bug#1062: 23.0.60; minibuffer-completion-help incomplete

bug#1062: 23.0.60; minibuffer-completion-help incomplete

[Stephen Berman]

GNU Emacs 23.0.60.5 (i686-pc-linux-gnu, GTK+ Version 2.12.9) of
2008-10-01 on escher

1. emacs -Q
2. Let some directory contain the files "tested" and "Testing" and no other
   files beginning with "test" or "Test".
3. M-x set-variable RET read-file-name-completion-ignore-case RET t RET
4. C-x C-f <directory/in/step/2>/test
5. Now type TAB or "?" and a *Completions* buffer pops that shows only
   "tested" as a possible completion.  However, typing "i" next and then
   TAB does complete to "Testing".

If before doing step 4 completion-styles is set to (emacs22), then after
the first TAB or ? in step 5 the *Completions* buffer shows both
"tested" and "Testing".

The following patch makes the *Completions* buffer show all possible
completions in step 5 also with the default completion-style basic, but
I don't know if it has undesirable consequences elsewhere.

Steve Berman

*** emacs/lisp/minibuffer.el.~1.58.~	2008-09-04 16:16:19.000000000 +0200
--- emacs/lisp/minibuffer.el	2008-10-01 11:42:40.000000000 +0200
***************
*** 1316,1322 ****
                     "" (list (substring beforepoint (car bounds))
                              'point
                              (substring afterpoint 0 (cdr bounds)))))
!          (all (completion-pcm--all-completions prefix pattern table pred)))
      (completion-hilit-commonality
       (if (consp all) (nconc all (car bounds)) all)
       point)))
--- 1316,1324 ----
                     "" (list (substring beforepoint (car bounds))
                              'point
                              (substring afterpoint 0 (cdr bounds)))))
!          (all (if completion-ignore-case
! 		  (all-completions (substring string 0 point) table pred)
! 		(completion-pcm--all-completions prefix pattern table pred))))
      (completion-hilit-commonality
       (if (consp all) (nconc all (car bounds)) all)
       point)))

bug#1062: marked as done (23.0.60; minibuffer-completion-help incomplete)

[Emacs bug Tracking System]

[-- Attachment #1: Type: text/plain, Size: 854 bytes --]


Your message dated Sat, 08 Nov 2008 11:57:14 -0500
with message-id <87bpwq6tqt.fsf@cyd.mit.edu>
and subject line Re: 23.0.60; minibuffer-completion-help incomplete
has caused the Emacs bug report #1062,
regarding 23.0.60; minibuffer-completion-help incomplete
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 don@donarmstrong.com
immediately.)


-- 
1062: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=1062
Emacs Bug Tracking System
Contact don@donarmstrong.com with problems

[-- Attachment #2: Type: message/rfc822, Size: 4057 bytes --]

From: Stephen Berman <stephen.berman@gmx.net>
To: emacs-pretest-bug@gnu.org
Subject: 23.0.60; minibuffer-completion-help incomplete
Date: Wed, 01 Oct 2008 12:00:52 +0200
Message-ID: <87od24ljmz.fsf@escher.local.home>

GNU Emacs 23.0.60.5 (i686-pc-linux-gnu, GTK+ Version 2.12.9) of
2008-10-01 on escher

1. emacs -Q
2. Let some directory contain the files "tested" and "Testing" and no other
   files beginning with "test" or "Test".
3. M-x set-variable RET read-file-name-completion-ignore-case RET t RET
4. C-x C-f <directory/in/step/2>/test
5. Now type TAB or "?" and a *Completions* buffer pops that shows only
   "tested" as a possible completion.  However, typing "i" next and then
   TAB does complete to "Testing".

If before doing step 4 completion-styles is set to (emacs22), then after
the first TAB or ? in step 5 the *Completions* buffer shows both
"tested" and "Testing".

The following patch makes the *Completions* buffer show all possible
completions in step 5 also with the default completion-style basic, but
I don't know if it has undesirable consequences elsewhere.

Steve Berman

*** emacs/lisp/minibuffer.el.~1.58.~	2008-09-04 16:16:19.000000000 +0200
--- emacs/lisp/minibuffer.el	2008-10-01 11:42:40.000000000 +0200
***************
*** 1316,1322 ****
                     "" (list (substring beforepoint (car bounds))
                              'point
                              (substring afterpoint 0 (cdr bounds)))))
!          (all (completion-pcm--all-completions prefix pattern table pred)))
      (completion-hilit-commonality
       (if (consp all) (nconc all (car bounds)) all)
       point)))
--- 1316,1324 ----
                     "" (list (substring beforepoint (car bounds))
                              'point
                              (substring afterpoint 0 (cdr bounds)))))
!          (all (if completion-ignore-case
! 		  (all-completions (substring string 0 point) table pred)
! 		(completion-pcm--all-completions prefix pattern table pred))))
      (completion-hilit-commonality
       (if (consp all) (nconc all (car bounds)) all)
       point)))



[-- Attachment #3: Type: message/rfc822, Size: 1784 bytes --]

From: Chong Yidong <cyd@stupidchicken.com>
To: Stephen Berman <stephen.berman@gmx.net>
Cc: 1062-done@emacsbugs.donarmstrong.com
Subject: Re: 23.0.60; minibuffer-completion-help incomplete
Date: Sat, 08 Nov 2008 11:57:14 -0500
Message-ID: <87bpwq6tqt.fsf@cyd.mit.edu>

> 1. emacs -Q
> 2. Let some directory contain the files "tested" and "Testing" and no
> other
>    files beginning with "test" or "Test".
> 3. M-x set-variable RET read-file-name-completion-ignore-case RET t RET
> 4. C-x C-f <directory/in/step/2>/test
> 5. Now type TAB or "?" and a *Completions* buffer pops that shows only
>    "tested" as a possible completion.  However, typing "i" next and then
>    TAB does complete to "Testing".
>
> The following patch makes the *Completions* buffer show all possible
> completions in step 5 also with the default completion-style basic, but
> I don't know if it has undesirable consequences elsewhere.

The bug was actually in the C code: file-name-all-completions did not
handle completion-regexp-list properly.  I've checked in a fix.  Thanks
for the bug report.