[RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark? (was: Fwd: Org-cite: Replace basic follow-processor with transient menu?)

[RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark? (was: Fwd: Org-cite: Replace basic follow-processor with transient menu?)

[Ihor Radchenko]

TL;DR: We are in the process of designing a more unified selection
interface for Org mode and want to see if there is some way to unify
context-menu-mode, transient, which-key and embark together. The idea is
to (1) avoid too many customizations; (2) allow users to decide how to
choose between multiple options - by mouse, keyboard, and using
customizable UIs.

Dear all,

I have raised the topic of refactoring Org mode menu systems during
EmacsConf in https://emacsconf.org/2024/talks/org-update/
The initial idea was replacing the self-written menu code in Org with
built-in transient.el.

Later, during OrgMeetup a number of people raised concerns that
transient may sometimes be an overkill, and that some people may prefer
alternative UIs. In particular, embark and context-menu-mode were
mentioned.

(I am CCing the discussion participants and potentially interested
maintainers)

In Org mode (although not only in Org mode, looking at the success of
embark.el), we often have a UI model where users call an "action"
command (like org-ctrl-c-ctrl-c or org-open-at-point) followed by
interactive selection among multiple actions.

For example, org-open-at-point on a heading with multiple links inside
will raise a section buffer listing all the links - pressing a number
will open the corresponding link.

Another example (see the example patch below), which is a
work-in-progress patch for Org citation system, is "following" a
citation. To "follow" citation may mean multiple things including but
not limited to: (1) going to citation record in the bibliography file;
(2) following URL; (3) downloading .pdf file for a citation; etc.
 **The list of "follow" actions may be customized by users**

The general UI flow in these scenarios will be:

1. User calls "action" with cursor at certain syntax element
2. Action menu is displayed, showing the available actions/targets (dynamically built)
3. User selects the action/target

This UI flow can be implemented using context menus, which-key popups,
transient menus, and also using embark (where the way menu is displayed
can be customized).

All the 4 approaches represent different UI models with various
strengths and weaknesses:

- transient has a very flexible layout builder where the menu items can
  be arranged granularly, but intercepts the main loop disrupting
  certain keyboard-based workflows
- which-key does not stand on the way and integrates well into Emacs'
  key binding model, but provides little flexibility for menu layout
- embark stays in the middle between which-key and transient, making use
  of transient keymaps and allowing a custom menu renderer
- context-menu-mode provides mouse experience

I am wondering if we can work out some universal API to plug the
described action->menu->selection model into the UI that user prefers.

Tentatively, I am thinking about the following:

For a given Emacs "prefix" command (e.g. org-open-at-point), we define a
set of customizations:

1. List of possible actions: ((name1 . action1 props) (name2 . action2 ...) ...)
   PROPS is a plist defining extra properties like key-binding, display
   string, maybe something else to be used in the future.
2. Menu interface to use (transient, context-menu, embark, which-key)
3. Layout settings for the specific interfaces. For example, transient
   layout definition.

WDYT?

Best,
Ihor

Tor-björn Claesson <tclaesson@gmail.com> writes:

