23.0.50; PC-do-completion should fix case if possible

23.0.50; PC-do-completion should fix case if possible

[Denis Howe]

[-- Attachment #1: Type: text/plain, Size: 541 bytes --]

In GNU Emacs 23.0.50.2 (i686-pc-linux-gnu, GTK+ Version 2.10.11), I do

  (setq read-file-name-completion-ignore-case t)
  (partial-completion-mode)

then type

  C-x C-f c

in a directory containing directory "Cat" and file "Cattle".  Normal
completion will correct the case of my "c", completing to "Cat",
matching the files.  Partial completion (complete.el) completes to
"cat", which is wrong.

Attached is a patch against the latest from CVS.

-- 
Denis Howe <denis.howe@gmail.com>
http://foldoc.org/
T: 020 8450 9448  M: 07950 686 615

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: complete.patch --]
[-- Type: text/x-patch; name="complete.patch", Size: 756 bytes --]

diff -w GNU/complete.el Fixed/complete.el
734,735d733
< 		  ;; Retain capitalization of user input even if
< 		  ;; completion-ignore-case is set.
742a741
>                                            (or
744,747c743
< 					       (downcase (following-char))))
< 				      ;; same char (modulo case); no action
< 				      (forward-char 1)
< 				    (if (and (< (point) end)
---
>                                                 (downcase (following-char)))
750c746
< 							PC-delims-list)))
---
>                                                        PC-delims-list))))
751a748
>                                       ;; or input char by prefix char
763c760
< 					(setq end (1+ end)))))
---
>                                       (setq end (1+ end))))

[-- Attachment #3: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: 23.0.50; PC-do-completion should fix case if possible

[Richard Stallman]

Would someone who uses and understands complete.el please check this
patch and DTRT?  It is small enough that we could install it without
legal papers.

Denis, in the future, would you please use `diff -c' to make patches,
and please send a change log entry with each change?

X-Spam-Status: No, score=0.0 required=5.0 tests=UNPARSEABLE_RELAY 
	autolearn=failed version=3.1.0
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=beta;
	h=domainkey-signature:received:received:message-id:date:from:to:subject:mime-version:content-type;
	bh=4BExZXVqaRJrSWgOgUgeuAUzVjgNTId0uZePscIqdcE=;
	b=kWfgiP5Gy3M5rDrv9rNmlxc94usWzR14ipPv6CPAdT6quGDEvqn0wA7eMQm2q0NAAkS4zYemoAujhR9ab1j7o5QkyRVgIOIjC9w9AEjTN4XHavbG2YbCu+kyRxYHxDDZ79HfS26GjD9kTVH23owtNyauJimDgF7VjbLQx1UBIxU=
Date: Tue, 2 Oct 2007 15:22:03 +0100
From: "Denis Howe" <denis.howe@gmail.com>
To: emacs-pretest-bug@gnu.org
MIME-Version: 1.0
Content-Type: multipart/mixed; 
	boundary="----=_Part_3563_14697534.1191334923617"
Cc: 
Subject: 23.0.50; PC-do-completion should fix case if possible

------=_Part_3563_14697534.1191334923617
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

In GNU Emacs 23.0.50.2 (i686-pc-linux-gnu, GTK+ Version 2.10.11), I do

  (setq read-file-name-completion-ignore-case t)
  (partial-completion-mode)

then type

  C-x C-f c

in a directory containing directory "Cat" and file "Cattle".  Normal
completion will correct the case of my "c", completing to "Cat",
matching the files.  Partial completion (complete.el) completes to
"cat", which is wrong.

Attached is a patch against the latest from CVS.

-- 
Denis Howe <denis.howe@gmail.com>
http://foldoc.org/
T: 020 8450 9448  M: 07950 686 615

------=_Part_3563_14697534.1191334923617
Content-Type: text/x-patch; name="complete.patch"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="complete.patch";
 filename="complete.patch"
X-Attachment-Id: f_f7ai1dz9
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------=_Part_3563_14697534.1191334923617
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel
------=_Part_3563_14697534.1191334923617--

Re: 23.0.50; PC-do-completion should fix case if possible

[Denis Howe]

This is an improved version of my report sent yesterday.

In the latest complete.el from CVS (v1.79),
function "PC-do-completion" fails to correct
the case of the input being completed when:

- "completion-ignore-case" is true and
- "read-file-name-completion-ignore-case" is true and
- all possible completions have the same case

Ordinary completion preserves input case only where the
potential completions differ in case.  If they agree then ordinary
completion changes the input case to that of the prefix.

For example, evaluate:

(partial-completion-mode 0)
(setq completion-ignore-case t)
(completing-read "Input: " '(("Den".1)("Denis".2)) nil t "d")

Hitting TAB runs ordinary "minibuffer-complete"
which completes "d" to "Den" (case corrected).

