in reply to Is there a "better" way to cycle an array?

G'day rje,

Here's a solution using state, which therefore requires 5.10.0 or later:

$ perl -Mstrict -Mwarnings -E ' sub cycle { state %i; no warnings "uninitialized"; $_[0]->[($i{$_[0]} % @{$_[0]}, ++$i{$_[0]})[0]]; } my $a2d = ["a" .. "d"]; my $e2h = ["e" .. "h"]; for (1..3) { print cycle($a2d) for 1..3; print cycle($e2h) for 1..3; } ' abcefgdabhefcdaghe

For earlier Perls, this uses a closure and retains the same level of privacy for %i:

$ perl -Mstrict -Mwarnings -E ' { my %i; sub cycle { no warnings "uninitialized"; $_[0]->[($i{$_[0]} % @{$_[0]}, ++$i{$_[0]})[0]]; } } my $a2d = ["a" .. "d"]; my $e2h = ["e" .. "h"]; for (1..3) { print cycle($a2d) for 1..3; print cycle($e2h) for 1..3; } ' abcefgdabhefcdaghe

If you only ever want to cycle through a single array (I wasn't sure if that's what your OP was suggesting), then this would be faster (but has distinct limitations):

$ perl -Mstrict -Mwarnings -E ' sub cycle { state $i = -1; $_[0]->[++$i, $i %= @{$_[0]}]; } my $stuff = ["a" .. "d"]; print cycle($stuff) for 1..30; ' abcdabcdabcdabcdabcdabcdabcdab

Obviously, that can also be modified for earlier Perls in the same manner as the previous example.

-- Ken