> From 7e9e0c64fbda2dcb67d8c8421d1c9923ca93e8b4 Mon Sep 17 00:00:00 2001
> From: =?UTF-8?q?Tor-bj=C3=B6rn=20Claesson?= <torb@barbar.claesson.fi>
> Date: Tue, 12 Nov 2024 11:09:16 +0200
> Subject: [PATCH] lisp/oc-basic.el: Transient menu for following citations
>
> * lisp/oc-basic.el (require 'transient): Pull in transient.
> (require 'org-element): Pull in org-element.
> (org-cite-basic-follow-ask): New customization option. should
> `org-cite-basic-follow' prompt the user for an action?
> (org-cite-basic-follow-actions): New customization option, that
> specifies the contents of the transient menu.
> (org-cite-basic--get-key): New function. Get citation key from
> citation or citation reference.
> (org-cite-basic-follow): New function.  Displays a menu asking how to
> follow a citation if `org-cite-basic-follow-ask' is
> non-nil. Otherwise, it retains the default behaviour of opening the
> bibliography entry. This can be inversed with a negative prefix argument.
> (org-cite-basic-follow--parse-suffix-specification and
> org-cite-basic-follow--setup): Helper functions for
> `org-cite-basic-follow'.
> (org-cite-register-processor 'basic): Update the basic citation
> processor to follow citations using `org-cite-basic-follow'.
>
> * etc/ORG_NEWS (Menu for choosing how to follow citations): Describe
> the new feature
> (New option ~org-cite-basic-follow-ask~): Describe this new
> customization option.
> (New option ~org-cite-basic-follow-actions~): Describe this new
> customization option, which specifies the layout of the
> `org-cite-basic-follow' transient menu.
>
> This change was co-authored with much support from Ihor Radchenko and
> Jonas Bernoulli, thanks!
> ---
>  etc/ORG-NEWS     |  22 +++++++++
>  lisp/oc-basic.el | 115 +++++++++++++++++++++++++++++++++++++++++++----
>  2 files changed, 128 insertions(+), 9 deletions(-)
>
> diff --git a/etc/ORG-NEWS b/etc/ORG-NEWS
> index de4f11b25..bacc38be2 100644
> --- a/etc/ORG-NEWS
> +++ b/etc/ORG-NEWS
> @@ -114,6 +114,15 @@ The keybindings in the repeat-maps can be changed by customizing
>  
>  See the new [[info:org#Repeating commands]["Repeating commands"]] section in Org mode manual.
>  
> +*** New transient menu when following citations
> +
> +Following citations with the org-cite-basic citation backend can now present a
> +transient menu. To show this menu, set ~org-cite-basic-follow-ask~ to non-nil. 
> +This behaviour can be reversed with a -4 prefix.
> +
> +The contents of this menu can be customized in
> +~org-cite-basic-follow-actions~.
> +
>  ** New and changed options
>  
>  # Chanes deadling with changing default values of customizations,
> @@ -158,6 +167,19 @@ English.  The default value is ~t~ as the CSL standard assumes that
>  English titles are specified in sentence-case but the bibtex
>  bibliography format requires them to be written in title-case.
>  
> +*** New option ~org-cite-basic-follow-ask~
> +
> +When this option is non-nil, following a citation with the basic citation
> +backend will present a transient menu with choices for how to follow the
> +citation.
> +If nil, following a citation will open its bibliography entry.
> +
> +This behaviour can be reversed with a -4 prefix argument.
> +
> +*** New option ~org-cite-basic-follow-actions~
> +
> +This option specifies the options presented by ~org-cite-basic-follow~.
> +
>  ** New functions and changes in function arguments
>  
>  # This also includes changes in function behavior from Elisp perspective.
> diff --git a/lisp/oc-basic.el b/lisp/oc-basic.el
> index e207a1997..fc55d3a32 100644
> --- a/lisp/oc-basic.el
> +++ b/lisp/oc-basic.el
> @@ -74,6 +74,8 @@
>  (require 'map)
>  (require 'oc)
>  (require 'seq)
> +(require 'transient)
> +(require 'org-element)
>  
>  (declare-function org-open-at-point "org" (&optional arg))
>  (declare-function org-open-file "org" (path &optional in-emacs line search))
> @@ -140,6 +142,39 @@
>    :type 'face
>    :safe #'facep)
>  
> +(defcustom org-cite-basic-follow-ask nil
> +  "Should `org-cite-basic' ask how to follow citations?
> +
> +When this option is nil, `org-cite-basic-follow' opens the bibliography entry. 
> +Otherwise, `org-cite-basic-follow' will display a transient menu prompting the 
> +user for an action.  The contents of this menu can be customized in 
> +`org-cite-basic-follow-actions'."
> +  :group 'org-cite
> +  :package-version '(Org . "9.8")
> +  :type 'boolean)
> +
> +(defcustom org-cite-basic-follow-actions
> +  '[["Open"
> +     ("b" "bibliography entry" (org-cite-basic-goto !citation !prefix))]
> +    ["Copy"
> +     ("d" "DOI"
> +      (kill-new
> +       (org-cite-basic--get-field
> +        'doi
> +        (org-cite-basic--get-key !citation))))]]
> +  "Actions in the `org-cite-basic-follow' transient menu.
> +
> +This option uses the same syntax as `transient-define-prefix', see Info node
> +`(transient)Binding Suffix and Infix Commands'.  In addition, it is possible 
> +to specify a function call for the COMMAND part, where !citation (the citation
> +object to be followed) and !prefix (any prefix argument to the follower) can be
> +used to access those values. For example:
