bug#4176: 23.1; partial completions

bug#4176: 23.1; partial completions

[Eli Barzilay]

With the default settings, enter

  C-h f select-window TAB

and Emacs will tell you that this is the only completion.

Trying this with `sele-wind TAB' does a partial completion to
`select-window' and moves the cursor to the right place -- a
second TAB pops up the completions (showing both functions) saying
"complete but not unique" *but* moves the cursor to the end.

It would be very nice if the cursor was always at the right place, and
it would not say that this is the only completion.  Doing

  (setq completion-styles '(partial-completion))

is fixes the message, but the cursor still moves to the end.

Another related problem -- the doc string for `completion-styles'
doesn't say anything about what's allowed (should mention
`completion-styles-alist').  Also, the functions that are used in
`completion-styles-alist' are not documented -- which leaves the user
with no explanation at all about the available completion styles.



In GNU Emacs 23.1.1 (x86_64-unknown-linux-gnu, GTK+ Version 2.10.14)
 of 2009-08-01 on winooski.ccs.neu.edu
Windowing system distributor `The X.Org Foundation', version 11.0.10300000
configured using `configure  '--prefix=/home/eli/bin/local/emacs-dir''

Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: POSIX
  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
  value of $XMODIFIERS: nil
  locale-coding-system: iso-latin-1-unix
  default-enable-multibyte-characters: t

Major mode: Emacs-Lisp

Minor modes in effect:
  whitespace-mode: t
  cua-mode: t
  tooltip-mode: t
  mouse-wheel-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
  temp-buffer-resize-mode: t
  line-number-mode: t
  transient-mark-mode: t

Recent input:
l e c t <tab> <tab> <tab> <tab> <tab> <tab> <backspace> 
<backspace> <backspace> <backspace> <backspace> <backspace> 
<backspace> <backspace> <backspace> <backspace> <backspace> 
- w i <tab> <tab> <C-left> <end> <tab> <tab> <tab> 
<tab> <tab> <tab> <C-left> <left> l <end> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> C-g C-g C-g <up> <left> 
<left> <delete> <up> ; <f2> <down> <down> <right> C-x 
C-e C-h f s e l e c <tab> - w i n <tab> <tab> <tab> 
<tab> <C-left> <C-left> <C-right> <tab> <C-left> <left> 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<C-left> <left> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> <C-left> <C-left> <C-right> 
<tab> <tab> <tab> <tab> <tab> <tab> C-g <up> <left> 
<left> ; <left> <up> <delete> <up> <delete> <f2> <down> 
<down> <down> <end> C-x C-e <up> C-h f s e l e - w 
i n d <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <tab> <C-left> <left> <tab> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <left> <left> <C-left> <left> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> C-g C-g C-h f s e l - w i n 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> C-g 
C-g <up> <C-left> <up> ; <f2> <down> <down> <down> 
<end> C-x C-e C-h f s e l e c t - w i n d o <C-left> 
<C-left> <C-right> <backspace> <backspace> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> <tab> C-g C-g C-h f s 
e l - w i n <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> C-g C-g M-x r e p o r t - w i n <backspace> <backspace> 
e <backspace> <backspace> e m a - b u <tab> <retur
n>

bug#4176: Acknowledgement (23.1; partial completions)

[Eli Barzilay]

After a quick look, I tried to redefine `minibuffer-complete' and just
comment out the second (goto-char (field-end)).  This seems to work
wrt wrongly moving the cursor to the end which seems to resolve both
problems when combined with

  (setq completion-styles '(partial-completion))

However,

1. Since this code was put there explicitly, there must be some issue
   I don't see in removing it.

2. Even after the quick code browse I have no idea what wonderful
   features I'm losing by removing `basic' and `emacs22' from
   `completion-styles'.  I would be happy if someone can enlighten me.

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                    http://barzilay.org/                   Maze is Life!

bug#4176: Acknowledgement (23.1; partial completions)

[Stefan Monnier]

> After a quick look, I tried to redefine `minibuffer-complete' and just
> comment out the second (goto-char (field-end)).  This seems to work
> wrt wrongly moving the cursor to the end which seems to resolve both
[...]
> 1. Since this code was put there explicitly, there must be some issue
>    I don't see in removing it.

I don't have time to dig into this right now, but the cursor placement
is fairly tricky indeed, and there are often different valid choices
depending on your particular preferences.

