Is there anything 'Bad (tm)' about creating a selective typeglob alias in one package to a lexical that is definitely going to go out of scope in a second package? It seems to work, but I feel like I may be missing something... For example, consider two packages, A and B. B has a subroutine which executes the following (B knows about A): sub foobar { my $lexval = 'foo'; *A::bar = \$lexval; } My concern is that B::foobar is going to get called possibly multiple times (it is okay that $A::bar is getting overwritten each time B::foobar gets called.) Since $A::bar ends up holding the only reference to $lexval once B::foobar exits, is the refcount decremented and the first instance of $lexval cleaned up properly when B::foobar gets called a second time, overwriting $A::bar? --Chris