From: Simon Strandgaard <neoneye@...>
Date: 2006-08-10T05:39:17+09:00
Subject: Re: Codegolf - Writing a Brainf*ck interpreter

On 8/9/06, Daniel Martin <martin@snowplow.org> wrote:
> "Simon Strandgaard" <neoneye@gmail.com> writes:
> > well here is my 141 bytes long version :-)
[snip]
> Note that both this and the earlier one you posted don't conform to
> the interface of the codegolf challenge: namely, that bf program will
> be presented on stdin, and separated from input by a !.

aha.. I did'nt really use the rules from codegolf. I looked it
up at wikipedia. I should had seen this.


> A correct bf interpreter using the traditional 8-bit cell should
> output nothing.  Yours... doesn't.  True, the codegolf people didn't
> say you had to use an 8-bit cell, so maybe this is ok.

Ok, I have to rethink the impl to restrict it to 8 bit.
Now I see why people is using strings for this.


> However, I hadn't known about putc, and seeing your code gives me some
> ideas.
>
> Note that even with your version, you can save two bytes by NOT doing
> the .gsub bit, and expanding X into d[a].  Doing that substitution may
> save you 15 characters out of the %w block, but
> .gsub(/X/,'d[i]')
> is seventeen characters long.

Argh.. I misscalculated..
5 times d[i]  = 20 bytes
5 times X + 17 = 22 bytes

Thanks for your hints.


> Anyway, here's where I'm at, at 168 characters:
>
> j='d="\0"*8**5;a=0'
> while i=STDIN.getc
> j+='
> '+%w{1 a-=1 a+=1 while(d[a]>0)
> end putc(d[a]) d[a]=STDIN.getc||exit
> d[a]+=1 d[a]-=1}["
> <>[].,+-".index(i)||break]
> end
> eval j

Idea for improvement:

while(d[a]>0)
end
#=> 13 bytes

(
)while(d[a]>0)
#=> 12 bytes

--
Simon Strandgaard