The rule I tried to follow (in the partia-completion code) is that the
cursor should be placed to the rightmost position that still offers ways
to persue completion.  In your `select-window' example, it seems that
the behavior doesn't follow this rule, so it might indeed be
a plain bug.

> 2. Even after the quick code browse I have no idea what wonderful
>    features I'm losing by removing `basic' and `emacs22' from
>    `completion-styles'.  I would be happy if someone can enlighten me.

You lose the fact that the default Emacs-23 completion is a compromise
between the non-partial-completion and the partial-completion behavior.
If you used partial-completion and liked it, you probably want to do
(setq completion-styles '(partial-completion)) like you did.
But remember that enabling partial-completion by default in Emacs-23
didn't get accepted without a fight.


        Stefan

bug#4176: marked as done (23.1; partial completions)

[Emacs bug Tracking System]

[-- Attachment #1: Type: text/plain, Size: 865 bytes --]

Your message dated Tue, 18 Aug 2009 15:11:54 -0400
with message-id <jwv1vn9uev7.fsf-monnier+emacsbugreports@gnu.org>
and subject line Re: bug#4176: 23.1; partial completions
has caused the Emacs bug report #4176,
regarding 23.1; partial completions
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@emacsbugs.donarmstrong.com
immediately.)


-- 
4176: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=4176
Emacs Bug Tracking System
Contact owner@emacsbugs.donarmstrong.com with problems

[-- Attachment #2: Type: message/rfc822, Size: 5584 bytes --]

From: Eli Barzilay <eli@barzilay.org>
To: bug-gnu-emacs@gnu.org
Subject: 23.1; partial completions
Date: Mon, 17 Aug 2009 10:30:10 -0400
Message-ID: <19081.26994.931953.186612@winooski.ccs.neu.edu>

With the default settings, enter

  C-h f select-window TAB

and Emacs will tell you that this is the only completion.

Trying this with `sele-wind TAB' does a partial completion to
`select-window' and moves the cursor to the right place -- a
second TAB pops up the completions (showing both functions) saying
"complete but not unique" *but* moves the cursor to the end.

It would be very nice if the cursor was always at the right place, and
it would not say that this is the only completion.  Doing

  (setq completion-styles '(partial-completion))

is fixes the message, but the cursor still moves to the end.

Another related problem -- the doc string for `completion-styles'
doesn't say anything about what's allowed (should mention
`completion-styles-alist').  Also, the functions that are used in
`completion-styles-alist' are not documented -- which leaves the user
with no explanation at all about the available completion styles.



In GNU Emacs 23.1.1 (x86_64-unknown-linux-gnu, GTK+ Version 2.10.14)
 of 2009-08-01 on winooski.ccs.neu.edu
Windowing system distributor `The X.Org Foundation', version 11.0.10300000
configured using `configure  '--prefix=/home/eli/bin/local/emacs-dir''

Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: POSIX
  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
  value of $XMODIFIERS: nil
  locale-coding-system: iso-latin-1-unix
  default-enable-multibyte-characters: t

Major mode: Emacs-Lisp

Minor modes in effect:
  whitespace-mode: t
  cua-mode: t
  tooltip-mode: t
  mouse-wheel-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
  temp-buffer-resize-mode: t
  line-number-mode: t
  transient-mark-mode: t

Recent input:
l e c t <tab> <tab> <tab> <tab> <tab> <tab> <backspace> 
<backspace> <backspace> <backspace> <backspace> <backspace> 
<backspace> <backspace> <backspace> <backspace> <backspace> 
- w i <tab> <tab> <C-left> <end> <tab> <tab> <tab> 
<tab> <tab> <tab> <C-left> <left> l <end> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> C-g C-g C-g <up> <left> 
<left> <delete> <up> ; <f2> <down> <down> <right> C-x 
C-e C-h f s e l e c <tab> - w i n <tab> <tab> <tab> 
<tab> <C-left> <C-left> <C-right> <tab> <C-left> <left> 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<C-left> <left> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> <C-left> <C-left> <C-right> 
<tab> <tab> <tab> <tab> <tab> <tab> C-g <up> <left> 
<left> ; <left> <up> <delete> <up> <delete> <f2> <down> 
<down> <down> <end> C-x C-e <up> C-h f s e l e - w 
i n d <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <tab> <C-left> <left> <tab> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <left> <left> <C-left> <left> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> C-g C-g C-h f s e l - w i n 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> C-g 
C-g <up> <C-left> <up> ; <f2> <down> <down> <down> 
<end> C-x C-e C-h f s e l e c t - w i n d o <C-left> 
<C-left> <C-right> <backspace> <backspace> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> <tab> <tab> <tab> <tab> <tab> C-g C-g C-h f s 
e l - w i n <tab> <tab> <tab> <tab> <tab> <tab> <tab> 
<tab> C-g C-g M-x r e p o r t - w i n <backspace> <backspace> 
e <backspace> <backspace> e m a - b u <tab> <retur
n>