Enable partial completion:

(partial-completion-mode 1)

Re-evaluate the above completing-read.

Hitting TAB now runs "PC-complete" which
completes "d" to "den" (fails to correct input case).

I've tested the attached patch thoroughly.
Please contact me if you want any more background on this.

------------------------------------------------------------

ChangeLog entry:

2007-10-03  Denis Howe  <denis.howe@gmail.com>

	* complete.el (PC-do-completion): Correct case when unambiguous.

------------------------------------------------------------

diff -cw GNU/complete.el Fix/complete.el
*** GNU/complete.el	2007-10-03 11:33:55.000000000 +0100
--- Fix/complete.el	2007-09-19 17:15:12.000000000 +0100
***************
*** 731,738 ****
  		     (setq prefix (PC-try-completion
  				   (PC-chunk-after basestr skip) poss)))
  		(let ((first t) i)
- 		  ;; Retain capitalization of user input even if
- 		  ;; completion-ignore-case is set.
  		  (if (eq mode 'word)
  		      (setq prefix (PC-chop-word prefix basestr)))
  		  (goto-char (+ beg (length dirname)))
--- 731,736 ----
***************
*** 740,754 ****
  				(setq i 0) ; index into prefix string
  				(while (< i (length prefix))
  				  (if (and (< (point) end)
  					   (eq (downcase (aref prefix i))
! 					       (downcase (following-char))))
! 				      ;; same char (modulo case); no action
! 				      (forward-char 1)
! 				    (if (and (< (point) end)
  					     (and (looking-at " ")
                                                    (memq (aref prefix i)
! 							PC-delims-list)))
  					;; replace " " by the actual delimiter
  					(progn
  					  (delete-char 1)
  					  (insert (substring prefix i (1+ i))))
--- 738,751 ----
                                  (setq i 0) ; index into prefix string
                                  (while (< i (length prefix))
                                    (if (and (< (point) end)
+                                            (or
                                              (eq (downcase (aref prefix i))
!                                                 (downcase (following-char)))
                                              (and (looking-at " ")
                                                   (memq (aref prefix i)
!                                                        PC-delims-list))))
                                        ;; replace " " by the actual delimiter
+                                       ;; or input char by prefix char
                                        (progn
                                          (delete-char 1)
                                          (insert (substring prefix i (1+ i))))
***************
*** 760,766 ****
                                                 (setq end (1- end))))
  					(setq improved t)
                                          (insert (substring prefix i (1+ i)))
! 					(setq end (1+ end)))))
  				  (setq i (1+ i)))
  				(or pt (setq pt (point)))
  				(looking-at PC-delim-regex))
--- 757,763 ----
                                               (setq end (1- end))))
                                        (setq improved t)
                                        (insert (substring prefix i (1+ i)))
!                                       (setq end (1+ end))))
                                    (setq i (1+ i)))
                                  (or pt (setq pt (point)))
                                  (looking-at PC-delim-regex))

------------------------------------------------------------

bug#1976: 23.0.50; PC-do-completion should fix case if possible

[Denis Howe]

[-- Attachment #1: Type: text/plain, Size: 4835 bytes --]

My 2007-10-03 attempt to get this patch applied got no further than RMS's
request for someone to DTRT.  It took several hours of my time to debug and
test, and the bug still bites me every time I re-install, so PLEASE could
someone apply it?  Thanks.
------------------------------------------------------------
In the latest complete.el from CVS (v1.94), function "PC-do-completion"
fails to correct the case of the input being completed when:

- "completion-ignore-case" is true and
- "read-file-name-completion-ignore-case" is true and
- all possible completions have the same case

Ordinary completion preserves input case only where the potential
completions differ in case.  If they agree then ordinary completion changes
the input case to that of the prefix.

For example, evaluate:

(partial-completion-mode 0)
(setq completion-ignore-case t)
(completing-read "Input: " '(("Den".1)("Denis".2)) nil t "d")

Hitting TAB runs ordinary "minibuffer-complete" which completes "d" to "Den"
(case corrected).

Enable partial completion:

(partial-completion-mode 1)

Re-evaluate the above completing-read.

Hitting TAB now runs "PC-complete" which completes "d" to "den" (fails to
correct input case).

I've tested the attached patch thoroughly.  Please contact me if you want
any more background on this.

------------------------------------------------------------

ChangeLog entry:

2009-01-21  Denis Howe <denis.howe@gmail.com>

       * complete.el (PC-do-completion): Correct case when unambiguous.

