bug#1757: 23.0.60; completion DWIM too clever

bug#1757: 23.0.60; completion DWIM too clever

[Richard M Stallman]

I typed C-x b *d TAB and it offered me as completions `diary' and `old',
neither of which starts with a `*'.

I typed C-x b *di TAB and it completed to `diary', which is not
even a completion of what I typed.

It is extremely annoying to discard the user's input characters.



In GNU Emacs 23.0.60.15 (mipsel-unknown-linux-gnu, GTK+ Version 2.12.11)
 of 2008-12-22 on lemote-yeeloong
configured using `configure  'CFLAGS=-O0 -g -Wno-pointer-sign' 'mipsel-unknown-linux-gnu' 'build_alias=mipsel-unknown-linux-gnu' 'host_alias=mipsel-unknown-linux-gnu' 'target_alias=mipsel-unknown-linux-gnu''

Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: en_US.UTF-8
  value of $XMODIFIERS: nil
  locale-coding-system: utf-8-unix
  default-enable-multibyte-characters: t

Major mode: Diff

Minor modes in effect:
  diff-auto-refine-mode: t
  gpm-mouse-mode: t
  tooltip-mode: t
  tool-bar-mode: t
  menu-bar-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  global-auto-composition-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  line-number-mode: t
  transient-mark-mode: t
  abbrev-mode: t

Recent input:
ESC : b u f f e r - u n d o - l i s t RET C-x b RET 
TAB TAB ESC < C-s p m a i l - v i e w - b u f f e r 
C-s C-s C-s C-s C-s C-s C-s C-s C-s C-s C-n C-n C-n 
C-n C-n C-n C-n C-a C-p C-p C-p C-o TAB ( s e t q SPC 
b u f f e r - u n d o - l i s t SPC t ) C-x C-s C-x 
b RET ESC p u C-x C-s C-x b RET ESC C-x C-x b RET C-x 
k RET ESC x p m a i l RET y ESC : b u f f e r - s w 
a p p e d - w i t h RET C-d C-d C-d C-d C-d C-d p C-d 
C-d C-d x C-x C-s ESC : b u f f e r ESC p RET C-x d 
e m a c s - c v s / l i s p / m a i l / p m a i l * 
. e l * RET C-u C-n C-n C-n C-@ C-u C-n C-n C-n C-n 
C-n C-n = C-g C-x C-x = RET ESC C-v C-x 1 C-u C-n C-u 
C-n C-n C-p C-n C-@ C-p = RET ESC C-v ESC C-v ESC C-v 
ESC C-v ESC C-v ESC C-v C-x 1 C-x b * d i TAB DEL DEL 
DEL DEL DEL * d i TAB DEL DEL DEL DEL DEL * C-g C-x 
C-b C-x o C-v C-n C-n C-n f C-x 1 ESC x r e p o r t 
SPC e m a v s DEL DEL c s RET

Recent messages:
0 new messages read
#<buffer  *message-viewer PMAIL*>
Expunging deleted messages...done
Saving file /home/rms/PMAIL...
Wrote /home/rms/PMAIL
#<buffer  *message-viewer PMAIL*>
Mark set
Source file `/home/rms/emacs-cvs/lisp/diff.el' newer than byte-compiled file
Quit
Mark activated
Quit

bug#1757: 23.0.60; completion DWIM too clever

[Drew Adams]

> I typed C-x b *d TAB and it offered me as completions `diary' 
> and `old', neither of which starts with a `*'.
> 
> I typed C-x b *di TAB and it completed to `diary', which is not
> even a completion of what I typed.
> 
> It is extremely annoying to discard the user's input characters.

I couldn't agree more. The is (the formerly optional) partial completion at
work. It is now the default behavior to first try the traditional completion,
and if that finds no matches, automatically try partial completion. IMO, the
default behavior should remain as it was in Emacs 22 (21, 20...), and the new
behavior should be optional.

With the traditional behavior, if there are no buffers with prefix `*', you are
told so immediately: [No match]. With the new, partial-completion behavior, you
are given possible completions that do not complete `*' in the normal way (as a
literal prefix).

To get the traditional behavior, you now need to customize option
`completion-styles', to be `(basic)', not `(basic partial-completion)'. This
significant change was made with no discussion, AFAIK.

See bug #1512:
http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=1512

bug#1757: 23.0.60; completion DWIM too clever

[Lennart Borgman]

On Thu, Jan 1, 2009 at 6:59 PM, Drew Adams <drew.adams@oracle.com> wrote:
>> I typed C-x b *d TAB and it offered me as completions `diary'
>> and `old', neither of which starts with a `*'.
>>
>> I typed C-x b *di TAB and it completed to `diary', which is not
>> even a completion of what I typed.
>>
>> It is extremely annoying to discard the user's input characters.
>
> I couldn't agree more. The is (the formerly optional) partial completion at
> work. It is now the default behavior to first try the traditional completion,
> and if that finds no matches, automatically try partial completion.

