dabbrev-expand for the minibuffer

dabbrev-expand for the minibuffer

[Dan Nicolaescu]

A lot of times when using dabbrev-expand in the minibuffer, the most
desirable expansion is something from the minibuffer history.
But dabbrev-expand does not search the minibuffer history...
If for example you use M-x compile a lot with a lot of very complex
command lines, having dabbrev-expand look in the minibuffer history is
very useful.
(This is not unlike what you suffer from when using bash: no dabbrev-expand at all).

So a long time ago I wrote this hack to put the minibuffer history in a
buffer and tell dabbrev to search it too.

(setq dabbrev-select-buffers-function 'my-minibuffer-setup-dabbrev-buffers)

(defun my-minibuffer-setup-dabbrev-buffers ()
  (let ((buf (get-buffer-create " *dabbrev history* ")))
    (save-excursion 
      (set-buffer buf)
      (erase-buffer)
      ;; dabbrev looks for text only in buffers.  Put the minibuffer
      ;; history in a buffer, so that dabbrev can look at it.
      (mapc (lambda (arg) 
            (insert arg "\n")) 
                (if minibuffer-history-variable 
                    (eval minibuffer-history-variable)
                          minibuffer-history)))
    ;; Put the minibuffer history at the beginning of the buffer list
    ;; that dabbrev will look at.
    (append (list buf) (dabbrev--select-buffers))))

I am not sure what is the proper way to integrate this functionality,
it would be great if someone that knows dabbrev could either add
something similar to this code, or do it in a better way and provide it
by default.  

Thanks

        --dan

Re: dabbrev-expand for the minibuffer

[Leo]

On 2008-02-23 22:59 +0000, Dan Nicolaescu wrote:
> A lot of times when using dabbrev-expand in the minibuffer, the most
> desirable expansion is something from the minibuffer history.
> But dabbrev-expand does not search the minibuffer history...
> If for example you use M-x compile a lot with a lot of very complex
> command lines, having dabbrev-expand look in the minibuffer history is
> very useful.
> (This is not unlike what you suffer from when using bash: no dabbrev-expand at all).
>
> So a long time ago I wrote this hack to put the minibuffer history in a
> buffer and tell dabbrev to search it too.
>
> (setq dabbrev-select-buffers-function 'my-minibuffer-setup-dabbrev-buffers)
>
> (defun my-minibuffer-setup-dabbrev-buffers ()
>   (let ((buf (get-buffer-create " *dabbrev history* ")))
>     (save-excursion 
>       (set-buffer buf)
>       (erase-buffer)
>       ;; dabbrev looks for text only in buffers.  Put the minibuffer
>       ;; history in a buffer, so that dabbrev can look at it.
>       (mapc (lambda (arg) 
>             (insert arg "\n")) 
>                 (if minibuffer-history-variable 
>                     (eval minibuffer-history-variable)
>                           minibuffer-history)))
>     ;; Put the minibuffer history at the beginning of the buffer list
>     ;; that dabbrev will look at.
>     (append (list buf) (dabbrev--select-buffers))))
>
> I am not sure what is the proper way to integrate this functionality,
> it would be great if someone that knows dabbrev could either add
> something similar to this code, or do it in a better way and provide it
> by default.  

It looks like the best way to integrate this is with hippie-expand,
probably not dabbrev-expand. Just a thought.

>
> Thanks
>
>         --dan

-- 
.:  Leo  :.  [ sdl.web AT gmail.com ]  .:  [ GPG Key: 9283AA3F ]  :.

          Use the best OS -- http://www.fedoraproject.org/

Re: dabbrev-expand for the minibuffer

[Juri Linkov]

> I am not sure what is the proper way to integrate this functionality,
> it would be great if someone that knows dabbrev could either add
> something similar to this code, or do it in a better way and provide it
> by default.

Good idea.  dabbrev.el has already two customizable options:

           dabbrev-check-all-buffers
           dabbrev-check-other-buffers

so it could also provide a new option:

           dabbrev-check-minibuffers

with the default value = t, to search the history of minibuffers.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: dabbrev-expand for the minibuffer

[Bastien]

Juri Linkov <juri@jurta.org> writes:

>> I am not sure what is the proper way to integrate this functionality,
>> it would be great if someone that knows dabbrev could either add
>> something similar to this code, or do it in a better way and provide it
>> by default.
>
> Good idea.  dabbrev.el has already two customizable options:
>
>            dabbrev-check-all-buffers
>            dabbrev-check-other-buffers
>
> so it could also provide a new option:
>
>            dabbrev-check-minibuffers

I would also add `dabbrev-check-kill-ring'.  

