Yes, because INIT and BEGIN happen before runtime, and $alpha is set at
runtime! Hence my original BEGIN example using a constant to set the first
value.
Another reason to prefer "state" over those phasers... unless you want a
counter over the lifetime of the process, which is valid.
-y
On Tue, Sep 1, 2020 at 1:17 PM William Michels <w...@caa.columbia.edu>
wrote:
> I tried combining Larry's code and Yary's code, variously using
> "state" or "INIT" or "BEGIN". This is what I saw:
>
> ~$ raku -e 'for <AA NN> -> $alpha { for (1..14) { print (state $ =
> $alpha)++ ~ " " } }'
> AA AB AC AD AE AF AG AH AI AJ AK AL AM AN NN NO NP NQ NR NS NT NU NV
> NW NX NY NZ OA
>
> ~$ raku -e 'for <AA NN> -> $alpha { for (1..14) { print (INIT $ =
> $alpha)++ ~ " " } }'
> 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
>
> ~$ raku -e 'for <AA NN> -> $alpha { for (1..14) { print (BEGIN $ =
> $alpha)++ ~ " " } }'
> 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
>
> Expected? --Bill.
>
> On Tue, Sep 1, 2020 at 11:44 AM yary <not....@gmail.com> wrote:
> >
> >
> > Thanks, that's cool, and shows me something I was wondering about
> >
> > On Tue, Sep 1, 2020 at 11:36 AM Larry Wall <la...@wall.org> wrote:
> >>
> >> If you want to re-initialize a state variable, it's probably better to
> make
> >> it explicit with the state declarator:
> >>
> >> $ raku -e "for <a b> { for (1..2) { say (state $ = 'AAA')++ } }"
> >> AAA
> >> AAB
> >> AAA
> >> AAB
> >
> >
> > $ raku -e 'for <AA OO> -> $alpha { for (1..3) { say (state $ = $alpha)++
> } }'
> > AA
> > AB
> > AC
> > OO
> > OP
> > OQ
> >
>
