GNU bug report logs - #21684
25.0.50; undo in query-replace w/o exit

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 21684 in the body.
You can then email your comments to 21684 AT debbugs.gnu.org in the normal way.

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

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

From: Tino Calancha <f92capac <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: 25.0.50; undo in query-replace w/o exit
Date: Wed, 14 Oct 2015 22:26:22 +0900 (JST)

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

In GNU Emacs 25.0.50.1 (x86_64-unknown-linux-gnu, GTK+ Version 2.24.23)
 of 2015-10-14
Repository revision: 38f99a02b83e9e408970cb7abdb685725b2004f9


Quite often, performing query-replace, i just hit the [space] quicky
with no much attention. If i replace one string by mistake, i can
comeback to it with "^", but if i edit the string with "E" to reset it 
to its original value, the interactive query-replace session ends.

I found useful if we can undo such replacements without exiting the 
query-replace.

The attached patch implement such behaviour for `query-replace'.
Having this also working for `query-replace-regexp' would be very nice,
but harder to implement.

[replace.patch (text/plain, attachment)]

Message #8 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Juri Linkov <juri <at> linkov.net>
To: Tino Calancha <f92capac <at> gmail.com>
Cc: 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Wed, 14 Oct 2015 19:23:05 +0300

> Quite often, performing query-replace, i just hit the [space] quicky
> with no much attention. If i replace one string by mistake, i can
> comeback to it with "^", but if i edit the string with "E" to reset it to
> its original value, the interactive query-replace session ends.
>
> I found useful if we can undo such replacements without exiting the
> query-replace.
>
> The attached patch implement such behaviour for `query-replace'.
> Having this also working for `query-replace-regexp' would be very nice,
> but harder to implement.

This would be a useful feature.  But I wonder why it's harder to
implement for regexps?  If you need, you could add all previous
replacements strings to the same stack at the end of data elements in
the stack like we did when necessary by adding new data at the end of the
similar isearch stack elements in isearch-cmds.  This would simplify
restoring the previous replaced strings from the stack.

PS: an alternative solution would be to call ordinary (undo) since
perform-replace places undo boundaries with (undo-boundary), but
it seems this would be a less reliable approach?

Message #11 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <f92capac <at> gmail.com>
To: 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Sat, 17 Oct 2015 22:04:44 +0900 (JST)

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

> I wonder why it's harder to implement for regexps?
Basically for my lack of expertise in this code: i was trying
to extend my patch to support regexp without success.

> you could add all previous replacements strings
> to the same stack at the end of data elements
> like we did when necessary by adding new data at
> the end of the similar isearch stack elements in isearch-cmds.
Thank you, it sounds a good idea!!

See new patch (replace-2.patch): AFAICS is working as i wanted.
Please try it and feel free to give me any comments if you
find some issues to fix/improve.

[replace-2.patch (text/plain, attachment)]

Message #14 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Juri Linkov <juri <at> linkov.net>
To: Tino Calancha <f92capac <at> gmail.com>
Cc: 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Sun, 18 Oct 2015 00:49:46 +0300

> See new patch (replace-2.patch): AFAICS is working as i wanted.
> Please try it and feel free to give me any comments if you
> find some issues to fix/improve.

Thank you, your patch is working well.  One issue to improve I noticed that
when a previous replacement is skipped with "n" then undo "U" acts exactly
as backup "^" by going to the previous place, and doing no undo actually.
Maybe better would be for undo "U" to only undo and skip non-replacement
entries in the stack?  (Implementation-wise this means looping until
finding a previous non-replacement entry in the stack).

A minor comment is that to save cons cells you could just put both new
elements shallow as elements 3 and 4 in the stack instead of adding
a list of them.

Another minor suggestions is for better names you could rename
the prefix -solved-regexp to -replaced, i.e. replace
search-string-solved-regexp, next-replacement-solved-regexp with
search-string-replaced, next-replacement-replaced.

Also we need this feature to be documented in etc/NEWS.

Have you already signed FSF copyright assignment papers?

Your patch for bug#21663 is short enough to be installed without the
copyright paperwork, but your changes for bug#21684 require it.

