From: "W. Kent Starr" <elderburn@...>
Date: 2001-03-13T11:06:25+09:00
Subject: [ruby-talk:12537] Re: FEATURE REQUEST: 'my' local variables

On Monday 12 March 2001 09:04, Dave Thomas wrote:
> ts <decoux@moulon.inra.fr> writes:
> > >>>>> "G" == GOTO Kentaro <gotoken@math.sci.hokudai.ac.jp> writes:
> >
> > G> For example, loop{|a| ... } means as same as the current model.  In
> > G> loop{|a;b| b = a; .... }, b is a shadowed counter.
> >
> >  Perhaps you can write  loop{(a;b) .... } where a and b are new local
> > variables in the current scope
>
> Doesn't that break existing code, though?
>
>    n = 99
>    b = a.collect {(n)}
>
>
> I'm wondering: was the idea of having a different block type for scope
> protection just too stupid? I've seen no comments about
> ruby-talk:12393.
>


Yes and no :-)

Ok, x = 5; loop {{ |y| x = 3; x * y }} protects "x" or...if I _wnat_ to 
change x with behavior of the block, i use loop { |y| x = 3; x *y }
but...

suppose I have x = 5; z = 10; loop { |y| x = 3; z = 5; x + (y * z) }
what if I want to change x but keep z as before?
loop {{ |y| x = 3; z = 5 ...})  _doesn't_ do it!

Or, I just noticed I read this wrong. (n) _would_ work, except it also might 
be confusing, although (as above example)

x = 5; z = 3; loop { |y| (x)=3; z = 5; x + ((z) * y} _woeld_ work, but is it 
also not confusing. Or perhaps... loop { |y| (x = 3); z = 5; x + (z * y)} ?

This would require a core change and to avoid that "conventions" are better 
_if_ we _all_ act in accordance with the "conventions". For me in my code the 
localization problem does not reallyy matter; if I want to keep x, I just 
_don't_ use x withing loop {  }; but if _my_ code is used within _your_ code 
as a require or whatever? _That_ I thnk is where the real problem lies.

Regards,

Kent Starr
elderburn@mindspring.com