GNU bug report logs -
#26525
`sit-for' in Flyspell slows typing down, hogs 90% of CPU
Previous Next
To reply to this bug, email your comments to 26525 AT debbugs.gnu.org.
Toggle the display of automated, internal messages from the tracker.
Report forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Sat, 15 Apr 2017 20:29:02 GMT)
Full text and
rfc822 format available.
Acknowledgement sent
to
Clément Pit-Claudel <cpitclaudel <at> gmail.com>:
New bug report received and forwarded. Copy sent to
bug-gnu-emacs <at> gnu.org.
(Sat, 15 Apr 2017 20:29:03 GMT)
Full text and
rfc822 format available.
Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
Hi bug-gnu-emacs,
I have attached a profile in which Flyspell's flyspell-check-word-p consumes 89% of CPU time while typing text in a buffer. This seems to be due to a call to sit-for in flyspell-check-word-p. What is this sit-for there for?
- flyspell-post-command-hook 6077 89%
- flyspell-check-word-p 6073 89%
- sit-for 6067 89%
- read-event 3826 56%
+ redisplay_internal (C function) 128 1%
+ timer-event-handler 91 1%
+ fstar-subp-filter 3 0%
- redisplay 2205 32%
+ redisplay_internal (C function) 100 1%
+ flyspell-get-not-casechars 2 0%
+ flyspell-check-pre-word-p 4 0%
+ command-execute 220 3%
+ ... 131 1%
+ timer-event-handler 116 1%
+ redisplay_internal (C function) 114 1%
+ company-post-command 102 1%
+ flycheck-perform-deferred-syntax-check 5 0%
+ eldoc-pre-command-refresh-echo-area 3 0%
sml/generate-position-help 1 0%
I haven't spent much time investigating this issue yet, and my current repro isn't exactly reliable, but I was hoping that someone might be able to provide insight into Flyspell's behaviour before I dive to deep :) I did verify that disabling Flyspell fixed the issue.
Thanks!
Clément.
[flyspell.prof (text/plain, attachment)]
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Sat, 15 Apr 2017 20:35:01 GMT)
Full text and
rfc822 format available.
Message #8 received at 26525 <at> debbugs.gnu.org (full text, mbox):
The relevant piece of code seems to be this:
(cond
((get this-command 'flyspell-deplacement)
(not (eq flyspell-previous-command this-command)))
((get this-command 'flyspell-delayed)
;; The current command is not delayed, that
;; is that we must check the word now.
(and (not unread-command-events)
(sit-for flyspell-delay)))
(t t)))
But the condition and the comment seem at odds, and it doesn't say much about why it calls `sit-for`.
Clément.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Mon, 17 Apr 2017 07:14:02 GMT)
Full text and
rfc822 format available.
Message #11 received at 26525 <at> debbugs.gnu.org (full text, mbox):
> From: Clément Pit-Claudel <cpitclaudel <at> gmail.com>
> Date: Sat, 15 Apr 2017 16:33:56 -0400
>
> The relevant piece of code seems to be this:
>
> (cond
> ((get this-command 'flyspell-deplacement)
> (not (eq flyspell-previous-command this-command)))
> ((get this-command 'flyspell-delayed)
> ;; The current command is not delayed, that
> ;; is that we must check the word now.
> (and (not unread-command-events)
> (sit-for flyspell-delay)))
> (t t)))
>
> But the condition and the comment seem at odds, and it doesn't say much about why it calls `sit-for`.
I'm not quite sure what needs to be explained. I think the doc string
of flyspell-delay and flyspell-default-delayed-commands tell the whole
story; if something is unclear there, please say what is unclear.
If you want to see this feature in action, turn on Flyspell, then type
something like "ssss", and then wait _without_ typing any non-word
character. You will see that the mis-spelled word is marked only
after 3 sec, the default value of flyspell-delay. Then contrast that
with the same word with a space typed after it.
IOW, this is a feature designed to avoid annoying users who type
slowly or make short breaks after typing only part of a word.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Mon, 17 Apr 2017 13:30:03 GMT)
Full text and
rfc822 format available.
Message #14 received at 26525 <at> debbugs.gnu.org (full text, mbox):
On 2017-04-17 03:14, Eli Zaretskii wrote:
>> But the condition and the comment seem at odds, and it doesn't say much about why it calls `sit-for`.
