glibc-2.0.95 causes ls -al & ps -aux to segfault.

[Byron Faber]

Ok, My fault, at least in that it didn't pass make check:

=========================
Starting process a
Starting process b
gcc -nostdlib -nostartfiles -o ex2  -Wl,-dynamic-linker=/lib/ld-linux.so.2 -g 
../csu/crt1.o ../csu/crti.o `gcc --print-file-name=crtbegin.o` ex2.o 
libpthread.so  -Wl,-rpath-link=..:../math:../elf:../nss:../nis:../db2:../rt:../
resolv:. ../libc.so.6 ../libc_nonshared.a -lgcc `gcc 
--print-file-name=crtend.o` ../csu/crtn.o
../elf/ld-linux.so.2 --library-path ..:../math:../elf:../nss:../nis:../db2:../r
t:../resolv:. ./ex2 > ex2.out
make[1]: *** [ex2.out] Error 139
make[1]: Leaving directory `/usr2/src/gnu/glibc-2.0.95/linuxthreads'
make: *** [linuxthreads/tests] Error 2
=========================

It dies in the linuxthreads check.  (wierd)


> > Was `make check' successful?  Do you consider this as a compiler or a
> > glibc2 bug?

I think its a glibc2 bug, given that it fails if I use gcc 2.8.1.


> > Can I you localize the problem with gdb/strace?

I'll give this a shot.


> > Is this a problem of ls in general or of `ls -l'?  In that case the
> > lookup of user/group from /etc/passwd might be somehow broken.  What
> > does /etc/nsswitch.conf look like?

Its attached.  [I haven't touched this, so unless glibc touches it
on install, its the standard redhat 5.0 file].


> How did you install glibc 2.0.95 on RedHat 5.1? It is kind of
> tricky. It took me a while to figure it out. Here is my
> 
> # ls -l /lib
> 

I was naive enough to get glibc-2.0.94 installed and working by doing
a 'make check;make install', and then recompiling a few binaries
that dropped dead because of the libg++ problems introduced by glibc 2.1.

I'll take a look at gdb and ls.

Thanks for your input,
Byron