Segfault caused by derived-type with allocatable component in private clause

[Jerry DeLisle]

Steve Kargl wrote:
> All,
> 
> I just submitted a PR 38724 with the code following my sig.  This is
> derived (pun intended) from code posted by Gib Bogle in c.l.f.
> 
> http://groups.google.com/group/comp.lang.fortran/browse_frm/thread/fc36db02a922ff07#
> 
> pault replied in the clf thread that the code worked with gfortran.
> To the contrary, I've found that the code leads to segfault.  Can
> some others compile the code and run a test.
> 
> 
I also see the segfault. Can't get a useful back trace.

[Thread debugging using libthread_db enabled]
[New Thread 0x1e648c0 (LWP 8647)]
[New Thread 0x422be950 (LWP 8650)]
[New Thread 0x42cbf950 (LWP 8651)]
[New Thread 0x436c0950 (LWP 8652)]

Program received signal SIGSEGV, Segmentation fault.
0x00000000004491d5 in free ()
(gdb) bt
#0  0x00000000004491d5 in free ()
#1  0x0000000000400b1f in MAIN__.omp_fn.0 (.omp_data_i=0x0) at omptest.f90:24
#2  0x00000000004008d3 in test_omp () at omptest.f90:22
#3  0x0000000000400c4a in main (argc=1, argv=0x7fff7e774ab8)
     at ../../../gcc44/libgfortran/fmain.c:21
#4  0x00000000004201d2 in __libc_start_main ()
#5  0x00000000004001b9 in _start ()
(gdb)

Jerry