>
> I'm not quite sure what needs to be explained. I think the doc string
> of flyspell-delay and flyspell-default-delayed-commands tell the whole
> story; if something is unclear there, please say what is unclear.
The code reads (cond ((get this-command 'flyspell-delayed) …)), but the comment says "The current command is not delayed". Does a non-nil flycheck-delayed really mean that the command is *not* delayed?
> If you want to see this feature in action, turn on Flyspell, then type
> something like "ssss", and then wait _without_ typing any non-word
> character. You will see that the mis-spelled word is marked only
> after 3 sec, the default value of flyspell-delay. Then contrast that
> with the same word with a space typed after it.
Thanks! I don't expect that this feature is intended to slow down typing, though, right? In the example I posted it makes typing painfully slow.
Clément.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Mon, 17 Apr 2017 14:26:02 GMT)
Full text and
rfc822 format available.
Message #17 received at 26525 <at> debbugs.gnu.org (full text, mbox):
> Cc: 26525 <at> debbugs.gnu.org
> From: Clément Pit-Claudel <cpitclaudel <at> gmail.com>
> Date: Mon, 17 Apr 2017 09:29:20 -0400
>
> On 2017-04-17 03:14, Eli Zaretskii wrote:
> >> But the condition and the comment seem at odds, and it doesn't say much about why it calls `sit-for`.
> >
> > I'm not quite sure what needs to be explained. I think the doc string
> > of flyspell-delay and flyspell-default-delayed-commands tell the whole
> > story; if something is unclear there, please say what is unclear.
>
> The code reads (cond ((get this-command 'flyspell-delayed) …)), but the comment says "The current command is not delayed". Does a non-nil flycheck-delayed really mean that the command is *not* delayed?
I think the comment should be moved 2 lines below its current place,
that's all.
> > If you want to see this feature in action, turn on Flyspell, then type
> > something like "ssss", and then wait _without_ typing any non-word
> > character. You will see that the mis-spelled word is marked only
> > after 3 sec, the default value of flyspell-delay. Then contrast that
> > with the same word with a space typed after it.
>
> Thanks! I don't expect that this feature is intended to slow down typing, though, right? In the example I posted it makes typing painfully slow.
I didn't see any example in your OP, only a profile. What did I miss?
And no, this is not supposed to slow down typing in any significant
way, as long as you type, because sit-for is supposed to return as
soon as input is available. Your complaint was about CPU usage, not
about time delays, AFAIU.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Mon, 17 Apr 2017 15:37:01 GMT)
Full text and
rfc822 format available.
Message #20 received at 26525 <at> debbugs.gnu.org (full text, mbox):
On 2017-04-17 10:25, Eli Zaretskii wrote:
>> Thanks! I don't expect that this feature is intended to slow down typing, though, right? In the example I posted it makes typing painfully slow.
> I didn't see any example in your OP, only a profile. What did I miss?
>
> And no, this is not supposed to slow down typing in any significant
> way, as long as you type, because sit-for is supposed to return as
> soon as input is available. Your complaint was about CPU usage, not
> about time delays, AFAIU.
Indeed, there was only a profile; I don't have a good repro yet. In that case, flyspell was slowing things down a lot, and the complaint was indeed about delays.
Sorry for being unclear.
Clément.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Tue, 17 Oct 2017 12:09:01 GMT)
Full text and
rfc822 format available.
Message #23 received at 26525 <at> debbugs.gnu.org (full text, mbox):
> I have attached a profile in which Flyspell's flyspell-check-word-p
> consumes 89% of CPU time while typing text in a buffer. […]
This happened to me in the past then stopped for months. Now it happens
again and is really annoying. Emacs regularly freezes for around 5
seconds when I manipulate text.
--
Damien Cassou
http://damiencassou.seasidehosting.st
"Success is the ability to go from one failure to another without
losing enthusiasm." --Winston Churchill
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Thu, 13 Aug 2020 01:42:01 GMT)
Full text and
rfc822 format available.
