From: Robert Feldt <feldt@...>
Date: 2003-05-29T20:37:07+09:00
Subject: Re: RCR: unpack/pack Bignum

On Thu, 29 May 2003 nobu.nokada@softhome.net wrote:

> I guess "N(2)" is equivalent to "n", so the behavior should be
> consist with the current one.  But it differs for Fixnum and
> Bignum.  Which is better?
>
>   $ ruby -e 'p [0x010203].pack("n")'
>   "\002\003"
>
>   $ ruby -e 'p [0x0102030405].pack("n")'
>   -e:1:in `pack': bignum too big to convert into `int' (RangeError)
>           from -e:1
>
The former seems better to me. So we truncate MSB bytes until it fits.

> > How can this be used when I have a Bignum that I don't know how large it
> > will be? 'N(*)'?
>
> I expect '(*)' would mean to convert all remained data like as
> 'a*' does, though you need to know the exact template to pack/
> unpack fundamentally.
>
Yes, but this should work then

  s = [large_bignum_or_really_any_int].pack("N(*)")
  num = s.unpack("N(#{s.length})").first
  # now num == large_bitnum_or_really_any_int

?

If we can agree on something along those lines I think it would be
great. Very useful, imho. Any opinion from matz or others?

/Robert