> +(org-cite-basic-goto !citation !prefix) or
> +(lambda () (message (org-element-property :key !citation)))"
> +  :group 'org-cite
> +  :package-version '(Org . "9.8")
> +  :type 'sexp)
> +
>  \f
>  ;;; Internal variables
>  (defvar org-cite-basic--bibliography-cache nil
> @@ -326,6 +361,16 @@ INFO is the export state, as a property list."
>                  (map-keys entries))
>                (org-cite-basic--parse-bibliography)))
>  
> +(defun org-cite-basic--get-key (citation-or-citation-reference)
> +  "Return citation key for CITATION."
> +  (if (org-element-type-p citation-or-citation-reference 'citation-reference)
> +      (org-element-property :key citation-or-citation-reference)
> +    (pcase (org-cite-get-references citation-or-citation-reference t)
> +      (`(,key) key)
> +      (keys
> +       (or (completing-read "Select citation key: " keys nil t)
> +           (user-error "Aborted"))))))
> +
>  (defun org-cite-basic--get-entry (key &optional info)
>    "Return BibTeX entry for KEY, as an association list.
>  When non-nil, INFO is the export state, as a property list."
> @@ -805,14 +850,7 @@ export state, as a property list."
>  When DATUM is a citation reference, open bibliography entry referencing
>  the citation key.  Otherwise, select which key to follow among all keys
>  present in the citation."
> -  (let* ((key
> -          (if (org-element-type-p datum 'citation-reference)
> -              (org-element-property :key datum)
> -            (pcase (org-cite-get-references datum t)
> -              (`(,key) key)
> -              (keys
> -               (or (completing-read "Select citation key: " keys nil t)
> -                   (user-error "Aborted"))))))
> +  (let* ((key (org-cite-basic--get-key datum))
>           (file
>            (pcase (seq-find (pcase-lambda (`(,_ . ,entries))
>                               (gethash key entries))
> @@ -832,6 +870,65 @@ present in the citation."
>         (bibtex-set-dialect)
>         (bibtex-search-entry key)))))
>  
> +(transient-define-prefix org-cite-basic-follow (citation-object &optional prefix)
> +  "Follow citation.
> +
> +If `org-cite-basic-follow-ask' is non-nil, this transient will present
> +a menu prompting the user for an action. 
> +Otherwise, it will open the bibliography entry for the citation at point.  
> +This behaviour is inverted when the transient is called with a -4 prefix
> +argument.
> +
> +The contents of the menu are defined in the variable
> +`org-cite-basic-follow-actions'."
> +  [:class transient-columns
> +          :setup-children org-cite-basic-follow--setup
> +          :pad-keys t]
> +  (interactive
> +    (list (let ((obj (org-element-context)))
> +           (pcase (org-element-type obj)
> +             ((or 'citation 'citation-reference) obj)
> +             (_ (user-error "No citation at point"))))))
> +  (if (xor org-cite-basic-follow-ask
> +           (equal prefix '(-4)))
> +      (transient-setup 'org-cite-basic-follow nil nil
> +                       :scope (list :citation citation-object :prefix prefix))
> +    (org-cite-basic-goto citation-object prefix)))
> +
> +(defun org-cite-basic-follow--parse-suffix-specification (specification)
> +  "Handle special syntax for `org-cite-basic-follow-actions'."
> +  (pcase specification
> +    (`(,key ,desc (lambda ,args . ,fn-args) . ,other)
> +     `(,key ,desc
> +            (lambda ,args
> +              ,(unless (and (listp (car fn-args))
> +                            (equal (caar fn-args)
> +                                   'interactive))
> +                 '(interactive))
> +              (let ((!citation (plist-get (transient-scope) :citation))
> +                    (!prefix (plist-get (transient-scope) :prefix)))
> +                ,@fn-args))
> +            ,@other))
> +    (`(,key ,desc (,fn . ,fn-args) . ,other)
> +     `(,key ,desc
> +            (lambda ()
> +	      (interactive)
> +              (let ((!citation (plist-get (transient-scope) :citation))
> +                    (!prefix (plist-get (transient-scope) :prefix)))
> +	        (,fn ,@fn-args)))
> +            ,@other))
> +    (other other)))
> +
> +(defun org-cite-basic-follow--setup (_)
> +  "Update `org-cite-basic-follow' when `org-cite-basic-follow-actions' changes."
> +  (transient-parse-suffixes
> +   'org-cite-basic-follow
> +   (cl-map 'vector
> +           (lambda (group)
> +             (cl-map 'vector #'org-cite-basic-follow--parse-suffix-specification
> +                     group))
> +           org-cite-basic-follow-actions)))
> +
>  \f
>  ;;; "Insert" capability
>  (defun org-cite-basic--complete-style (_)
> @@ -920,7 +1017,7 @@ Raise an error when no bibliography is set in the buffer."
>    :activate #'org-cite-basic-activate
>    :export-citation #'org-cite-basic-export-citation
>    :export-bibliography #'org-cite-basic-export-bibliography
> -  :follow #'org-cite-basic-goto
> +  :follow #'org-cite-basic-follow
>    :insert (org-cite-make-insert-processor #'org-cite-basic--complete-key
>                                            #'org-cite-basic--complete-style)
>    :cite-styles
> -- 
> 2.46.0
>

