On Sat Aug 30 08:31:30 2014, masak wrote:
> <masak> n: sub foo { my $s = 0; ($s += $_ for 1..3) }; say foo()
> <camelia> niecza v24-109-g48a8de3: OUTPUT«1 3 6»
> <masak> this is a data point, too. Niecza decontainerizes.
It's a tad more sutble:
15:07 < jnthn> masak: Only tricky thing is that it may not be as simple as
decont
15:08 < jnthn> m: say (1, 2, 3).map({ $['a' xx $_] }).perl
15:08 <+camelia> rakudo-moar e5612a: OUTPUT«($["a"], $["a", "a"], $["a", "a",
"a"]).Seq»
15:08 < jnthn> If I decont then we lose the itemization on those.
15:09 < masak> ...hm.
So we need to decide if we care to retain the meaning of the $ in such cases,
and perhaps rule on the sub case too:
15:09 < jnthn> So we more need the return from sub semantics (recont if
Iterable)
15:09 < masak> uh, ok
15:09 < masak> I had no idea subs did that :P
15:11 < jnthn> They did 'cus you filed a ticket when they didn't :P
15:12 * PerlJam wonders if this falls in the category of tormenting the
implementor ...
15:12 < jnthn> Back when we had lots of implicit flattening and [...] was an
item, it was rather more noticable.
15:12 < jnthn> Nowadays I'd probably get through spectest with a patch that
just deconts, but... :)
