bug#29063: C-g not quitting minibuffer

[Apocalypse Mystic]

timer-idle-list

working emacs:

([nil 0 0 500000 t jit-lock-context-fontify nil idle 0])

broken emacs:

([nil 0 0 500000 t jit-lock-context-fontify nil idle 0] [nil 0 0 500000 0.5 blink-cursor-start nil idle 0])

Although calling cancel-timer on the timers does not fix the problem, 

I have also noticed that, with toggle-debug-on-quit on, just typing C-g while in the normal writing buffer will always send me to the debugger on only one line:

Debugger entered--Lisp error: (quit)

  internal-timer-start-idle()

This happens in the broken emacs only, and C-g here does nothing in the working emacs. Likewise, in the broken emacs, while I am highlighting text with C-space, C-g will 50% of the time cancel the highlight and 50% of the time open the above debugger on the timer start function. In the working emacs, of course, it cancels the highlight 100% of the time.

On Sat, Nov 4, 2017 at 4:17 AM, Eli Zaretskii <eliz@gnu.org> wrote:

[Please keep the bug address on the CC list.]

> From: Apocalypse Mystic <apocalypsemystic@gmail.com>
> Date: Fri, 3 Nov 2017 21:53:28 -0400
>
> I just updated my other ubuntu system and it works fine, so I don't know that it will be easy to isolate that way.
> The one thing I do notice that is different between the two systems is that when I toggle-debug-on-quit, and
> then try to quit out of the M-x minibuffer, my broken emacs has one extra line in the debugger print out before
> everything else that is missing from the working emacs:
>
> Debugger entered--Lisp error: (quit)
> internal-timer-start-idle()
> ...
>
> I don't really know the emacs internals or timers that well, but might that mean something like that my C-g's
> are getting eaten by quitting out of a frequently restarting internal-timer-start-idle call? I'm not sure what calls it
> or if that helps any. It may just be normal functioning, but maybe it means something to someone who knows
> emacs better than I do. I notice it is in this description of the same issue:
>
> https://github.com/abo-abo/swiper/issues/922
>
> If I can provide any other information that might be helpful, just let me know.

I guess it would be interesting to know which idle timer was
interrupted by C-g in your case.