Re: master fbc9c59: Make goto-line-history buffer local only when so customized

Re: master fbc9c59: Make goto-line-history buffer local only when so customized

[Basil L. Contovounesios]

acm@muc.de (Alan Mackenzie) writes:

> branch: master
> commit fbc9c59b9eb02d49f426341ee32334784d224ce4
> Author: Alan Mackenzie <acm@muc.de>
> Commit: Alan Mackenzie <acm@muc.de>
>
>     Make goto-line-history buffer local only when so customized
>     
>     * lisp/simple.el (goto-line-history-local): New customizable option.
>     (goto-line-history): Define this simply with defvar, not defvar-local.
>     (goto-line-read-args): Handle goto-line-history-local, and changes to it.
>     
>     * doc/emacs/basic.texi (Moving Point): Add a paragraph about
>     goto-line-history-local.
>     
>     * etc/NEWS: Add an item under "Editing Changes in Emacs 28.1".

[...]

> diff --git a/etc/NEWS b/etc/NEWS
> index b96bcd9..7665d47 100644
> --- a/etc/NEWS
> +++ b/etc/NEWS
> @@ -345,6 +345,11 @@ trying to be non-destructive.
>  This command opens a new buffer called "*Memory Report*" and gives a
>  summary of where Emacs is using memory currently.
>  
> ++++
> +** The history list for the 'goto-line' command is now a single list
> +for all buffers by default.  You can configure a separate list for
> +each buffer by customizing the user option 'goto-line-history-local'.

I think this contradicts a preceding entry:

** Input history for 'goto-line' is now local to every buffer.
Each buffer will keep a separate history of line numbers used with
'goto-line'.  This should help making faster the process of finding
line numbers that were previously jumped to.

Thanks,

-- 
Basil

Re: master fbc9c59: Make goto-line-history buffer local only when so customized

[Alan Mackenzie]

Hello, Basil.

On Fri, Feb 26, 2021 at 12:19:14 +0000, Basil L. Contovounesios wrote:
> acm@muc.de (Alan Mackenzie) writes:

> > branch: master
> > commit fbc9c59b9eb02d49f426341ee32334784d224ce4
> > Author: Alan Mackenzie <acm@muc.de>
> > Commit: Alan Mackenzie <acm@muc.de>

> >     Make goto-line-history buffer local only when so customized

> >     * lisp/simple.el (goto-line-history-local): New customizable option.
> >     (goto-line-history): Define this simply with defvar, not defvar-local.
> >     (goto-line-read-args): Handle goto-line-history-local, and changes to it.

> >     * doc/emacs/basic.texi (Moving Point): Add a paragraph about
> >     goto-line-history-local.

> >     * etc/NEWS: Add an item under "Editing Changes in Emacs 28.1".

> [...]

> > diff --git a/etc/NEWS b/etc/NEWS
> > index b96bcd9..7665d47 100644
> > --- a/etc/NEWS
> > +++ b/etc/NEWS
> > @@ -345,6 +345,11 @@ trying to be non-destructive.
> >  This command opens a new buffer called "*Memory Report*" and gives a
> >  summary of where Emacs is using memory currently.

> > ++++
> > +** The history list for the 'goto-line' command is now a single list
> > +for all buffers by default.  You can configure a separate list for
> > +each buffer by customizing the user option 'goto-line-history-local'.

> I think this contradicts a preceding entry:

> ** Input history for 'goto-line' is now local to every buffer.
> Each buffer will keep a separate history of line numbers used with
> 'goto-line'.  This should help making faster the process of finding
> line numbers that were previously jumped to.

Well, I think "contradict" is not quite the right word.  Whether the list
is buffer local or not is now customisable, which it wasn't before.  The
default is somewhat arbitrary, as it always is in these things, with some
people proclaiming a particular setting "obviously" should be the
default, others saying the opposite is "obvious".  That the list, before
that previous patch, wasn't buffer local points to the current default.

Or, have I misunderstood what you're saying?

