bug#66096: High CPU usage, basically runaway emacs with visit to bidi_cache_search on and off?

["ISHIKAWA,chiaki" <ishikawa@yk.rim.or.jp>]

On 2023/09/20 22:24, Eli Zaretskii wrote:
>> Date: Wed, 20 Sep 2023 16:09:03 +0900
>> Cc: 66096@debbugs.gnu.org, "ishikawa, chiaki" <ishikawa@yk.rim.or.jp>
>> From: "ISHIKAWA,chiaki" <ishikawa@yk.rim.or.jp>
>>
>> I have the following code snippet to monitor GC issues on my PC for
>> quite sometime.
>> But I have not seen this particular problem before.
>> The long pause I have seen before was strictly in GC-related routines.
>>
>> (setq my-gc-statistics (make-vector 30 nil))
>>
>> ;;; The element is
>> ;;; (append (memory-use-counts) (list gc-elapsed gcs-done))
>> ;;; Each time the following function is called, the
>> ;;; elements in the array is shifted toward the end.
>> ;;; Use (message "%S" my-gc-statistics) to force the
>> ;;; recording of my-gc-statistics value in *Messages* buffer for later
>> analysis.
>>
>>
>> (defun update-my-gc-statistics ()
>>     (let ((i 28))
>>       (progn
>>        ;;; very unlike Lisp
>>        (while (<= 0 i)
>>          (progn (aset my-gc-statistics (+ 1 i) (aref my-gc-statistics i))
>>                (setq i (- i 1) )))
>>        (aset my-gc-statistics 0
>>              (append (memory-use-counts) (list gc-elapsed gcs-done)))
>>        ;;; print the latest one last so that I can see the glimpse in the
>> narrow
>>        ;;; output window.
>>        (message "%S\n%S" (current-time-string) (pp (reverse
>> my-gc-statistics))))))
>>
>> (setq post-gc-hook 'update-my-gc-statistics)
>>
>> For example, in one place, I see
>> #0  0x0000556f1905be9e in string_char_and_length
>>       (length=<synthetic pointer>, p=0x556f20d8a3d8 " 70337 417360
>> 2.323839139 50)\n (99058573 3875 59193441 34949 11084445 70337 417401
>> 2.391973008 51)]\n\"") at
>> /home/ishikawa/repos/emacs-29.1/src/character.h:375
>>
>> And that string is probably   from (append (memory-use-counts) (list
>> gc-elapsed gcs-done))
> I tried to run with this, and I don't see any such infloop.  I see the
> GC results displayed in the echo-area once in a while.
>
> I guess some other customizations cause this, or maybe some special
> circumstances with different fonts you use.  Customizations that
> affect the echo-area display and resizing are of main interest.  If
> you can try to reproduce this in "emacs -Q", with only the
> post-gc-hook defined and as little of your other customizations as
> possible, it could help.
>
>> Also I see recursive_edit at the bottom of the stakctrace.
>> Not sure why. Maybe I was doing some error recovery of Japanese input?
> No, this is normal: Emacs enters one level of recursive-edit when it
> starts.

Thank you for the comment.

Yeah, I noticed strange font-related functions in a stacktrace.
In any case, if I see the issue with a file being edited,
I will report it.
I have re-compiled emacs with the flags suggested in etc/DEBUG for now.


Chiaki