set-transient-map and a question

set-transient-map and a question

[T.V Raman]

set-transient-map is used in text-scale-adjust here:
https://git.savannah.gnu.org/cgit/emacs.git/tree/lisp/face-remap.el?h=master#n423...

But it looks like it passes nil as the keep-pressed arg which sounds
counter to the documentation for set-transient-map; but
text-scale-adjust does work as intended  -- so am I reading the
set-transient-map doc incorrectly?
-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

--

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

Re: set-transient-map and a question

[Robert Pluim]

>>>>> On Mon, 10 Jul 2023 18:12:18 -0700, "T.V Raman" <raman@google.com> said:

    T> set-transient-map is used in text-scale-adjust here:
    T> https://git.savannah.gnu.org/cgit/emacs.git/tree/lisp/face-remap.el?h=master#n423...

    T> But it looks like it passes nil as the keep-pressed arg which sounds
    T> counter to the documentation for set-transient-map; but
    T> text-scale-adjust does work as intended  -- so am I reading the
    T> set-transient-map doc incorrectly?

All the bindings in that map in `text-scale-adjust' call
`text-scale-adjust', which then results in `set-transient-mapʼ being
called again, which means you donʼt need any logic to de-activate the
transient keymap when someone presses a non-text-scale-adjust key.

(Iʼve just realized `text-scale-adjust' actually uses a minor-mode,
which seems like overkill, but it works)

Robert
-- 

Re: set-transient-map and a question

[T.V Raman]

interesting; so what is the point of keep-pressed as an argument to
set-transient-map?


Robert Pluim writes:
 > >>>>> On Mon, 10 Jul 2023 18:12:18 -0700, "T.V Raman" <raman@google.com> said:
 > 
 >     T> set-transient-map is used in text-scale-adjust here:
 >     T> https://git.savannah.gnu.org/cgit/emacs.git/tree/lisp/face-remap.el?h=master#n423...
 > 
 >     T> But it looks like it passes nil as the keep-pressed arg which sounds
 >     T> counter to the documentation for set-transient-map; but
 >     T> text-scale-adjust does work as intended  -- so am I reading the
 >     T> set-transient-map doc incorrectly?
 > 
 > All the bindings in that map in `text-scale-adjust' call
 > `text-scale-adjust', which then results in `set-transient-mapʼ being
 > called again, which means you donʼt need any logic to de-activate the
 > transient keymap when someone presses a non-text-scale-adjust key.
 > 
 > (Iʼve just realized `text-scale-adjust' actually uses a minor-mode,
 > which seems like overkill, but it works)
 > 
 > Robert
 > -- 

-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

--

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

Re: set-transient-map and a question

[Robert Pluim]

>>>>> On Tue, 11 Jul 2023 07:09:13 -0700, "T.V Raman" <raman@google.com> said:

    T> interesting; so what is the point of keep-pressed as an argument to
    T> set-transient-map?

As far as I can tell itʼs used for things related to mouse-dragging
and modal commands like `indent-rigidly', where keys not in the
transient map exit the mode.

Robert
-- 

Re: set-transient-map and a question

[T.V Raman]

Interesting. I played with  it a bit more, and setting keep-pressed to
t in my code got the other part that was failing for me, the on-exit
arg being passed as a function.

It works now, ie I have it playing a sound on exit and that now works;
earlier it was playing even when the transient hadn't exitted.


Robert Pluim writes:
 > >>>>> On Tue, 11 Jul 2023 07:09:13 -0700, "T.V Raman" <raman@google.com> said:
 > 
 >     T> interesting; so what is the point of keep-pressed as an argument to
 >     T> set-transient-map?
 > 
 > As far as I can tell itʼs used for things related to mouse-dragging
 > and modal commands like `indent-rigidly', where keys not in the
 > transient map exit the mode.
 > 
 > Robert
 > -- 

-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

--

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

Re: set-transient-map and a question

[Robert Pluim]

>>>>> On Tue, 11 Jul 2023 10:16:56 -0700, "T.V Raman" <raman@google.com> said:

    T> Interesting. I played with  it a bit more, and setting keep-pressed to
    T> t in my code got the other part that was failing for me, the on-exit
    T> arg being passed as a function.

Based on glancing at the implementation, `on-exit' requires
`keep-pred' non-nil, which isnʼt 100% clear from the docstring.

And I think the `keep-pred' as a function logic is inverted. But
thatʼs not something I have time to look at right noww.