> Thanks,

> -- 
> Basil

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: master fbc9c59: Make goto-line-history buffer local only when so customized

[Basil L. Contovounesios]

Alan Mackenzie <acm@muc.de> writes:

>> > ++++
>> > +** The history list for the 'goto-line' command is now a single list
>> > +for all buffers by default.  You can configure a separate list for
>> > +each buffer by customizing the user option 'goto-line-history-local'.
>
>> I think this contradicts a preceding entry:
>
>> ** Input history for 'goto-line' is now local to every buffer.
>> Each buffer will keep a separate history of line numbers used with
>> 'goto-line'.  This should help making faster the process of finding
>> line numbers that were previously jumped to.
>
> Well, I think "contradict" is not quite the right word.  Whether the list
> is buffer local or not is now customisable, which it wasn't before.  The
> default is somewhat arbitrary, as it always is in these things, with some
> people proclaiming a particular setting "obviously" should be the
> default, others saying the opposite is "obvious".  That the list, before
> that previous patch, wasn't buffer local points to the current default.
>
> Or, have I misunderstood what you're saying?

I think so.  My point is that the older entry says goto-line has
buffer-local history by default, whereas the newer entry says goto-line
does not have buffer-local history by default.

The older entry came with the following change:

Make goto-line keep a separate input history per buffer
7c5d6a2afc 2019-12-24 17:40:15 +0100
https://git.savannah.gnu.org/cgit/emacs.git/commit/?id=7c5d6a2afc6c23a7fff8456f506ee2aa2d37a3b9

The newer entry came with the following change:

Make goto-line-history buffer local only when so customized
fbc9c59b9e 2021-02-17 21:15:51 +0000
https://git.savannah.gnu.org/cgit/emacs.git/commit/?id=fbc9c59b9eb02d49f426341ee32334784d224ce4

The latter change reverts some parts of the former, and makes the
behaviour customisable, but the older NEWS entry was not updated to
reflect this.  I was hoping you would merge the two NEWS entries or
simply delete the older one, since it no longer accurately represents
the default, and is duplicated by the newer entry.

This part of (info "(elisp) Minibuffer History") also needs updating:

 -- Variable: goto-line-history
     A history list for arguments to ‘goto-line’.  This variable is
     buffer local.

Thanks,

-- 
Basil

Re: master fbc9c59: Make goto-line-history buffer local only when so customized

[Alan Mackenzie]

Hello, Basil.

On Fri, Feb 26, 2021 at 13:15:02 +0000, Basil L. Contovounesios wrote:
> Alan Mackenzie <acm@muc.de> writes:

> >> > ++++
> >> > +** The history list for the 'goto-line' command is now a single list
> >> > +for all buffers by default.  You can configure a separate list for
> >> > +each buffer by customizing the user option 'goto-line-history-local'.

> >> I think this contradicts a preceding entry:

> >> ** Input history for 'goto-line' is now local to every buffer.
> >> Each buffer will keep a separate history of line numbers used with
> >> 'goto-line'.  This should help making faster the process of finding
> >> line numbers that were previously jumped to.

> > Well, I think "contradict" is not quite the right word.  Whether the list
> > is buffer local or not is now customisable, which it wasn't before.  The
> > default is somewhat arbitrary, as it always is in these things, with some
> > people proclaiming a particular setting "obviously" should be the
> > default, others saying the opposite is "obvious".  That the list, before
> > that previous patch, wasn't buffer local points to the current default.

> > Or, have I misunderstood what you're saying?

> I think so.  My point is that the older entry says goto-line has
> buffer-local history by default, whereas the newer entry says goto-line
> does not have buffer-local history by default.

Yes, I had misunderstood, sorry.  I was under the mistaken impression
that the previous change to the input history was in Emacs 27.  So, yes,
you're correct, the two entries in NEWS do contradict eachother, and
need merging into a single entry.

I don't have time to do this today, I'll try and do it over the weekend.

[ .... ]

> The latter change reverts some parts of the former, and makes the
> behaviour customisable, but the older NEWS entry was not updated to
> reflect this.  I was hoping you would merge the two NEWS entries or
> simply delete the older one, since it no longer accurately represents
> the default, and is duplicated by the newer entry.

> This part of (info "(elisp) Minibuffer History") also needs updating:

>  -- Variable: goto-line-history
>      A history list for arguments to ‘goto-line’.  This variable is
>      buffer local.

