info-lookup-alist mangled, breaks C-h S

info-lookup-alist mangled, breaks C-h S

[T.V Raman]

This appears to happen when packages from elpa get loaded best I can
tell, but it's proven difficult/impossible to track down.

Running C-h S  info-lookup-symbol
throws an error about an invalid symbol that appears to come from an
entry of the form
(emacs-lisp-mode . "woman") that appears inside info-lookup-alist --
but it's impossible to track down who/what is putting that entry in
there.



-- 

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: info-lookup-alist mangled, breaks C-h S

[Daniel Martín]

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

> This appears to happen when packages from elpa get loaded best I can
> tell, but it's proven difficult/impossible to track down.
>
> Running C-h S  info-lookup-symbol
> throws an error about an invalid symbol that appears to come from an
> entry of the form
> (emacs-lisp-mode . "woman") that appears inside info-lookup-alist --
> but it's impossible to track down who/what is putting that entry in
> there.

You could try M-x debug-on-variable-change RET info-lookup-alist RET,
and then load your ELPA packages to debug which one is changing that
variable in an incorrect way.

Re: info-lookup-alist mangled, breaks C-h S

[T.V Raman]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=gb18030, Size: 1111 bytes --]

Daniel Mart¨ªn <mardani29@yahoo.es> writes:

tried that already -- leads down various twisty ratholes since too many
things in emacs core itself touch that variable, and some of that code
was beyond my abilities, eg see the code in tramp-integration.el

-- that one appears to be an earlier offendor with respect to adding an
invalid entry.
> "T.V Raman" <raman@google.com> writes:
>
>> This appears to happen when packages from elpa get loaded best I can
>> tell, but it's proven difficult/impossible to track down.
>>
>> Running C-h S  info-lookup-symbol
>> throws an error about an invalid symbol that appears to come from an
>> entry of the form
>> (emacs-lisp-mode . "woman") that appears inside info-lookup-alist --
>> but it's impossible to track down who/what is putting that entry in
>> there.
>
> You could try M-x debug-on-variable-change RET info-lookup-alist RET,
> and then load your ELPA packages to debug which one is changing that
> variable in an incorrect way.
>

-- 

Thanks,

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

Re: info-lookup-alist mangled, breaks C-h S

[Tassilo Horn]

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

> tried that already -- leads down various twisty ratholes since too many
> things in emacs core itself touch that variable, and some of that code
> was beyond my abilities, eg see the code in tramp-integration.el
>
> -- that one appears to be an earlier offendor with respect to adding an
> invalid entry.

FWIW, the problematic entry doesn't seem to come from some external
package, at least I can reproduce the error using

  1. emacs -Q
  2. M-x fundamental-mode RET in *stratch*
  3. C-h S

with GNU Emacs 29.0.50 of 2022-01-03.  The step 2 above is needed
because `C-h S' in an emacs-lisp-mode or lisp-interaction-mode buffer
works just fine whereas I get the error in fundamental-mode or
Info-mode.  I can't reproduce with emacs 27 or 28.

I'd suggest you do M-x report-emacs-bug.

Bye,
Tassilo

Re: info-lookup-alist mangled, breaks C-h S

[Eli Zaretskii]

> From: Tassilo Horn <tsdh@gnu.org>
> Date: Mon, 03 Jan 2022 21:02:05 +0100
> Cc: emacs-devel@gnu.org, Daniel Martín <mardani29@yahoo.es>
> 
> "T.V Raman" <raman@google.com> writes:
> 
> > tried that already -- leads down various twisty ratholes since too many
> > things in emacs core itself touch that variable, and some of that code
> > was beyond my abilities, eg see the code in tramp-integration.el
> >
> > -- that one appears to be an earlier offendor with respect to adding an
> > invalid entry.
> 
> FWIW, the problematic entry doesn't seem to come from some external
> package, at least I can reproduce the error using
> 
>   1. emacs -Q
>   2. M-x fundamental-mode RET in *stratch*
>   3. C-h S
> 
> with GNU Emacs 29.0.50 of 2022-01-03.  The step 2 above is needed
> because `C-h S' in an emacs-lisp-mode or lisp-interaction-mode buffer
> works just fine whereas I get the error in fundamental-mode or
> Info-mode.  I can't reproduce with emacs 27 or 28.

Does the patch below give good results?

diff --git a/lisp/info-look.el b/lisp/info-look.el
index e6f267d..9224aa3 100644
--- a/lisp/info-look.el
+++ b/lisp/info-look.el
@@ -347,7 +347,10 @@ info-lookup-select-mode
 
 (defun info-lookup-change-mode (topic)
   (let* ((completions (mapcar (lambda (arg)
-				(cons (symbol-name (car arg)) (car arg)))
+                                (let ((mode-spec (car arg)))
+                                  (and (consp mode-spec)
+                                       (setq mode-spec (car mode-spec)))
+				  (cons (symbol-name mode-spec) mode-spec)))
 			      (info-lookup->topic-value topic)))
 	 (mode (completing-read
 		(format "Use %s help mode: " topic)

