unofficial mirror of help-gnu-emacs@gnu.org
 help / color / mirror / Atom feed
* typing slows down dramatically when typing digits in emails
@ 2022-12-21 13:53 Eric S Fraga
  2022-12-22  4:09 ` Stefan Monnier via Users list for the GNU Emacs text editor
  0 siblings, 1 reply; 7+ messages in thread
From: Eric S Fraga @ 2022-12-21 13:53 UTC (permalink / raw)
  To: help-gnu-emacs

Hello,

When composing an email in message mode in Emacs, completion works just
fine except when I type any digits (0-9) where there is an appreciable
lag.  Using the profiler, I get the following summary:

--8<---------------cut here---------------start------------->8---
        5482  82% - ...
        5311  79%  - completion--capf-wrapper
        5311  79%   - apply
        5311  79%    - corfu--capf-wrapper-advice
        5311  79%     - corfu--capf-wrapper
        5311  79%      - completion-try-completion
        5311  79%       - completion--nth-completion
        5311  79%        - completion--some
        5311  79%         - #<compiled -0x187b0d97893b0bbf>
        5311  79%          - orderless-try-completion
        5311  79%           - orderless-filter
        5311  79%            - orderless--prefix+pattern
        5311  79%             - completion-boundaries
        5311  79%              - #<compiled 0x1f4100e4f6213ad7>
        5311  79%               - complete-with-action
        5311  79%                - #<compiled 0xfe6876f217f45e4>
        4776  71%                 - cape--dabbrev-list
        4776  71%                  - dabbrev--find-all-expansions
        4704  70%                   - dabbrev--find-expansion
        4671  70%                    - dabbrev--try-find
        2959  44%                     - dabbrev--search
          36   0%                      - dabbrev--goto-start-of-abbrev
          28   0%                         #<compiled 0xae9961fbf742f>
           4   0%                        #<compiled 0xae9967326d42f>
           4   0%                      + internal--syntax-propertize
           9   0%                    + dabbrev--make-friend-buffer-list
--8<---------------cut here---------------end--------------->8---

It comes down to dabbrev... maybe.

Is there any setting that I might have missed that would improve this?
I am supposedly using company mode, by the way, I think (at least, I
have company indicated in my mode line) and completion candidates only
popup after 5 keystrokes yet lag takes place as soon as I type a single
digit.

Thank you,
eric

-- 
Eric S Fraga via gnus (Emacs 30.0.50 2022-12-21) on Debian 11.5




^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: typing slows down dramatically when typing digits in emails
  2022-12-21 13:53 typing slows down dramatically when typing digits in emails Eric S Fraga
@ 2022-12-22  4:09 ` Stefan Monnier via Users list for the GNU Emacs text editor
  2022-12-22  4:34   ` Emanuel Berg
  2022-12-22 10:22   ` typing slows down dramatically when typing digits in emails Eric S Fraga
  0 siblings, 2 replies; 7+ messages in thread
From: Stefan Monnier via Users list for the GNU Emacs text editor @ 2022-12-22  4:09 UTC (permalink / raw)
  To: help-gnu-emacs

>         5311  79%             - completion-boundaries
>         5311  79%              - #<compiled 0x1f4100e4f6213ad7>
>         5311  79%               - complete-with-action
>         5311  79%                - #<compiled 0xfe6876f217f45e4>
>         4776  71%                 - cape--dabbrev-list
>         4776  71%                  - dabbrev--find-all-expansions
>         4704  70%                   - dabbrev--find-expansion
>         4671  70%                    - dabbrev--try-find
>         2959  44%                     - dabbrev--search

`completion-boundaries` should not need to perform a `dabbrev--search`.
Which version of Emacs is that?
It would be useful to figure out what is this
#<compiled 0x1f4100e4f6213ad7> function (presumably a completion-table)
and how it calls `complete-with-action` because normally it should call
it with an argument of the form `(boundaries . N)` which should cause
`complete-with-action` to return nil immediately.


        Stefan




^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: typing slows down dramatically when typing digits in emails
  2022-12-22  4:09 ` Stefan Monnier via Users list for the GNU Emacs text editor
@ 2022-12-22  4:34   ` Emanuel Berg
  2022-12-22  4:45     ` always latest Emacs (was: Re: typing slows down dramatically when typing digits in emails) Emanuel Berg
  2022-12-22 10:22   ` typing slows down dramatically when typing digits in emails Eric S Fraga
  1 sibling, 1 reply; 7+ messages in thread
From: Emanuel Berg @ 2022-12-22  4:34 UTC (permalink / raw)
  To: help-gnu-emacs

Stefan Monnier via Users list for the GNU Emacs text editor wrote:

> Which version of Emacs is that?

