Front page | perl.perl6.language |
Postings from May 2003
arcadi shehter <fearcadi@figaro.weizmann.ac.il> writes:
> I was reading
> http://www.linux-mag.com/downloads/2003-04/perl6/03.04.perl6.2.txt
> and in particular , this
>
> print -sum(@probs >>*<< >>log<<(@probs))/log(2);
>
> And I have several questions
>
> * are all these are correct ( and equivalent ) ?
>
> @y = >>log<< @x # (1)
> @y = >>log<<(@x) # (2)
> @y = >>log<< .(@x) # (3)
> @y = log >>.<< (@x)# (4)
> @y = log >>.(@x)<< # (5) wild and probably wrong guess
>
> it seems that there is no way to insert space between vectorized sub
> and its argument if only 2 is OK.
>
> But log is unary operator so probably (1) and (2) apply to it for that
> reason .
Personally I'd argue that you should only be able to vectorize
things that are used in 'operator form'. So I'd say that only 1 and
2 are definitely legal. I'm wary of >>.<< being considered legal
syntax though.
> What if I have a sub accepting 2 scalar args
>
> sub pow($base , $exp=2) {...}
>
> will both these work ?:
>
> #list of squares
> @y = >>pow<<(@base,$exp)
> #or:
> @y = pow >>.<< (@base,$exp)
>
>
> #list of powers
> @y = >>pow<<($base,@exp)
I would hope not. But:
@y = >>pow.assuming(exp => $exp)<<(@base);
and
@y = >>pow.assuming(base => $base)<<(@exp)
should work...
> Also :
> is these equivalent to @c = @a >>+<< @b :
>
> @c = >>infix:+<< (@a,@b)
>
> or it should be
>
> @c = >>infix:+<< .(@a,@b)
>
> or it should be
>
> @c = infix:>>+<< .(@a,@b)
Ick... I haven't a clue. The infix:>>+<< option looks the most
promising though, since it implies that one could write:
multi infix:+ ($a, $b) {...}
multi infix:>>+<< ($a, @b) {...}
multi infix:>>+<< (@a, $b) {...}
multi infix:>>+<< (@a, @a) {...}
But I'm not sure either way here, after all, in most cases you could
probably abstract out the vectorizing behaviour:
multi outfix:>><< (&op, $arg) { &op($arg) }
multi outfix:>><< (&op, @arg) { map {&op($_)} @arg }
multi outfix:>><< (&op, $a1, @a2) { map { &op($a1, $_) } @a2 }
...
Hmm... I'm pretty sure those signatures are wrong though...
--
Piers
Thread Previous
|
Thread Next