Re: info-lookup-alist mangled, breaks C-h S

[Eli Zaretskii]

Lars, your commit f839c7f, viz.:

  Author:     Lars Ingebrigtsen <larsi@gnus.org>
  AuthorDate: Sat Oct 9 14:44:21 2021 +0200
  Commit:     Lars Ingebrigtsen <larsi@gnus.org>
  CommitDate: Sat Oct 9 14:44:21 2021 +0200

      Make `C-h S' work on symbols from most of the manuals in Emacs

      * lisp/info-look.el (lambda): Add `C-h S' lookup for all the
      "misc" manuals in Emacs (bug#42753).

introduced a new form for HELP-MODE in info-look up-alist:

+(mapc
+ (lambda (elem)
+   (let* ((prefix (car elem)))
+     (info-lookup-add-help
+      :mode (cons 'emacs-lisp-mode prefix)  <<<<<<<<<<<<<<<<<

Now HELP-MODE could be a cons cell, not just a symbol.  But the doc
string of info-lookup-alist was not updated, and therefore the
semantics of this form is not documented.  What does that cons cell
value mean?

Re: info-lookup-alist mangled, breaks C-h S

[T.V Raman]

Will file the bug report from a  bare-boned emacs later today.

I suspect info-look could do with a major overhaul; some of the list
surgery that modules like tramp-integration.el do to that list look
scarily unmaintainable
Tassilo Horn writes:
 > "T.V Raman" <raman@google.com> writes:
 > 
 > > tried that already -- leads down various twisty ratholes since too many
 > > things in emacs core itself touch that variable, and some of that code
 > > was beyond my abilities, eg see the code in tramp-integration.el
 > >
 > > -- that one appears to be an earlier offendor with respect to adding an
 > > invalid entry.
 > 
 > FWIW, the problematic entry doesn't seem to come from some external
 > package, at least I can reproduce the error using
 > 
 >   1. emacs -Q
 >   2. M-x fundamental-mode RET in *stratch*
 >   3. C-h S
 > 
 > with GNU Emacs 29.0.50 of 2022-01-03.  The step 2 above is needed
 > because `C-h S' in an emacs-lisp-mode or lisp-interaction-mode buffer
 > works just fine whereas I get the error in fundamental-mode or
 > Info-mode.  I can't reproduce with emacs 27 or 28.
 > 
 > I'd suggest you do M-x report-emacs-bug.
 > 
 > Bye,
 > Tassilo

-- 

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: info-lookup-alist mangled, breaks C-h S

[Tassilo Horn]

Eli Zaretskii <eliz@gnu.org> writes:

Hi Eli,

>> FWIW, the problematic entry doesn't seem to come from some external
>> package, at least I can reproduce the error using
>> 
>>   1. emacs -Q
>>   2. M-x fundamental-mode RET in *stratch*
>>   3. C-h S
>> 
>> with GNU Emacs 29.0.50 of 2022-01-03.  The step 2 above is needed
>> because `C-h S' in an emacs-lisp-mode or lisp-interaction-mode buffer
>> works just fine whereas I get the error in fundamental-mode or
>> Info-mode.  I can't reproduce with emacs 27 or 28.
>
> Does the patch below give good results?

Yes, it does, thanks!

Bye,
Tassilo

> diff --git a/lisp/info-look.el b/lisp/info-look.el
> index e6f267d..9224aa3 100644
> --- a/lisp/info-look.el
> +++ b/lisp/info-look.el
> @@ -347,7 +347,10 @@ info-lookup-select-mode
>  
>  (defun info-lookup-change-mode (topic)
>    (let* ((completions (mapcar (lambda (arg)
> -				(cons (symbol-name (car arg)) (car arg)))
> +                                (let ((mode-spec (car arg)))
> +                                  (and (consp mode-spec)
> +                                       (setq mode-spec (car mode-spec)))
> +				  (cons (symbol-name mode-spec) mode-spec)))
>  			      (info-lookup->topic-value topic)))
>  	 (mode (completing-read
>  		(format "Use %s help mode: " topic)

Re: info-lookup-alist mangled, breaks C-h S

[T.V Raman]

OK, that's great, I'll wait for that patch to land and wont load the
system with a bug report:-)
Happy New Year all!