I often kill a whole region and expect dabbrev (or hippie-expand) to
provide completion on words I've been writing in the current session,
even if I deleted them.

-- 
Bastien

Re: dabbrev-expand for the minibuffer

[Johan Bockgård]

Dan Nicolaescu <dann@ics.uci.edu> writes:

> (This is not unlike what you suffer from when using bash: no
> dabbrev-expand at all).

    `dynamic-complete-history (M-<TAB>)'
         Attempt completion on the text before point, comparing the text
         against lines from the history list for possible completion
         matches.

-- 
Johan Bockgård

Re: dabbrev-expand for the minibuffer

[Dan Nicolaescu]

bojohan+news@dd.chalmers.se (Johan Bockgård) writes:

  > Dan Nicolaescu <dann@ics.uci.edu> writes:
  > 
  > > (This is not unlike what you suffer from when using bash: no
  > > dabbrev-expand at all).
  > 
  >     `dynamic-complete-history (M-<TAB>)'
  >          Attempt completion on the text before point, comparing the text
  >          against lines from the history list for possible completion
  >          matches.

Nah, this is far away from dabbrev-expand.

Re: dabbrev-expand for the minibuffer

[Johan Bockgård]

Dan Nicolaescu <dann@ics.uci.edu> writes:

> bojohan+news@dd.chalmers.se (Johan Bockgård) writes:
>
>   > Dan Nicolaescu <dann@ics.uci.edu> writes:
>   >
>   > > (This is not unlike what you suffer from when using bash: no
>   > > dabbrev-expand at all).
>   >
>   >     `dynamic-complete-history (M-<TAB>)'
>   >          Attempt completion on the text before point, comparing the text
>   >          against lines from the history list for possible completion
>   >          matches.
>
> Nah, this is far away from dabbrev-expand.

Well, it completes against input history (not output).  This is
comparable to dabbrev-expand for minibuffer history.

-- 
Johan Bockgård

Re: dabbrev-expand for the minibuffer

[Dan Nicolaescu]

bojohan+news@dd.chalmers.se (Johan Bockgård) writes:

  > Dan Nicolaescu <dann@ics.uci.edu> writes:
  > 
  > > bojohan+news@dd.chalmers.se (Johan Bockgård) writes:
  > >
  > >   > Dan Nicolaescu <dann@ics.uci.edu> writes:
  > >   >
  > >   > > (This is not unlike what you suffer from when using bash: no
  > >   > > dabbrev-expand at all).
  > >   >
  > >   >     `dynamic-complete-history (M-<TAB>)'
  > >   >          Attempt completion on the text before point, comparing the text
  > >   >          against lines from the history list for possible completion
  > >   >          matches.
  > >
  > > Nah, this is far away from dabbrev-expand.
  > 
  > Well, it completes against input history (not output).  This is
  > comparable to dabbrev-expand for minibuffer history.

Yeah, just in about the same way next-complete-history-element is
comparable to dabbrev-expand.

Re: dabbrev-expand for the minibuffer

[Johan Bockgård]

Dan Nicolaescu <dann@ics.uci.edu> writes:

> Yeah, just in about the same way next-complete-history-element is
> comparable to dabbrev-expand.

No, that is a misunderstanding.  dynamic-complete-history works anywhere
in a line, not just at the beginning.

-- 
Johan Bockgård

Re: dabbrev-expand for the minibuffer

[Dan Nicolaescu]

bojohan+news@dd.chalmers.se (Johan Bockgård) writes:

  > Dan Nicolaescu <dann@ics.uci.edu> writes:
  > 
  > > Yeah, just in about the same way next-complete-history-element is
  > > comparable to dabbrev-expand.
  > 
  > No, that is a misunderstanding.  dynamic-complete-history works anywhere
  > in a line, not just at the beginning.

Yeah, and that is the only difference.  The rest is similar, and not really
that similar to dabbrev-expand.  
Can we please stop wasting time on this?

Re: dabbrev-expand for the minibuffer

[Juri Linkov]

>> I am not sure what is the proper way to integrate this functionality,
>> it would be great if someone that knows dabbrev could either add
>> something similar to this code, or do it in a better way and provide it
>> by default.
>
> It looks like the best way to integrate this is with hippie-expand,
> probably not dabbrev-expand. Just a thought.

It would be easier to implement this in hippie-expand because it has all
necessary infrastructure.  This could be implemented like an existing
method `try-expand-dabbrev-from-kill' added to `hippie-expand-try-functions-list',
but using `minibuffer-history' instead of `kill-ring'.