Hmm.  OK.  But what's this variable doing in the elisp manual, as
opposed to the emacs manual?  It's purely a user convenience.

> Thanks,

> -- 
> Basil

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: master fbc9c59: Make goto-line-history buffer local only when so customized

[Basil L. Contovounesios]

Alan Mackenzie <acm@muc.de> writes:

> Yes, I had misunderstood, sorry.  I was under the mistaken impression
> that the previous change to the input history was in Emacs 27.  So, yes,
> you're correct, the two entries in NEWS do contradict eachother, and
> need merging into a single entry.
>
> I don't have time to do this today, I'll try and do it over the weekend.

Thanks.

>> This part of (info "(elisp) Minibuffer History") also needs updating:
>
>>  -- Variable: goto-line-history
>>      A history list for arguments to ‘goto-line’.  This variable is
>>      buffer local.
>
> Hmm.  OK.  But what's this variable doing in the elisp manual, as
> opposed to the emacs manual?  It's purely a user convenience.

I think it's there because (info "(elisp) Minibuffer History") already
lists several built-in history variables for standard user commands.

And (info "(emacs) Moving Point") already explains that goto-line has
its own history variable, and links to the Elisp manual, so I don't see
any problem, or how a history variable constitutes a user convenience.

But I also have no strong opinions on whether or how this history
variable should be documented, so feel free to change it as you or
others see fit.

-- 
Basil

Re: master fbc9c59: Make goto-line-history buffer local only when so customized

[Alan Mackenzie]

Hello, Basil.

On Sat, Feb 27, 2021 at 14:27:39 +0000, Basil L. Contovounesios wrote:
> Alan Mackenzie <acm@muc.de> writes:

> > Yes, I had misunderstood, sorry.  I was under the mistaken impression
> > that the previous change to the input history was in Emacs 27.  So, yes,
> > you're correct, the two entries in NEWS do contradict eachother, and
> > need merging into a single entry.

> > I don't have time to do this today, I'll try and do it over the weekend.

> Thanks.

I think it's now OK.

> >> This part of (info "(elisp) Minibuffer History") also needs updating:

> >>  -- Variable: goto-line-history
> >>      A history list for arguments to ‘goto-line’.  This variable is
> >>      buffer local.

I've updated this, too.

> > Hmm.  OK.  But what's this variable doing in the elisp manual, as
> > opposed to the emacs manual?  It's purely a user convenience.

> I think it's there because (info "(elisp) Minibuffer History") already
> lists several built-in history variables for standard user commands.

> And (info "(emacs) Moving Point") already explains that goto-line has
> its own history variable, and links to the Elisp manual, so I don't see
> any problem, or how a history variable constitutes a user convenience.

> But I also have no strong opinions on whether or how this history
> variable should be documented, so feel free to change it as you or
> others see fit.

I think I came over as more combative than I intended, sorry.  I was
merely curious about the reasons, not particularly wanting to
restructure anything.  Thanks for explaining.

> -- 
> Basil

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: master fbc9c59: Make goto-line-history buffer local only when so customized

[Basil L. Contovounesios]

Alan Mackenzie <acm@muc.de> writes:

> I think it's now OK.

Indeed, thanks.

>> > Hmm.  OK.  But what's this variable doing in the elisp manual, as
>> > opposed to the emacs manual?  It's purely a user convenience.
>
>> I think it's there because (info "(elisp) Minibuffer History") already
>> lists several built-in history variables for standard user commands.
>
>> And (info "(emacs) Moving Point") already explains that goto-line has
>> its own history variable, and links to the Elisp manual, so I don't see
>> any problem, or how a history variable constitutes a user convenience.
>
>> But I also have no strong opinions on whether or how this history
>> variable should be documented, so feel free to change it as you or
>> others see fit.
>
> I think I came over as more combative than I intended, sorry.  I was
> merely curious about the reasons, not particularly wanting to
> restructure anything.  Thanks for explaining.

Your curiosity came across just fine to me, and I didn't think you were
combative at all.  I too was merely stating facts as I saw them in case
they helped you make a judgement, so I hope I didn't come across as
combative either.

Thanks,

-- 
Basil