This is part of The Pile, a partial archive of some open source mailing lists and newsgroups.
To: "raptor" <raptor@unacs.bg> From: merlyn@stonehenge.com (Randal L. Schwartz) Subject: Re: my OR our that is the question ?! Date: 10 Aug 2001 05:04:43 -0700 >>>>> "raptor" == raptor <raptor@unacs.bg> writes: raptor> - closure makes a new instance of of itself and get the latest value of raptor> outer-lexical variable in its private space. raptor> - there is no reason why normal sub() can get/alias! this value every time, raptor> this doens't broke the closure behaviour (i.e. normal sub() doesn't need to raptor> make its own private copy but use the outer-scope-lexical ) or I'm wrong raptor> again. You're mixing two concepts: a subroutine may or may not be named - those that don't are called "anonymous subroutines" a subroutine may or may not see lexical variables that can go out of scope before the lifetime of the subroutine's scope - that that do are called "closures". An anonymous subroutine is not necessarily a closure. A closure is not necessarily an anonymous subroutine. A named subroutine may have visibility of limited-lifetime variables, and that's what's happening here... a closure-behavior on the first activation of the subroutine. Now, since the subroutine has a slot in the global symbol table (it's named!), it has a lifetime of the entire program. So the first time the subroutine is activated (called) in the program, it wakes up, sees some limited lifetime variables, and says "those are MINE MINE all MINE", and never lets go of them. But it stays alive forever, so even on subsequent activations, it has to hold the first scope it saw to be a proper closure. Probably a bit fuzzy, but it makes sense. ===