Use ido for *all* completion?

Use ido for *all* completion?

[Mark Plaksin]

Is there a way to use ido's magical completion for everything?  For
example, I'd like it to complete when answering the prompt M-x gives.

CVS Emacs finally broke lightning completion[1] for me so I decided to look
at ido.  I like it!  You can configure Lightning completion so that it
completes for all minibuffer prompts.  It would be great if ido did that
too.  Can it?

Thanks!

[1] http://www.math.washington.edu/~palmieri/Emacs/light.html  Hasn't been
updated since 2002.

Re: Use ido for *all* completion?

[Oliver Scholz]

Mark Plaksin <happy@usg.edu> writes:

> Is there a way to use ido's magical completion for everything?  For
> example, I'd like it to complete when answering the prompt M-x gives.
>
> CVS Emacs finally broke lightning completion[1] for me so I decided to look
> at ido.  I like it!  You can configure Lightning completion so that it
> completes for all minibuffer prompts.  It would be great if ido did that
> too.  Can it?

I don't think that ido can do this (yet I don't /know/ either that it
can not.) However, there's M-x icomplete-mode

    Oliver
-- 
19 Ventôse an 213 de la Révolution
Liberté, Egalité, Fraternité!

RE: Use ido for *all* completion?

[Drew Adams]

AFAIK, you can use `ido-completing-read' as a general replacement for
`completing-read'.

I don't know if you could just do a `defalias' (or would want to, even if
you could), but at least for your own commands you can choose to use
`ido-completing-read' instead of `completing-read'.

It looks like the signatures of the two functions are a bit different, so I
doubt that you could just `defalias' completing-read to
ido-completing-read - the latter does not accept an inherit-input-method
argument, and its second argument must be a list of strings, not an obarray
or an alist whose cars are strings.

CCing ido's author, in case he has an idea about this. - Drew

    > Is there a way to use ido's magical completion for everything?  For
    > example, I'd like it to complete when answering the prompt M-x gives.
    >
    > CVS Emacs finally broke lightning completion[1] for me so I
    decided to look
    > at ido.  I like it!  You can configure Lightning completion so that it
    > completes for all minibuffer prompts.  It would be great if
    ido did that
    > too.  Can it?

    I don't think that ido can do this (yet I don't /know/ either that it
    can not.) However, there's M-x icomplete-mode

Re: Use ido for *all* completion?

[Oliver Scholz]

[Top posting fixed]

"Drew Adams" <drew.adams@oracle.com> writes:
>
>     > Is there a way to use ido's magical completion for everything?  For
>     > example, I'd like it to complete when answering the prompt M-x gives.
>     >
>     > CVS Emacs finally broke lightning completion[1] for me so I
>     decided to look
>     > at ido.  I like it!  You can configure Lightning completion so that it
>     > completes for all minibuffer prompts.  It would be great if
>     ido did that
>     > too.  Can it?
>
>     I don't think that ido can do this (yet I don't /know/ either that it
>     can not.) However, there's M-x icomplete-mode

> AFAIK, you can use `ido-completing-read' as a general replacement for
> `completing-read'.

[...]

> CCing ido's author, in case he has an idea about this. - Drew

Excuse me, did you actually take the time to try M-x icomplete-mode?

    Oliver
-- 
19 Ventôse an 213 de la Révolution
Liberté, Egalité, Fraternité!

Re: Use ido for *all* completion?

[Stefan Monnier]

> CVS Emacs finally broke lightning completion[1] for me so I decided to look
> at ido.

Please report it via M-x report-emacs-bug.


        Stefan

Re: Use ido for *all* completion?

[Mark Plaksin]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> CVS Emacs finally broke lightning completion[1] for me so I decided to look
>> at ido.
>
> Please report it via M-x report-emacs-bug.

Do you mean I should report that I'd like ido-like completion for
everything?  I didn't realize report-emacs-bug was for feature requests
too.  (Lightning completion isn't part of Emacs so I suppose you're not
telling me to report the fact that it stopped working.)

RE: Use ido for *all* completion?

