K Stol writes:
>  (Just another question concerning GC)

>  

>  If I remember correctly, Java has a separate garbage collecting

>  thread. I though Parrot has a synchronous garbage collector, right?

>  So, the garbage collector (or DoD for that matter) will run in between

>  execution of ops. If that is so, and DoD/GC is done on each scope

>  exit, then that will slow down quite a bit, right? Or is that just my

>  imagination?


That's what I'm worried about.  In my code in particular, scopes go in
and out about half as much as statements do.  That's why I want a really
efficient C<sweep 0> , and the reason for my proposal.  I'd also like

Perl 6 to be smart enough to realize when I'm using "pure" functions and
have not changed anything that would affect DESTROY-able objects.

Luke

>  Klaas-Jan

>  

>  ----- Original Message -----

>  From: "Leopold Toetsch" <lt@[...].at>

>  To: "Benjamin Goldberg" <ben.goldberg@[...].com>

>  Cc: <perl6-internals@[...].org>

>  Sent: Tuesday, August 19, 2003 12:20 AM

>  Subject: Re: What the heck is: timely destruction

>  

>  

>  > Benjamin Goldberg <ben.goldberg@[...].com> wrote:

>  >

>  > > Every time we come to a scope end, we do a garbage collection run.

>  >

>  > ... only if there are objects around, that were marked to need timely

>  > destruction. If the HLL will decide that each object might need timely

>  > destructions we will get into troubles, at least with the current DOD

>  > scheme.

>  >

>  > > What we'd like is a way (and there've been a couple proposed) to make it

>  > > so that the sweep at the end of scope can *quickly* determine that all

>  > > objects needing timely destruction are still alive/reachabe/in-scope,

>  > > and abort early (and thus use less time).

>  >

>  > Yes. Or that all such objects are destroyed already.

>  >

>  > leo

>  >

>  

>