GNU bug report logs - #77089
[PATCH] Simplify Eshell history de-duplication

Previous Next

To reply to this bug, email your comments to 77089 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Morgan Smith <Morgan.J.Smith <at> outlook.com>
To: bug-gnu-emacs <at> gnu.org
Subject: [PATCH] Simplify Eshell history de-duplication
Date: Mon, 17 Mar 2025 18:33:50 -0400

[Message part 1 (text/plain, inline)]

Tags: patch

Hello!

I was noticing a perceptible delay in starting eshell of 0.838 seconds.
Turns out it was because I had `eshell-hist-ignoredups' set to 'erase
and I had a 5000 line shell-history file.

This patch speeds it up for me to 0.011 seconds.

[0001-Simplify-Eshell-history-de-duplication.patch (text/patch, attachment)]

Message #8 received at 77089 <at> debbugs.gnu.org (full text, mbox):

From: John Wiegley <johnw <at> gnu.org>
To: Morgan Smith <Morgan.J.Smith <at> outlook.com>
Cc: 77089 <at> debbugs.gnu.org
Subject: Re: bug#77089: [PATCH] Simplify Eshell history de-duplication
Date: Mon, 17 Mar 2025 18:35:21 -0700

>>>>> "MS" == Morgan Smith <Morgan.J.Smith <at> outlook.com> writes:

MS> I was noticing a perceptible delay in starting eshell of 0.838 seconds.
MS> Turns out it was because I had `eshell-hist-ignoredups' set to 'erase and
MS> I had a 5000 line shell-history file.

MS> This patch speeds it up for me to 0.011 seconds.

This is great, I didn’t realize this about ‘delete-duplicate-lines':

  Identical lines need not be adjacent, unless the argument
  ADJACENT is non-nil (interactively, with a C-u C-u prefix).
  This is a more efficient mode of operation, and may be useful
  on large regions that have already been sorted.

So it should accomplish just what the original intended to do, but relying on
a likely better optimized function.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

Message #11 received at 77089 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Morgan Smith <Morgan.J.Smith <at> outlook.com>,
 Jim Porter <jporterbugs <at> gmail.com>
Cc: 77089 <at> debbugs.gnu.org, johnw <at> gnu.org
Subject: Re: bug#77089: [PATCH] Simplify Eshell history de-duplication
Date: Tue, 18 Mar 2025 14:47:51 +0200

> Cc: John Wiegley <johnw <at> gnu.org>
> From: Morgan Smith <Morgan.J.Smith <at> outlook.com>
> Date: Mon, 17 Mar 2025 18:33:50 -0400
> 
> I was noticing a perceptible delay in starting eshell of 0.838 seconds.
> Turns out it was because I had `eshell-hist-ignoredups' set to 'erase
> and I had a 5000 line shell-history file.
> 
> This patch speeds it up for me to 0.011 seconds.

Thanks, I'm adding Jim to the discussion.

> >From 68160b6c84c2f8946caca5a6be274434f4e3d053 Mon Sep 17 00:00:00 2001
> From: Morgan Smith <Morgan.J.Smith <at> outlook.com>
> Date: Mon, 17 Mar 2025 18:19:21 -0400
> Subject: [PATCH] Simplify Eshell history de-duplication
> 
> This also increases performance significantly for large history
> files when `eshell-hist-ignoredups' is set to 'erase.
> 
> * lisp/eshell/em-hist.el (eshell-read-history): Run
> `delete-duplicate-lines' on the shell history text instead of
> checking for duplicates at every insertion.
> ---
>  lisp/eshell/em-hist.el | 17 ++++++-----------
>  1 file changed, 6 insertions(+), 11 deletions(-)
> 
> diff --git a/lisp/eshell/em-hist.el b/lisp/eshell/em-hist.el
> index 3f779f95acd..78ceed4bd6a 100644
> --- a/lisp/eshell/em-hist.el
> +++ b/lisp/eshell/em-hist.el
> @@ -458,22 +458,17 @@ eshell-read-history
>  	     (ignore-dups eshell-hist-ignoredups))
>  	(with-temp-buffer
>  	  (insert-file-contents file)
> +          (when ignore-dups
> +            (delete-duplicate-lines (point-min) (point-max) t
> +                                    (not (eq ignore-dups 'erase))))
>  	  ;; Watch for those date stamps in history files!
>  	  (goto-char (point-max))
>  	  (while (and (< count size)
>  		      (re-search-backward "^[ \t]*\\([^#\n].*\\)[ \t]*$"
>  					  nil t))
> -	    (let ((history (match-string 1)))
> -              (when (or (ring-empty-p ring)
> -                        (null ignore-dups)
> -                        (and (not (string-equal
> -                                   (ring-ref ring (1- (ring-length ring)))
> -                                   history))
> -                             (not (and (eq ignore-dups 'erase)
> -                                       (ring-member ring history)))))
> -                (ring-insert-at-beginning
> -		 ring (subst-char-in-string ?\177 ?\n history))
> -                (setq count (1+ count))))))
> +            (ring-insert-at-beginning
> +             ring (subst-char-in-string ?\177 ?\n (match-string 1)))
> +            (setq count (1+ count))))
>  	(setq eshell-history-ring ring
>  	      eshell-history-index nil
>                eshell-hist--new-items 0))))))
> -- 
> 2.48.1
> 

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.