Message #17 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <f92capac <at> gmail.com>
To: 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Sun, 18 Oct 2015 18:45:46 +0900 (JST)

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

>Maybe better would be for undo "U" to only undo and skip non-replacement
>entries in the stack?  (Implementation-wise this means looping until
>finding a previous non-replacement entry in the stack).
I agree, its better undo 'just' undo. Fixed in the new patch
(replace-3.patch).

>A minor comment is that to save cons cells you could just put both new
>elements shallow as elements 3 and 4 in the stack instead of adding
>a list of them.
Sure, thank you for remind me this. Applied.

>Another minor suggestions is for better names you could rename
>the prefix -solved-regexp to -replaced
Done. Thank you.

>Also we need this feature to be documented in etc/NEWS.
>Have you already signed FSF copyright assignment papers?
I didn't. I would like to sign. Maybe someone could guide me how to
complete such process.

[replace-3.patch (text/plain, attachment)]

Message #20 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Richard Stallman <rms <at> gnu.org>
To: Tino Calancha <f92capac <at> gmail.com>
Cc: 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Sun, 18 Oct 2015 11:58:55 -0400

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > >Maybe better would be for undo "U" to only undo and skip non-replacement
  > >entries in the stack?  (Implementation-wise this means looping until
  > >finding a previous non-replacement entry in the stack).
  > I agree, its better undo 'just' undo. Fixed in the new patch
  > (replace-3.patch).

What change in functionality is being considered here?

-- 
Dr Richard Stallman
President, Free Software Foundation (gnu.org, fsf.org)
Internet Hall-of-Famer (internethalloffame.org)
Skype: No way! See stallman.org/skype.html.

Message #23 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <f92capac <at> gmail.com>
To: Richard Stallman <rms <at> gnu.org>
Cc: Tino Calancha <f92capac <at> gmail.com>, 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Mon, 19 Oct 2015 10:20:25 +0900 (JST)

> What change in functionality is being considered here?

Hi Richard,

the purpose is provide 'undo' into `query-replace'/`query-replace-regexp'.
Original code do not support such functionality in a convenient way.

For instance, in the scratch buffer:
M-x query-replace RET f RET @ RET [space] RET [space] RET C-/
(C-/ undo last replacement but end session: [space] would insert one space)

Indeed, you may correct last replacement with existing code without exit current
search replace, as follows:

I)
M-x query-replace RET f RET @ RET [space] RET [space] "^" C-r
(now you can edit last replacement; comeback to query-replace session with C-M-c)
C-M-c [space] [space] ...

This seems too complicated/slow to just undo last 1-2 replacements.

In my experience as an user 99% of the times, the situation is that i hit [space]
many times and fast, and then i realize i just want to undo last 1-2 
replacements and keep going. It would be useful if that can be done just 
with one key.

The patch replace-3.patch in this thread implements such interactive undo and binds
the new funtionality to keys "u" and "U".

Every time you hit "u" the latest replacement is undo: now you can hit "n" to skip
it (tipically) or [space] to replace it again, or maybe "u" to undo 
another previous replacement (if any). The query-replace session is not 
exit.

Tino

Message #26 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Juri Linkov <juri <at> linkov.net>
To: Tino Calancha <f92capac <at> gmail.com>
Cc: 21684 <at> debbugs.gnu.org, Richard Stallman <rms <at> gnu.org>
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Tue, 20 Oct 2015 01:04:34 +0300

> The patch replace-3.patch in this thread implements such interactive undo
> and binds the new funtionality to keys "u" and "U".

Please note that usually the upper-case "U" applies to all
elements in the set such as e.g. in Dired the lower-case "u"
is bound to a single-mark dired-unmark whereas the upper-case "U"
is bound to dired-unmark-all-marks, etc.

For query-replace this means that better would be to bind the
upper-case "U" to a new operation undo-all that could be easy to
implement in the same branch with (or (eq def 'undo) (eq def 'undo-all))
and looping until the top of the stack when (eq def 'undo-all).

Message #29 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <f92capac <at> gmail.com>
To: 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Tue, 20 Oct 2015 21:02:55 +0900 (JST)

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

> Please note that usually the upper-case "U" applies to all
> elements in the set.
> For query-replace this means that better would be to bind the
> upper-case "U" to a new operation undo-all

Good point. Thank you very much.
Included in new patch (replace-4.patch).