[Drew Adams]

    > AFAIK, you can use `ido-completing-read' as a general replacement for
    > `completing-read'.
    > CCing ido's author, in case he has an idea about this.

    Excuse me, did you actually take the time to try M-x icomplete-mode?

Why? I replied to a question about `ido-completing-read' with some info
about `ido-completing-read'.

You said:

    I don't think that ido can do this (yet I don't /know/ either that it
    can not.)

I replied to that, as well as to the original question:

    > Is there a way to use ido's magical completion for everything?

Re: Use ido for *all* completion?

[Romain Francoise]

Oliver Scholz <alkibiades@gmx.de> writes:

> Excuse me, did you actually take the time to try M-x icomplete-mode?

Did you take the time to try ido?  :-)

They do entirely different things, icomplete merely shows the available
completions; ido does substring/flexible/regexp matching, allows you to
rotate through the list of available completions, etc.

-- 
Romain Francoise <romain@orebokech.com> | I'm closer than ever to
it's a miracle -- http://orebokech.com/ | finding the hidden track.

Re: Use ido for *all* completion?

[Oliver Scholz]

Romain Francoise <romain@orebokech.com> writes:

> Oliver Scholz <alkibiades@gmx.de> writes:
>
>> Excuse me, did you actually take the time to try M-x icomplete-mode?
>
> Did you take the time to try ido?  :-)

Yes.

    Oliver
-- 
19 Ventôse an 213 de la Révolution
Liberté, Egalité, Fraternité!

Re: Use ido for *all* completion?

[Stefan Monnier]

>>> CVS Emacs finally broke lightning completion[1] for me so I decided to
>>> look at ido.
>> 
>> Please report it via M-x report-emacs-bug.

> Do you mean I should report that I'd like ido-like completion for
> everything?

Well, you could do that too, but I mostly meant that you should report that
Emacs-CVS breaks lightning.el.

> (Lightning completion isn't part of Emacs so I suppose you're not
> telling me to report the fact that it stopped working.)

While we don't guarantee backward compatibility, we do try to maintain it,
so we like to be told when we break it,


        Stefan

Re: Use ido for *all* completion?

[Romain Francoise]

Oliver Scholz <alkibiades@gmx.de> writes:

> Yes.

And don't you agree that while they both show available completions,
icomplete is nowhere near ido when it comes to functionality?  Or am I
missing something?

-- 
Romain Francoise <romain@orebokech.com> | Sometimes I don't know where
it's a miracle -- http://orebokech.com/ | this dirty road is taking me,
                                        | sometimes I can't even see
                                        | the reason why.

Re: Use ido for *all* completion?

[Oliver Scholz]

Romain Francoise <romain@orebokech.com> writes:

> Oliver Scholz <alkibiades@gmx.de> writes:
>
>> Yes.
>
> And don't you agree that while they both show available completions,
> icomplete is nowhere near ido when it comes to functionality?  Or am I
> missing something?

I don't recall having seen any list of required functionality. I do
remember that somebody wanted to use ido for other sorts of
completion. Such things most of the time mean: people want to have a
particular functionality and express that as "can X also do Y?" So
before we jump on anything without consideration and merge
gnu.emacs.help and emacs-devel, it would be quite more methodical to
find the least difficult way to achieve the required functionality.

    Oliver
-- 
20 Ventôse an 213 de la Révolution
Liberté, Egalité, Fraternité!

Re: Use ido for *all* completion?

[Mark Plaksin]

Oliver Scholz <alkibiades@gmx.de> writes:

> Romain Francoise <romain@orebokech.com> writes:
>
>> Oliver Scholz <alkibiades@gmx.de> writes:
>>
>>> Yes.
>>
>> And don't you agree that while they both show available completions,
>> icomplete is nowhere near ido when it comes to functionality?  Or am I
>> missing something?
>
> I don't recall having seen any list of required functionality. I do
> remember that somebody wanted to use ido for other sorts of
> completion. Such things most of the time mean: people want to have a
> particular functionality and express that as "can X also do Y?" So
> before we jump on anything without consideration and merge
> gnu.emacs.help and emacs-devel, it would be quite more methodical to
> find the least difficult way to achieve the required functionality.

