bug#21676: 25.0.50; crash in code_convert_string

[Eli Zaretskii <eliz@gnu.org>]

> From: Mark Oteiza <mvoteiza@udel.edu>
> Date: Mon, 12 Oct 2015 18:09:23 -0400
> 
> I have no idea why this happened or how to reproduce.
> Perhaps related to bug#21602? In any case, here's a backtrace.

No, it doesn't seem to be related to 21602.  And it's not even a crash
in code_convert_string, it's an abort during GC, due to some
corruption in memory allocation data:

> #9  0x0000000000556d9d in shut_down_emacs (sig=6, stuff=0) at emacs.c:2018
> No locals.
> #10 0x00000000005548a8 in terminate_due_to_signal (sig=6, backtrace_limit=40)
>     at emacs.c:382
> No locals.
> #11 0x000000000057a231 in handle_fatal_signal (sig=6) at sysdep.c:1593
> No locals.
> #12 0x000000000057a202 in deliver_thread_signal (sig=6, 
>     handler=0x57a217 <handle_fatal_signal>) at sysdep.c:1567
>         old_errno = 6
> #13 0x000000000057a268 in deliver_fatal_thread_signal (sig=6) at sysdep.c:1605
> No locals.
> #14 <signal handler called>
> No symbol table info available.
> #15 0x00007f20158995f8 in raise () from /usr/lib/libc.so.6
> No symbol table info available.
> #16 0x00007f201589aa7a in abort () from /usr/lib/libc.so.6
> No symbol table info available.
> #17 0x00007f20158d805a in __libc_message () from /usr/lib/libc.so.6
> No symbol table info available.
> #18 0x00007f20158dd9a6 in malloc_printerr () from /usr/lib/libc.so.6
> No symbol table info available.
> #19 0x00007f20158de18e in _int_free () from /usr/lib/libc.so.6
> No symbol table info available.
> #20 0x00000000005d028b in xfree (block=0x40b8da0) at alloc.c:765
> No locals.
> #21 0x000000000067cebe in ftcrfont_close (font=0x462fd30) at ftcrfont.c:184
>         ftcrfont_info = 0x462fd30
>         i = 2
> #22 0x00000000005d2926 in cleanup_vector (vector=0x462fd30) at alloc.c:2971
>         drv = 0xc98da0
> #23 0x00000000005d2a6e in sweep_vectors () at alloc.c:3023
>         total_bytes = 48
>         free_this_block = false
>         nbytes = 48
>         block = 0x462fd00
>         bprev = 0x46630f8
>         lv = 0x661008 <balance_intervals+31>
>         lvprev = 0xc636f0
>         vector = 0x462fd00
>         next = 0x462fd30
> #24 0x00000000005d8e14 in gc_sweep () at alloc.c:6708
> No locals.
> #25 0x00000000005d6b5e in garbage_collect_1 (end=0x7ffeedb5a2f0) at alloc.c:5536
>         nextb = 0x0
>         stack_top_variable = 0 '\000'
>         i = 582
>         message_p = false
>         count = 26
>         start = {
>           tv_sec = 1444684638, 
>           tv_nsec = 913808806
>         }
>         retval = 0
>         tot_before = 0
>         total = {140732886524664, 83643444, 0, 3, 13157648, 7, 0, 140732886524608, 
>           5572695, 2, 140732886524800}
> #26 0x00000000005d718e in Fgarbage_collect () at alloc.c:5720
>         end = 0x7ffeedb5a2f0
> #27 0x00000000005526ca in maybe_gc () at lisp.h:4517
> No locals.

Emacs then tried to auto-save the session, and segfaulted during that,
probably because doing that during (a failed) GC is almost certainly
going to crash.