------------------------------------------------------------

cvs diff -cw complete.el
Index: complete.el
===================================================================
RCS file: /sources/emacs/emacs/lisp/complete.el,v
retrieving revision 1.94
diff -c -w -r1.94 complete.el
*** complete.el    9 Jan 2009 04:44:19 -0000    1.94
--- complete.el    21 Jan 2009 14:21:17 -0000
***************
*** 737,744 ****
               (setq prefix (PC-try-completion
                     (PC-chunk-after basestr skip) poss)))
          (let ((first t) i)
-           ;; Retain capitalization of user input even if
-           ;; completion-ignore-case is set.
            (if (eq mode 'word)
                (setq prefix (PC-chop-word prefix basestr)))
            (goto-char (+ beg (length dirname)))
--- 737,742 ----
***************
*** 746,760 ****
                  (setq i 0) ; index into prefix string
                  (while (< i (length prefix))
                    (if (and (< (point) end)
                         (eq (downcase (aref prefix i))
!                            (downcase (following-char))))
!                       ;; same char (modulo case); no action
!                       (forward-char 1)
!                     (if (and (< (point) end)
                           (and (looking-at " ")
                                                    (memq (aref prefix i)
!                             PC-delims-list)))
                      ;; replace " " by the actual delimiter
                      (progn
                        (delete-char 1)
                        (insert (substring prefix i (1+ i))))
--- 744,757 ----
                  (setq i 0) ; index into prefix string
                  (while (< i (length prefix))
                    (if (and (< (point) end)
+                                            (or
                         (eq (downcase (aref prefix i))
!                                                 (downcase
(following-char)))
                           (and (looking-at " ")
                                                    (memq (aref prefix i)
!                                                        PC-delims-list))))
                      ;; replace " " by the actual delimiter
+                                       ;; or input char by prefix char
                      (progn
                        (delete-char 1)
                        (insert (substring prefix i (1+ i))))
***************
*** 766,772 ****
                                                 (setq end (1- end))))
                      (setq improved t)
                                          (insert (substring prefix i (1+
i)))
!                     (setq end (1+ end)))))
                    (setq i (1+ i)))
                  (or pt (setq pt (point)))
                  (looking-at PC-delim-regex))
--- 763,769 ----
                                                 (setq end (1- end))))
                      (setq improved t)
                                          (insert (substring prefix i (1+
i)))
!                                       (setq end (1+ end))))
                    (setq i (1+ i)))
                  (or pt (setq pt (point)))
                  (looking-at PC-delim-regex))

------------------------------------------------------------
End

-- Denis Howe <denis.howe@gmail.com>

[-- Attachment #2: Type: text/html, Size: 11236 bytes --]

bug#1976: 23.0.50; PC-do-completion should fix case if possible

[Stefan Monnier]

> My 2007-10-03 attempt to get this patch applied got no further than RMS's
> request for someone to DTRT.  It took several hours of my time to debug and
> test, and the bug still bites me every time I re-install, so PLEASE could
> someone apply it?  Thanks.

BTW, the default completion in Emacs-23 is very close to
partial-completion-mode.  More to the point, if you remove `basic' from
`completion-styles', it should be pretty much identical except for the
following known differences:
- no <foo.h> style completion for file names.
- M-x lch TAB doesn't know how to expand to `list-command-history'.
And even if you don't remove `basic', the behavior is fairly close.
Also it should fix the problem you're reporting and provides
a few improvements.

Since partial-completion-mode is slated for obsolescence (tho not
marked obsolete yet), I'm interested to hear if you find some other
missing feature.


        Stefan

bug#1976: 23.0.50; PC-do-completion should fix case if possible

[Denis Howe]

[-- Attachment #1: Type: text/plain, Size: 354 bytes --]

2009/1/21 Stefan Monnier <monnier@iro.umontreal.ca>

> the default completion in Emacs-23 is very close to partial-completion-mode


OMG, so it is!   I've been using partial-completion-mode so long I can't
remember why I turned it on.  What's even stupider is that I've never
actually used the "partial" feature.  Thanks very much for the
enlightenment.

[-- Attachment #2: Type: text/html, Size: 676 bytes --]

bug#1976: marked as done (23.0.50; PC-do-completion should fix case if possible)

[Emacs bug Tracking System]

[-- Attachment #1: Type: text/plain, Size: 892 bytes --]


Your message dated Wed, 28 Jan 2009 10:36:04 -0500
with message-id <87bptr1lzv.fsf@cyd.mit.edu>
and subject line Re: 23.0.50; PC-do-completion should fix case if possible
has caused the Emacs bug report #1976,
regarding 23.0.50; PC-do-completion should fix case if possible
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.)


