[#18974] Perl/Python/Ruby common backend (Perl6) — ptkwt@...1.aracnet.com (Phil Tomson)

There is a thread about using .NET's CLR as a backend for Ruby, but how

17 messages 2001/08/01

[#19064] ANN: Code Amelioration Contest (presented by Ruby Conference 2001) — David Alan Black <dblack@...>

17 messages 2001/08/03
[#19184] Re: ANN: Code Amelioration Contest (presented by Ruby Conference 2001) — John Carter <john.carter@...> 2001/08/06

On Fri, 3 Aug 2001, David Alan Black wrote:

[#19185] Re: ANN: Code Amelioration Contest (presented by Ruby Conference 2001) — David Alan Black <dblack@...> 2001/08/06

Hello --

[#19186] Re: ANN: Code Amelioration Contest (presented by Ruby Conference 2001) — John Carter <john.carter@...> 2001/08/06

On Mon, 6 Aug 2001, David Alan Black wrote:

[#19125] My 1st look @ ruby: No prototypes and problem with String#gsub — stesch@... (Stefan Scholl)

My first ruby program:

23 messages 2001/08/04

[#19192] Some remarks from a nembie in Ruby — Renaud HEBERT <renaud.hebert@...>

After having read the book "Programming Ruby: The Pragmatic Programmer's

38 messages 2001/08/06

[#19269] Re: Perl/Python/Ruby common backend (Parrot, can Ruby play too?) — ptkwt@...1.aracnet.com (Phil Tomson)

In article <72X97.12093$9i1.972452@e420r-atl1.usenetserver.com>,

50 messages 2001/08/07
[#19349] Re: Perl/Python/Ruby common backend (Parrot, can Ruby play too?) — Mathieu Bouchard <matju@...> 2001/08/08

[#19456] Re: Perl/Python/Ruby common backend (Parrot, can Ruby play too?) — Harry Ohlsen <harryo@...> 2001/08/09

Ned Konz wrote:

[#19451] Re: Help! I'm still confused about threadin g in the ML — "Morris, Chris" <chris.morris@...>

> Is there an Outlook option to turn on In-Reply-To or References

14 messages 2001/08/09
[#19453] Re: Help! I'm still confused about threadin g in the ML — Dave Thomas <Dave@...> 2001/08/09

"Morris, Chris" <chris.morris@snelling.com> writes:

[#19506] the way class variables work — David Alan Black <dblack@...>

Hello --

51 messages 2001/08/10
[#19511] Re: the way class variables work — Chris Uzdavinis <chris@...> 2001/08/11

David Alan Black <dblack@candle.superlink.net> writes:

[#19524] order and freedom in Ruby (was: Re: Re: the way class variables work) — David Alan Black <dblack@...> 2001/08/11

Hello --

[#19517] Why not?: Assigning to self — furufuru@... (Ryo Furue)

Hi there,

55 messages 2001/08/11
[#19689] Re: Why not?: Assigning to self — Ron Jeffries <ronjeffries@...> 2001/08/14

On 13 Aug 2001 20:59:54 -0700, furufuru@ccsr.u-tokyo.ac.jp (Ryo Furue)

[#19694] Re: Why not?: Assigning to self — Ned Konz <ned@...> 2001/08/14

On Tuesday 14 August 2001 05:09 am, Ron Jeffries wrote:

[#19695] Re: Why not?: Assigning to self — ts <decoux@...> 2001/08/14

>>>>> "N" == Ned Konz <ned@bike-nomad.com> writes:

[#19696] Re: Why not?: Assigning to self — Ned Konz <ned@...> 2001/08/14

On Tuesday 14 August 2001 07:51 am, you wrote:

[#19697] Re: Why not?: Assigning to self — ts <decoux@...> 2001/08/14

>>>>> "N" == Ned Konz <ned@bike-nomad.com> writes:

[#19700] Re: Why not?: Assigning to self — Ned Konz <ned@...> 2001/08/14

On Tuesday 14 August 2001 08:27 am, you wrote:

[#19701] Re: Why not?: Assigning to self — ts <decoux@...> 2001/08/14

>>>>> "N" == Ned Konz <ned@bike-nomad.com> writes:

