develooper Front page | perl.perl5.porters | Postings from August 2003

Re: [perl #10030] DESTROY not called on code reference objects

Thread Previous | Thread Next
From:
Kurt Starsinic
Date:
August 19, 2003 20:33
Subject:
Re: [perl #10030] DESTROY not called on code reference objects
Message ID:
20030820033148.GA27701@verizon.net
On Aug 19, Randal L. Schwartz wrote:
> >>>>> "Kurt" == Kurt Starsinic <kstar@cpan.org> writes:
> Kurt>     It seems very strange to me to say that an anonymous sub with
> Kurt> no private variables is "not a closure."  Is not the (shorthand)
> Kurt> definition of a closure "a subroutine, associated with all its 
> Kurt> bindings," even when "all" == 0?
> 
> Well, we need a term for that then.  I've been adovcating "closure"
> as when "all > 0", whether named or anonymous.
> 
> But if you have a better term (that doesn't take more than three
> words) for the subroutines that generate a different CV when created
> as opposed to those that don't, I'm all for it.

    I suggest that it's a closure when the language supports
associating bindings to the subroutine.  In Perl, this is always
and only the case for anonymous subroutines, whether or not the
number of private bindings > 0.

    So it seems to me that, in Perl, a named subroutine is never a
closure, and an anonymous subroutine is always a closure.  The fact
that perl optimizes the special case of an anonymous subroutine with 0
private bindings is an implementation detail of interest to a tiny few.
This behavior is not guaranteed, and based on Simon's bug, I suspect
that it will change.

> If "all subroutines are closures", then the term loses its usefulness
> within the Perl world.

    I'm with you 100% there, Randal.

    - Kurt


Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About