Laziness — lazy, eager & infinite lists

Work with lists that compute elements on demand, including infinite ones.

Many Raku list producers are lazy: they compute elements only as you ask for them. That makes infinite lists ordinary values — you can define "all the squares" or "all the primes" and then take just the first few. A lazy list does no work until something pulls on it, so an unbounded source never runs away.

Taking from an infinite list #

An infinite range like 1..Inf (or 1..∞) is lazy; mapping or grepping over it stays lazy, so indexing or .head only forces the elements you actually read.

my @squares = (1..Inf).map(* ** 2);
say @squares[^5];
say (1..Inf).grep(*.is-prime).head(5);

(1 4 9 16 25)
(2 3 5 7 11)

Neither line computes more than it needs: @squares[^5] forces five squares, and .head(5) stops the prime filter after the fifth prime.

Testing laziness — .is-lazy #

.is-lazy reports whether a list still has unrealised, on-demand elements. A bounded range knows its size and is not lazy; an unbounded one is.

say (1..10).is-lazy;
say (1..Inf).is-lazy;

False
True

Finding without a bound #

Because the search is lazy, .first can scan an infinite list and stop at the first hit — no upper limit needed.

say (1..Inf).first(* > 1000);

1001

Self-referential sequences #

The sequence operator ... (see The sequence operator) builds a lazy series from a closure — even one that reads its own earlier terms, like Fibonacci — and you take a prefix the same way.

my \fib = (1, 1, * + * ... *);
say fib[^10];

(1 1 2 3 5 8 13 21 34 55)

Notes #

• eager forces a lazy list fully into memory — use it when you want all the work done now (and never on a truly infinite list, which would loop forever).
• Indexing (@a[100]), .head(n), .first, and for all pull lazily; .elems, .sort, .reverse, and say @a (the whole list) force evaluation.
• Ranges (1..Inf) and the sequence operator (1, 2, 4 ... *) are the common lazy sources; gather/take (gather / take) produces values on demand too.
• .head/.tail and slices are the safe way to look at an infinite list — printing the whole thing would never finish.