From @moritz

Rakudo r33209​:

my @​a = 1; say (^@​a).elems; # fine
1
my @​a = <jan>; say (^@​a).elems; # ouch
100
my @​a = 'jan'; say (^@​a).elems;
100
my @​a = 'ja'; say (^@​a).elems;
10
my @​a = 'janu'; say (^@​a).elems; # ouch * 10
1000
my @​a = <jan feb>; say (^@​a).elems; # 2 * ouch
200

Moritz

From @moritz

Part of the problem is also that ^@​a returns a List, not a Range.

We now have 6 failing tests in t/spec/S03-operators/context-forcers.t (4
skip, 2 TODO).

@moritz - Status changed from 'new' to 'open'

From @pmichaud

Currently Rakudo is treating ^@​list as though it's
prefix​:<^> on a List, which S03 says

  If [prefix​:<^> is] applied to a list, it generates a
  multidimensional set of subscripts.

  for ^(3,3) { ... } # (0,0)(0,1)(0,2)(1,0)(1,1)(1,2)(2,0)(2,1)(2,2)

So, Rakudo is currently seeing ^@​list as following this definition,
and trying to generate the subscripts (perhaps wrongly).

There's still some ambiguity in how (or if) we should support
both interpretations [1], so we'll want to get that resolved before
we can fix prefix​:<^> here.

Thanks,

Pm

[1] http://irclog.perlgeek.de/perl6/2008-11-26#i_720703

From @moritz

Patrick R. Michaud wrote​:

Currently Rakudo is treating ^@​list as though it's
prefix​:<^> on a List, which S03 says

If \[prefix&#8203;:\<^> is\] applied to a list, it generates a
multidimensional set of subscripts\.

for ^\(3,3\) \{ \.\.\. \} \# \(0,0\)\(0,1\)\(0,2\)\(1,0\)\(1,1\)\(1,2\)\(2,0\)\(2,1\)\(2,2\)

So, Rakudo is currently seeing ^@​list as following this definition,
and trying to generate the subscripts (perhaps wrongly).

Yes, wrongly​:
08​:48 < moritz_> rakudo​: say (^(3,3)).perl
08​:48 < p6eval> rakudo 33212​: OUTPUT[[0, 1, 2, 0, 1, 2]␤]
08​:51 < moritz_> rakudo​: say (^(10,3)).perl
08​:51 < p6eval> rakudo 33212​: OUTPUT[[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0,
1, 2]␤]

It counts up first the first index, then the second.

I see how the specced makes sense for a List of Ints, but not for any
other list - any ideas from the design team?

Moritz

There's still some ambiguity in how (or if) we should support
both interpretations [1], so we'll want to get that resolved before
we can fix prefix​:<^> here.

[1] http://irclog.perlgeek.de/perl6/2008-11-26#i_720703

From @TimToady

On Wed, Nov 26, 2008 at 08​:54​:50AM +0100, Moritz Lenz wrote​:
:
:
: Patrick R. Michaud wrote​:
: > Currently Rakudo is treating ^@​list as though it's
: > prefix​:<^> on a List, which S03 says
: >
: > If [prefix​:<^> is] applied to a list, it generates a
: > multidimensional set of subscripts.
: >
: > for ^(3,3) { ... } # (0,0)(0,1)(0,2)(1,0)(1,1)(1,2)(2,0)(2,1)(2,2)
: >
: > So, Rakudo is currently seeing ^@​list as following this definition,
: > and trying to generate the subscripts (perhaps wrongly).
:
: Yes, wrongly​:
: 08​:48 < moritz_> rakudo​: say (^(3,3)).perl
: 08​:48 < p6eval> rakudo 33212​: OUTPUT[[0, 1, 2, 0, 1, 2]␤]
: 08​:51 < moritz_> rakudo​: say (^(10,3)).perl
: 08​:51 < p6eval> rakudo 33212​: OUTPUT[[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0,
: 1, 2]␤]
:
: It counts up first the first index, then the second.
:
:
: I see how the specced makes sense for a List of Ints, but not for any
: other list - any ideas from the design team?

My guess is that the list overloading will simply vanish into thin air,
and you'll have to say something like

  ^«(3,3)

to get the current Parrot meaning, and

  [X] ^«(3,3)

or

  ^3 X ^3

to get the specced list meaning. But other viewpoints are welcome...

Larry

From @pmichaud

On Wed, Nov 26, 2008 at 06​:21​:22PM -0800, Larry Wall wrote​:

On Wed, Nov 26, 2008 at 08​:54​:50AM +0100, Moritz Lenz wrote​:
: Patrick R. Michaud wrote​:
: > Currently Rakudo is treating ^@​list as though it's
: > prefix​:<^> on a List, which S03 says
: > for ^(3,3) { ... } # (0,0)(0,1)(0,2)(1,0)(1,1)(1,2)(2,0)(2,1)(2,2)
:
: I see how the specced makes sense for a List of Ints, but not for any
: other list - any ideas from the design team?

My guess is that the list overloading will simply vanish into thin air,
and you'll have to say something like

^«\(3,3\)

to get the current Parrot meaning, and

\[X\] ^«\(3,3\)

or

^3 X ^3

to get the specced list meaning. But other viewpoints are welcome...

+1 to the idea that ^@​list is the same as ^+@​list . Now implemented
as such in Rakudo, and unfudged the corresponding tests.

Closing ticket -- thanks!

Pm

@pmichaud - Status changed from 'open' to 'resolved'