New interactive spec

New interactive spec

[Didier Verna]

       Hello,

[please Cc: me; I'm not on the list ]

I've recently modified the behavior of switch-to-buffer in XEmacs so
that the completion mechanism doesn't include the current buffer. In
doing so, I figured it would be generally useful to have an interactive
spec for that (meaning "a buffer except the current one"). I was
thinking g/G because that's the next thing available to us, but I'm
wondering if GNU Emacs already uses it, or if you'd have another
suggestion.


Thank you !

-- 
Resistance is futile. You will be jazzimilated.

Scientific site:   http://www.lrde.epita.fr/~didier
Music (Jazz) site: http://www.didierverna.com

Re: New interactive spec

[Tassilo Horn]

Didier Verna <didier@xemacs.org> writes:

Hi Didier,

> I've recently modified the behavior of switch-to-buffer in XEmacs so
> that the completion mechanism doesn't include the current buffer. In
> doing so, I figured it would be generally useful to have an
> interactive spec for that (meaning "a buffer except the current
> one"). I was thinking g/G because that's the next thing available to
> us, but I'm wondering if GNU Emacs already uses it, or if you'd have
> another suggestion.

"G" is already taken, but "g" would be free.

  G -- Possibly nonexistent file name, defaulting to just directory
       name.

Bye,
Tassilo

Re: New interactive spec

[Stefan Monnier]

>> I've recently modified the behavior of switch-to-buffer in XEmacs so
>> that the completion mechanism doesn't include the current buffer. In
>> doing so, I figured it would be generally useful to have an
>> interactive spec for that (meaning "a buffer except the current
>> one"). I was thinking g/G because that's the next thing available to
>> us, but I'm wondering if GNU Emacs already uses it, or if you'd have
>> another suggestion.

> "G" is already taken, but "g" would be free.

>   G -- Possibly nonexistent file name, defaulting to just directory
>        name.

How important is it to have a letter for it?


        Stefan

Re: New interactive spec

[Juri Linkov]

>>> I've recently modified the behavior of switch-to-buffer in XEmacs so
>>> that the completion mechanism doesn't include the current buffer. In
>>> doing so, I figured it would be generally useful to have an
>>> interactive spec for that (meaning "a buffer except the current
>>> one"). I was thinking g/G because that's the next thing available to
>>> us, but I'm wondering if GNU Emacs already uses it, or if you'd have
>>> another suggestion.
>
>> "G" is already taken, but "g" would be free.
>
>>   G -- Possibly nonexistent file name, defaulting to just directory
>>        name.
>
> How important is it to have a letter for it?

I see a tendency in the opposite direction - replacing letters
with equivalent Lisp code to be able to add more functionality.

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

Re: New interactive spec

[Didier Verna]

[ Added xemacs-beta to the discussion ]

Stefan Monnier <monnier@IRO.UMontreal.CA> wrote:

>>> I've recently modified the behavior of switch-to-buffer in XEmacs so
>>> that the completion mechanism doesn't include the current buffer. In
>>> doing so, I figured it would be generally useful to have an
>>> interactive spec for that (meaning "a buffer except the current
>>> one"). I was thinking g/G because that's the next thing available to
>>> us, but I'm wondering if GNU Emacs already uses it, or if you'd have
>>> another suggestion.
>
>> "G" is already taken, but "g" would be free.
>
>>   G -- Possibly nonexistent file name, defaulting to just directory
>>        name.
>
> How important is it to have a letter for it?

  Is this question directed to Tassilo (about G) or to me ?

If it's to me: it is not critical to have a letter (I'd like both upcase
and downcase BTW) for it. It just seems that it could be generally
useful. In the core of XEmacs I can see several functions that could use
it: append|prepend|copy-to-buffer, clone-indirect-buffer[-other-window]
and view-buffer[-other-window].

But now that I think about it, I'm starting to realize that every
interactive call using b and providing the current buffer as the default
one could avoid completing to the current buffer altogether because you
just need to type RETURN to get it. So maybe I should modify the b
specification instead...

-- 
Resistance is futile. You will be jazzimilated.

Scientific site:   http://www.lrde.epita.fr/~didier
Music (Jazz) site: http://www.didierverna.com

Re: New interactive spec

[Miles Bader]

Didier Verna <didier@xemacs.org> writes:
> But now that I think about it, I'm starting to realize that every
> interactive call using b and providing the current buffer as the default
> one could avoid completing to the current buffer altogether because you
> just need to type RETURN to get it. So maybe I should modify the b
> specification instead...

I don't think that's a very good idea -- I find that I _often_ try to
type/complete a name even if it's available as a default via RET, and it
would be very disconcerting if valid names were not available for
completion simply because they were also available using a different
mechanism.

-Miles

-- 
Defenceless, adj. Unable to attack.

Re: New interactive spec

[Didier Verna]

Miles Bader <miles@gnu.org> wrote:

> I don't think that's a very good idea -- I find that I _often_ try to
> type/complete a name even if it's available as a default via RET, and
> it would be very disconcerting if valid names were not available for
> completion simply because they were also available using a different
> mechanism.

  OK, but I'm not talking about /any/ buffer available as the default.
Only the /current/ buffer. Do you think it really happens that when
prompted with the /current/ buffer as a default value, you try to type
and complete it anyway ?

-- 
Resistance is futile. You will be jazzimilated.

Scientific site:   http://www.lrde.epita.fr/~didier
Music (Jazz) site: http://www.didierverna.com

Re: New interactive spec

[Per Starbäck]

2009/11/11 Didier Verna <didier@xemacs.org>:

>  OK, but I'm not talking about /any/ buffer available as the default.
> Only the /current/ buffer. Do you think it really happens that when
> prompted with the /current/ buffer as a default value, you try to type
> and complete it anyway ?

I think I do sometimes at least, for example with ediff-buffers where
I most of the time already display the two buffers, but sometimes have
"A" and sometimes "B" as current buffer when invoking ediff-buffers.

I'm still surprised now and then by the new completion in emacs 23
that avoids current buffer sometimes.

Re: New interactive spec

[Juri Linkov]

>>  OK, but I'm not talking about /any/ buffer available as the default.
>> Only the /current/ buffer. Do you think it really happens that when
>> prompted with the /current/ buffer as a default value, you try to type
>> and complete it anyway ?
>
> I think I do sometimes at least, for example with ediff-buffers where
> I most of the time already display the two buffers, but sometimes have
> "A" and sometimes "B" as current buffer when invoking ediff-buffers.
>
> I'm still surprised now and then by the new completion in emacs 23
> that avoids current buffer sometimes.

Could you please provide a detailed test case.  I tried to reproduce
what you've described above, but I see no problem in Emacs 23.

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