Strange behavior with major mode specific key bindings

Strange behavior with major mode specific key bindings

[Tim Johnson]

I have customized a key binding for jinja2 mode.

;; In my init.el I have the following
(require 'jinja2-mode)
(require 'tj-mode-custom)
(add-to-list 'auto-mode-alist '("\\.html$" . jinja2-mode))
(add-hook 'jinja2-mode-hook 'tj-jinja2-mode-keys)

;; In  tj-mode-custom the following

(defun tj-jinja2-mode-keys ()
   "Bindings and for jinja2"
   (define-prefix-command 'jinja2-mode-map)
   (local-set-key (kbd "C-c") 'jinja2-mode-map)
   (define-key jinja2-mode-map "a" 'jinja2-close-tag))

If an .html file is loaded at emacs startup, the following key sequence:

C-c a

invokes jinja2-close-tag

BUT if I open an .html file that is on disk but not preloaded at 
startup, OR if I create a new .html file, the single character "a" will 
execute the same function.

In other words, now both "C-c a" and "a" are bound to 'jinja2-close-tag, 
clobbering 'insert-char

I can observe that the newly opened files are in jinja2 mode.

I do not have similar issues with elpy mode

I am using GNU Emacs 26.3 (build 1, x86_64-pc-linux-gnu, GTK+ Version 
2.24.30)

On ubuntu 16.04

thanks


-- 
Tim
tj49.com

Re: Strange behavior with major mode specific key bindings

[Emanuel Berg via Users list for the GNU Emacs text editor]

Tim Johnson wrote:

> (defun tj-jinja2-mode-keys ()
>   "Bindings and for jinja2"
>   (define-prefix-command 'jinja2-mode-map)
>   (local-set-key (kbd "C-c") 'jinja2-mode-map)
>   (define-key jinja2-mode-map "a" 'jinja2-close-tag))

... why is it called the same thing? :O

And why set locally for a mode?

And isnt' C-c already a prefix command?

???

This works for me. Just change the Gnus map for the
real, intended, unquoted keymap, and change
the function.

Doesn't that work, no hook needed?

  (define-key gnus-group-mode-map "\C-ca" #'gnus-group-unsubscribe-current-group)

-- 
underground experts united
http://user.it.uu.se/~embe8573
https://dataswamp.org/~incal

Re: Strange behavior with major mode specific key bindings

[Tim Johnson]

On 4/24/20 2:36 PM, Emanuel Berg via Users list for the GNU Emacs text 
editor wrote:
> Tim Johnson wrote:
>
>> (defun tj-jinja2-mode-keys ()
>>    "Bindings and for jinja2"
>>    (define-prefix-command 'jinja2-mode-map)
>>    (local-set-key (kbd "C-c") 'jinja2-mode-map)
>>    (define-key jinja2-mode-map "a" 'jinja2-close-tag))
> ... why is it called the same thing? :O
>
> And why set locally for a mode?
>
> And isnt' C-c already a prefix command?
>
> ???
>
> This works for me. Just change the Gnus map for the
> real, intended, unquoted keymap, and change
> the function.
I followed your pattern and it works. Thank you very much.
> Doesn't that work, no hook needed?
Hook **is** needed, else another function is executed with  "c-c a"
>
>    (define-key gnus-group-mode-map "\C-ca" #'gnus-group-unsubscribe-current-group)

Good work, quick reply. Much gratitude.

cheers

>
-- 
Tim
tj49.com

Re: Strange behavior with major mode specific key bindings

[Emanuel Berg via Users list for the GNU Emacs text editor]

Tim Johnson wrote:

> I followed your pattern and it works. Thank you

np :)

> Hook **is** needed, else another function is executed
> with  "c-c a"

Wierd. Maybe another hook resets it :)

-- 
underground experts united
http://user.it.uu.se/~embe8573
https://dataswamp.org/~incal

Re: Strange behavior with major mode specific key bindings

[Tim Johnson]

On 4/24/20 3:07 PM, Emanuel Berg via Users list for the GNU Emacs text 
editor wrote:
> Tim Johnson wrote:
>
>> I followed your pattern and it works. Thank you
> np :)
>
>> Hook **is** needed, else another function is executed
>> with  "c-c a"
> Wierd. Maybe another hook resets it :)
>
Its org-agenda (found in global-map),

-- 
Tim
tj49.com

Re: Strange behavior with major mode specific key bindings

[Emanuel Berg via Users list for the GNU Emacs text editor]

Tim Johnson wrote:

>>> Hook **is** needed, else another function is
>>> executed with  "c-c a"
>>
>> Wierd. Maybe another hook resets it :)
>>
>
> Its org-agenda (found in global-map),

Case closed :)

-- 
underground experts united
http://user.it.uu.se/~embe8573
https://dataswamp.org/~incal