-- 
1976: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=1976
Emacs Bug Tracking System
Contact owner@emacsbugs.donarmstrong.com with problems

[-- Attachment #2: Type: message/rfc822, Size: 19792 bytes --]

[-- Attachment #2.1.1: Type: text/plain, Size: 4835 bytes --]

My 2007-10-03 attempt to get this patch applied got no further than RMS's
request for someone to DTRT.  It took several hours of my time to debug and
test, and the bug still bites me every time I re-install, so PLEASE could
someone apply it?  Thanks.
------------------------------------------------------------
In the latest complete.el from CVS (v1.94), function "PC-do-completion"
fails to correct the case of the input being completed when:

- "completion-ignore-case" is true and
- "read-file-name-completion-ignore-case" is true and
- all possible completions have the same case

Ordinary completion preserves input case only where the potential
completions differ in case.  If they agree then ordinary completion changes
the input case to that of the prefix.

For example, evaluate:

(partial-completion-mode 0)
(setq completion-ignore-case t)
(completing-read "Input: " '(("Den".1)("Denis".2)) nil t "d")

Hitting TAB runs ordinary "minibuffer-complete" which completes "d" to "Den"
(case corrected).

Enable partial completion:

(partial-completion-mode 1)

Re-evaluate the above completing-read.

Hitting TAB now runs "PC-complete" which completes "d" to "den" (fails to
correct input case).

I've tested the attached patch thoroughly.  Please contact me if you want
any more background on this.

------------------------------------------------------------

ChangeLog entry:

2009-01-21  Denis Howe <denis.howe@gmail.com>

       * complete.el (PC-do-completion): Correct case when unambiguous.

------------------------------------------------------------

cvs diff -cw complete.el
Index: complete.el
===================================================================
RCS file: /sources/emacs/emacs/lisp/complete.el,v
retrieving revision 1.94
diff -c -w -r1.94 complete.el
*** complete.el    9 Jan 2009 04:44:19 -0000    1.94
--- complete.el    21 Jan 2009 14:21:17 -0000
***************
*** 737,744 ****
               (setq prefix (PC-try-completion
                     (PC-chunk-after basestr skip) poss)))
          (let ((first t) i)
-           ;; Retain capitalization of user input even if
-           ;; completion-ignore-case is set.
            (if (eq mode 'word)
                (setq prefix (PC-chop-word prefix basestr)))
            (goto-char (+ beg (length dirname)))
--- 737,742 ----
***************
*** 746,760 ****
                  (setq i 0) ; index into prefix string
                  (while (< i (length prefix))
                    (if (and (< (point) end)
                         (eq (downcase (aref prefix i))
!                            (downcase (following-char))))
!                       ;; same char (modulo case); no action
!                       (forward-char 1)
!                     (if (and (< (point) end)
                           (and (looking-at " ")
                                                    (memq (aref prefix i)
!                             PC-delims-list)))
                      ;; replace " " by the actual delimiter
                      (progn
                        (delete-char 1)
                        (insert (substring prefix i (1+ i))))
--- 744,757 ----
                  (setq i 0) ; index into prefix string
                  (while (< i (length prefix))
                    (if (and (< (point) end)
+                                            (or
                         (eq (downcase (aref prefix i))
!                                                 (downcase
(following-char)))
                           (and (looking-at " ")
                                                    (memq (aref prefix i)
!                                                        PC-delims-list))))
                      ;; replace " " by the actual delimiter
+                                       ;; or input char by prefix char
                      (progn
                        (delete-char 1)
                        (insert (substring prefix i (1+ i))))
***************
*** 766,772 ****
                                                 (setq end (1- end))))
                      (setq improved t)
                                          (insert (substring prefix i (1+
i)))
!                     (setq end (1+ end)))))
                    (setq i (1+ i)))
                  (or pt (setq pt (point)))
                  (looking-at PC-delim-regex))
--- 763,769 ----
                                                 (setq end (1- end))))
                      (setq improved t)
                                          (insert (substring prefix i (1+
i)))
!                                       (setq end (1+ end))))
                    (setq i (1+ i)))
                  (or pt (setq pt (point)))
                  (looking-at PC-delim-regex))

------------------------------------------------------------
End

-- Denis Howe <denis.howe@gmail.com>

[-- Attachment #2.1.2: Type: text/html, Size: 11236 bytes --]

[-- Attachment #3: Type: message/rfc822, Size: 1453 bytes --]

From: Chong Yidong <cyd@stupidchicken.com>
To: Denis Howe <denis.howe@gmail.com>
Cc: 1976-done@emacsbugs.donarmstrong.com
Subject: Re: 23.0.50; PC-do-completion should fix case if possible
Date: Wed, 28 Jan 2009 10:36:04 -0500
Message-ID: <87bptr1lzv.fsf@cyd.mit.edu>

> > the default completion in Emacs-23 is very close to
> > partial-completion-mode
>
> OMG, so it is!  I've been using partial-completion-mode so long I
> can't remember why I turned it on.  What's even stupider is that I've
> never actually used the "partial" feature.  Thanks very much for the
> enlightenment.

In any case, I've checked in your patch to complete.el.  Thanks.