On 12/30/11 1:23 AM, Eli Zaretskii wrote:
>> Date: Thu, 29 Dec 2011 13:50:32 -0800
>> From: Daniel Colascione <dancol@dancol.org>
>> CC: emacs-devel@gnu.org
>>
>>>>>> -  (if (eq system-type 'windows-nt)
>>>>>> +  (if (featurep 'w32)
>>>>>>        '(("iso8859-1" "ms-oemlatin")
>>>>>>  	("gb2312.1980" "gb2312" "gbk" "gb18030")
>>>>>>  	("jisx0208.1990" "jisx0208.1983" "jisx0208.1978")
>>>>>
>>>>> Why not use window-system (the function) here?
>>>>
>>>> Isn't it a little early for window-system at this point?
>>>
>>> A defcustom is evaluated more than once.  So I'm a bit fuzzy about
>>> "too early" ;-)
>>
>> When we initialize for the first time, we won't have a GUI frame to
>> check, yes? (And then we dump the table into pure storage.)
> 
> In a release tarball, this first time is normally on a GNU/Linux box,
> and so the w32-specific setting will never be chosen.
> 
> The important initialization is when the dumped Emacs is started on a
> Windows system.  At that time, we do have a GUI frame, I think.
> 
>> If there's
>> no facility to force reevaluation of the mapping table, we'll end up
>> using the wrong one when we *do* finally come up with a GUI frame.
> 
> But there _is_ such a facility: defcustoms are reevaluated upon
> startup (and then again when you actually use "M-x customize", or have
> custom-related forms in your .emacs).  

~/software/emacs/cyg.w32/src
$ ./emacs --daemon
("./emacs")
...
Restarting server

~/software/emacs/cyg.w32/src
$ cd ../lib-src/

~/software/emacs/cyg.w32/lib-src
$ ./emacsclient --eval '(emacs-version)'
"GNU Emacs 24.0.92.71 (i686-pc-cygwin)\n of 2011-12-29 on dfaew"

~/software/emacs/cyg.w32/lib-src
$ ./emacsclient --eval '(window-system)'
nil

~/software/emacs/cyg.w32/lib-src
$ ./emacsclient --eval 'window-system'
nil

~/software/emacs/cyg.w32/lib-src
$ ./emacsclient --eval "(featurep 'w32)"
t

> My point, however, is that it is a good idea to use window-system (the
> function) where it could happen that different frames have different
> types, in the current mult-tty Emacs.  If nothing else, using
> window-system consistently contributes to the clarity of the code and
> shows the overall design to newcomers, who would otherwise become
> confused if they see inconsistent tests in each and every place.

I agree that using window-system where appropriate is best.  That said,
window-system isn't always appropriate.  It answers the question "what
window system am I using right now?".  The featurep query answers "what
window system is this copy of Emacs capable of using?".  Sometimes, the
latter question is of greater interest.