[ruby-list:38838] ruby-dev summary: 21928-22011

高橋征義です。

そんなわけで、ruby-dev summary先週分です。

# とりあえず [ruby-dev:21938] 1.8.1 schedule? は
# 入れないことにしました……。


[ruby-dev:21932] rough / tabs.rb

  Minero Aoki reviewed tabs.rb in rough directory in CVS
because he wanted a standard library to treat tabs. But
he have some complaints of that, so he rewrote new one.

  * much faster than original in difficult case
  * nearly fast in normal case (some tabs in line head)
  * bang method is slower than non-bang method

He thought:

  * is it much suitable to name expand_tabs when it's in String class?
  * someone may want to put under the Text module, not top class.
    (Minero himself doesn't case of it)
  * may want aliases tabify/untabify
  * may want detab/entab

    FYI:
     google(expandtab)    =   13900
     google(expandtabs)   =    3380
     google(expand tab)   = 1340000
     google(unexpandtab)  =       0
     google(unexpandtabs) =      18
     google(unexpand tab) =    3290
     google(untabify)     =    8260
     google(tabify)       =    4680
     google(detab)        =    7340
     google(entab)        =    3130


   [ruby-dev:21933] new hash literal

Nobu Nakata saw Matz's slides
(http://www.rubyist.net/~matz/slides/rc2003/mgp00027.html)
and made a patch for a new hash literal like {key: value}.


  [ruby-dev:21955] block local parameters in Ruby2

Shugo Maeda saw Matz's slides of RC2003 and found a notation:

  * Block parameters will be block local
    even if variables with same names exist

But he thought a sample like below should raise an error.

  def foo
    x = 1
    iter do |x|
      p x #  x is block local  -> error!
    end
    p x #  x is method local
  end


Maeda and Matz feel that the style like that is
ugly, but some people doesn't. So Matz will allow
such nested variables with warning.


  [ruby-dev:21990] environment variables in ruby.1

Minero Aoki added a new section 'environment variables'
to man page of Ruby. Now we can see it in CVS.


  [ruby-dev:22000] purge pthread at configure

U. Nakamura reported that he cannot pass 'make test'
with -lpthread option in his NetBSD/i386-current,
so he proposed to make an option to disable pthread.
Then a new option '--enable-pthread' was added.


高橋征義 (TAKAHASHI Masayoshi)   E-mail: [e-mail]