OTOH, dabbrev.el has the TODO item "Check the kill-ring when everything
else fails."  Implementing this means practically copying code from
hippie-exp.el to dabbrev.el.  I'm not sure if two identical packages
is the goal?

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: dabbrev-expand for the minibuffer

[Leo]

On 2008-02-27 22:49 +0000, Juri Linkov wrote:
>>> I am not sure what is the proper way to integrate this functionality,
>>> it would be great if someone that knows dabbrev could either add
>>> something similar to this code, or do it in a better way and provide it
>>> by default.
>>
>> It looks like the best way to integrate this is with hippie-expand,
>> probably not dabbrev-expand. Just a thought.
>
> It would be easier to implement this in hippie-expand because it has all
> necessary infrastructure.  This could be implemented like an existing
> method `try-expand-dabbrev-from-kill' added to `hippie-expand-try-functions-list',
> but using `minibuffer-history' instead of `kill-ring'.
>
> OTOH, dabbrev.el has the TODO item "Check the kill-ring when everything
> else fails."  Implementing this means practically copying code from
> hippie-exp.el to dabbrev.el.  I'm not sure if two identical packages
> is the goal?

The TODO even suggests it is for hippie-exp.

BTW, it may be better rebind M-/ to hippie-expand with
hippie-expand-try-functions-list set to a proper value so that the
change is transparent to end users.

-- 
.:  Leo  :.  [ sdl.web AT gmail.com ]  .:  [ GPG Key: 9283AA3F ]  :.

          Use the best OS -- http://www.fedoraproject.org/

Re: dabbrev-expand for the minibuffer

[Miles Bader]

Juri Linkov <juri@jurta.org> writes:
> It would be easier to implement this in hippie-expand because it has all
> necessary infrastructure.

What exactly is the relationship between dabbrev and hippie-expand?

They seem like two implementations of the same thing; is there a reason
to have both?

-Miles

-- 
Do not taunt Happy Fun Ball.

Re: dabbrev-expand for the minibuffer

[Jason Rumney]

Miles Bader wrote:
> Juri Linkov <juri@jurta.org> writes:
>   
>> It would be easier to implement this in hippie-expand because it has all
>> necessary infrastructure.
>>     
>
> What exactly is the relationship between dabbrev and hippie-expand?
>
> They seem like two implementations of the same thing; is there a reason
> to have both?
>   

They are quite different. dabbrev completes based on existing text in 
the buffer (or multiple buffers with a prefix arg). hippie-expand 
completes using multiple methods, which may or may not include dabbrev.

Re: dabbrev-expand for the minibuffer

[Miles Bader]

Jason Rumney <jasonr@gnu.org> writes:
> They are quite different. dabbrev completes based on existing text in
> the buffer (or multiple buffers with a prefix arg). hippie-expand
> completes using multiple methods, which may or may not include dabbrev.

So is hippie-expand pretty much a superset of dabbrev?   If so, wouldn't
it make sense to only have hippie-expand, perhaps renamed, with a
default configuration that only includes dabbrev functionality?

That would avoid the current argument over whether dabbrev should be
expanded, or a current dabbrev binding switched to hippie-expand
(implementing features by changing local bindings is generally a bad
method because it's rather brittle).

-Miles

-- 
Erudition, n. Dust shaken out of a book into an empty skull.

Re: dabbrev-expand for the minibuffer

[Jason Rumney]

Miles Bader wrote:
> So is hippie-expand pretty much a superset of dabbrev?
Only in so much as it uses dabbrev in addition to several other 
expansion methods. If dabbrev was removed, its functionality would no 
longer be available to hippie-expand.

Re: dabbrev-expand for the minibuffer

[Leo]

On 2008-02-28 08:21 +0000, Jason Rumney wrote:
> Miles Bader wrote:
>> So is hippie-expand pretty much a superset of dabbrev?
> Only in so much as it uses dabbrev in addition to several other
> expansion methods. If dabbrev was removed, its functionality would
> no longer be available to hippie-expand.

It looks like hippie-exp is entirely independent from dabbrev.

-- 
.:  Leo  :.  [ sdl.web AT gmail.com ]  .:  [ GPG Key: 9283AA3F ]  :.

          Use the best OS -- http://www.fedoraproject.org/

Re: dabbrev-expand for the minibuffer

[Richard Stallman]

    They seem like two implementations of the same thing; is there a reason
    to have both?

dabbrev is a "standard" feature, documented in the manual, with a
standard key binding.  hippie-expand is more obscure.

It may well be that hippie-expand can do everything dabbrev does and
more.  But the crucial question is whether it is an upward-compatible
dabbrev replacement convenent for people like me that use dabbrev.
(I don't know the answer.)

