From @ronaldxs

There are similar bug reports but I thought this looked different enough ...

.\perl6 -e "my $m = '3' ~~ /<digit>|<alpha>/; unless $m<alpha> { say 'at least its false'}; say $m<alpha>.perl"
at least its false
Null PMC access in find_method('perl')
  in<anon> at line 1317​:src\gen\core.pm
  in 'Any​::join' at line 1​:src\metamodel\RoleToInstanceApplier.nqp
  in 'Mu​::attribs' at line 1318​:src\gen\core.pm
  in 'Mu​::perl' at line 1322​:src\gen\core.pm
  in main program body at line 1

In the example above if you use $m<digit> instead of $m<alpha> things
work as expected but if you pick a non-existent rule name, say
$m<alphax>, you get pretty much the same output as above.

Ron

From @ronaldxs

Seems at least partially fixed in nom but still may be related concerns
like

nom​: my $m = '34' ~~ /<digit>+|<alpha>+/; say $m<alpha>.perl; say
Nil.perl; my $x = $m<alpha> // 4; say 'x is ', $x
nom​: OUTPUT«()␤Nil␤x is ␤»

See http://irclog.perlgeek.de/perl6/2011-08-07#i_4239512

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

From @pmichaud

On Sun Aug 07 12​:25​:15 2011, ronaldxs wrote​:

Seems at least partially fixed in nom but still may be related
concerns
like

nom​: my $m = '34' ~~ /<digit>+|<alpha>+/; say $m<alpha>.perl; say
Nil.perl; my $x = $m<alpha> // 4; say 'x is ', $x
nom​: OUTPUT«()␤Nil␤x is ␤»

See http://irclog.perlgeek.de/perl6/2011-08-07#i_4239512

Since <alpha> is quantified here, it's guaranteed to be a (possibly
empty) Array in the match object -- so nom has this one correct.

In the non-quantified case, nom is currently returning "Any", which
seems reasonable​:

  21​:00 <pmichaud> nom​: my $m = '34' ~~ /<digit>|<alpha>/; say
$m<alpha>.perl
  21​:00 <p6eval> nom​: OUTPUT«Any␤»

I'll likely close this ticket when nom becomes master, unless there's
something wrong with my analysis above.

Pm

From @coke

On Sun Aug 07 14​:03​:47 2011, pmichaud wrote​:

On Sun Aug 07 12​:25​:15 2011, ronaldxs wrote​:

I'll likely close this ticket when nom becomes master, unless there's
something wrong with my analysis above.

Pm

Then is now; just needs tests.

--
Will "Coke" Coleda

From @moritz

now tested in t/spec/S05-capture/subrule.t

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