GNU bug report logs - #76617
completing-read should accept spaces in require-match nil mode

Previous Next

To reply to this bug, email your comments to 76617 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Daniel Colascione <dancol <at> dancol.org>
To: "Eshel Yaron via Bug reports for GNU Emacs,  the Swiss army knife of
 text editors" <bug-gnu-emacs <at> gnu.org>
Subject: completing-read should accept spaces in require-match nil mode
Date: Thu, 27 Feb 2025 16:10:34 -0500

It's sometimes useful (especially when require-match nil) to enter text
with spaces in completing read, but right now, that's inconvenient (you
have to use C-q SPC) because we bind SPC to minibuffer-complete-word,
which won't insert a space unless space matches some possible
completion, and in require-match nil mode, there might not be a
completion with a space in it. 

Message #8 received at 76617 <at> debbugs.gnu.org (full text, mbox):

From: Drew Adams <drew.adams <at> oracle.com>
To: Daniel Colascione <dancol <at> dancol.org>, "76617 <at> debbugs.gnu.org"
 <76617 <at> debbugs.gnu.org>
Subject: RE: [External] : bug#76617: completing-read should accept spaces in
 require-match nil mode
Date: Thu, 27 Feb 2025 22:18:27 +0000

> It's sometimes useful (especially when require-match nil) to enter text
> with spaces in completing read, but right now, that's inconvenient (you
> have to use C-q SPC) because we bind SPC to minibuffer-complete-word,
> which won't insert a space unless space matches some possible
> completion, and in require-match nil mode, there might not be a
> completion with a space in it.

Indeed.  Good luck!

Since 2005 (20 years ago), Icicles has made SPC self-inserting
during completion.  I tried, in vain, to get vanilla Emacs to
do the same.  Eventually Emacs did so for file-name completion
(only).

SPC and ? should be self-inserting.  How many more decades till
this sinks in?

For word completion, it's sufficient to bind M-SPC instead of
SPC (also done in 2005).

Message #11 received at 76617 <at> debbugs.gnu.org (full text, mbox):

From: Richard Stallman <rms <at> gnu.org>
To: Drew Adams <drew.adams <at> oracle.com>
Cc: dancol <at> dancol.org, 76617 <at> debbugs.gnu.org
Subject: Re: bug#76617: completing-read should accept spaces in require-match
 nil mode
Date: Sun, 02 Mar 2025 10:07:55 -0500

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

Perhaps SPC and ? should self-insert when they do not extend the
completion text at all.  Would that behavior be convenient?

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Message #14 received at 76617 <at> debbugs.gnu.org (full text, mbox):

From: Daniel Colascione <dancol <at> dancol.org>
To: Richard Stallman <rms <at> gnu.org>
Cc: 76617 <at> debbugs.gnu.org, Drew Adams <drew.adams <at> oracle.com>
Subject: Re: bug#76617: completing-read should accept spaces in
 require-match nil mode
Date: Sun, 02 Mar 2025 10:53:04 -0500

Richard Stallman <rms <at> gnu.org> writes:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
> Perhaps SPC and ? should self-insert when they do not extend the
> completion text at all.  Would that behavior be convenient?

Only when require-match is nil, IMHO.

Message #17 received at 76617 <at> debbugs.gnu.org (full text, mbox):

From: Drew Adams <drew.adams <at> oracle.com>
To: "rms <at> gnu.org" <rms <at> gnu.org>
Cc: "dancol <at> dancol.org" <dancol <at> dancol.org>,
 "76617 <at> debbugs.gnu.org" <76617 <at> debbugs.gnu.org>
Subject: RE: [External] : Re: bug#76617: completing-read should accept spaces
 in require-match nil mode
Date: Sun, 2 Mar 2025 19:39:11 +0000

> Perhaps SPC and ? should self-insert when they do not extend the
> completion text at all.  Would that behavior be convenient?

I don't think so, but mine's just one opinion.

I think that would lead to confusion, hurting
more than helping.

Plus, with some completion "frameworks" (I'd
call them just packages that enhance/extend
completion possibilities) you can switch the
kind of completion/matching on the fly, with
the same pattern (input in the minibuffer).

And in that case SPC or ? could, after such
a switch, suddenly lead to completions, or
vice versa.

I really think that keys that have actions
such as what SPC and ? have had should not be
keys for printable (i.e., insertable) chars.
There's no more reason to have `?' not be
inserting than there is for `#' or `@' to not
be inserting.

Better to use unprintable keys for non-insert
actions.  In my code I use `M-SPC' in place
of `SPC' and `M-?' in place of `?', for those
actions.

In general, in editable buffers, users shouldn't
need to use `C-q' to insert printable chars.

Just one opinion.

Message #20 received at 76617 <at> debbugs.gnu.org (full text, mbox):

From: Richard Stallman <rms <at> gnu.org>
To: Daniel Colascione <dancol <at> dancol.org>
Cc: 76617 <at> debbugs.gnu.org, drew.adams <at> oracle.com
Subject: Re: bug#76617: completing-read should accept spaces in
 require-match nil mode
Date: Fri, 02 May 2025 17:39:58 -0400

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > > Perhaps SPC and ? should self-insert when they do not extend the
  > > completion text at all.  Would that behavior be convenient?

  > Only when require-match is nil, IMHO.

To self-insert when REQUIRE-MATCH is nil seems like a good idea.

What about with other values of REQUIRE-MATCH, in the case where
adding the character would result in a match match something in the
completion list?

For instance, if the minibuffer contains `foo' and `foo?' is one
of the completions, them typing ? twice could insert ?

And something similar for SPC.


-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Message #23 received at 76617 <at> debbugs.gnu.org (full text, mbox):

From: Drew Adams <drew.adams <at> oracle.com>
To: "rms <at> gnu.org" <rms <at> gnu.org>, Daniel Colascione <dancol <at> dancol.org>
Cc: "76617 <at> debbugs.gnu.org" <76617 <at> debbugs.gnu.org>
Subject: RE: [External] : Re: bug#76617: completing-read should accept spaces
 in require-match nil mode
Date: Fri, 2 May 2025 23:43:20 +0000

>   > > Perhaps SPC and ? should self-insert when they do not extend the
>   > > completion text at all.  Would that behavior be convenient?
> 
>   > Only when require-match is nil, IMHO.
> 
> To self-insert when REQUIRE-MATCH is nil seems like a good idea.
> 
> What about with other values of REQUIRE-MATCH, in the case where
> adding the character would result in a match match something in the
> completion list?
> 
> For instance, if the minibuffer contains `foo' and `foo?' is one
> of the completions, them typing ? twice could insert ?
> 
> And something similar for SPC.

IMO there should be _no_ such restrictions, conditional
or otherwise.

Printable/self-inserting chars should generally
self-insert (i.e., without C-q) during completion.
RET/C-m and TAB/C-i are reasonable exceptions
(especially given their history).  But even C-j should
self-insert.

`completing-read' is completely general.  We shouldn't
assume anything about what the set of required-match
completions might be at any time - any chars, any order,
etc.

A given `completing-read' _call_ can, if need be, impose
such a restriction by let-binding a relevant minibuffer
keymap.

It took decades to get `read-file-name' to allow SPC to
self-insert.  It's taken decades more for SPC and ? to
be taken seriously also for `completing-read' insertion.
Time to realize that `completing-read' candidates can
contain any chars.  By default, we should self-insert
all printable chars.

(One opinion.)

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.