Message #26 received at 26525 <at> debbugs.gnu.org (full text, mbox):
Clément Pit-Claudel <cpitclaudel <at> gmail.com> writes:
> On 2017-04-17 10:25, Eli Zaretskii wrote:
>>> Thanks! I don't expect that this feature is intended to slow down typing, though, right? In the example I posted it makes typing painfully slow.
>> I didn't see any example in your OP, only a profile. What did I miss?
>>
>> And no, this is not supposed to slow down typing in any significant
>> way, as long as you type, because sit-for is supposed to return as
>> soon as input is available. Your complaint was about CPU usage, not
>> about time delays, AFAIU.
>
> Indeed, there was only a profile; I don't have a good repro yet. In that case, flyspell was slowing things down a lot, and the complaint was indeed about delays.
> Sorry for being unclear.
That was 3 years ago. Are you still seeing this on a recent version of
Emacs, such as the recently released 27.1?
Best regards,
Stefan Kangas
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Fri, 28 Aug 2020 11:50:01 GMT)
Full text and
rfc822 format available.
Message #29 received at 26525 <at> debbugs.gnu.org (full text, mbox):
Damien Cassou <damien <at> cassou.me> writes:
>> I have attached a profile in which Flyspell's flyspell-check-word-p
>> consumes 89% of CPU time while typing text in a buffer. […]
>
> This happened to me in the past then stopped for months. Now it happens
> again and is really annoying. Emacs regularly freezes for around 5
> seconds when I manipulate text.
That was 3 years ago. Are you still seeing this on a recent version of
Emacs, such as the recently released 27.1?
Best regards,
Stefan Kangas
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Fri, 28 Aug 2020 12:37:02 GMT)
Full text and
rfc822 format available.
Message #32 received at 26525 <at> debbugs.gnu.org (full text, mbox):
Stefan Kangas <stefan <at> marxist.se> writes:
> That was 3 years ago. Are you still seeing this on a recent version of
> Emacs, such as the recently released 27.1?
This hasn't happened for some time.
--
Damien Cassou
"Success is the ability to go from one failure to another without
losing enthusiasm." --Winston Churchill
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Fri, 28 Aug 2020 18:41:02 GMT)
Full text and
rfc822 format available.
Message #35 received at 26525 <at> debbugs.gnu.org (full text, mbox):
On 8/28/20 8:35 AM, Damien Cassou wrote:
> Stefan Kangas <stefan <at> marxist.se> writes:
>> That was 3 years ago. Are you still seeing this on a recent version of
>> Emacs, such as the recently released 27.1?
>
> This hasn't happened for some time.
>
There was a report of a similar issue on reddit recently, but it hasn't happened to me of late.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Sat, 29 Aug 2020 16:28:02 GMT)
Full text and
rfc822 format available.
Message #38 received at 26525 <at> debbugs.gnu.org (full text, mbox):
tags 26525 + unreproducible
thanks
Clément Pit-Claudel <cpitclaudel <at> gmail.com> writes:
> On 8/28/20 8:35 AM, Damien Cassou wrote:
>> Stefan Kangas <stefan <at> marxist.se> writes:
>>> That was 3 years ago. Are you still seeing this on a recent version of
>>> Emacs, such as the recently released 27.1?
>>
>> This hasn't happened for some time.
>>
>
> There was a report of a similar issue on reddit recently, but it hasn't happened to me of late.
You wouldn't happen to have a link to the relevant thread? I tried
searching the web but couldn't find anything.
I'm marking this bug as unreproducible for now. Please report back if
you see this again or can provide any more information.
Added tag(s) unreproducible.
Request was from
Stefan Kangas <stefan <at> marxist.se>
to
control <at> debbugs.gnu.org.
(Sat, 29 Aug 2020 16:28:02 GMT)
Full text and
rfc822 format available.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Sat, 29 Aug 2020 18:11:01 GMT)
Full text and
rfc822 format available.
Message #43 received at 26525 <at> debbugs.gnu.org (full text, mbox):
On 8/29/20 12:27 PM, Stefan Kangas wrote:
> You wouldn't happen to have a link to the relevant thread? I tried
> searching the web but couldn't find anything.
https://www.reddit.com/r/emacs/comments/i96lbo/flyspell_consuming_most_of_my_resources_why_oo/ , sorry for not including it the first time.