-- 
Ihor Radchenko // yantar92,
Org mode maintainer,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>

Re: [RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark? (was: Fwd: Org-cite: Replace basic follow-processor with transient menu?)

[Psionic K]

[-- Attachment #1: Type: text/plain, Size: 4497 bytes --]

> intercepts the main loop
This is optional, per transient menu (prefix) and the commands within it.
A common technique, used by magit and others, is to have entry point
commands in the regular keymap so that many commands can be started without
traversing several menus.  If you want normal un-shadowed bindings active
at the same time, the prefix has a slot called `transient-non-suffix' that
is similar to the `:suppress' option in keymaps or setting a `t'
`undefined' binding in a keymap.  However the results of mixing self-insert
and modal or modal and another modal are generally bad.

The biggest drawbacks of transient are well known in Github issues:
- Which-key discovery of Transient keymaps
- Transient binding generation based on keymaps
- Normalizing how to obtain arguments when being called independently as an
interactive command versus being called as a Transient suffix

In the short term, to punch the first two problems in the face, override
the `:setup-children' method.  If you know what keymap you are borrowing
bindings from, you can synchronize it at display time.

Regarding the normalization with interactive, if you are not using
transient infixes and instead lean on the :info class and dynamic
:descriptions, you can display state and store it using normal buffer-local
and defvar techniques, providing visual feedback for what might be hidden
states after the user gets more familiar.  The commands work with or
without a prefix currently active.  In this usage model, you only use
Transient for its flow control, display, and layout.  I find the infix
system to be somewhat of a distraction if you are not actually building a
CLI wrapper, but you can definitely make suffixes and descriptions "smart"
by reading a scope from the prefix and making custom infixes that also
obtain more state when displayed.  A custom infix for storing org elements
or objects could certainly be a thing.