[-- Attachment #3: Type: message/rfc822, Size: 5199 bytes --]

From: Stefan Monnier <monnier@IRO.UMontreal.CA>
To: Eli Barzilay <eli@barzilay.org>
Subject: Re: bug#4176: 23.1; partial completions
Date: Tue, 18 Aug 2009 15:11:54 -0400
Message-ID: <jwv1vn9uev7.fsf-monnier+emacsbugreports@gnu.org>

>>>>> "Eli" == Eli Barzilay <eli@barzilay.org> writes:

> With the default settings, enter
>   C-h f select-window TAB

> and Emacs will tell you that this is the only completion.

> Trying this with `sele-wind TAB' does a partial completion to
> `select-window' and moves the cursor to the right place -- a
> second TAB pops up the completions (showing both functions) saying
> "complete but not unique" *but* moves the cursor to the end.

Removing the second (goto-char (field-end)) in minibuffer-complete was
indeed a correct solution, in the end.  I've installed a sligtly
different patch so that the point-motion is consolidated inside
completion--do-completion.  See below the patch I installed.

> It would be very nice if the cursor was always at the right place, and
> it would not say that this is the only completion.  Doing

>   (setq completion-styles '(partial-completion))

> is fixes the message, but the cursor still moves to the end.

The "sole completion" message is correct, given the default value of
completion-styles.

> Another related problem -- the doc string for `completion-styles'
> doesn't say anything about what's allowed (should mention
> `completion-styles-alist').

Thanks, it does mention it now.

> Also, the functions that are used in `completion-styles-alist' are not
> documented -- which leaves the user with no explanation at all about
> the available completion styles.

The functions are low-level, so the user shouldn't be expected to look
at them and/or their docstrings to find out what the style does.
The name of the style should hopefully be sufficient.


        Stefan


--- minibuffer.el.~1.77.~	2009-08-04 14:03:35.000000000 -0400
+++ minibuffer.el	2009-08-18 15:07:35.000000000 -0400
@@ -308,7 +308,8 @@
 ALL-COMPLETIONS is the function that lists the completions.")
 
 (defcustom completion-styles '(basic partial-completion emacs22)
-  "List of completion styles to use."
+  "List of completion styles to use.
+The available styles are listed in `completion-styles-alist'."
   :type `(repeat (choice ,@(mapcar (lambda (x) (list 'const (car x)))
                                    completion-styles-alist)))
   :group 'minibuffer
@@ -388,7 +389,9 @@
     (cond
      ((null comp)
       (ding) (minibuffer-message "No match") (minibuffer--bitset nil nil nil))
-     ((eq t comp) (minibuffer--bitset nil nil t)) ;Exact and unique match.
+     ((eq t comp)
+      (goto-char (field-end))
+      (minibuffer--bitset nil nil t)) ;Exact and unique match.
      (t
       ;; `completed' should be t if some completion was done, which doesn't
       ;; include simply changing the case of the entered string.  However,
@@ -462,11 +465,9 @@
 
       (case (completion--do-completion)
         (#b000 nil)
-        (#b001 (goto-char (field-end))
-               (minibuffer-message "Sole completion")
+        (#b001 (minibuffer-message "Sole completion")
                t)
-        (#b011 (goto-char (field-end))
-               (minibuffer-message "Complete, but not unique")
+        (#b011 (minibuffer-message "Complete, but not unique")
                t)
         (t     t)))))
 
@@ -694,11 +695,9 @@
   (interactive)
   (case (completion--do-completion 'completion--try-word-completion)
     (#b000 nil)
-    (#b001 (goto-char (field-end))
-           (minibuffer-message "Sole completion")
+    (#b001 (minibuffer-message "Sole completion")
            t)
-    (#b011 (goto-char (field-end))
-           (minibuffer-message "Complete, but not unique")
+    (#b011 (minibuffer-message "Complete, but not unique")
            t)
     (t     t)))