bug#1787: 23.0.60; `concat' with longer strings aborts Emacs ungracefully

[Markus Triska]

On OSX and Ubuntu jaunty machines with 1 GB RAM, doing:

    $ emacs -Q --eval "(let ((v \"x\")) (while t (setq v (concat v v))))"

aborts Emacs with the following backtrace:

   (gdb) bt
   #0  0x9010d3db in abort ()
   #1  0x0013a561 in make_uninit_multibyte_string (nchars=-536870912, 
nbytes=-536870912) at alloc.c:2503
   #2  0x0013a590 in make_uninit_string (length=-536870912) at alloc.c:2486
   #3  0x00159210 in concat (nargs=2, args=0xbfffec90, target_type=Lisp_String, 
last_special=0) at fns.c:619
   #4  0x00152c10 in Feval (form=41853805) at eval.c:2348
   #5  0x00152cbc in Fsetq (args=41853797) at eval.c:551
   #6  0x00152b58 in Feval (form=41853789) at eval.c:2322
   #7  0x00152e34 in Fprogn (args=41853837) at eval.c:449
   #8  0x00155199 in Fwhile (args=41853781) at eval.c:1111
   #9  0x00152b58 in Feval (form=41853773) at eval.c:2322
   #10 0x00152e34 in Fprogn (args=41853845) at eval.c:449
   #11 0x001552a5 in Flet (args=41853765) at eval.c:1089
   #12 0x00152b58 in Feval (form=41853733) at eval.c:2322
   #13 0x00153621 in Ffuncall (nargs=2, args=0xbffff0b0) at eval.c:3050
   #14 0x0018cb8b in Fbyte_code (bytestr=2175051, vector=2175068, maxdepth=10) 
at bytecode.c:678
   #15 0x00152f2b in funcall_lambda (fun=2175020, nargs=1, 
arg_vector=0xbffff254) at eval.c:3231
   #16 0x0015340e in Ffuncall (nargs=2, args=0xbffff250) at eval.c:3101
   #17 0x0018cb8b in Fbyte_code (bytestr=2161403, vector=2161420, maxdepth=7) 
at bytecode.c:678
   #18 0x00152f2b in funcall_lambda (fun=2161380, nargs=0, 
arg_vector=0xbffff3e4) at eval.c:3231
   #19 0x0015340e in Ffuncall (nargs=1, args=0xbffff3e0) at eval.c:3101
   #20 0x0018cb8b in Fbyte_code (bytestr=2158059, vector=2158076, maxdepth=6) 
at bytecode.c:678
   #21 0x00152f2b in funcall_lambda (fun=2158036, nargs=0, 
arg_vector=0xbffff500) at eval.c:3231
   #22 0x00153196 in apply_lambda (fun=2158036, args=58721289, eval_flag=1) at 
eval.c:3155
   #23 0x001528e0 in Feval (form=35544669) at eval.c:2435
   #24 0x000e3a9c in top_level_2 () at keyboard.c:1358
   #25 0x00151798 in internal_condition_case (bfun=0xe3a80 <top_level_2>, 
handlers=58760953, hfun=0xea6ef <cmd_error>) at eval.c:1511
   #26 0x000e3ae4 in top_level_1 () at keyboard.c:1366
   #27 0x001513ea in internal_catch (tag=58757025, func=0xe3aa2 <top_level_1>, 
arg=58721289) at eval.c:1247
   #28 0x000e37ff in command_loop () at keyboard.c:1321
   #29 0x000e38cf in recursive_edit_1 () at keyboard.c:942
   #30 0x000e3a17 in Frecursive_edit () at keyboard.c:1004
   #31 0x000e282f in main (argc=4, argv=0xbffff97c) at emacs.c:1786

   Lisp Backtrace:
   "concat" (0xbfffec90)
   "setq" (0xbfffede8)
   "while" (0xbfffeed8)
   "let" (0xbffff008)
   "eval" (0xbffff0b4)
   "command-line-1" (0xbffff254)
   "command-line" (0xbffff3e4)
   "normal-top-level" (0xbffff500)

To me, throwing an error seems to be much preferable in this case.

In GNU Emacs 23.0.60.1 (i386-apple-darwin8.11.1, GTK+ Version 2.12.9)
 of 2008-12-25 on mt-computer.local
Windowing system distributor `The XFree86 Project, Inc', version 11.0.40400000
Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: en_GB.UTF-8
  value of $XMODIFIERS: nil
  locale-coding-system: nil
  default-enable-multibyte-characters: t