tar 1.13 segfaults reading remote tape device w/o -f option

[Michal Szymanski]

Hello,

I've found a strange behavior of Gnu tar 1.13 (used as provided for
RedHat Linux 7.2, versions 1.13-19, 1.13-25.

If I put a remote tape drive designation into TAPE variable, like:

> setenv TAPE transit:/dev/nst0

and then invoke:

> tar x

It fails immediately returning Segmentation Fault (core dumped)
The output of "gdb /bin/tar core" is given below.

But, quite strangely, if I use "-f" option, everything is working FINE:

> tar xf $TAPE      (reads the remote tape archive as expected).

I'm not sure whether this is related to 'tar' itself or somehow to the
system, as the same procedure using tar 1.13.17 on a Solaris machine
works fine, both with and without "-f $TAPE" option.

with best regards, Michal.

-- 
  Michal Szymanski (address@hidden)
  Warsaw University Observatory, Warszawa, POLAND
--

Core was generated by `tar x'.
Program terminated with signal 11, Segmentation fault.
Reading symbols from /lib/i686/libc.so.6...done.
Loaded symbols for /lib/i686/libc.so.6
Reading symbols from /lib/ld-linux.so.2...done.
Loaded symbols for /lib/ld-linux.so.2
#0  0x400b7d24 in __memchr () at __memchr:-1
-1      __memchr: No such file or directory.
        in __memchr
(gdb) where
#0  0x400b7d24 in __memchr () at __memchr:-1
#1  0xbffff374 in ?? ()
#2  0x08055b15 in strcpy () at strcpy:-1
#3  0x0805dac6 in strcpy () at strcpy:-1
#4  0x4004b657 in __libc_start_main (main=0x805d94c <strcpy+80480>,
argc=2, 
    ubp_av=0xbffff374, init=0x8049794, fini=0x80656a0 <__umoddi3+928>, 
    rtld_fini=0x4000dcd4 <_dl_fini>, stack_end=0xbffff36c)
    at ../sysdeps/generic/libc-start.c:129