[replace-4.patch (text/plain, attachment)]

Message #32 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <f92capac <at> gmail.com>
To: 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: Acknowledgement (25.0.50; undo in query-replace w/o
 exit)
Date: Tue, 20 Oct 2015 22:01:20 +0900 (JST)

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

In previous patch 'undo-all just show the total number of undid 
replacements if the first element was replaced.

New patch (replace-5.patch) fix that: undo-all always show the accumulated
number of undone performed.

[replace-5.patch (text/plain, attachment)]

Message #35 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Richard Stallman <rms <at> gnu.org>
To: Tino Calancha <f92capac <at> gmail.com>
Cc: f92capac <at> gmail.com, 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Mon, 26 Oct 2015 00:09:13 -0400

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

Thanks.  It seems useful.

(Please forgive my delay.)

-- 
Dr Richard Stallman
President, Free Software Foundation (gnu.org, fsf.org)
Internet Hall-of-Famer (internethalloffame.org)
Skype: No way! See stallman.org/skype.html.

Message #38 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Lars Ingebrigtsen <larsi <at> gnus.org>
To: Tino Calancha <f92capac <at> gmail.com>
Cc: 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Tue, 23 Feb 2016 19:40:26 +1100

Tino Calancha <f92capac <at> gmail.com> writes:

> Good point. Thank you very much.
> Included in new patch (replace-4.patch).

Looks good.  Could you also add documentation for this (to the manual)
and a NEWS entry?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Message #41 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <f92capac <at> gmail.com>
To: Lars Ingebrigtsen <larsi <at> gnus.org>
Cc: Tino Calancha <f92capac <at> gmail.com>, 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Wed, 24 Feb 2016 01:46:14 +0900 (JST)

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

> Looks good.  Could you also add documentation for this (to the manual)
> and a NEWS entry?
Sure. Added NEWS entry and updated emacs manual.

[0001-Undo-in-query-replace-without-exit.patch (text/plain, attachment)]

[0002-Add-NEWS-and-documentation-for-last-commit.patch (text/plain, attachment)]

Message #44 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Tino Calancha <f92capac <at> gmail.com>
Cc: 21684 <at> debbugs.gnu.org, larsi <at> gnus.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Tue, 23 Feb 2016 19:54:15 +0200

> Date: Wed, 24 Feb 2016 01:46:14 +0900 (JST)
> From: Tino Calancha <f92capac <at> gmail.com>
> Cc: Tino Calancha <f92capac <at> gmail.com>, 21684 <at> debbugs.gnu.org
> 
> > Looks good.  Could you also add documentation for this (to the manual)
> > and a NEWS entry?
> Sure. Added NEWS entry and updated emacs manual.

Thanks.  A minor nit below:

> --- a/etc/NEWS
> +++ b/etc/NEWS
> @@ -435,6 +435,11 @@ is intended for adding to 'kill-emacs-query-functions'.
>  in favor of the global `M-s h' bindings introduced in Emacs-23.1.
>  They'll disappear soon.
>  
> ++++
> +** New bindings for 'query-replace-map'.
> +`undo', undo the last replacement; bind to `u'.
> +`undo-all', undo all replacements; Bind to `U'.

"Bound", not "bind"; and the capitalization needs to be fixed.

Message #47 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Lars Ingebrigtsen <larsi <at> gnus.org>
To: Tino Calancha <f92capac <at> gmail.com>
Cc: 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Wed, 24 Feb 2016 12:36:37 +1100

Tino Calancha <f92capac <at> gmail.com> writes:

>> Looks good.  Could you also add documentation for this (to the manual)
>> and a NEWS entry?
> Sure. Added NEWS entry and updated emacs manual.

Thanks; I've now applied this to the trunk.  Seems to be working great,
and it's a feature I've missed for years.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Message #50 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <f92capac <at> gmail.com>
To: 21684 <at> debbugs.gnu.org
Cc: Tino Calancha <f92capac <at> gmail.com>
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Wed, 24 Feb 2016 22:13:45 +0900 (JST)

> Thanks; I've now applied this to the trunk.  Seems to be working great,
> and it's a feature I've missed for years.
Great! It has being my pleasure.

