bug#6604: 23.2; Crashes in GC on MS-Windows

[Lennart Borgman]

On Sat, Jul 10, 2010 at 5:52 PM, Eli Zaretskii <eliz@gnu.org> wrote:
> Emacs 23.2 started crashing on me a few days ago.  So far it crashed
> twice in a couple of days.  Both times the crash is inside GC, at the
> same place.  I attach one such backtrace below; it is not a GDB
> backtrace because I wasn't running under GDB at the time.
>
> I have no built a non-optimized binary and run it under GDB, so I
> could debug it when it crashes.  But if someone sees similar crashes,
> please post the data.

I do not know if this is related at all, but I just got a crash of a
type that I have never seen before. This is with my patched version of
Emacs, but I do not think I have any patches that are related to what
I see here.

If there is anything I can  investigate more please tell me. (I will
be back later today/tonight.)

Program received signal SIGSEGV, Segmentation fault.
0x77c46fa3 in msvcrt!memcpy () from C:\WINDOWS\system32\msvcrt.dll
(gdb) bt
#0  0x77c46fa3 in msvcrt!memcpy () from C:\WINDOWS\system32\msvcrt.dll
#1  0x011ea61f in face_at_buffer_position (w=0x4b85200, pos=222,
    region_beg=0, region_end=0, endptr=0x82f30c, limit=223, mouse=1,
    base_face_id=-1) at xfaces.c:6139
#2  0x0106702c in mouse_face_from_buffer_pos (window=79188485,
    dpyinfo=0x1405fc0, mouse_charpos=222, start_charpos=210, end_charpos=237,
    before_string=45475866, after_string=45475866, display_string=45475866)
    at xdisp.c:24067
#3  0x01068e80 in note_mouse_highlight (f=0x3ff7800, x=340, y=232)
    at xdisp.c:24900
#4  0x011dae66 in note_mouse_movement (frame=0x3ff7800, msg=0x82f590)
    at w32term.c:3055
#5  0x011dd613 in w32_read_socket (terminal=0x2d97900, expected=0,
    hold_quit=0x82f600) at w32term.c:4236
#6  0x0100e559 in read_avail_input (expected=0) at keyboard.c:6969
#7  0x0100e463 in gobble_input (expected=0) at keyboard.c:6887
#8  0x0100e416 in get_input_pending (addr=0x13fbb20, flags=1)
    at keyboard.c:6850
#9  0x01014af6 in detect_input_pending_run_timers (do_display=1)
    at keyboard.c:10515
#10 0x01088be4 in wait_reading_process_output (time_limit=30, microsecs=0,
    read_kbd=-1, do_display=1, wait_for_cell=45475866, wait_proc=0x0,
    just_wait_proc=0) at process.c:4935
#11 0x010c0670 in sit_for (timeout=120, reading=1, do_display=1)
    at dispnew.c:6067
#12 0x010084d1 in read_char (commandflag=1, nmaps=20, maps=0x82f9d0,
    prev_event=45475866, used_mouse_menu=0x82fb6c, end_time=0x0)
    at keyboard.c:2787
#13 0x01011d56 in read_key_sequence (keybuf=0x82fc60, bufsize=30,
    prompt=45475866, dont_downcase_last=0, can_return_switch_frame=1,
    fix_current_buffer=1) at keyboard.c:9329
#14 0x010066db in command_loop_1 () at keyboard.c:1618
#15 0x0102083f in internal_condition_case (bfun=0x10063b9 <command_loop_1>,
    handlers=45533426, hfun=0x1005daa <cmd_error>) at eval.c:1473
#16 0x0100611e in command_loop_2 (ignore=45475866) at keyboard.c:1341
#17 0x01020330 in internal_catch (tag=45531594,
    func=0x10060fb <command_loop_2>, arg=45475866) at eval.c:1217
#18 0x010060d4 in command_loop () at keyboard.c:1320
#19 0x010059c6 in recursive_edit_1 () at keyboard.c:943
#20 0x01005b2a in Frecursive_edit () at keyboard.c:1005
#21 0x01002872 in main (argc=1, argv=0xa94330) at emacs.c:1791
(gdb)



I have added a command to give information about sources which says:

This Emacs was built from sources in bazaar identified as:

Nick 'trunk' info:
  revision id: albinus@detlef-20100709192438-t85hy1rf4qo7fgnt
        revno: 100765
         date: 2010-07-09 21:24:38 +0200

Nick 'emacsw32' (from 'trunk') info:
  revision id: lennart.borgman@gmail.com-20100710131757-bm881dgw0t8rqnjf
        revno: 99331
         date: 2010-07-10 15:17:57 +0200