From: matz@... (Yukihiro Matsumoto)
Date: 2002-05-15T05:56:31+09:00
Subject: Re: Useless patch...

Hi,

In message "Useless patch..."
    on 02/05/14, Michal Rokos <m.rokos@sh.cvut.cz> writes:

|	yesterday I read through array.c and made some comments for my
|	self.
|
|	Here's the diff as a result. There can be maybe some good bits
|	in it...
|
|	If you have a time read it with me...
<snip>
|	Comments are very welcomed.
|	(But who would read it? :-))

I did.  Comments:

  * I'm not sure whether we don't need size adjustment in
    rb_ary_shift() any longer.  Since we have bigger memory these
    days, maybe it will be a problem.
  * you don't need RTEST around rb_range_beg_len().
  * In the source code of Ruby, Qtrue and Qfalse are also serve as C
    boolean (int), so you don't have to replace them.
  * I use -2 indentation for labels.
  * rb_scan_args() is slow enough to avoid in frequently used
    execution path like in rb_ary_aref().
  * return value from inspect_ensure is not used, so that constant
    value might be better, how about Qfalse?
  * In sort_2, you don't need FIX2INT() for Fixnum values.
  * you don't need RTEST around rb_equal().

Other changes are OK.

You *must* have CVS account now, since you broke the promise.  Remember?
Send your prefered account name and crypted password to <cvs-admin@ruby-lang.org>.

							matz.