I think deeper user customization is an area that is weak with transient,
but only because the user actually needs to have a vision for how they want
to build up stateful commands.  If you're just doing prefix maps, transient
and hydra are equivalent concepts.  Transient becomes differentiated when
you consider commands that build up state for other commands.  Executing
slightly modified command sentences in rapid succession is not something
the user customizes casually.  Complex commands only make sense when the
context they depend on is populated, which naturally decides the menu flow.

> I am wondering if we can work out some universal API to plug the
> described action->menu->selection model into the UI that user prefers.
>
> Tentatively, I am thinking about the following:
>
> For a given Emacs "prefix" command (e.g. org-open-at-point), we define a
> set of customizations:
>
> 1. List of possible actions: ((name1 . action1 props) (name2 . action2
...) ...)
>    PROPS is a plist defining extra properties like key-binding, display
>    string, maybe something else to be used in the future.
> 2. Menu interface to use (transient, context-menu, embark, which-key)
> 3. Layout settings for the specific interfaces. For example, transient
>    layout definition.

Well, I'm sure you know that transient has more decisions encoded in the
layout than the other options.  If the data going in is a least common
denominator, you need supplementary data elsewhere to achieve a good result.

What I fear is a system like org-speed-keys which relies on an override of
`org-self-insert' and is yet another orthogonal system.  I much prefer the
Lispy style of integration, which uses a keymap.  Using keymaps, even if
they are not active, to generate transient key bindings via :setup-children
is the best way to have certain integration with other Emacs tools.

How people can collaborate with me on general questions of design is to
open issues on the Transient Showcase.  Either I can point to an existing
example or make a new one.  I've been giving some thought to how to
demonstrate an idea more generally of composing multiple commands and
manipulating the composition to dispatch complex commands in rapid
succession with minor differences.  I personally have my own org speed keys
solution that I've been developing for yet another more complex package I
call Afterburner.  These projects can become stuck in design hell when I
don't have the prodding from other problem analysis, so please, bother me.

https://github.com/positron-solutions/transient-showcase

[-- Attachment #2: Type: text/html, Size: 4850 bytes --]

Re: [RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark? (was: Fwd: Org-cite: Replace basic follow-processor with transient menu?)

[Panayotis Manganaris]

Hello. I'm very much an elisp amateur, but one of my main motives for
learning has been to make use of emacs' many beautiful UI systems.

Therefore, I'll provide my comments from a designer perspective and
avoid implementation detail.

Ihor Radchenko <yantar92@posteo.net> writes:

> TL;DR: We are in the process of designing a more unified selection
> interface for Org mode

Thanks to Tor-björn, Jonas, Ihor, and everyone else for this work. I've
learned much by following this thread.

> a number of people raised concerns that transient may sometimes be an
> overkill

I agree with others that transient is overkill for most of the menus in
org-mode. A counter argument, going by the patches exchanged here, might
be that transient menus are a clean and scalable approach.

I've not tried the in-dev transient follow processor, but I imagine it
is similar to the old org-ref (John Kitchin) hydra menu. I used org-ref
for years, and enjoyed how its menu united many features, including
- a system analogous to org-cite follow processors for navigating to sources/notes
- a system akin to reftex-mode for org
- a system for downloading new references and making new bibliography entries

With this level of complexity, a transient menu seems like the modern
weapon of choice for a similar feature set. Only, wait, I stopped using
org-ref.

Long story short, I concluded that treating org-mode like a nexus for
these intertwined features was, at least, encouraging my bad habit of
treating org-mode as an "everything app."

I have since delegated the task of populating my personal library to a
dedicated reference manager. I started using latex-mode and reftex to
write papers, thus gaining more typesetting control and making version
management much easier. I still use org to take notes, but now enjoy a
much more stable and focused configuration.