Message #53 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Lars Ingebrigtsen <larsi <at> gnus.org>
Cc: Tino Calancha <f92capac <at> gmail.com>, 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Tue, 31 May 2016 18:32:08 -0400

Lars Ingebrigtsen wrote:

> Thanks; I've now applied this to the trunk.  Seems to be working great,
> and it's a feature I've missed for years.

Was a copyright assignment ever completed?
I don't see one on file.

http://debbugs.gnu.org/21684#14

Juri Linkov wrote:

> Have you already signed FSF copyright assignment papers?
>
> Your patch for bug#21663 is short enough to be installed without the
> copyright paperwork, but your changes for bug#21684 require it.

Message #56 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Tino Calancha <f92capac <at> gmail.com>
To: Glenn Morris <rgm <at> gnu.org>
Cc: Tino Calancha <f92capac <at> gmail.com>, 21684 <at> debbugs.gnu.org,
 Lars Ingebrigtsen <larsi <at> gnus.org>
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Wed, 1 Jun 2016 11:44:24 +0900 (JST)

On Tue, 31 May 2016, Glenn Morris wrote:

> Was a copyright assignment ever completed?
> I don't see one on file.
>
> http://debbugs.gnu.org/21684#14

Yes it was, around November 2015: copyright assignment  #1056152

Tino

Message #59 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Tino Calancha <f92capac <at> gmail.com>
Cc: 21684 <at> debbugs.gnu.org, Lars Ingebrigtsen <larsi <at> gnus.org>
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Wed, 01 Jun 2016 13:45:56 -0400

Tino Calancha wrote:

>> Was a copyright assignment ever completed?
>> I don't see one on file.
>>
>> http://debbugs.gnu.org/21684#14
>
> Yes it was, around November 2015: copyright assignment  #1056152

Thanks. The semi-public file that records these things must not be
up-to-date (not for the first time). Not your problem. I'll ask the
relevant people.

Message #62 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Lars Ingebrigtsen <larsi <at> gnus.org>
To: Glenn Morris <rgm <at> gnu.org>
Cc: Tino Calancha <f92capac <at> gmail.com>, 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Wed, 01 Jun 2016 19:53:23 +0200

Glenn Morris <rgm <at> gnu.org> writes:

> Tino Calancha wrote:
>
>>> Was a copyright assignment ever completed?
>>> I don't see one on file.
>>>
>>> http://debbugs.gnu.org/21684#14
>>
>> Yes it was, around November 2015: copyright assignment  #1056152
>
> Thanks. The semi-public file that records these things must not be
> up-to-date (not for the first time). Not your problem. I'll ask the
> relevant people.

It was the last time I fetched it, but the name is in a slightly
different format.  :-)

Search for "Calancha".  (I'm not sure I'm allowed to post the full name
from the assignment here.)

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Message #65 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Lars Ingebrigtsen <larsi <at> gnus.org>
Cc: Tino Calancha <f92capac <at> gmail.com>, 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Wed, 01 Jun 2016 13:56:42 -0400

Lars Ingebrigtsen wrote:

> It was the last time I fetched it, but the name is in a slightly
> different format.  :-)
>
> Search for "Calancha".  (I'm not sure I'm allowed to post the full name
> from the assignment here.)

I don't think my grep-fu is to blame here. I generally try a broad search.
I've asked assign <at> gnu. There seem to be several missing entries AFAICS.

Message #70 received at 21684 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Lars Ingebrigtsen <larsi <at> gnus.org>
Cc: Tino Calancha <f92capac <at> gmail.com>, 21684 <at> debbugs.gnu.org
Subject: Re: bug#21684: 25.0.50; undo in query-replace w/o exit
Date: Thu, 02 Jun 2016 12:44:25 -0400

Glenn Morris wrote:

> I've asked assign <at> gnu. There seem to be several missing entries AFAICS.

The file that records the info was indeed out-of-date (malfunctioning
cron job) and has now been fixed.

Previous Next

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