We should have a Gnus posting style that inserts the Emacs
version last in each post that goes to a GNU Emacs list, and
that would be computed each time, so after that, to quote the
Captain of Titanic, "now nothing can go wrong" ... and we
would be saving time as well!

(info "(gnus) Posting Styles")

;;; -*- lexical-binding: t -*-
;;
;; this file:
;;   https://dataswamp.org/~incal/emacs-init/meta.el
;;
;; also see:
;;   https://dataswamp.org/~incal/emacs-init/w3m/w3m-version.el
;;
;; example:
;;   GNU Emacs 29.0.50 (build 1, x86_64-pc-linux-gnu, cairo version 1.16.0)
;;   of 2022-05-22 [commit 714970f5967f2153bb95e35823dbd917e0e5b60b]

(defun emacs-version-commit (&optional here)
  (interactive "P")
  (let ((ver (replace-regexp-in-string " \\(of\\)" "\\1"
               (format "%s [commit %s]"
                 (emacs-version)
                 (emacs-repository-get-version) ))))
    (when here
      (insert ver) )
    (message ver) ))

(defalias 'ever #'emacs-version-commit)

-- 
underground experts united
https://dataswamp.org/~incal




^ permalink raw reply	[flat|nested] 7+ messages in thread

* always latest Emacs (was: Re: typing slows down dramatically when typing digits in emails)
  2022-12-22  4:34   ` Emanuel Berg
@ 2022-12-22  4:45     ` Emanuel Berg
  0 siblings, 0 replies; 7+ messages in thread
From: Emanuel Berg @ 2022-12-22  4:45 UTC (permalink / raw)
  To: help-gnu-emacs

>> Which version of Emacs is that?
>
> We should have a Gnus posting style that inserts the Emacs
> version last ...

To automate even further, take a look at this for my
#2 favorite program, mpv.

Can you do something similar with Emacs?

Because this version thing is passé anyway ... there should be
only one version, the latest, which everyone should
have by definition.

#! /bin/zsh
#
# this file:
#   https://dataswamp.org/~incal/conf/.zsh/mpv-install
#
# commands:
#   https://fostips.com/install-mpv-media-player-debian/

mpv-install-setup () {
    sudo apt install apt-transport-https

    local key=uvt_key.gpg
    local url=https://non-gnu.uvt.nl/debian/$key
    curl $url --output $key
    sudo mv $key /etc/apt/trusted.gpg.d

    sudo sh -c 'echo "deb https://non-gnu.uvt.nl/debian $(lsb_release -sc) uvt" >> /etc/apt/sources.list.d/non-gnu-uvt.list'
    sudo apt update
}

mpv-install () {
    sudo apt install -t "o=UvT" mpv
}

-- 
underground experts united
https://dataswamp.org/~incal




^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: typing slows down dramatically when typing digits in emails
  2022-12-22  4:09 ` Stefan Monnier via Users list for the GNU Emacs text editor
  2022-12-22  4:34   ` Emanuel Berg
@ 2022-12-22 10:22   ` Eric S Fraga
  2022-12-22 21:29     ` cape and boundaries (was: typing slows down dramatically when typing digits in emails) Stefan Monnier via Users list for the GNU Emacs text editor
  1 sibling, 1 reply; 7+ messages in thread
From: Eric S Fraga @ 2022-12-22 10:22 UTC (permalink / raw)
  To: help-gnu-emacs

On Wednesday, 21 Dec 2022 at 23:09, Stefan Monnier via Users list for the GNU Emacs text editor wrote:
> Which version of Emacs is that?

It's Emacs from git, up to date as of yesterday morning.

> It would be useful to figure out what is this
> #<compiled 0x1f4100e4f6213ad7> function (presumably a completion-table)
> and how it calls `complete-with-action` because normally it should call
> it with an argument of the form `(boundaries . N)` which should cause
> `complete-with-action` to return nil immediately.

Any suggestion on how to do this?  I guess I would need to disable
native compilation and re-build?

-- 
Eric S Fraga via gnus (Emacs 30.0.50 2022-12-21) on Debian 11.5




^ permalink raw reply	[flat|nested] 7+ messages in thread

* cape and boundaries (was: typing slows down dramatically when typing digits in emails)
  2022-12-22 10:22   ` typing slows down dramatically when typing digits in emails Eric S Fraga
@ 2022-12-22 21:29     ` Stefan Monnier via Users list for the GNU Emacs text editor
  2022-12-23 12:14       ` cape and boundaries Eric S Fraga
  0 siblings, 1 reply; 7+ messages in thread
From: Stefan Monnier via Users list for the GNU Emacs text editor @ 2022-12-22 21:29 UTC (permalink / raw)
  To: help-gnu-emacs; +Cc: Daniel Mendler

[-- Attachment #1: Type: text/plain, Size: 1686 bytes --]

