From @masak

Rakudo r34703​:

$ perl6 -e 'my $year; $year //= 2009'
get_iter() not implemented in class 'Integer'

This looks like a misparse, interpreting '//= 2009' as '// =2009'.

From dave.whipp@gmail.com

<dwhipp> rakudo​: my $a; $a //= 1
<p6eval> rakudo 34768​: OUTPUT«get_iter() not implemented in class
'Integer'␤current instr.​: 'prefix​:=' pc 17718
(src/builtins/misc.pir​:17)␤»

From @jnthn

On Wed Dec 31 08​:44​:02 2008, masak wrote​:

Rakudo r34703​:

$ perl6 -e 'my $year; $year //= 2009'
get_iter() not implemented in class 'Integer'

This looks like a misparse, interpreting '//= 2009' as '// =2009'.

Tried just adding it to generate as a meta-op, then realized that we're
using pasttype def_or for // so there's no infix​://. So I hand-rolled a
//= for now, but are we really going to get away with using pasttype for
//? I'm thinking that // will want to be overloadable, so needs to
become a multi infix​:<//>? If there's agreement I'll re-do it that way
(then we can probably have [//] too...)

Anyway, it's testable now. OH HAI, moritz...

Jonathan

The RT System itself - Status changed from 'new' to 'open'

From @moritz

On Mon Jan 19 14​:18​:43 2009, jnthn@​jnthn.net wrote​:

Anyway, it's testable now. OH HAI, moritz...

Jonathan

Jonathan found (and corrected) the tests in S03-operators/assign.t
himself ;-)

Cheers,
Moritz

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

From @pmichaud

On Mon, Jan 19, 2009 at 02​:18​:46PM -0800, jnthn@​jnthn.net via RT wrote​:

On Wed Dec 31 08​:44​:02 2008, masak wrote​:

Rakudo r34703​:

$ perl6 -e 'my $year; $year //= 2009'
get_iter() not implemented in class 'Integer'

This looks like a misparse, interpreting '//= 2009' as '// =2009'.

Tried just adding it to generate as a meta-op, then realized that we're
using pasttype def_or for // so there's no infix​://. So I hand-rolled a
//= for now, but are we really going to get away with using pasttype for
//? I'm thinking that // will want to be overloadable, so needs to
become a multi infix​:<//>?

It also needs to short-circuit, which is why we currently use a
pasttype and not a simple multi.

Beyond that, many of the infixes are switching from left
associative to list associative precedence, which means we'll
be revisiting this when the dust settles anyway.

Pm