maybe don't set completion-ignore-case to t on w32 ?

maybe don't set completion-ignore-case to t on w32 ?

[John Paul Wallington]

Is it necessary to set `completion-ignore-case' to t in w32-fns.el for
the purpose of ignoring case on file-name completion now that we have
`read-file-name-completion-ignore-case' that defaults to t on
MS-Windows systems?

I'm using an MS-Windows system here, and M-x cust[TAB] gives `Custom-'
prefixed commands where I would only get `customize-' prefixed commands
with `completion-ignore-case' set to t.

Proposed change:

--- w32-fns.el	21 Jan 2007 03:53:10 +0000	1.64
+++ w32-fns.el	04 Mar 2007 20:39:57 +0000	
@@ -43,9 +43,6 @@
 (define-key function-key-map [M-backspace] [?\M-\177])
 (define-key function-key-map [C-M-backspace] [\C-\M-delete])
 
-;; Ignore case on file-name completion
-(setq completion-ignore-case t)
-
 ;; Map all versions of a filename (8.3, longname, mixed case) to the
 ;; same buffer.
 (setq find-file-visit-truename t)

Re: maybe don't set completion-ignore-case to t on w32 ?

[Eli Zaretskii]

> Date: Sun, 04 Mar 2007 20:55:08 +0000
> From: jpw@pobox.com (John Paul Wallington)
> 
> Is it necessary to set `completion-ignore-case' to t in w32-fns.el for
> the purpose of ignoring case on file-name completion now that we have
> `read-file-name-completion-ignore-case' that defaults to t on
> MS-Windows systems?

I don't know if it's necessary, but I suggest that we not change
anything at this point that does not fix a clear bug.

> I'm using an MS-Windows system here, and M-x cust[TAB] gives `Custom-'
> prefixed commands where I would only get `customize-' prefixed commands
> with `completion-ignore-case' set to t.

I cannot reproduce that, for me "M-x cust[TAB]" yields `customize'
(sic, without the trailing dash that you show above), and tapping on
TAB 2 more times shows a completion buffer where all completions begin
with "customize", with a lower-case `c'.

Do you see this problem in "emacs -Q"?

Re: maybe don't set completion-ignore-case to t on w32 ?

[John Paul Wallington]

Eli Zaretskii <eliz@gnu.org> writes:

>> Is it necessary to set `completion-ignore-case' to t in w32-fns.el for
>> the purpose of ignoring case on file-name completion now that we have
>> `read-file-name-completion-ignore-case' that defaults to t on
>> MS-Windows systems?
>
> I don't know if it's necessary, but I suggest that we not change
> anything at this point that does not fix a clear bug.

Fair enough.  I think it would be a good change to make after the
release though.

>> I'm using an MS-Windows system here, and M-x cust[TAB] gives `Custom-'
>> prefixed commands where I would only get `customize-' prefixed commands
>> with `completion-ignore-case' set to t.
>
> I cannot reproduce that, for me "M-x cust[TAB]" yields `customize'
> (sic, without the trailing dash that you show above), and tapping on
> TAB 2 more times shows a completion buffer where all completions begin
> with "customize", with a lower-case `c'.
> Do you see this problem in "emacs -Q"?

I think that's because only `customize-' prefixed stuff is defined
after emacs -Q.  Doing emacs -Q then M-x customize-variable varname
[RET] loads cus-edit at which point M-x cust [TAB} shows both
`customize-' and `Custom-' prefixed commands.