Is there a good reason not to simplify it like this ? As you already mentioned, the first splice modifies @a ...

sub rm($@) {
   my ($e,@a) = @_;
   splice(@a,$e,1);
   return (@a);
}
[download]

Torsten