ruby-dev suumary 26957-27102

[SASADA Koichi]

Hi. This is ruby-dev summary 26957-27102.



[ruby-dev:26965] patch for RDoc about Module#class_variable_get

sheepman pointed out that RDoc about Module#class_variable_get and
Module#class_variable_set were wrong (described as public method).
Matz answered that these methods should be public, so Ruby 1.9 was
changed these visibility, and on Ruby 1.8 these problem were fixed.


[ruby-dev:26967] Tempfile.new("z").extend(M)

Tanaka Akira asked that following behaviour is not coherent (different
behaviour between "kind_of?" and "===").

% ./ruby -v -rtempfile -e '
tempfile = Tempfile.new("z")
module M
end
tempfile.extend M

p M === tempfile
p tempfile.kind_of?(M)
'
ruby 1.9.0 (2005-09-06) [i686-linux]
false
true

Matz answered that this is because of using Delegator. He planned to
leave this problem until anyone proposes any good idea.


[ruby-dev:26975] [proposal] ANSI style function

H.Yamamoto proposed that Ruby source code should use ANSI style function
definition instead of K&R style. Matz agreed this proposal. After
that, many messages about this changeover were posted to ruby-dev.


[ruby-dev:27077] select pipe support on native Win32

U.Nakamura announced that he enables "select()" on native win32 port.
For example, a script of [ruby-talk:82507] works.


[ruby-dev:27078] console pipe support on native Win32

U.Nakamura also announced supporting "select()" on console I/O.
Following example works:

Thread.new {
loop {
puts "thread"
sleep 1
}
}
loop {
p gets
}

Known problem:

(1) pipe and console were considerd always writable.

(2) On console, if characters except line break were input in waiting,
this ruby thread will be blocked. It's impossible to solve this problem
without native thread.

(3) Only few tests on Win9x


[ruby-dev:27082] Re: ruby-1.8.3 status for release

Masayoshi Takahashi summarised release process of ruby-1.8.3. This
process will lead us to release 1.8.3 at the end of Sep.
Matz decided to fix the exact release date (preview3: Sep-19, release:
Sep-21 pm 12:00 (JST)). Masayoshi rescheduled and proposed release plan
for Sep-21 release.

And now you can get ruby-1.8.3 ([ruby-talk:05844] Ruby 1.8.3 released).


ruby-dev summary index:
http://i.loveruby.net/en/ruby-dev-summary.html

 

[David A. Black]

Hi --

And now you can get ruby-1.8.3 ([ruby-talk:05844] Ruby 1.8.3 released).

s/talk/core/


David

 

[Paul Brannan]

[ruby-dev:27078] console pipe support on native Win32

U.Nakamura also announced supporting "select()" on console I/O.
Following example works:

Thread.new {
loop {
puts "thread"
sleep 1
}
}
loop {
p gets
}

Known problem:

(1) pipe and console were considerd always writable.

(2) On console, if characters except line break were input in waiting,
this ruby thread will be blocked. It's impossible to solve this problem
without native thread.

(3) Only few tests on Win9x

Would any of this be solved by using WFMO instead of select on win32?
It would be nice to have a pluggable IO multiplexing system (e.g. so
ruby could use /dev/poll or /dev/epoll or kqueue on platforms that
support it).

I don't know how difficult such a change is or what the consequences
might be. It has been discussed before in [ruby-talk:47186].

Paul

 

[nobu.nokada]

Hi,

At Wed, 21 Sep 2005 22:48:41 +0900,
Paul Brannan wrote in [ruby-talk:156944]:

Would any of this be solved by using WFMO instead of select on win32?

WFMO doesn't work with sockets directly.

It would be nice to have a pluggable IO multiplexing system (e.g. so
ruby could use /dev/poll or /dev/epoll or kqueue on platforms that
support it).

I don't know how difficult such a change is or what the consequences
might be. It has been discussed before in [ruby-talk:47186].

It should be a quite big change.