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Wed, 02 Sep 2020 23:17:02 GMT)
Full text and
rfc822 format available.
Message #46 received at 26525 <at> debbugs.gnu.org (full text, mbox):
> I'm marking this bug as unreproducible for now. Please report back
> if you see this again or can provide any more information.
I just bumped into this bug on 27.1 with the same symptoms - typing lag
after activating flyspell-mode.
I took a couple of profiles - the first when lag was apparent and
another after clearing the lag by toggling flyspell-mode.
Both profiles show a similar result, so the "high cpu" in sit-for might
be a red-herring.
Profile 1 - with lag
- flyspell-post-command-hook 1146 92%
- flyspell-check-word-p 1120 90%
- sit-for 1117 90%
- redisplay 551 44%
+ redisplay_internal (C function) 1 0%
read-event 390 31%
- flyspell-get-not-casechars 1 0%
ispell-get-decoded-string 1 0%
+ flyspell-word 8 0%
+ flyspell-check-pre-word-p 3 0%
+ command-execute 65 5%
+ ... 22 1%
+ undo-auto--add-boundary 1 0%
Profile 2 - without lag
- flyspell-post-command-hook 3507 94%
- flyspell-check-word-p 3432 92%
- sit-for 3414 92%
- redisplay 1660 44%
+ redisplay_internal (C function) 3 0%
+ read-event 1317 35%
+ ispell-get-otherchars 6 0%
+ flyspell-get-not-casechars 5 0%
+ flyspell-word 17 0%
+ flyspell-check-pre-word-p 13 0%
+ command-execute 93 2%
+ ... 87 2%
+ timer-event-handler 3 0%
undo-auto--add-boundary 2 0%
tooltip-hide 1 0%
Nick
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Thu, 03 Sep 2020 10:20:01 GMT)
Full text and
rfc822 format available.
Message #49 received at 26525 <at> debbugs.gnu.org (full text, mbox):
Nick H <nick <at> tenpoint.co.nz> writes:
> I just bumped into this bug on 27.1 with the same symptoms - typing lag
> after activating flyspell-mode.
What is your ispell/flyspell configuration?
Do you have any ideas about how to reproduce it, or could you say
something about the circumstances under which you saw this?
Information forwarded
to
bug-gnu-emacs <at> gnu.org:
bug#26525; Package
emacs.
(Fri, 04 Sep 2020 01:02:02 GMT)
Full text and
rfc822 format available.
Message #52 received at 26525 <at> debbugs.gnu.org (full text, mbox):
On Thu, 2020-09-03 at 03:19 -0700, Stefan Kangas wrote:
> Nick H <nick <at> tenpoint.co.nz> writes:
>
> > I just bumped into this bug on 27.1 with the same symptoms - typing
> > lag after activating flyspell-mode.
>
> What is your ispell/flyspell configuration?
Nothing unusual:
grep -C2 -E \(ispell\|flyspell\) ~/.config/emacs/init.el
; spelling
(setq ispell-choices-win-default-height 2)
(setq ispell-dictionary "en_NZ")
(setq ispell-personal-dictionary (concat nh-etc "hunspell-dict"))
(setq ispell-program-name "hunspell")
(setq ispell-dictionary-alist
'(("en_NZ" "[[:alpha:]]" "[^[:alpha:]]" "[']" t nil t utf-8)))
(with-eval-after-load 'flyspell
(progn
(define-key
flyspell-mouse-map [down-mouse-3] #'flyspell-correct-word)
(define-key
flyspell-mouse-map [mouse-3] nil)))
> Do you have any ideas about how to reproduce it, or could you say
> something about the circumstances under which you saw this?
The lead-up was pretty mundane - I visited an elisp file, set the
region to a block of text, M-x write-region, C-x C-f the resulting
file, M-x vi-l-m, a few edits, M-x flys-buf, M-x flys-mod --> lag
I tried to reproduce the problem, but no success. Sorry.
Removed tag(s) moreinfo.
Request was from
Lars Ingebrigtsen <larsi <at> gnus.org>
to
control <at> debbugs.gnu.org.
(Mon, 30 Nov 2020 12:46:02 GMT)
Full text and
rfc822 format available.
This bug report was last modified 4 years and 200 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.