[Bug libfortran/64732] [5 Regression] julia build failure: double free or corruption in libgfortran.so.3

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64732

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kargl at gcc dot gnu.org

--- Comment #4 from kargl at gcc dot gnu.org ---
(In reply to Markus Trippelsdorf from comment #3)
> (In reply to ktkachov from comment #2)
> > (In reply to Richard Biener from comment #1)
> > > Waiting for some kind of a testcase (or at least a pointer to what Julia is).
> > 
> > Judging by the .jl suffix, I'd think it's this:
> > http://julialang.org/
> 
> Yes.
> 
> Here's another backtrace:
> 
> ======= Backtrace: =========
> /lib/libc.so.6(+0x73fbe)[0x7f3682997fbe]
> /lib/libc.so.6(+0x7973d)[0x7f368299d73d]
> /lib/libc.so.6(+0x79f4b)[0x7f368299df4b]
> /usr/lib/gcc/x86_64-pc-linux-gnu/5.0.0/libgfortran.so.
> 3(+0xe6aa6)[0x7f366f954aa6]
> /usr/lib/gcc/x86_64-pc-linux-gnu/5.0.0/libgfortran.so.
> 3(+0xe6bb2)[0x7f366f954bb2]
> /usr/lib/gcc/x86_64-pc-linux-gnu/5.0.0/libgfortran.so.
> 3(+0x1a7f9)[0x7f366f8887f9]
> /lib64/ld-linux-x86-64.so.2(+0xf927)[0x7f368399a927]
> /lib/libc.so.6(+0x3686a)[0x7f368295a86a]
> /lib/libc.so.6(+0x368b5)[0x7f368295a8b5]
> /lib/libc.so.6(__libc_start_main+0xf7)[0x7f3682943fd7]
> /var/tmp/julia/usr/bin/julia[0x4014d5]
> 
> addr2line shows:
> 
> libgfortran/io/unit.c:718 : 718   free_format_hash_table (u);
> libgfortran/io/unit.c:764
> libgfortran/runtime/main.c:292
> 
> Don't know how to come up with a small testcase yet.

Try building julia with -g and then using a debugger to
get an actual backtrace.  But, I suspect that julia isn't
properly initializing the gfortran runtime library and you
have conflicting IO subsystems (ie., julia's IO subsystem 
and gfortran's IO subsystem).