Re: dabbrev-expand for the minibuffer

[Leo]

On 2008-02-28 16:41 +0000, Richard Stallman wrote:
>     They seem like two implementations of the same thing; is there a reason
>     to have both?
>
> dabbrev is a "standard" feature, documented in the manual, with a
> standard key binding.  hippie-expand is more obscure.
> 
> It may well be that hippie-expand can do everything dabbrev does and
> more.  But the crucial question is whether it is an upward-compatible
> dabbrev replacement convenent for people like me that use dabbrev.
> (I don't know the answer.)

The most convincing way is to try the following setup and see if you
have noticed any difference.

--8<---------------cut here---------------start------------->8---
(global-set-key (kbd "M-/") 'hippie-expand)
(setq hippie-expand-try-functions-list
      '(try-expand-dabbrev
        try-expand-dabbrev-all-buffers
        try-expand-dabbrev-from-kill))
--8<---------------cut here---------------end--------------->8---

Best,
-- 
.:  Leo  :.  [ sdl.web AT gmail.com ]  .:  [ GPG Key: 9283AA3F ]  :.

          Use the best OS -- http://www.fedoraproject.org/

Re: dabbrev-expand for the minibuffer

[Richard Stallman]

    The most convincing way is to try the following setup and see if you
    have noticed any difference.

I will try it.

Re: dabbrev-expand for the minibuffer

[Richard Stallman]

When I tried

    (global-set-key (kbd "M-/") 'hippie-expand)
    (setq hippie-expand-try-functions-list
          '(try-expand-dabbrev
            try-expand-dabbrev-all-buffers
            try-expand-dabbrev-from-kill))

I found it broke the feature that you can type M-SPC M-/ to pull
in successive words.  I use that often, so this proved
unacceptable.

Maybe that feature can be put into hippie-expand.

RE: dabbrev-expand for the minibuffer

[Drew Adams]

> From: Richard Stallman Sent: Saturday, March 01, 2008 5:32 AM
> When I tried
>
>     (global-set-key (kbd "M-/") 'hippie-expand)
>     (setq hippie-expand-try-functions-list
>           '(try-expand-dabbrev
>             try-expand-dabbrev-all-buffers
>             try-expand-dabbrev-from-kill))
> 
> I found it broke the feature that you can type M-SPC M-/ to pull
> in successive words.  I use that often, so this proved
> unacceptable.
> 
> Maybe that feature can be put into hippie-expand.

Interesting; never thought of doing that (M-SPC M-/). It might help users to
mention this possibility in the Emacs manual.

BTW, was there any followup on adding this to hippie-expand? The thread seems to
have petered out.

Re: dabbrev-expand for the minibuffer

[Lennart Borgman (gmail)]

Drew Adams wrote:
>> From: Richard Stallman Sent: Saturday, March 01, 2008 5:32 AM
>> When I tried
>>
>>     (global-set-key (kbd "M-/") 'hippie-expand)
>>     (setq hippie-expand-try-functions-list
>>           '(try-expand-dabbrev
>>             try-expand-dabbrev-all-buffers
>>             try-expand-dabbrev-from-kill))
>>
>> I found it broke the feature that you can type M-SPC M-/ to pull
>> in successive words.  I use that often, so this proved
>> unacceptable.
>>
>> Maybe that feature can be put into hippie-expand.
> 
> Interesting; never thought of doing that (M-SPC M-/). It might help users to
> mention this possibility in the Emacs manual.


Should it really be `M-SPC M-/'? That does not work for me. However 
`M-\' does.

Re: dabbrev-expand for the minibuffer

[Lennart Borgman (gmail)]

Lennart Borgman (gmail) wrote:
> Drew Adams wrote:
>>> From: Richard Stallman Sent: Saturday, March 01, 2008 5:32 AM
>>> When I tried
>>>
>>>     (global-set-key (kbd "M-/") 'hippie-expand)
>>>     (setq hippie-expand-try-functions-list
>>>           '(try-expand-dabbrev
>>>             try-expand-dabbrev-all-buffers
>>>             try-expand-dabbrev-from-kill))
>>>
>>> I found it broke the feature that you can type M-SPC M-/ to pull
>>> in successive words.  I use that often, so this proved
>>> unacceptable.
>>>
>>> Maybe that feature can be put into hippie-expand.
>>
>> Interesting; never thought of doing that (M-SPC M-/). It might help 
>> users to
>> mention this possibility in the Emacs manual.
> 
> 
> Should it really be `M-SPC M-/'? That does not work for me. However 
> `M-\' does.

Eh, sorry forget it. I use M-/ too ;-)