I tell this story as an (anecdotal) case study for how a powerful menu
can be less productive in the long run. Org-ref's hydra menu is not to
blame for my bad decisions but, at least in-part, I used it (and even
extended it) because a greater man put it there.

Scalable design is an invitation to scale. Perhaps the wise know better.
Anyway, the simple tasks of navigating from a citation link to a copy of
the original document, its URL, or an associated note is achievable with
a completing-read (helm, counsel, vertico, etc.) interface.

Transient is overkill.

> This UI flow can be implemented using context menus [...] embark

I like org's context menus, but I like embark better. I do think it
would be very generous of org to delegate its own contextual actions to
embark if so configured.

I'll elaborate:

As is, I like to run org-babel code block functions through embark, I
only wish there were more of them to run. I find this to be far superior
in every way to binding equivalent key chords for the mode.

Also, for me, embark-bindings and embark-prefix-help-command have
completely replaced which-key for discovering key bindings on-the-fly.

> I am wondering if we can work out some universal API to plug the
> described action->menu->selection model into the UI that user prefers.

I support this idea in principle.

> 1. List of possible actions: ((name1 . action1 props) (name2 . action2 ...) ...)
>    PROPS is a plist defining extra properties like key-binding, display
>    string, maybe something else to be used in the future.

I would focus on associating just enough metadata with functions to
enable using them with embark. I really have no idea what the best way
to do this is. This seems to be the biggest missing piece.

Maybe in the short term we just crowd-source some embark configuration
on WORG.

Menus should of course be useful without embark. If the goal here is to
deprecate many of org's ad-hoc interfaces, I think the natural default
replacement for most is completing-read. I may be completely wrong.

> 2. Menu interface to use (transient, context-menu, embark, which-key)

It seems like Tor-björn and Ihor have already laid the groundwork for
using transient in this way. So, I'll just say bravo. Regardless of my
opinion on transient overuse, it looks like great work.

On the other hand, e.g. if the org export dispatch menu isn't transient,
it should be. Seems obvious to me.

> 3. Layout settings for the specific interfaces. For example, transient
>    layout definition.

It is important to avoid making more org-mode-specific UI abstractions.
Whatever the solution, it is important that the UI implementations
tapped to take over help share the maintenance burden.

In this way, I reckon most interface customization can be done in hooks.

> WDYT?

Ideally org-mode uses established APIs rather than making another.

Thanks for reading, and thanks for the chance to contribute.

Re: [RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark? (was: Fwd: Org-cite: Replace basic follow-processor with transient menu?)

[Ihor Radchenko]

Psionic K <psionik@positron.solutions> writes:

>> intercepts the main loop

> This is optional, per transient menu (prefix) and the commands within it.
> A common technique, used by magit and others, is to have entry point
> commands in the regular keymap so that many commands can be started without
> traversing several menus.  If you want normal un-shadowed bindings active
> at the same time, the prefix has a slot called `transient-non-suffix' that
> is similar to the `:suppress' option in keymaps or setting a `t'
> `undefined' binding in a keymap.  However the results of mixing self-insert
> and modal or modal and another modal are generally bad.

Thanks for the info!
So, we can have something like

:transient-non-suffix 'leave

and then pressing something that is not bound to a suffix or infix will
run the parent keymap command, automatically leaving transient state.

> - Normalizing how to obtain arguments when being called independently as an
> interactive command versus being called as a Transient suffix

I think it is addressed in the example patch I shared. There, we pass
around the original function arguments via macro expansion (!arg-name)

  ["Open" ("b" "bibliography entry" (org-cite-basic-goto !citation !prefix))]

> In the short term, to punch the first two problems in the face, override
> the `:setup-children' method.  If you know what keymap you are borrowing
> bindings from, you can synchronize it at display time.

This is also partially solved. We do use :setup-children, although the
initial implementation simply reads user customization into menu layout.

I believe that we can read a keymap in similar way and generate
transient layout automatically.

