bug#52543: Add option to eliminate all duplications from the kill ring

bug#52543: Add option to eliminate all duplications from the kill ring

[ndame via Bug reports for GNU Emacs, the Swiss army knife of text editors]

There is  kill-do-not-save-duplicates, but it deletes
only subsequent duplicates.

There should be an other variable which prevents any duplicates
even if there are, say,  10 or 20 items between them.

I expected kill-do-not-save-duplicates to do that, that variable
seems misnamed. It should be kill-do-not-save-subsequent-duplicates
instead.

bug#52543: Add option to eliminate all duplications from the kill ring

[Lars Ingebrigtsen]

ndame <laszlomail@protonmail.com> writes:

> There is  kill-do-not-save-duplicates, but it deletes
> only subsequent duplicates.
>
> There should be an other variable which prevents any duplicates
> even if there are, say,  10 or 20 items between them.
>
> I expected kill-do-not-save-duplicates to do that, that variable
> seems misnamed. It should be kill-do-not-save-subsequent-duplicates
> instead.

This has previously been discussed, and we decided to not alter how this
works, because it'd be too surprising.  (And there are options to filter
the kill ring that users that want that can use.)  So I'm closing this
bug report.

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

bug#52543: Add option to eliminate all duplications from the kill ring

[Juri Linkov]

>> There is  kill-do-not-save-duplicates, but it deletes
>> only subsequent duplicates.
>>
>> There should be an other variable which prevents any duplicates
>> even if there are, say,  10 or 20 items between them.
>>
>> I expected kill-do-not-save-duplicates to do that, that variable
>> seems misnamed. It should be kill-do-not-save-subsequent-duplicates
>> instead.
>
> This has previously been discussed, and we decided to not alter how this
> works, because it'd be too surprising.  (And there are options to filter
> the kill ring that users that want that can use.)  So I'm closing this
> bug report.

Indeed:

  (setq kill-transform-function (lambda (s) (unless (member s kill-ring) s)))

bug#52543: Add option to eliminate all duplications from the kill ring

[Juri Linkov]

>>> There is  kill-do-not-save-duplicates, but it deletes
>>> only subsequent duplicates.
>>>
>>> There should be an other variable which prevents any duplicates
>>> even if there are, say,  10 or 20 items between them.
>>>
>>> I expected kill-do-not-save-duplicates to do that, that variable
>>> seems misnamed. It should be kill-do-not-save-subsequent-duplicates
>>> instead.
>>
>> This has previously been discussed, and we decided to not alter how this
>> works, because it'd be too surprising.  (And there are options to filter
>> the kill ring that users that want that can use.)  So I'm closing this
>> bug report.
>
> Indeed:
>
>   (setq kill-transform-function (lambda (s) (unless (member s kill-ring) s)))

Here is the same as a choice to make it easier for everyone
to satisfy this frequent need - like history-delete-duplicates
and comint-input-ignoredups in Emacs, and ignoredups in shell.
If this is ok, more documentation could be added:

diff --git a/lisp/simple.el b/lisp/simple.el
index 959e7ea94c..466bc9c12d 100644
--- a/lisp/simple.el
+++ b/lisp/simple.el
@@ -5118,6 +5118,8 @@ kill-transform-function
 be put on the kill ring).  It should return a string or nil.  If
 the latter, the string is not put on the kill ring."
   :type '(choice (const :tag "No transform" nil)
+                 (const :tag "Ignore duplicates"
+                        (lambda (s) (unless (member s kill-ring) s)))
                  function)
   :group 'killing
   :version "28.1")
-- 

bug#52543: Add option to eliminate all duplications from the kill ring

[Juri Linkov]

>>> There is  kill-do-not-save-duplicates, but it deletes
>>> only subsequent duplicates.
>>>
>>> There should be an other variable which prevents any duplicates
>>> even if there are, say,  10 or 20 items between them.
>>>
>>> I expected kill-do-not-save-duplicates to do that, that variable
>>> seems misnamed. It should be kill-do-not-save-subsequent-duplicates
>>> instead.
>>
>> This has previously been discussed, and we decided to not alter how this
>> works, because it'd be too surprising.  (And there are options to filter
>> the kill ring that users that want that can use.)  So I'm closing this
>> bug report.
>
> Indeed:
>
>   (setq kill-transform-function (lambda (s) (unless (member s kill-ring) s)))

Sorry, this was wrong.  It should delete all previous strings,
and add a new to the top of the kill-ring:

  (setq kill-transform-function (lambda (s) (setq kill-ring (delete s kill-ring)) s))

But it would be strange to modify kill-ring in the transform function.
So maybe kill-do-not-save-duplicates should support a special choice
'delete-duplicates' after all.

bug#52543: Add option to eliminate all duplications from the kill ring

[Lars Ingebrigtsen]

Juri Linkov <juri@linkov.net> writes:

> But it would be strange to modify kill-ring in the transform function.
> So maybe kill-do-not-save-duplicates should support a special choice
> 'delete-duplicates' after all.

I think some people wouldn't want to remove the duplicates (and add to
the front), and some wouldn't want to add at all (I think that was the
request in the other bug report?), and some want to take text properties
into account, and some don't, so I think leaving this to the user and
the transform function is the right solution.  But we don't have to make
any pre-made functions for this.

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