Tassilo Horn writes:
 > Eli Zaretskii <eliz@gnu.org> writes:
 > 
 > Hi Eli,
 > 
 > >> FWIW, the problematic entry doesn't seem to come from some external
 > >> package, at least I can reproduce the error using
 > >> 
 > >>   1. emacs -Q
 > >>   2. M-x fundamental-mode RET in *stratch*
 > >>   3. C-h S
 > >> 
 > >> with GNU Emacs 29.0.50 of 2022-01-03.  The step 2 above is needed
 > >> because `C-h S' in an emacs-lisp-mode or lisp-interaction-mode buffer
 > >> works just fine whereas I get the error in fundamental-mode or
 > >> Info-mode.  I can't reproduce with emacs 27 or 28.
 > >
 > > Does the patch below give good results?
 > 
 > Yes, it does, thanks!
 > 
 > Bye,
 > Tassilo
 > 
 > > diff --git a/lisp/info-look.el b/lisp/info-look.el
 > > index e6f267d..9224aa3 100644
 > > --- a/lisp/info-look.el
 > > +++ b/lisp/info-look.el
 > > @@ -347,7 +347,10 @@ info-lookup-select-mode
 > >  
 > >  (defun info-lookup-change-mode (topic)
 > >    (let* ((completions (mapcar (lambda (arg)
 > > -				(cons (symbol-name (car arg)) (car arg)))
 > > +                                (let ((mode-spec (car arg)))
 > > +                                  (and (consp mode-spec)
 > > +                                       (setq mode-spec (car mode-spec)))
 > > +				  (cons (symbol-name mode-spec) mode-spec)))
 > >  			      (info-lookup->topic-value topic)))
 > >  	 (mode (completing-read
 > >  		(format "Use %s help mode: " topic)

-- 

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: info-lookup-alist mangled, breaks C-h S

[Eli Zaretskii]

> From: Tassilo Horn <tsdh@gnu.org>
> Cc: raman@google.com, emacs-devel@gnu.org, mardani29@yahoo.es
> Date: Tue, 04 Jan 2022 17:19:33 +0100
> 
> >>   1. emacs -Q
> >>   2. M-x fundamental-mode RET in *stratch*
> >>   3. C-h S
> >> 
> >> with GNU Emacs 29.0.50 of 2022-01-03.  The step 2 above is needed
> >> because `C-h S' in an emacs-lisp-mode or lisp-interaction-mode buffer
> >> works just fine whereas I get the error in fundamental-mode or
> >> Info-mode.  I can't reproduce with emacs 27 or 28.
> >
> > Does the patch below give good results?
> 
> Yes, it does, thanks!

Thanks for testing.  Will install, once Lars helps me figure out how
to fix the doc string of info-lookup-alist.

Re: info-lookup-alist mangled, breaks C-h S

[Eli Zaretskii]

> Date: Tue, 04 Jan 2022 18:46:59 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org, mardani29@yahoo.es, raman@google.com
> 
> > >>   1. emacs -Q
> > >>   2. M-x fundamental-mode RET in *stratch*
> > >>   3. C-h S
> > >> 
> > >> with GNU Emacs 29.0.50 of 2022-01-03.  The step 2 above is needed
> > >> because `C-h S' in an emacs-lisp-mode or lisp-interaction-mode buffer
> > >> works just fine whereas I get the error in fundamental-mode or
> > >> Info-mode.  I can't reproduce with emacs 27 or 28.
> > >
> > > Does the patch below give good results?
> > 
> > Yes, it does, thanks!
> 
> Thanks for testing.  Will install, once Lars helps me figure out how
> to fix the doc string of info-lookup-alist.

I went ahead and installed it.

Re: info-lookup-alist mangled, breaks C-h S

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> +(mapc
> + (lambda (elem)
> +   (let* ((prefix (car elem)))
> +     (info-lookup-add-help
> +      :mode (cons 'emacs-lisp-mode prefix)  <<<<<<<<<<<<<<<<<
>
> Now HELP-MODE could be a cons cell, not just a symbol.  But the doc
> string of info-lookup-alist was not updated, and therefore the
> semantics of this form is not documented.  What does that cons cell
> value mean?

The doc string change you've done looks correct to me.

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

Re: info-lookup-alist mangled, breaks C-h S

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: emacs-devel@gnu.org
> Date: Wed, 05 Jan 2022 16:33:48 +0100
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > +(mapc
> > + (lambda (elem)
> > +   (let* ((prefix (car elem)))
> > +     (info-lookup-add-help
> > +      :mode (cons 'emacs-lisp-mode prefix)  <<<<<<<<<<<<<<<<<
> >
> > Now HELP-MODE could be a cons cell, not just a symbol.  But the doc
> > string of info-lookup-alist was not updated, and therefore the
> > semantics of this form is not documented.  What does that cons cell
> > value mean?
> 
> The doc string change you've done looks correct to me.

Great, thanks.

Re: info-lookup-alist mangled, breaks C-h S

[T.V Raman]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=gb18030, Size: 884 bytes --]

Lars Ingebrigtsen <larsi@gnus.org> writes:


I still wonder whether the complex list surgery performed by modules
like tramp-integration is truly needed in current Emacs, or if it's a
throw-back from the distant past that has been amplified by
copy/paste/edit cycles.

> Eli Zaretskii <eliz@gnu.org> writes:
>
>> +(mapc
>> + (lambda (elem)
>> +   (let* ((prefix (car elem)))
>> +     (info-lookup-add-help
>> +      :mode (cons 'emacs-lisp-mode prefix)  <<<<<<<<<<<<<<<<<
>>
>> Now HELP-MODE could be a cons cell, not just a symbol.  But the doc
>> string of info-lookup-alist was not updated, and therefore the
>> semantics of this form is not documented.  What does that cons cell
>> value mean?
>
> The doc string change you've done looks correct to me.

-- 

Thanks,

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