Robert
-- 

Re: set-transient-map and a question

[T.V Raman]

Thanks; I have it doing what I want now -- but perhaps it might be
worth logging a bug for the docs? I live purely in emacs -- but can
report the bug using email, but it might be clearer if filed directly
in the bug-tracker? (not sure)



Robert Pluim writes:
 > >>>>> On Tue, 11 Jul 2023 10:16:56 -0700, "T.V Raman" <raman@google.com> said:
 > 
 >     T> Interesting. I played with  it a bit more, and setting keep-pressed to
 >     T> t in my code got the other part that was failing for me, the on-exit
 >     T> arg being passed as a function.
 > 
 > Based on glancing at the implementation, `on-exit' requires
 > `keep-pred' non-nil, which isnʼt 100% clear from the docstring.
 > 
 > And I think the `keep-pred' as a function logic is inverted. But
 > thatʼs not something I have time to look at right noww.
 > 
 > Robert
 > -- 

-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

--

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

Re: set-transient-map and a question

[Bob Rogers]

   From: "T.V Raman" <raman@google.com>
   Date: Tue, 11 Jul 2023 10:53:20 -0700

   Thanks; I have it doing what I want now -- but perhaps it might be
   worth logging a bug for the docs? I live purely in emacs -- but can
   report the bug using email, but it might be clearer if filed directly
   in the bug-tracker? (not sure)

Filing a bug via email is 100% equivalent; that's the way I always do
it.

					-- Bob Rogers
					   http://www.rgrjr.com/

Re: set-transient-map and a question

[Robert Pluim]

>>>>> On Tue, 11 Jul 2023 11:10:10 -0700, Bob Rogers <rogers@rgrjr.com> said:

    Bob>    From: "T.V Raman" <raman@google.com>
    Bob>    Date: Tue, 11 Jul 2023 10:53:20 -0700

    Bob>    Thanks; I have it doing what I want now -- but perhaps it might be
    Bob>    worth logging a bug for the docs? I live purely in emacs -- but can
    Bob>    report the bug using email, but it might be clearer if filed directly
    Bob>    in the bug-tracker? (not sure)

    Bob> Filing a bug via email is 100% equivalent; that's the way I always do
    Bob> it.

I donʼt think a bug is needed here. How does this sound:

diff --git a/lisp/subr.el b/lisp/subr.el
index fac5fd58092..6ac040fe5d3 100644
--- a/lisp/subr.el
+++ b/lisp/subr.el
@@ -6343,7 +6343,8 @@ set-transient-map
 if it returns non-nil, then MAP stays active.
 
 Optional arg ON-EXIT, if non-nil, specifies a function that is
-called, with no arguments, after MAP is deactivated.
+called, with no arguments, after MAP is deactivated (this happens
+only if KEEP-PRED or TIMEOUT are non-nil).
 
 Optional arg MESSAGE, if non-nil, requests display of an informative
 message after activating the transient map.  If MESSAGE is a string,

Robert
-- 

Re: set-transient-map and a question

[T.V Raman]

this is perfect, Thanks Robert!



Robert Pluim writes:
 > >>>>> On Tue, 11 Jul 2023 11:10:10 -0700, Bob Rogers <rogers@rgrjr.com> said:
 > 
 >     Bob>    From: "T.V Raman" <raman@google.com>
 >     Bob>    Date: Tue, 11 Jul 2023 10:53:20 -0700
 > 
 >     Bob>    Thanks; I have it doing what I want now -- but perhaps it might be
 >     Bob>    worth logging a bug for the docs? I live purely in emacs -- but can
 >     Bob>    report the bug using email, but it might be clearer if filed directly
 >     Bob>    in the bug-tracker? (not sure)
 > 
 >     Bob> Filing a bug via email is 100% equivalent; that's the way I always do
 >     Bob> it.
 > 
 > I donʼt think a bug is needed here. How does this sound:
 > 
 > diff --git a/lisp/subr.el b/lisp/subr.el
 > index fac5fd58092..6ac040fe5d3 100644
 > --- a/lisp/subr.el
 > +++ b/lisp/subr.el
 > @@ -6343,7 +6343,8 @@ set-transient-map
 >  if it returns non-nil, then MAP stays active.
 >  
 >  Optional arg ON-EXIT, if non-nil, specifies a function that is
 > -called, with no arguments, after MAP is deactivated.
 > +called, with no arguments, after MAP is deactivated (this happens
 > +only if KEEP-PRED or TIMEOUT are non-nil).
 >  
 >  Optional arg MESSAGE, if non-nil, requests display of an informative
 >  message after activating the transient map.  If MESSAGE is a string,
 > 
 > Robert
 > -- 

-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

--

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

Re: set-transient-map and a question

[Robert Pluim]

>>>>> On Tue, 11 Jul 2023 13:53:03 -0700, "T.V Raman" <raman@google.com> said:

    T> this is perfect, Thanks Robert!

Except, based on testing, itʼs not true :-)

Iʼll work on the docstring some more.

Robert
-- 

Re: set-transient-map and a question

[T.V Raman]

:-) I didn't test that because I wasn't binding the modifier letters;
one of my current goals with emacs/emacspeak is to minimize chording
-- if interested, see the article on xcape in the  Emacspeak blog:
https://emacspeak.blogspot.com/2023/02/enhance-emacs-ergonomics-under-x-happy.html

Robert Pluim writes:
 > >>>>> On Tue, 11 Jul 2023 13:53:03 -0700, "T.V Raman" <raman@google.com> said:
 > 
 >     T> this is perfect, Thanks Robert!
 > 
 > Except, based on testing, itʼs not true :-)
 > 
 > Iʼll work on the docstring some more.
 > 
 > Robert
 > -- 

-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

--

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
♉ Id: kg:/m/0285kf1  🦮

Re: set-transient-map and a question

[Michael Heerdegen]

Robert Pluim <rpluim@gmail.com> writes:

>  Optional arg ON-EXIT, if non-nil, specifies a function that is
> -called, with no arguments, after MAP is deactivated.
> +called, with no arguments, after MAP is deactivated (this happens
> +only if KEEP-PRED or TIMEOUT are non-nil).

