From: mhm26@... (matt)
Date: 2003-11-22T02:42:16+09:00
Subject: Re: break in yield/block

matz@ruby-lang.org (Yukihiro Matsumoto) wrote in message news:<1069395225.856100.31032.nullmailer@picachu.netlab.jp>...
> Hi,
> 
> In message "Re: break in yield/block"
>     on 03/11/20, matt <mhm26@drexel.edu> writes:
> 
> |I figured that was why break was terminating in #[] - but why not have
> |yield act as a `method' in this sense?  My guess is that in this
> |respect, they should be the same.
> 
> They cannot be same.  I would surprise if calling Proc#[] terminates
> meow().  But I prepared the way to make meow behave as you expect:

I was actually desiring the other way - that break from yield act like
break from Proc#[].


>   def meow(doyield, &block)
>     puts "in meow #{doyield}"
>     if doyield
>       yield
>     else
>       block[]
>     end
>     puts "leaving meow"
>   rescue LocalJumpError
>     return
>   end
> 
>   b = Proc.new {
>     puts "in proc"
>     break
>     puts "leaving proc"
>   }
> 
>   meow(true, &b)
>   puts
>   meow(false, &b)
> 
> Tip: do not use "proc" nor "lambda" when you want objectified Block.
> 
> 							matz.

Is there a reason that Proc#[] is different wrt this depending on
whether or not proc/lambda was used, and are there any other
differences?

I still don't see why the break works the way it does (well, I can
understand the reasoning) - but if I break inside a block, then I want
that block to terminate, not the method.

My current plans are to just do something like this
  method_invoke(*args) {|*block_args|
    callcc {|block_break|
      code - yadda yadda yadda
      block_break.call if want break / return from block
      other code...
    }
  }

Any better ideas?