bug#28714: Shouldn't minibuffer-complete add a space after sole completions?

bug#28714: Shouldn't minibuffer-complete add a space after sole completions?

[積丹尼 Dan Jacobson]

X-Debbugs-Cc: monnier@iro.umontreal.ca

Shouldn't minibuffer-complete add a space after sole completions?

That is how bash does it.

Then the user wouldn't be worrying that there is more than one
completion and need to press TAB again to finally see "[Sole completion]".

bug#28714: Shouldn't minibuffer-complete add a space after sole completions?

[Stefan Monnier]

> Shouldn't minibuffer-complete add a space after sole completions?

Depends in which context: for example when completing for `M-x` or `C-x
C-f`, adding a space at the end would be an error since with the space
it would not correspond to an existing command or file name any more.

> That is how bash does it.

So does Emacs too during completion of shell commands, AFAIK.


        Stefan

bug#28714: Shouldn't minibuffer-complete add a space after sole completions?

[積丹尼 Dan Jacobson]

$ touch /tmp/abbbbb /tmp/abbbbbbbbbbbbbb
Find file or URL:
 / [self-insert-command]
 t [self-insert-command]
 m [self-insert-command]
 p [self-insert-command]
 / [self-insert-command]
 a [self-insert-command]
 b [self-insert-command]
 <tab> [minibuffer-complete]
 <tab> [minibuffer-complete]
 b b [self-insert-command]
 <tab> [minibuffer-complete]
 <tab> [minibuffer-complete]

Now try the same thing in bash outside of emacs.

The user at every moment is aware how many completions there are, and it
only takes a total of two tabs vs. four in emacs.

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

SM> Depends in which context: for example when completing for `M-x` or `C-x
SM> C-f`, adding a space at the end would be an error since with the space
SM> it would not correspond to an existing command or file name any more.

How about perhaps just adding the space for the user's visual pleasure,
but not using it when processing the command / filename?

Another idea might be not requiring a second TAB before telling the user
[Sole completion] etc. but that would seem too chatty.

Anyway if he really wants a space he could use "\ ".

bug#28714: Shouldn't minibuffer-complete add a space after sole completions?

[Eli Zaretskii]

> From: 積丹尼 Dan Jacobson
> 	<jidanni@jidanni.org>
> Date: Sat, 07 Oct 2017 05:54:44 +0800
> Cc: 28714@debbugs.gnu.org
> 
> $ touch /tmp/abbbbb /tmp/abbbbbbbbbbbbbb
> Find file or URL:
>  / [self-insert-command]
>  t [self-insert-command]
>  m [self-insert-command]
>  p [self-insert-command]
>  / [self-insert-command]
>  a [self-insert-command]
>  b [self-insert-command]
>  <tab> [minibuffer-complete]
>  <tab> [minibuffer-complete]
>  b b [self-insert-command]
>  <tab> [minibuffer-complete]
>  <tab> [minibuffer-complete]
> 
> Now try the same thing in bash outside of emacs.
> 
> The user at every moment is aware how many completions there are, and it
> only takes a total of two tabs vs. four in emacs.

"M-x icomplete-mode RET", then repeat your experiment.  Which is
better now, Emacs or Bash?

And if icomplete is still not good enough, there's a plethora of
fancier completion packages out there, both bundled and unbundled.

bug#28714: Shouldn't minibuffer-complete add a space after sole completions?

[Stefan Monnier]

> Another idea might be not requiring a second TAB before telling the user
> [Sole completion] etc. but that would seem too chatty.

I'd welcome a patch which adds that to minibuffer.el (optionnally, as
usual in Emacs).


        Stefan

bug#28714: Shouldn't minibuffer-complete add a space after sole completions?

[積丹尼 Dan Jacobson]

>>>>> "EZ" == Eli Zaretskii <eliz@gnu.org> writes:

EZ> "M-x icomplete-mode RET", then repeat your experiment.  Which is
EZ> better now, Emacs or Bash?

Waaa... icomplete-mode is way too chatty.

bug#28714: Shouldn't minibuffer-complete add a space after sole completions?

[Lars Ingebrigtsen]

積丹尼 Dan Jacobson <jidanni@jidanni.org> writes:

> $ touch /tmp/abbbbb /tmp/abbbbbbbbbbbbbb
> Find file or URL:
>  / [self-insert-command]
>  t [self-insert-command]
>  m [self-insert-command]
>  p [self-insert-command]
>  / [self-insert-command]
>  a [self-insert-command]
>  b [self-insert-command]
>  <tab> [minibuffer-complete]
>  <tab> [minibuffer-complete]
>  b b [self-insert-command]
>  <tab> [minibuffer-complete]
>  <tab> [minibuffer-complete]
>
> Now try the same thing in bash outside of emacs.

So you basically want it to say "Sole completion" (or something) after
the next-to-last TAB?  I guess that would make sense (as an option), and
is basically the same thing that's requested in bug#51650 (but there
it's about things that do have several completions).

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

bug#28714: Shouldn't minibuffer-complete add a space after sole completions?

[積丹尼 Dan Jacobson]

>>>>> "LI" == Lars Ingebrigtsen <larsi@gnus.org> writes:
LI> 積丹尼 Dan Jacobson <jidanni@jidanni.org> writes:

>> $ touch /tmp/abbbbb /tmp/abbbbbbbbbbbbbb
>> Find file or URL:
>> / [self-insert-command]
>> t [self-insert-command]
>> m [self-insert-command]
>> p [self-insert-command]
>> / [self-insert-command]
>> a [self-insert-command]
>> b [self-insert-command]
>> <tab> [minibuffer-complete]
>> <tab> [minibuffer-complete]
>> b b [self-insert-command]
>> <tab> [minibuffer-complete]
>> <tab> [minibuffer-complete]
>> 
>> Now try the same thing in bash outside of emacs.

LI> So you basically want it to say "Sole completion" (or something) after
LI> the next-to-last TAB?  I guess that would make sense (as an option), and
LI> is basically the same thing that's requested in bug#51650 (but there
LI> it's about things that do have several completions).

Yeah. Currently it takes two tabs for it to finally cough up "Sole
completion". What sense is that? It should do that with just one tab.