> What I fear is a system like org-speed-keys which relies on an override of
> `org-self-insert' and is yet another orthogonal system.  I much prefer the
> Lispy style of integration, which uses a keymap.  Using keymaps, even if
> they are not active, to generate transient key bindings via :setup-children
> is the best way to have certain integration with other Emacs tools.

May you please elaborate?

-- 
Ihor Radchenko // yantar92,
Org mode maintainer,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>

Re: [RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark? (was: Fwd: Org-cite: Replace basic follow-processor with transient menu?)

[Ihor Radchenko]

Panayotis Manganaris <panos.manganaris@gmail.com> writes:

> Menus should of course be useful without embark. If the goal here is to
> deprecate many of org's ad-hoc interfaces, I think the natural default
> replacement for most is completing-read. I may be completely wrong.

For the existing interfaces, we are not going to change things in
incompatible ways. The idea is to keep the existing menus working as
similarly as possible compared to the current ad-hoc menus.
So, no, we are not going to replace things with completing-read.
Not by default.

But what I hope to see is a way for users to customize the UI to their
preference. Without touching the defaults.

> On the other hand, e.g. if the org export dispatch menu isn't transient,
> it should be. Seems obvious to me.

Yes. That's where we will have to use transient.
This discussion mostly concerns simpler cases when user needs to choose
from a list.

-- 
Ihor Radchenko // yantar92,
Org mode maintainer,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>

Re: [RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark?

[Philip Kaludercic]

Can I please be removed from the CCs in this thread?

Ihor Radchenko <yantar92@posteo.net> writes:

> Psionic K <psionik@positron.solutions> writes:
>
>>> intercepts the main loop
>
>> This is optional, per transient menu (prefix) and the commands within it.
>> A common technique, used by magit and others, is to have entry point
>> commands in the regular keymap so that many commands can be started without
>> traversing several menus.  If you want normal un-shadowed bindings active
>> at the same time, the prefix has a slot called `transient-non-suffix' that
>> is similar to the `:suppress' option in keymaps or setting a `t'
>> `undefined' binding in a keymap.  However the results of mixing self-insert
>> and modal or modal and another modal are generally bad.
>
> Thanks for the info!
> So, we can have something like
>
> :transient-non-suffix 'leave
>
> and then pressing something that is not bound to a suffix or infix will
> run the parent keymap command, automatically leaving transient state.
>
>> - Normalizing how to obtain arguments when being called independently as an
>> interactive command versus being called as a Transient suffix
>
> I think it is addressed in the example patch I shared. There, we pass
> around the original function arguments via macro expansion (!arg-name)
>
>   ["Open" ("b" "bibliography entry" (org-cite-basic-goto !citation !prefix))]
>
>> In the short term, to punch the first two problems in the face, override
>> the `:setup-children' method.  If you know what keymap you are borrowing
>> bindings from, you can synchronize it at display time.
>
> This is also partially solved. We do use :setup-children, although the
> initial implementation simply reads user customization into menu layout.
>
> I believe that we can read a keymap in similar way and generate
> transient layout automatically.
>
>> What I fear is a system like org-speed-keys which relies on an override of
>> `org-self-insert' and is yet another orthogonal system.  I much prefer the
>> Lispy style of integration, which uses a keymap.  Using keymaps, even if
>> they are not active, to generate transient key bindings via :setup-children
>> is the best way to have certain integration with other Emacs tools.
>
> May you please elaborate?

Re: [RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark? (was: Fwd: Org-cite: Replace basic follow-processor with transient menu?)

[Karthik Chikmagalur]

> Regarding the normalization with interactive, if you are not using
> transient infixes and instead lean on the :info class and dynamic
> :descriptions, you can display state and store it using normal buffer-local
> and defvar techniques, providing visual feedback for what might be hidden
> states after the user gets more familiar.  The commands work with or
> without a prefix currently active.  In this usage model, you only use
> Transient for its flow control, display, and layout.  I find the infix
> system to be somewhat of a distraction if you are not actually building a
> CLI wrapper, but you can definitely make suffixes and descriptions "smart"
> by reading a scope from the prefix and making custom infixes that also
> obtain more state when displayed.  A custom infix for storing org elements
> or objects could certainly be a thing.

