Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

[Michael Albinus]

Eli Zaretskii <eliz@gnu.org> writes:

Hi,

> diff --git a/lisp/minibuffer.el b/lisp/minibuffer.el
> index 3beb3c06a18..6fae62b3904 100644
> --- a/lisp/minibuffer.el
> +++ b/lisp/minibuffer.el
> @@ -1141,7 +1141,15 @@ and DOC describes the way this style of completion works.")
>      ;; and simply add "bar" to the end of the result.
>      emacs22)
>    "List of completion styles to use.
> -The available styles are listed in `completion-styles-alist'.
> +An element should be a symbol which is listed in
> +`completion-styles-alist'.
> +
> +An element can also be a list of the form
> +(STYLE ((VARIABLE VALUE) ...))
> +STYLE must be a symbol listed in `completion-styles-alist', followed by
> +a `let'-style list of variable/value pairs.  VARIABLE will be bound to
> +VALUE (without evaluating it) while the style is handling completion.
> +This allows repeating the same style with different configurations.

The :type of completion-styles should be adapted in order to reflect
this change. Furthermore, this warrants a change of the :version, I
believe.

Just being curious: why must it be a form (STYLE ((VARIABLE VALUE) ...)) ?
Wouldn't it be sufficient to use (STYLE (VARIABLE VALUE) ...) ?

> +(defcustom completion-pcm-leading-wildcard nil
> +  "If non-nil, partial-completion completes as if there's a leading wildcard.
> +
> +If nil (the default), partial-completion requires a matching completion
> +alternative to have the same beginning as the first \"word\" in the
> +minibuffer text, where \"word\" is determined by
> +`completion-pcm-word-delimiters'.
> +
> +If non-nil, partial-completion allows any string of characters to occur
> +at the beginning of a completion alternative, as if a wildcard such as
> +\"*\" was present at the beginning of the minibuffer text.  This makes
> +partial-completion behave more like the substring completion style."
> +  :version "30.1"
> +  :type 'boolean)

This doesn't look right, it should be :version "31.1".

Best regards, Michael.

Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

[Eli Zaretskii]

Spencer, would you please take care of these gotchas?

> From: Michael Albinus <michael.albinus@gmx.de>
> Cc: Spencer Baugh <sbaugh@janestreet.com>
> Date: Sat, 24 Aug 2024 12:15:49 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> Hi,
> 
> > diff --git a/lisp/minibuffer.el b/lisp/minibuffer.el
> > index 3beb3c06a18..6fae62b3904 100644
> > --- a/lisp/minibuffer.el
> > +++ b/lisp/minibuffer.el
> > @@ -1141,7 +1141,15 @@ and DOC describes the way this style of completion works.")
> >      ;; and simply add "bar" to the end of the result.
> >      emacs22)
> >    "List of completion styles to use.
> > -The available styles are listed in `completion-styles-alist'.
> > +An element should be a symbol which is listed in
> > +`completion-styles-alist'.
> > +
> > +An element can also be a list of the form
> > +(STYLE ((VARIABLE VALUE) ...))
> > +STYLE must be a symbol listed in `completion-styles-alist', followed by
> > +a `let'-style list of variable/value pairs.  VARIABLE will be bound to
> > +VALUE (without evaluating it) while the style is handling completion.
> > +This allows repeating the same style with different configurations.
> 
> The :type of completion-styles should be adapted in order to reflect
> this change. Furthermore, this warrants a change of the :version, I
> believe.
> 
> Just being curious: why must it be a form (STYLE ((VARIABLE VALUE) ...)) ?
> Wouldn't it be sufficient to use (STYLE (VARIABLE VALUE) ...) ?
> 
> > +(defcustom completion-pcm-leading-wildcard nil
> > +  "If non-nil, partial-completion completes as if there's a leading wildcard.
> > +
> > +If nil (the default), partial-completion requires a matching completion
> > +alternative to have the same beginning as the first \"word\" in the
> > +minibuffer text, where \"word\" is determined by
> > +`completion-pcm-word-delimiters'.
> > +
> > +If non-nil, partial-completion allows any string of characters to occur
> > +at the beginning of a completion alternative, as if a wildcard such as
> > +\"*\" was present at the beginning of the minibuffer text.  This makes
> > +partial-completion behave more like the substring completion style."
> > +  :version "30.1"
> > +  :type 'boolean)
> 
> This doesn't look right, it should be :version "31.1".
> 
> Best regards, Michael.
> 
> 

Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

[Spencer Baugh via Emacs development discussions.]

