit the same issue with pam-1.1.5 in a crosscompilation environment. Would be nice if you reconsidered creating a patch for this issue or do that flex RDEPEND thing.

[James Le Cuirot (chewi)]

I also just hit this while cross-compiling. I maintain my own set of cross-compile patches and scripts that I plan on making public soon but I just want to register my interest in seeing this fixed. This env snippet did the trick for me. Note that a second unneeded binary in doc/specs is now also affected.

pre_src_prepare() {
    sed -i "/SUBDIRS/s/\bpam_conv1\b//g" conf/Makefile.am || die
    sed -i "/SUBDIRS/s/\bspecs\b//g" doc/Makefile.am || die
}

[7-email (7-email)]

*** Bug 507668 has been marked as a duplicate of this bug. ***

[7-email (7-email)]

Still an issue with latest ~arch pam :( This is a blocker for my hardened-amd64-nomultilib catalyst build.

[Gienah (gienah)]

(In reply to Fernando (likewhoa) from comment #30)
> Still an issue with latest ~arch pam :( This is a blocker for my
> hardened-amd64-nomultilib catalyst build.

This builds for me:

sys-devel/flex-2.5.39-r1 USE="nls -static {-test}" ABI_X86="32%* (64%*) (-x32)"
sys-libs/pam-1.1.8-r2 USE="berkdb cracklib nls vim-syntax -audit -debug -nis (-selinux) {-test}" ABI_X86="32%* (64%*) (-x32)"

[FireBurn (fireburn)]

Incase anyone else has hit this problem too -- flex must be built without LTO on gcc 4.9

[David Heidelberg (okias)]

(In reply to Mike Lothian from comment #32)
> Incase anyone else has hit this problem too -- flex must be built without
> LTO on gcc 4.9

Thank you Mike, exactly this problem is present with flex compiled by GCC 4.9.1 with CFLAGS & CXXFLAGS "-flto".

After removing LTO, sys-libs/pam can be compiled without this issue.