Doesn't CLEARFUN call ON-EXIT even when KEEP-PRED and TIMEOUT are both
nil?

Michael.

Re: set-transient-map and a question

[T.V Raman]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=gb18030, Size: 478 bytes --]

From my experimentation, it appeared (needs confirmation) that if
kee-pressed was nil, the exit-func never got called -- even though the
function called itself recursively -- ie if keep-pressed is nil, and it
calls itself, the behavior is indistinguishable from keep-pressed = t. I
fell in this hole by mimicking the implementation of text-scale-adjust
as a starter.

-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
7©4 Id: kg:/m/0285kf1  •0Ü8

Re: set-transient-map and a question

[Michael Heerdegen]

"T.V Raman" <raman@google.com> writes:

> From my experimentation, it appeared (needs confirmation) that if
> kee-pressed was nil, the exit-func never got called

The code says that EXIT-FUN is always called.  Here is a simple test
case:

(set-transient-map emacs-lisp-mode-map nil
  (lambda () (message "!!!")))

Works as expected here: after hitting a random key I get the expected
`message'

> -- even though the function called itself recursively -- ie if
> keep-pressed is nil, and it calls itself, the behavior is
> indistinguishable from keep-pressed = t.

Note that nested transient maps are handled specially: the first (or
"outer") ones wait until the newest one is gone.  Maybe you hit that
case.

Michael.

Re: set-transient-map and a question

[Michael Heerdegen]

Michael Heerdegen <michael_heerdegen@web.de> writes:

> Note that nested transient maps are handled specially: the first (or
> "outer") ones wait until the newest one is gone.  Maybe you hit that
> case.

Note that this doesn't happen for `eq' maps.  But if your recursive
calls install a new transient map that is only `equal' to the first one
it's treated as an additional transient map hiding the old one.

Michael.

Re: set-transient-map and a question

[T.V Raman]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=gb18030, Size: 786 bytes --]

Michael Heerdegen <michael_heerdegen@web.de> writes:
For me, hitting c-x c-e at the end of that code while in the gnus
article buffer did not appear to work, it continued to call keys in the
buffer-local gnus article mode.

> Michael Heerdegen <michael_heerdegen@web.de> writes:
>
>> Note that nested transient maps are handled specially: the first (or
>> "outer") ones wait until the newest one is gone.  Maybe you hit that
>> case.
>
> Note that this doesn't happen for `eq' maps.  But if your recursive
> calls install a new transient map that is only `equal' to the first one
> it's treated as an additional transient map hiding the old one.
>
> Michael.
>
>

-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
7©4 Id: kg:/m/0285kf1  •0Ü8

Re: set-transient-map and a question

[Michael Heerdegen]

"T.V Raman" <raman@google.com> writes:

> For me, hitting c-x c-e at the end of that code while in the gnus
> article buffer did not appear to work, it continued to call keys in the
> buffer-local gnus article mode.

Works for me.  C-c C-e C-c C-b tries to compile the article buffer, for
example.  It was a silly example, of course, but it should work.

There are not many bindings in `emacs-lisp-mode-map'.  You really tried
one of those keys _immediately_ after evaluating the `set-transient-map'
expression (note that the transient map is gone after one entering one
key binding not present in the map)?

Michael.

Re: set-transient-map and a question

[Robert Pluim]

>>>>> On Wed, 12 Jul 2023 05:39:48 +0200, Michael Heerdegen <michael_heerdegen@web.de> said:

    Michael> "T.V Raman" <raman@google.com> writes:
    >> For me, hitting c-x c-e at the end of that code while in the gnus
    >> article buffer did not appear to work, it continued to call keys in the
    >> buffer-local gnus article mode.

    Michael> Works for me.  C-c C-e C-c C-b tries to compile the article buffer, for
    Michael> example.  It was a silly example, of course, but it should work.

    Michael> There are not many bindings in `emacs-lisp-mode-map'.  You really tried
    Michael> one of those keys _immediately_ after evaluating the `set-transient-map'
    Michael> expression (note that the transient map is gone after one entering one
    Michael> key binding not present in the map)?

I tried the following in "*scratch*", and I get "deactivated" in
*Messages* after hitting "+" (or indeed any key).

    (set-transient-map
     (let ((map (make-sparse-keymap)))
       (define-key map "+" (lambda () (interactive) (message "whoohoo")))
       map)
     nil (lambda () (message "deactivated")))

which means my proposed doc update is wrong.

Moral: test before documenting, not after :-)

Robert
--