[-- Attachment #1: Type: text/plain, Size: 2697 bytes --]

Eli Zaretskii <eliz@gnu.org> writes:
> Spencer, would you please take care of these gotchas?

The attached patch does so.

>> From: Michael Albinus <michael.albinus@gmx.de>
>> Cc: Spencer Baugh <sbaugh@janestreet.com>
>> Date: Sat, 24 Aug 2024 12:15:49 +0200
>> 
>> Eli Zaretskii <eliz@gnu.org> writes:
>> 
>> Hi,
>> 
>> > diff --git a/lisp/minibuffer.el b/lisp/minibuffer.el
>> > index 3beb3c06a18..6fae62b3904 100644
>> > --- a/lisp/minibuffer.el
>> > +++ b/lisp/minibuffer.el
>> > @@ -1141,7 +1141,15 @@ and DOC describes the way this style of completion works.")
>> >      ;; and simply add "bar" to the end of the result.
>> >      emacs22)
>> >    "List of completion styles to use.
>> > -The available styles are listed in `completion-styles-alist'.
>> > +An element should be a symbol which is listed in
>> > +`completion-styles-alist'.
>> > +
>> > +An element can also be a list of the form
>> > +(STYLE ((VARIABLE VALUE) ...))
>> > +STYLE must be a symbol listed in `completion-styles-alist', followed by
>> > +a `let'-style list of variable/value pairs.  VARIABLE will be bound to
>> > +VALUE (without evaluating it) while the style is handling completion.
>> > +This allows repeating the same style with different configurations.
>> 
>> The :type of completion-styles should be adapted in order to reflect
>> this change. Furthermore, this warrants a change of the :version, I
>> believe.

Fixed.

>> Just being curious: why must it be a form (STYLE ((VARIABLE VALUE) ...)) ?
>> Wouldn't it be sufficient to use (STYLE (VARIABLE VALUE) ...) ?

Both would work fine.

I thought the (style ((var val) (var val))) format would be slightly
more familiar for a novice, and slightly easier for us to change: we
could add (style ((var val)) something-else) later if we wanted.

But it would be easy to change it to (style (var val) (var val)) if
people prefer that.

>> > +(defcustom completion-pcm-leading-wildcard nil
>> > +  "If non-nil, partial-completion completes as if there's a leading wildcard.
>> > +
>> > +If nil (the default), partial-completion requires a matching completion
>> > +alternative to have the same beginning as the first \"word\" in the
>> > +minibuffer text, where \"word\" is determined by
>> > +`completion-pcm-word-delimiters'.
>> > +
>> > +If non-nil, partial-completion allows any string of characters to occur
>> > +at the beginning of a completion alternative, as if a wildcard such as
>> > +\"*\" was present at the beginning of the minibuffer text.  This makes
>> > +partial-completion behave more like the substring completion style."
>> > +  :version "30.1"
>> > +  :type 'boolean)
>> 
>> This doesn't look right, it should be :version "31.1".

Fixed. 


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-Update-completion-styles-defcustom-for-variable-over.patch --]
[-- Type: text/x-patch, Size: 4049 bytes --]

From 818332ad5069bd865a865f22ab5d5c2dda083049 Mon Sep 17 00:00:00 2001
From: Spencer Baugh <sbaugh@janestreet.com>
Date: Fri, 6 Sep 2024 13:12:52 -0400
Subject: [PATCH] Update completion-styles defcustom for variable overrides

In 69ec333eab0b801949d33ef5ae505addc9061793 I allowed
completion-styles to contain a list of bindings.  Now the defcustom
type also supports this.