[#19703] Re: Why not?: Assigning to self — Ned Konz <ned@...> 2001/08/14

On Tuesday 14 August 2001 09:05 am, Guy Decoux wrote:

[#19704] Re: Why not?: Assigning to self — ts <decoux@...> 2001/08/14

>>>>> "N" == Ned Konz <ned@bike-nomad.com> writes:

[#19708] Re: Why not?: Assigning to self — Ned Konz <ned@...> 2001/08/14

On Tuesday 14 August 2001 09:27 am, you wrote:

[#19709] Re: Why not?: Assigning to self — ts <decoux@...> 2001/08/14

>>>>> "N" == Ned Konz <ned@bike-nomad.com> writes:

[#19713] Re: Why not?: Assigning to self — Ned Konz <ned@...> 2001/08/14

On Tuesday 14 August 2001 09:45 am, you wrote:

[#19750] Re: Why not?: Assigning to self — matz@... (Yukihiro Matsumoto) 2001/08/15

Hi,

[#19819] Re: Why not?: Assigning to self — Ned Konz <ned@...> 2001/08/15

On Tuesday 14 August 2001 08:14 pm, matz wrote:

[#19852] Re: Why not?: Assigning to self — matz@... (Yukihiro Matsumoto) 2001/08/16

Hi,

[#19857] Re: Why not?: Assigning to self — "Florian G. Pflug" <fgp@...> 2001/08/16

On Thu, Aug 16, 2001 at 11:05:59AM +0900, Yukihiro Matsumoto wrote:

[#19858] Re: Why not?: Assigning to self — matz@... (Yukihiro Matsumoto) 2001/08/16

Hi,

[#19867] Re: Why not?: Assigning to self — "Pit Capitain" <pit@...> 2001/08/16

Just a followup at (my) current end of the thread:

[#19550] Forced garbage collection — Lars Christensen <larsch@...>

14 messages 2001/08/11
[#19562] Re: Forced garbage collection — "Nat Pryce" <nat.pryce@...13media.com> 2001/08/12

From: "Lars Christensen" <larsch@cs.auc.dk>

[#19551] /.ed again — Tobias Reif <tobiasreif@...>

Ruy gets slasdotted again ;)

19 messages 2001/08/11

[#19650] Ruby Newbie mailing list — Michael Pence <mikepence@...>

Hello all.

14 messages 2001/08/13
[#19656] RE: Ruby Newbie mailing list — "Louis Brothers" <lcb134@...> 2001/08/13

We had a similar discussion on the OmniWeb Objective-C mailing list not to

[#19659] Re: Ruby Newbie mailing list — Michael Pence <mikepence@...> 2001/08/13

I appreciate your references to Objectionable-C ;-)

[#19685] Compiling Ruby with cygwin and Tk support — Manuel Zabelt <ng@...>

Hello!

13 messages 2001/08/14

[#19718] General (GUI/license) questions — Ryan Tarpine <rtarpine@...>

First: Kero commented in the description of his new Ruby Agenda program

18 messages 2001/08/14

[#19755] "new" returning nil: how to report the failure of object creation — furufuru@... (Ryo Furue)

Hi there,

14 messages 2001/08/15

[#19758] The GUI poll is in, and the results are surprising — Dave Thomas <Dave@...>

40 messages 2001/08/15
[#19774] Re: The GUI poll is in, and the results are surprising — Lars Christensen <larsch@...> 2001/08/15

On Wed, 15 Aug 2001, Dave Thomas wrote:

[#19784] Re: The GUI poll is in, and the results aresurprising — "Lyle Johnson" <ljohnson@...> 2001/08/15

> Please don't forget what Ruby is all about in this discussion! I think

[#19824] Ruby GUI — "Hal E. Fulton" <hal9000@...>

The concept of a new GUI is somewhat appealing,

16 messages 2001/08/15

[#20033] Ruby Article — Joshua Drake <jd.nospam@...>

Hello,

38 messages 2001/08/20

[#20127] Another Possible RCR - Wrappers via Mixins — Stephen White <spwhite@...>

The main difference between mix-ins and multiple inheritence is (to my understanding) that parent classes do not call child code, but mix-ins do.

