bug#39977: 28.0.50; Unhelpful stack trace

[martin rudalics]

>> As far as xdisp.c is concerned it simply must not set selected_frame
>> to a dead frame.
>
> I don't think that's possible in xdisp.c cases you've shown.
>
>> Never ever.
>
> Why not?

Because it might shift the abort to the next instance of SELECTED_FRAME.

>> As far as frame.c is concerned, it should do something like in the
>> attached patch.
>
> We cannot punt like that in the display engine.

Why not?  At least one of the frame restorations is unprotected anyway
and might leave the temporarily selected frame selected.

> So you are saying that selecting such a frame will cause trouble to
> some other code, not to the display engine?

Not "will" but "may".  The problem is that it then might be harder
to find the cause.

With emacs -Q evaluate

(defvar foo
  '(:eval
    (when (> (length (frame-list)) 1)
      (delete-frame (next-frame)))))

(setq-default mode-line-format foo)

and do C-x 5 2.  The backtrace I get here is


#0  0x000000000063f7a3 in terminate_due_to_signal (sig=6, backtrace_limit=40) 
at ../../src/emacs.c:371
#1  0x000000000068ac8a in emacs_abort () at ../../src/sysdep.c:2448
#2  0x00000000004ee088 in select_window (window=XIL(0x1be5745), 
norecord=XIL(0x30), inhibit_point_swap=false) at ../../src/window.c:544
#3  0x00000000004ee2f9 in Fselect_window (window=XIL(0x1be5745), 
norecord=XIL(0x30)) at ../../src/window.c:630
#4  0x0000000000484c33 in gui_consider_frame_title (frame=XIL(0x1be5505)) at 
../../src/xdisp.c:12318
#5  0x00000000004974b6 in redisplay_window (window=XIL(0x1be5745), 
just_this_one_p=false) at ../../src/xdisp.c:18940
#6  0x000000000048cb00 in redisplay_window_0 (window=XIL(0x1be5745)) at 
../../src/xdisp.c:16179
#7  0x00000000007b10dd in internal_condition_case_1 (bfun=0x48cabe 
<redisplay_window_0>, arg=XIL(0x1be5745), handlers=XIL(0x7ffff40bafbb), hfun=0x48ca86 
<redisplay_window_error>) at ../../src/eval.c:1379
#8  0x000000000048ca58 in redisplay_windows (window=XIL(0x1be5745)) at 
../../src/xdisp.c:16159
#9  0x000000000048b486 in redisplay_internal () at ../../src/xdisp.c:15627
#10 0x0000000000489084 in redisplay () at ../../src/xdisp.c:14854
#11 0x0000000000650828 in read_char (commandflag=1, map=XIL(0x17ce993), 
prev_event=XIL(0), used_mouse_menu=0x7fffffffe13f, end_time=0x0) at 
../../src/keyboard.c:2493
#12 0x0000000000663705 in read_key_sequence (keybuf=0x7fffffffe2d0, 
prompt=XIL(0), dont_downcase_last=false, can_return_switch_frame=true, 
fix_current_buffer=true, prevent_redisplay=false) at ../../src/keyboard.c:9549
#13 0x000000000064ccee in command_loop_1 () at ../../src/keyboard.c:1350
#14 0x00000000007b1002 in internal_condition_case (bfun=0x64c872 <command_loop_1>, 
handlers=XIL(0x90), hfun=0x64be81 <cmd_error>) at ../../src/eval.c:1355
#15 0x000000000064c457 in command_loop_2 (ignore=XIL(0)) at 
../../src/keyboard.c:1091
#16 0x00000000007b04b6 in internal_catch (tag=XIL(0xd0e0), func=0x64c42a 
<command_loop_2>, arg=XIL(0)) at ../../src/eval.c:1116
#17 0x000000000064c3f5 in command_loop () at ../../src/keyboard.c:1070
#18 0x000000000064b968 in recursive_edit_1 () at ../../src/keyboard.c:714
#19 0x000000000064bb60 in Frecursive_edit () at ../../src/keyboard.c:786
#20 0x0000000000641f98 in main (argc=2, argv=0x7fffffffe7c8) at 
../../src/emacs.c:2035


which is almost the same as Madhu's.  So maybe the display engine should
simply set a global boolean inhibit_frame_changes while evaluating mode
lines or frame titles and have at least delete_frame not delete a frame
when that variable is set.  If we decide that fixing such a problem is
urgent.

martin