Crystal: [RFC] Endless ranges

Created on 10 Dec 2018 · 9Comments · Source: crystal-lang/crystal

Ruby 2.6 introduces endless ranges:

[1, 2, 3, 4, 5][2..] # => [3, 4, 5]

I think it's a small nice addition. It removes the need to write -1 and the confusion of 2..-1 vs 2...-1 (in the above snippet 2.. and 2... would give the same result).

Ruby implements this by having the end of the range be nil. We could do the same.

The only "ugly" thing about it is that you always have to write that range inside parentheses, brakcets, or another expression (you can't just write a = 1.. because the parser will wait for the end value, and a newline makes it still expect the end value). But one can write a = 1..nil so it's a minor issue (plus these ranges are mostly useful in nested positions).

I don't think it's too hard to implement this.

Thoughts?

(side note, I also thought about having ..3 meaning "from the start", but it's not that useful because one could always put a zero there, it's always zero, and then there's the confusion of which zero is it (int or float?))

Source

asterite

👍10 😕2 👎1

Most helpful comment

It seems Ruby is adding beginless ranges too :-)

asterite on 25 May 2019

😄3

All 9 comments

The only "ugly" thing about it is that you always have to write that range inside parentheses

Well, ranges often need parentheses already, like (a..b).to_a.

ysbaddaden on 10 Dec 2018

That's true. I guess the only corner case is in assignments, but you can still use parentheses or nil there.

asterite on 10 Dec 2018

I also thought about having ..3 meaning "from the start", but it's not that useful

Ruby tracker has such [proposition]:

case release_date
when ..1.year.ago
  puts "ancient"
when 1.year.ago..3.months.ago
  puts "old"
when 3.months.ago..Date.today
  puts "recent"
when Date.today..
  puts "upcoming"
end