Since the type is somewhat unusual (a value in the list can be either
a symbol or a list) I had to add a new widget to support it.

* lisp/minibuffer.el (completion--styles-type): Update to allow
setting variable overrides.
(completion-styles, completion-category-overrides)
(completion-pcm-leading-wildcard): Update :version.

* lisp/wid-edit.el (widget-single-or-list-to-internal,
single-or-list): Add.
---
 lisp/minibuffer.el | 12 +++++++-----
 lisp/wid-edit.el   | 24 ++++++++++++++++++++++++
 2 files changed, 31 insertions(+), 5 deletions(-)

diff --git a/lisp/minibuffer.el b/lisp/minibuffer.el
index 1efe71f10e3..d5f666dfecf 100644
--- a/lisp/minibuffer.el
+++ b/lisp/minibuffer.el
@@ -1119,8 +1119,10 @@ completion--update-styles-options
     widget))
 
 (defconst completion--styles-type
-  `(repeat :tag "insert a new menu to add more styles"
-           (choice :convert-widget completion--update-styles-options)))
+  '(repeat :tag "insert a new menu to add more styles"
+           (single-or-list
+            (choice :convert-widget completion--update-styles-options)
+            (repeat :tag "Variable overrides" (group variable sexp)))))
 
 (defconst completion--cycling-threshold-type
   '(choice (const :tag "No cycling" nil)
@@ -1154,7 +1156,7 @@ completion-styles
 Note that `completion-category-overrides' may override these
 styles for specific categories, such as files, buffers, etc."
   :type completion--styles-type
-  :version "23.1")
+  :version "31.1")
 
 (defvar completion-category-defaults
   '((buffer (styles . (basic substring)))
@@ -1205,7 +1207,7 @@ completion-category-overrides
 
 If a property in a category is specified by this variable, it
 overrides the default specified in `completion-category-defaults'."
-  :version "25.1"
+  :version "31.1"
   :type `(alist :key-type (choice :tag "Category"
 				  (const buffer)
                                   (const file)
@@ -3895,7 +3897,7 @@ completion-pcm-leading-wildcard
 at the beginning of a completion alternative, as if a wildcard such as
 \"*\" was present at the beginning of the minibuffer text.  This makes
 partial-completion behave more like the substring completion style."
-  :version "30.1"
+  :version "31.1"
   :type 'boolean)
 
 (defun completion-pcm--string->pattern (string &optional point)
diff --git a/lisp/wid-edit.el b/lisp/wid-edit.el
index e7e6351fcab..05c3f412bf6 100644
--- a/lisp/wid-edit.el
+++ b/lisp/wid-edit.el
@@ -3891,6 +3891,30 @@ widget-cons-match
   (and (consp value)
        (widget-group-match widget
 			   (widget-apply widget :value-to-internal value))))
+
+(defun widget-single-or-list-to-internal (widget val)
+  (if (listp val) val
+    (cons val (make-list (1- (length (widget-get widget :args))) nil))))
+
+(define-widget 'single-or-list 'group
+  "Either a single value (`nlistp') or a list of values (`listp').
+
+If the initial value is `nlistp', the first child widget gets
+that value and the other children get nil.
+
+If the first child's value is `nlistp' and the other children are
+nil, then `widget-value' just returns the first child's value."
+  ;; The internal value is always a list; only :value-to-internal and
+  ;; :match ever get called with the external value, which might be
+  ;; `nlistp'.
+  :value-to-external (lambda (_ val)
+                       (if (and (nlistp (car val))
+                                (cl-every #'null (cdr val)))
+                           (car val) val))
+  :value-to-internal #'widget-single-or-list-to-internal
+  :match (lambda (widget val)
+           (widget-group-match widget (widget-single-or-list-to-internal widget val))))
+
 \f
 ;;; The `lazy' Widget.
 ;;
-- 
2.39.3

Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

[Eli Zaretskii]

> Date: Fri, 06 Sep 2024 13:23:19 -0400
> From:  Spencer Baugh via "Emacs development discussions." <emacs-devel@gnu.org>
> Eli Zaretskii <eliz@gnu.org> writes:
> > Spencer, would you please take care of these gotchas?
> 
> The attached patch does so.

