Re : What about a 'series' type?

[Eric Torreborre]

There is a complete lazy list implementation from Reginald Braithwaite webs=
ite:=0A=0Ahttp://raganwald.com/source/lazy_lists.html=0A=0AIt supports a lo=
t of operations on this kind of "infinite" structures such as merge and car=
tesian product.=0A=0AThat one could make it to the core!=0A=0AEric.=0A =0A-=
--------------------------------------------------------------------------=
=0AEric TORREBORRE=0Atel: +81 (0)90 5580 3280=0Ae-mail: (e-mail address removed)=
om / (e-mail address removed)=0Ablog: http://etorreborre.blogspot.com=0A---=
------------------------------------------------------------------------=0A=
=0A----- Message d'origine ----=0ADe : Robert Dober <[email protected]=

=0A=C0 : ruby-talk ML <[email protected]>=0AEnvoy=E9 le : Jeudi, 7 J=

uin 2007, 20h29mn 46s=0AObjet : Re: What about a 'series' type?=0A=0AOn 6/7=
elieve that the core is the place to put such=0Athings, furthermore I belie=
ve it is too specific a feature, a more=0Ageneral approach might have bette=
r chances to be fit for the core; Yet=0AI do not think what follows is fit =
for the core either, but maybe you=0Afind it interesting or helpful:=0A=0Ac=
lass Lazy=0A def initialize init, op, *args=0A @init =3D init=0A @op=
=3D op=0A @args =3D args.dup=0A end=0A=0A def upto value=0A retu=
rn [] if value < 1=0A (2..value).inject([@init]){ |acc,| acc << acc.last=
send( @op, *@args ) }=0A end=0Aend # class Lazy=0A=0A505/6 > irb -r lazy.=
rb=0Airb(main):001:0> l =3D Lazy.new 1, :+, 2=0A=3D> #<Lazy:0xb7ddfa60 @arg=
s=3D[2], @init=3D1, @op=3D:+>=0Airb(main):002:0> l.upto 5=0A=3D> [1, 3, 5, =
7, 9]=0Airb(main):003:0> m =3D Lazy.new 2, :*, 3=0A=3D> #<Lazy:0xb7dd4908 @=
args=3D[3], @init=3D2, @op=3D:*>=0Airb(main):004:0> m.upto 4=0A=3D> [2, 6, =
18, 54]=0Airb(main):005:0>=0A=0ACheers=0ARobert=0A=0AP.S. Implementations w=
ithout #inject are theoretically possible =0AR.=0A-- =0AYou see things; a=
nd you say Why?=0ABut I dream things that never were; and I say Why not?=0A=
-- George Bernard Shaw=0A=0A=0A=0A=0A=0A=0A=0A=0A =0A=09=0A=09=09=0A__=
_________________________________________________________________________ =
=0AD=E9couvrez une nouvelle fa=E7on d'obtenir des r=E9ponses =E0 toutes vos=
questions ! =0AProfitez des connaissances, des opinions et des exp=E9rienc=
es des internautes sur Yahoo! Questions/R=E9ponses =0Ahttp://fr.answers.yah=
oo.com

 

[Robert Dober]

Pls no top post

There is a complete lazy list implementation from Reginald Braithwaite website:

http://raganwald.com/source/lazy_lists.html

It supports a lot of operations on this kind of "infinite" structures such as merge and cartesian product.

That one could make it to the core!

Not to core, given the dependencies it has, however I would be the
first to strongly advocate Facets to be put into the stdlib [No I am
not involved with Factes ]
Stdlib would be the target than for everything using Factes.

And maybe the following is of interest too:
http://blog.grayproductions.net/articles/2006/02/20/infinite-streams

Please do not look too closely at my toy code it was to show that the
concept should be abstracted about and others have shown/done that
much better

Cheers
Robert