Do you have an example of using buffer-local variables to store
transient state?  Directing me to an appropriate section of the
showcase should be plenty.

Karthik

Re: [RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark? (was: Fwd: Org-cite: Replace basic follow-processor with transient menu?)

[Karthik Chikmagalur]

>> intercepts the main loop
> This is optional, per transient menu (prefix) and the commands within it.
> A common technique, used by magit and others, is to have entry point
> commands in the regular keymap so that many commands can be started without
> traversing several menus.  If you want normal un-shadowed bindings active
> at the same time, the prefix has a slot called `transient-non-suffix' that
> is similar to the `:suppress' option in keymaps or setting a `t'
> `undefined' binding in a keymap.  However the results of mixing self-insert
> and modal or modal and another modal are generally bad.

From the manual page describing the `transient-non-suffix' slot:

> If the value is ‘t’, then non-suffixes can be invoked, when it is
> ‘nil’ (the default) then they cannot be invoked.
>
> The only other recommended value is ‘leave’.  If that is used, then
> non-suffixes can be invoked, but if one is invoked, then that exits the
> transient.

This is one of my issues with Transient -- here it is reimplementing the
idea of a regular transient keymap (the kind you can use with
`set-transient-map') but increasing the cognitive load of applying this
behavior.

I understand that Transient needs to implement behaviors not provided by
Elisp (like infixes), but fashioning whole-cloth replacements for
behaviors that are provided -- like the KEEP-PRED behavior of
`set-transient-map' -- makes writing Transients a very high effort
activity.

Karthik

Re: [RFC] The best way to choose an "action" at point: context-menu-mode, transient, which-key or embark? (was: Fwd: Org-cite: Replace basic follow-processor with transient menu?)

[Psionic K]

> Do you have an example of using buffer-local variables to store
> transient state?  Directing me to an appropriate section of the
> showcase should be plenty.

It's less magic than it sounds.  I basically bypassed the infix system
when making transients for Master of Ceremonies.
https://github.com/positron-solutions/moc

`moc-dispatch` is a kind of utility dashboard for screen casting.
Besides controls, it displays various states you might not see in the
minibuffer or elsewhere.  I used dynamic :description functions rather
than infixes to display these states.  It is much simpler to bring
Elisp states into the interface this way.  There's no need and there
can be no meaning of persisting the current state of Emacs.

> like the KEEP-PRED behavior of `set-transient-map'

There's more.  Calling (or declining to call) `setup-prefix' manually
in an interactive form or using the stack manipulation commands can
switch between menus and pure interactive flows.  It's possible to
construct flows that are more program-driven than simple KEEP-PRED or
repeat maps.  Obviously we can hack them both since they are functions
that can manipulate the downstream flow.  Transient's pre-commands are
a bit complex to understand when hacking in behavior but are related
to its menu stack, something transient maps don't need to express.

Instead of talking about Transient's second system behaviors, we can
identify and fix the worst ones.  Jonas is aware of them and generally
in favor of normalizing how transient meshes with interactive, as long
as it doesn't break magit and all the other dependents.

> > org-speed-keys
> May you please elaborate?

I discovered org speed keys because I was making my own speed-key
system and came across the shadowed bindings.  The bindings are always
shadowed, even when speed keys are off.  Unlike Lispy, where I
customize the shadowing using keymaps, org speed keys has
`org-speed-commands'.  From a discoverability standpoint, it breaks
some things.  The commands are opaque.  All I see is `org-self-insert'
whereas my own bindings have unique command names on every key,
another thing I copied from Lispy.  Even if they have DWIM or
situational behavior, it is easier to identify all behavior when
starting from an entry point that doesn't implement details of other
commands.