Thanks, installed.  But the new single-or-list feature should be
documented in the ELisp manual.

Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

[Eli Zaretskii]

> Date: Sat, 14 Sep 2024 11:47:42 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org
> 
> > Date: Fri, 06 Sep 2024 13:23:19 -0400
> > From:  Spencer Baugh via "Emacs development discussions." <emacs-devel@gnu.org>
> > Eli Zaretskii <eliz@gnu.org> writes:
> > > Spencer, would you please take care of these gotchas?
> > 
> > The attached patch does so.
> 
> Thanks, installed.  But the new single-or-list feature should be
> documented in the ELisp manual.

Ping! single-or-list is still undocumented, AFAICT.  We should
document it ASAP, as it is not good to have on the master branch
features which are not reflected in the documentation.

TIA

Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

[Spencer Baugh]

Eli Zaretskii <eliz@gnu.org> writes:
>> Date: Sat, 14 Sep 2024 11:47:42 +0300
>> From: Eli Zaretskii <eliz@gnu.org>
>> Cc: emacs-devel@gnu.org
>> 
>> > Date: Fri, 06 Sep 2024 13:23:19 -0400
>> > From:  Spencer Baugh via "Emacs development discussions." <emacs-devel@gnu.org>
>> > Eli Zaretskii <eliz@gnu.org> writes:
>> > > Spencer, would you please take care of these gotchas?
>> > 
>> > The attached patch does so.
>> 
>> Thanks, installed.  But the new single-or-list feature should be
>> documented in the ELisp manual.
>
> Ping! single-or-list is still undocumented, AFAICT.  We should
> document it ASAP, as it is not good to have on the master branch
> features which are not reflected in the documentation.
>
> TIA

I started writing documentation, but I came to the conclusion that most
users other than completion-styles should use the "group" widget
instead.  And documenting single-or-list would encourage using it
instead of group, which I don't think is correct.

Maybe this widget should just be moved to minibuffer.el?

Alternatively, if there are any other customizable variables or values
in Emacs which are sometimes a single value, and sometimes a list, then
this widget would be useful for those.  If we can think of any such
variables, then that would help justify the existence of single-or-list
(and I would convert their defcustoms to use single-or-list).  That
would help in writing documentation about when exactly to use it.

Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

[Eli Zaretskii]

> From: Spencer Baugh <sbaugh@janestreet.com>
> Cc: emacs-devel@gnu.org
> Date: Mon, 23 Sep 2024 15:36:10 -0400
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> >> Date: Sat, 14 Sep 2024 11:47:42 +0300
> >> From: Eli Zaretskii <eliz@gnu.org>
> >> Cc: emacs-devel@gnu.org
> >> 
> >> > Date: Fri, 06 Sep 2024 13:23:19 -0400
> >> > From:  Spencer Baugh via "Emacs development discussions." <emacs-devel@gnu.org>
> >> > Eli Zaretskii <eliz@gnu.org> writes:
> >> > > Spencer, would you please take care of these gotchas?
> >> > 
> >> > The attached patch does so.
> >> 
> >> Thanks, installed.  But the new single-or-list feature should be
> >> documented in the ELisp manual.
> >
> > Ping! single-or-list is still undocumented, AFAICT.  We should
> > document it ASAP, as it is not good to have on the master branch
> > features which are not reflected in the documentation.
> >
> > TIA
> 
> I started writing documentation, but I came to the conclusion that most
> users other than completion-styles should use the "group" widget
> instead.  And documenting single-or-list would encourage using it
> instead of group, which I don't think is correct.
> 
> Maybe this widget should just be moved to minibuffer.el?
> 
> Alternatively, if there are any other customizable variables or values
> in Emacs which are sometimes a single value, and sometimes a list, then
> this widget would be useful for those.  If we can think of any such
> variables, then that would help justify the existence of single-or-list
> (and I would convert their defcustoms to use single-or-list).  That
> would help in writing documentation about when exactly to use it.