Eric S Fraga [2022-12-22 10:22:29] wrote:
> On Wednesday, 21 Dec 2022 at 23:09, Stefan Monnier via Users list for the
> GNU Emacs text editor wrote:
>>>         5311  79%             - completion-boundaries
>>>         5311  79%              - #<compiled 0x1f4100e4f6213ad7>
>>>         5311  79%               - complete-with-action
>>>         5311  79%                - #<compiled 0xfe6876f217f45e4>
>>>         4776  71%                 - cape--dabbrev-list
>>>         4776  71%                  - dabbrev--find-all-expansions
>>>         4704  70%                   - dabbrev--find-expansion
>>>         4671  70%                    - dabbrev--try-find
>>>         2959  44%                     - dabbrev--search
>>
>>`completion-boundaries` should not need to perform a `dabbrev--search`.
>> Which version of Emacs is that?
>
> It's Emacs from git, up to date as of yesterday morning.
>
>> It would be useful to figure out what is this
>> #<compiled 0x1f4100e4f6213ad7> function (presumably a completion-table)
>> and how it calls `complete-with-action` because normally it should call
>> it with an argument of the form `(boundaries . N)` which should cause
>> `complete-with-action` to return nil immediately.
>
> Any suggestion on how to do this?  I guess I would need to disable
> native compilation and re-build?

Oh, I forgot that `complete-with-action` actually won't return immediately
if the completion table it's passed as argument is itself a function.
I'm not sure if the patch below will make much difference (there's
a chance the `dabbrev--search` will still take place, just via some
other call), but at least I think it's "Right".

Daniel, WDYT?


        Stefan

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: cape.patch --]
[-- Type: text/x-diff, Size: 1309 bytes --]

diff --git a/cape.el b/cape.el
index dfb1e7a95c..f10decb5ff 100644
--- a/cape.el
+++ b/cape.el
@@ -199,13 +199,17 @@ VALID is the input comparator, see `cape--input-valid-p'."
         (end (copy-marker end t))
         (table nil))
     (lambda (str pred action)
-      (let ((new-input (buffer-substring-no-properties beg end)))
-        (when (or (eq input 'init) (not (cape--input-valid-p input new-input valid)))
-          ;; NOTE: We have to make sure that the completion table is interruptible.
-          ;; An interruption should not happen between the setqs.
-          (setq table (funcall fun new-input)
-                input new-input)))
-      (complete-with-action action table str pred))))
+      (if (or (eq action 'metadata) (eq (car-safe action) 'boundaries))
+          nil
+        (let ((new-input (buffer-substring-no-properties beg end)))
+          (when (or (eq input 'init)
+                    (not (cape--input-valid-p input new-input valid)))
+            ;; NOTE: We have to make sure that the completion table is
+            ;; interruptible.  An interruption should not happen
+            ;; between the setqs.
+            (setq table (funcall fun new-input)
+                  input new-input)))
+        (complete-with-action action table str pred)))))
 
 ;;;; Capfs
 

^ permalink raw reply related	[flat|nested] 7+ messages in thread

* Re: cape and boundaries
  2022-12-22 21:29     ` cape and boundaries (was: typing slows down dramatically when typing digits in emails) Stefan Monnier via Users list for the GNU Emacs text editor
@ 2022-12-23 12:14       ` Eric S Fraga
  0 siblings, 0 replies; 7+ messages in thread
From: Eric S Fraga @ 2022-12-23 12:14 UTC (permalink / raw)
  To: help-gnu-emacs

On Thursday, 22 Dec 2022 at 16:29, Stefan Monnier via Users list for the GNU Emacs text editor wrote:
> I'm not sure if the patch below will make much difference (there's
> a chance the `dabbrev--search` will still take place, just via some
> other call), but at least I think it's "Right".

FTR, I've tried this patch.  There's still an appreciable lag when
typing a number but everything else continues working just fine.

Thank you,
eric
-- 
Eric S Fraga via gnus (Emacs 30.0.50 2022-12-21) on Debian 11.5




^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2022-12-23 12:14 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-12-21 13:53 typing slows down dramatically when typing digits in emails Eric S Fraga
2022-12-22  4:09 ` Stefan Monnier via Users list for the GNU Emacs text editor
2022-12-22  4:34   ` Emanuel Berg
2022-12-22  4:45     ` always latest Emacs (was: Re: typing slows down dramatically when typing digits in emails) Emanuel Berg
2022-12-22 10:22   ` typing slows down dramatically when typing digits in emails Eric S Fraga
2022-12-22 21:29     ` cape and boundaries (was: typing slows down dramatically when typing digits in emails) Stefan Monnier via Users list for the GNU Emacs text editor
2022-12-23 12:14       ` cape and boundaries Eric S Fraga

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).