Bug 445293

Thanks for report, but I'm not able to reproduce the problem on my machine with
the coreutils package you have specified. Is the rpm -V coreutils silent(or
without unexpected things)?
Could you please provide strace of the piped tail? I mean `ls -l | strace tail`
. Thanks.

The following command produces no output (on my machine):
rpm -V coreutils

The output to `ls -l | strace tail` is as follows:

execve("/usr/bin/tail", ["tail"], [/* 45 vars */]) = 0
brk(0)                                  = 0x9a1f000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY)      = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=101232, ...}) = 0
mmap2(NULL, 101232, PROT_READ, MAP_PRIVATE, 3, 0) = 0xb7f1c000
close(3)                                = 0
open("/lib/libc.so.6", O_RDONLY)        = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\360tD\0004\0\0\0"...,
512) = 512
fstat64(3, {st_mode=S_IFREG|0755, st_size=1692524, ...}) = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0xb7f1b000
mmap2(0x431000, 1410608, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) =
0x431000
mmap2(0x584000, 12288, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x153) = 0x584000
mmap2(0x587000, 9776, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS,
-1, 0) = 0x587000
close(3)                                = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0xb7f1a000
set_thread_area({entry_number:-1 -> 6, base_addr:0xb7f1a6c0, limit:1048575,
seg_32bit:1, contents:0, read_exec_only:0, limit_in_pages:1, seg_not_present:0,
useable:1}) = 0
mprotect(0x584000, 8192, PROT_READ)     = 0
mprotect(0x42d000, 4096, PROT_READ)     = 0
munmap(0xb7f1c000, 101232)              = 0
brk(0)                                  = 0x9a1f000
brk(0x9a40000)                          = 0x9a40000
open("/usr/lib/locale/locale-archive", O_RDONLY|O_LARGEFILE) = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=77962480, ...}) = 0
mmap2(NULL, 2097152, PROT_READ, MAP_PRIVATE, 3, 0) = 0xb7d1a000
close(3)                                = 0
fstat64(0, {st_mode=S_IFIFO|0600, st_size=0, ...}) = 0
read(0, "total 4\n-rw-rw-r-- 1 mike mike 0"..., 8192) = 61
read(0, "", 8192)                       = 0
fstat64(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 10), ...}) = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0xb7f34000
write(1, "total 4\n", 8)                = 8
write(1, "-rw-rw-r-- 1 mike mike 0 2008-05"..., 53) = 53
close(0)                                = 0
close(1)                                = 0
munmap(0xb7f34000, 4096)                = 0
close(2)                                = 0
exit_group(0)                           = ?

Thanks for rpm -V info and strace output - but it shows correct output with "no
cannot follow `-' by name" error message. I need strace output from the error
case - as this one shows correct behaviour of the "ls -l | tail" command... This
command works without troubles for me, therefore without strace of the failure
or an 100% reproducer I can't do anything with that report.

The output that I provided of `ls -l | strace tail` was performed in the same
directory as was done without strace.  As far as I know, I did provide the
strace of the case were I am having the error message regarding "-".  

Perhaps there is nothing wrong with coreutilities -- but there may be another
culprit: SELinux.

I have SELinux running, but in permissive mode. I've noticed that even then, it
should simply report things it would otherwise block. However, sometimes it will
block things, especially when I perform a yum update.

I have little understanding of strace, and even less of SELinx.  Any advice?

Ok, just a little explanation of strace is just providing system calls - and
this could be used for analyze of error case. Strace you provided doesn't show
the problem - it shows just common scenario of the `ls -l | tail` - and I'm not
able to reproduce the problem on my machine - therefore I need strace of the
errorneus case - you will see "tail: cannot follow `-' by name" in the strace log.

This issue is IMHO not a problem of SELinux - as there is no problem with access... 

I could reproduce the issue if you have alias tail="tail -F" - because if you
use follow by name (which is not default), that behaviour is expected. Could you
please confirm that alias suspection?

As you suspected, I do have an alias for tail, namely

alias tail='tail -s5 -n99 -F'

I have since fixed that, and now `ls -l | tail` works correctly.

Thanks for the help!

Ok, closing NOTABUG.