I'm not sure I understand: are you saying that your change to
minibuffer.el can be rewritten to not use single-or-list?  If so, I'm
okay with that, and in that case we should simply remove
single-or-list, until there are some users of it.

Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

[Spencer Baugh]

[-- Attachment #1: Type: text/plain, Size: 2187 bytes --]

On Tue, Sep 24, 2024, 7:22 AM Eli Zaretskii <eliz@gnu.org> wrote:

> > From: Spencer Baugh <sbaugh@janestreet.com>
> > Cc: emacs-devel@gnu.org
> > Date: Mon, 23 Sep 2024 15:36:10 -0400
> >
> > Eli Zaretskii <eliz@gnu.org> writes:
> > >> Date: Sat, 14 Sep 2024 11:47:42 +0300
> > >> From: Eli Zaretskii <eliz@gnu.org>
> > >> Cc: emacs-devel@gnu.org
> > >>
> > >> > Date: Fri, 06 Sep 2024 13:23:19 -0400
> > >> > From:  Spencer Baugh via "Emacs development discussions." <
> emacs-devel@gnu.org>
> > >> > Eli Zaretskii <eliz@gnu.org> writes:
> > >> > > Spencer, would you please take care of these gotchas?
> > >> >
> > >> > The attached patch does so.
> > >>
> > >> Thanks, installed.  But the new single-or-list feature should be
> > >> documented in the ELisp manual.
> > >
> > > Ping! single-or-list is still undocumented, AFAICT.  We should
> > > document it ASAP, as it is not good to have on the master branch
> > > features which are not reflected in the documentation.
> > >
> > > TIA
> >
> > I started writing documentation, but I came to the conclusion that most
> > users other than completion-styles should use the "group" widget
> > instead.  And documenting single-or-list would encourage using it
> > instead of group, which I don't think is correct.
> >
> > Maybe this widget should just be moved to minibuffer.el?
> >
> > Alternatively, if there are any other customizable variables or values
> > in Emacs which are sometimes a single value, and sometimes a list, then
> > this widget would be useful for those.  If we can think of any such
> > variables, then that would help justify the existence of single-or-list
> > (and I would convert their defcustoms to use single-or-list).  That
> > would help in writing documentation about when exactly to use it.
>
> I'm not sure I understand: are you saying that your change to
> minibuffer.el can be rewritten to not use single-or-list?  If so, I'm
> okay with that, and in that case we should simply remove
> single-or-list, until there are some users of it.
>

No, it's still necessary for my changes in minibuffer.el.  But I don't
think it's useful outside that.

>

[-- Attachment #2: Type: text/html, Size: 3561 bytes --]

Re: master 69ec333eab0: Allow customizing partial-completion to be more like substring

[Eli Zaretskii]

> From: Spencer Baugh <sbaugh@janestreet.com>
> Date: Tue, 24 Sep 2024 07:32:22 -0400
> Cc: emacs-devel@gnu.org
> 
>  > I started writing documentation, but I came to the conclusion that most
>  > users other than completion-styles should use the "group" widget
>  > instead.  And documenting single-or-list would encourage using it
>  > instead of group, which I don't think is correct.
>  > 
>  > Maybe this widget should just be moved to minibuffer.el?
>  > 
>  > Alternatively, if there are any other customizable variables or values
>  > in Emacs which are sometimes a single value, and sometimes a list, then
>  > this widget would be useful for those.  If we can think of any such
>  > variables, then that would help justify the existence of single-or-list
>  > (and I would convert their defcustoms to use single-or-list).  That
>  > would help in writing documentation about when exactly to use it.
> 
>  I'm not sure I understand: are you saying that your change to
>  minibuffer.el can be rewritten to not use single-or-list?  If so, I'm
>  okay with that, and in that case we should simply remove
>  single-or-list, until there are some users of it.
> 
> No, it's still necessary for my changes in minibuffer.el.  But I don't think it's useful outside that.

OK, then I think moving the widget to minibuffer.el is indeed enough.