Is perhaps the problem that partial completion does not respect the
user input (of *)?

bug#1757: 23.0.60; completion DWIM too clever

[Drew Adams]

> >> I typed C-x b *d TAB and it offered me as completions `diary'
> >> and `old', neither of which starts with a `*'.
> >>
> >> I typed C-x b *di TAB and it completed to `diary', which is not
> >> even a completion of what I typed.
> >>
> >> It is extremely annoying to discard the user's input characters.
> >
> > I couldn't agree more. The is (the formerly optional) partial
> > completion at work. It is now the default behavior to first try
> > the traditional completion, and if that finds no matches,
> > automatically try partial completion.
> 
> Is perhaps the problem that partial completion does not respect the
> user input (of *)?

It respects it as a _wildcard_. This is a feature of partial completion. See
`completion-pcm--string->pattern' and `completion-pcm--all-completions' in
minibuffer.el.

Stefan has decided that if your input cannot be completed literally, as a
prefix, it should be completed as a wildcard pattern, using partial completion.
That is the new default behavior.

This takes away the useful negative feedback that your input has no literal
prefix completion: [No match]. Instead of letting you immediately correct a
simple typo (e.g. accidentally hitting `*' instead of some other char), you can
end up needing to correct lots of characters (e.g. `diary', as a completion of
`*di').

The problem is not that partial completion has a bug wrt *. The problem is that
partial completion is being used as an automatic fallback by default. The
assumption is that you always want to complete one way or the other - either by
treating your input as a literal prefix, if possible, or by treating it as a
wildcard expression, if literal prefix matching fails.

This "extremely annoying" feature should be optional. Partial completion is one
thing. Automatically slipping from prefix completion to partial completion is
another thing. Either of these is annoying and inappropriate as the default
behavior. And the automatic change of completing behavior is arguably more
annoying and more confusing than plain partial completion.

bug#1757: 23.0.60; completion DWIM too clever

[Richard M Stallman]

    I couldn't agree more. The is (the formerly optional) partial completion at
    work. It is now the default behavior to first try the traditional completion,
    and if that finds no matches, automatically try partial completion. IMO, the
    default behavior should remain as it was in Emacs 22 (21, 20...), and the new
    behavior should be optional.

My first tendency is to dislike the new behavior, but now that I have
some idea what it does, I will try looking for some advantages in it.

However, a change like this calls for polling the users.

bug#1757: 23.0.60; completion DWIM too clever

[Richard M Stallman]

Using * as a wildcard in buffer name completion
is particularly a hassle given that * is used in many buffer names.

bug#1757: 23.0.60; completion DWIM too clever

[Stefan Monnier]

tag 1757 +wontfix
thanks

> I typed C-x b *di TAB and it completed to `diary', which is not
> even a completion of what I typed.

It's not a bug, it's a feature.


        Stefan

Processed: Re: bug#1757: 23.0.60; completion DWIM too clever

[Emacs bug Tracking System]

Processing commands for control@emacsbugs.donarmstrong.com:

> tag 1757 +wontfix
bug#1757: 23.0.60; completion DWIM too clever
There were no tags set.
Tags added: wontfix

> thanks
Stopping processing here.

Please contact me if you need assistance.

Don Armstrong
(administrator, Emacs bugs database)

bug#1757: 23.0.60; completion DWIM too clever

[Drew Adams]

> tags 1757 fixed
Bug #1757 [emacs] 23.0.60; completion DWIM too clever
Added tag(s) fixed.
> close 1757

How in the world can you consider this "fixed"?
The behavior reported as a bug is still there.
No fix was made.

Stefan claims it's a feature.
RMS said to poll the users before doing it.
That was several releases ago (and no poll).

You can claim the behavior is not a bug.  Or you can admit that it is a bug but
state that it won't be fixed.

But how could you possibly claim that, (a) yes, it is a bug and (b) it has been
fixed?  Makes no sense.

bug#1757: 23.0.60; completion DWIM too clever

[Lars Magne Ingebrigtsen]

"Drew Adams" <drew.adams@oracle.com> writes:

>> tags 1757 fixed
> Bug #1757 [emacs] 23.0.60; completion DWIM too clever
> Added tag(s) fixed.
>> close 1757
>
> How in the world can you consider this "fixed"?

Wrong key.  I meant to close it without "fixed".

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