This was the second try to compile sys-libs/glibc-2.3.5-r1 for updating my
2-CPU system with a fresh 2005.1 installation. It always fails there:

!!! ERROR: sys-libs/glibc-2.3.5-r1 failed.
!!! Function toolchain-glibc_src_test, Line 250, Exitcode 2
!!! linuxthreads glibc did not pass make check

I wanted to compile with -build -erandom -glibc-compat20 -glibc-omitfp
+hardened* -linuxthreads-tls (-multilib) +nls +nptl* -nptlonly +pic*
-profile (-selinux) +userlocales*

So, before I let it run for another day I want to ask if +nptlonly would do
the job (No, I won't alter my FEATURES!) and what does it mean for my
system to run without the linuxthreads fallback.

emerge info:

Portage 2.0.51.22-r2 (default-linux/x86/2005.1, gcc-3.3.6,
glibc-2.3.4.20041102-r1, 2.6.12-gentoo-r10 i686)
================================================== ===============
System uname: 2.6.12-gentoo-r10 i686 Pentium II (Deschutes)
Gentoo Base System version 1.6.12
dev-lang/python: 2.3.5
sys-apps/sandbox: 1.2.11
sys-devel/autoconf: 2.13, 2.59-r6
sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.5
sys-devel/binutils: 2.15.92.0.2-r10
sys-devel/libtool: 1.5.18-r1
virtual/os-headers: 2.6.11-r2
ACCEPT_KEYWORDS="x86"
AUTOCLEAN="yes"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-O0 -march=pentium2 -pipe"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3/share/config /usr/share/config /var/bind /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-O0 -march=pentium2 -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig distlocks fixpackages maketest sandbox sfperms strict
test userpriv usersandbox"
GENTOO_MIRRORS="ftp://ftp-stud.fht-esslingen.de/pub/Mirrors/gentoo/
ftp://ftp.du.se/pub/os/gentoo ftp://pandemonium.tiscali.de/pub/gentoo/
ftp://gentoo.ITDNet.net/gentoo
ftp://mirrors.sec.informatik.tu-darmstadt.de/gentoo/"
LANG="de_DE@euro"
LC_ALL="de_DE@euro"
MAKEOPTS="-j3"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
SYNC="rsync://rsync.de.gentoo.org/gentoo-portage"
USE="x86 X509 acl apache2 apm avi bash-completion berkdb bitmap-fonts bzip2
chroot crypt cups eds emboss encode exif foomaticdb fortran gd gdbm gif
gstreamer gtk2 hardened idn imlib ipv6 ithreads jpeg libg++ libwww
logrotate mad memlimit mikmod motif mp3 mpeg ncurses nls no-old-linux nomac
nptl objc ogg oggvorbis opengl oss pam pcre pdflib perl pic png posix
python qt quicktime readline samba sasl slp spell ssl symlink tcpd threads
truetype truetype-fonts type1-fonts unicode vorbis xml2 xmms xv zlib
fritzcapi_cards_fcpci userland_GNU kernel_linux elibc_glibc"
Unset: ASFLAGS, CTARGET, LDFLAGS, LINGUAS, PORTDIR_OVERLAY
--
http://hommingberger-gepardenforelle.malkusch.de/
Bewahrt die Existenz der Hommingberger Gepardenforelle!

Markus Malkusch wrote:
> This was the second try to compile sys-libs/glibc-2.3.5-r1 for updating my
> 2-CPU system with a fresh 2005.1 installation. It always fails there:
>
> !!! ERROR: sys-libs/glibc-2.3.5-r1 failed.
> !!! Function toolchain-glibc_src_test, Line 250, Exitcode 2
> !!! linuxthreads glibc did not pass make check

It should work for x86 (32bits), but is broken for PPC and it seems like they
don't really know what breaks the nplt+linuxthreads build.

The bad news for all who needs linuxthreads (mostly people using old
precompiled stuff, I guess) is that the linuxthreads will be dropped in glibc
2.3.6 and you will have nptl only.

IMHO it's better to use the bugzilla for software "bugs".


//Aho

J.O. Aho:

> It should work for x86 (32bits), but is broken for PPC and it seems like
> they don't really know what breaks the nplt+linuxthreads build.

Just right now I tried with nptlonly and it failed also on the same test:

!!! Function toolchain-glibc_src_test, Line 256, Exitcode 2
!!! nptl glibc did not pass make check

So it's nothing about linuxthreads or NPTL. I read at
<http://www.tldp.org/LDP/lfs/html/chapter05/glibc.html>:

| The math tests sometimes fail when running on systems where the CPU is not
| a relatively new genuine Intel or authentic AMD. Certain optimization
| settings are also known to be a factor here.

So, is this the case for those old Pentium II (Deschutes)? How can I run the
tests without the math test? I think there would come some more tests which
I don't want to miss.
--
http://hommingberger-gepardenforelle.malkusch.de/
Bewahrt die Existenz der Hommingberger Gepardenforelle!

Markus Malkusch wrote:

> | The math tests sometimes fail when running on systems where the CPU is not
> | a relatively new genuine Intel or authentic AMD. Certain optimization
> | settings are also known to be a factor here.
>
> So, is this the case for those old Pentium II (Deschutes)? How can I run the
> tests without the math test? I think there would come some more tests which
> I don't want to miss.

Try to use no optimization at all on the glibc build and see if it will
successed (modify your /etc/make.conf).


//Aho

J.O. Aho:

> Try to use no optimization at all on the glibc build and see if it will
> successed (modify your /etc/make.conf).

I had as you could have seen in my first post used "-O0 -march=pentium2
-pipe", but I decided to ignore this and installed the glibc without
testing.
--
http://hommingberger-gepardenforelle.malkusch.de/
Bewahrt die Existenz der Hommingberger Gepardenforelle!