I posted the original question.  What I really want is to type as little as
possible.  The lightning completion code I mentioned is great because it
completes automatically (you don't have to hit TAB) and because you can use
it for all minibuffer prompts.

I have looked at icomplete, partial-completion-mode, and ido.  Of the
three, ido is the best for me.  It's completion is similar to but much more
powerful than lightning completion's.  The only missing piece is (an option)
to make it active for all minibuffer prompts.

Re: Use ido for *all* completion?

[Peter Lee]

>>>> Mark Plaksin writes:

    > I have looked at icomplete, partial-completion-mode, and ido.  Of the
    > three, ido is the best for me.

There is also mcomplete.  Not sure of the differences between
mcomplete and those other modes you listed, but it does prefix and
substring completion.

Re: Use ido for *all* completion?

[Mark Plaksin]

Peter Lee <pete.a.lee@gmail.com> writes:

>>>>> Mark Plaksin writes:
>
>     > I have looked at icomplete, partial-completion-mode, and ido.  Of the
>     > three, ido is the best for me.
>
> There is also mcomplete.  Not sure of the differences between
> mcomplete and those other modes you listed, but it does prefix and
> substring completion.

mcomplete.el (http://homepage1.nifty.com/bmonkey/emacs/elisp/mcomplete.el)
has a nice ASCII drawing of how it integrates with the minibuffer.  Maybe
we can use this same method with ido.  I'll take a look at this one day
soon.  Here's the drawing;

;;; How this package works:

;; On loading this package,
;; `mcomplete-minibuffer-setup' is registered in `minibuffer-setup-hook'.
;;
;;
;; Some command tries to begin a minibuffer session
;;    |
;;    v
;; (run-hooks 'minibuffer-setup-hook) -+
;;    |                                |
;;    |                                v
;;    |           `mcomplete-minibuffer-setup'
;;    |              (when (mcomplete-p)
;;    |                 (run-hooks 'mcomplete-minibuffer-setup-hook))
;;    |                                |
;;    |                                v
;;    |           `mcomplete-setup-command-hooks'
;;    |              * registers `mcomplete-pre-command-hook' in
;;    |                the buffer local `pre-command-hook'.
;;    |              * registers `mcomplete-post-command-hook' in
;;    |                the buffer local `post-command-hook'.
;;    v
;; The minibuffer session begins
;;    |
;;    v
;; Some key is pressed in the minibuffer
;;    |
;;    v
;; (run-hooks 'pre-command-hook) -+
;;    |                           |
;;    |                           v
;;    |           (run-hooks 'mcomplete-pre-command-hook)
;;    |                           |
;;    |                           v
;;    |           `mcomplete-pre-command-tidy'
;;    |               * clears information display in the minibuffer.
;;    v
;; The command for the key is executed
;;    |
;;    v
;; (run-hooks 'post-command-hook) -+
;;                                 |
;;                                 v
;;                 (run-hooks 'mcomplete-post-command-hook)
;;                                 |
;;                                 v
;;                 `mcomplete-post-command-exhibit'
;;                    * displays information in the minibuffer.

Re: Use ido for *all* completion?

[Mark Plaksin]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>>>> CVS Emacs finally broke lightning completion[1] for me so I decided to
>>>> look at ido.
>>> 
>>> Please report it via M-x report-emacs-bug.
>
>> Do you mean I should report that I'd like ido-like completion for
>> everything?
>
> Well, you could do that too, but I mostly meant that you should report that
> Emacs-CVS breaks lightning.el.

Heh.  The problem with lightning completion turned out to be a bad test.  It
tested for Emacs 21 with
  (= emacs-major-version 21)
instead of
  (>= emacs-major-version 21)

Changing the defconst to this fixed the problem:
(defconst lc-emacs-21-p
  (and (boundp 'emacs-major-version)
       (not lc-xemacs-p)
       (>= emacs-major-version 21))
  "Non-nil if using GNU Emacs >= 21.")

I still plan to try to make ido work for all minibuffer completion.