Consider 'abc' ~~ m/b/;
By current spec this would 1) temporary set $_ = 'abc' 2) call m/b/, which matches against $_ 3) produce a Match object 4) calls .ACCEPTS($_) on the Match object 5) return False Likewise 'abc' ~~ .uc ends up comparing 'abc' to 'ABC and return False. I guess that's not what we want, so there might be a need for a rule preventing the .ACCEPTS call when $_ is used on the RHS. (The previous spec with syntactic forms kinda solved the problem, the $_ topicalization re-introduced it). Cheers, Moritz -- Moritz Lenz http://perlgeek.de/ | http://perl-6.de/ | http://sudokugarden.de/