15 messages 2001/08/22

[#20135] Bruce Eckel's criticism of Ruby — Ned Konz <ned@...>

Python.org links to http://www.mindview.net/Etc/notes.html#Ruby , saying

24 messages 2001/08/22

[#20183] ++ Operator — kamphausen@... (SKa)

Dear Community,

35 messages 2001/08/23
[#20234] Re: ++ Operator — Dave Thomas <Dave@...> 2001/08/24

matz@ruby-lang.org (Yukihiro Matsumoto) writes:

[#20236] Re: ++ Operator — matz@... (Yukihiro Matsumoto) 2001/08/24

Hi,

[#20209] In Ruby 0 is true but nil is false.. or how to shoot yourself?.. — Guillaume Cottenceau <gc@...>

I have a simple Audio-CD database (using CSV format). I was writing a

11 messages 2001/08/23

[#20254] File.readline(s) — Michael Husmann <michael.husmann@...>

I am reading a 55MB ASCII file by using File.readline(s) which takes on

14 messages 2001/08/24

[#20303] New Windows InstallShield version of Ruby — Andrew Hunt <andy@...>

19 messages 2001/08/24

[#20307] Backwards language — "Sean Middleditch" <elanthis@...>

Greetings,

30 messages 2001/08/24

[ruby-talk:20591] Re: Weird things with eval

From: David Alan Black <dblack@...>
Date: 2001-08-30 21:59:13 UTC
List: ruby-talk #20591
Hello --

The part I don't get, as a general matter, is why #send gets to bypass
the access rules.  Of course #send itself is a public method.  The
weird part is that if it sends the name of a private method, that
method acts public.

For example, #eval:

  irb(main):001:0> [].eval("puts 'hi'")
  NoMethodError: private method `eval' called for []:Array
	  from (irb):1
  irb(main):002:0> [].send(:eval,"puts 'hi'")
  hi

And that particular example, of course, relates closely to yours.
Anyway, with that in the background, here are some further comments.

For one thing, keep in mind that Class.send(:eval, str) is not the
same as Class.class_eval str.

  class Foo; end
  Foo.class_eval "def bar; puts 'bar'; end"
  Foo.new.bar  # => bar

  class Foo; end
  Foo.send(:eval, "def thing; end")
  [].thing
  -:3: private method `thing' called for []:Array (NoMethodError)

Note that the #eval in the second one executed at the top level, so
that every object (such as []) now has a private method 'thing'.
(That's what happens with top-level method definitions.)

So...

When you do this:

>   class Module
>     def foo
>       p self
>       Foo.__send__(:eval, "def bar; puts 'bar'; end")
>     end
>   end

and then call #foo, you're calling Kernel#eval in the context of class
Module.  The fact that Foo is the receiver doesn't actually matter.
(At least not as long as the 'send' trick works....)

So you could actually do this, and get, I think, exactly the same
result, visibly and internally:

  class Module
    def foo
      p self
      "some random object".send(:eval, "def bar; puts 'bar'; end")
    end
  end

or this:

  class Module
    def foo
      p self
      eval "def bar; puts 'bar'; end
    end
  end

or just forget about Foo and foo and do:

  class Module
    def bar
      puts 'bar'
    end
  end

which is really what you've ended up doing, at least in net effect.

And when you do this, at the top level:

>   class Foo
>   end
>
>   Foo.__send__(:eval, "def bar; puts 'bar'; end")

you could just as well replace the whole thing with:

  def bar; puts 'bar'; end

The status of Foo as receiver for #eval is irrelevant; you're eval'ing
at the top level.  Try adding these to your tests, along with Module
and f.bar:

  [].bar
  Hash.new("hi").bar
  123.bar

As you'll see, you given *every* object a private method 'bar',
because you've defined a top-level method 'bar'.  The whole Foo thing
is misleading.


On Fri, 31 Aug 2001, Paul Brannan wrote:

> So the questions are:
>
> 1) Why in (A) does bar get defined for Module but not for Foo?
> 2) Why in (B) does bar get defined for both Module and Foo, and why is it
> private?
> 3) Why in (D) does Ruby allow me to modify a frozen class, and why is the
> resulting method private?


I haven't organized my comments as well as you organized the question,
but I hope they help :-)


David

-- 
David Alan Black
home: dblack@candle.superlink.net
work: blackdav@shu.edu
Web:  http://pirate.shu.edu/~blackdav


In This Thread