chicken-hackers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Chicken-hackers] CR #1142: remove queues, mmap, evict and binary se


From: Peter Bex
Subject: Re: [Chicken-hackers] CR #1142: remove queues, mmap, evict and binary search from core libraries
Date: Sun, 20 Jul 2014 13:46:44 +0200
User-agent: Mutt/1.4.2.3i

On Sun, Jul 20, 2014 at 01:42:19PM +0200, Felix Winkelmann wrote:
> From: Peter Bex <address@hidden>
> > Excellent!  In the ticket you mention you've prepared eggs for this.
> > Can we preview these eggs somewhere?
> 
> I can commit them to SVN, if you like. Or should I use a branch for
> this?

Putting them in svn should be fine; if there's no tag it won't get
published.  Alternatively, you could upload them to your bitbucket or
github account for the time being.  Then when the CR is accepted
move them to svn (or simply publish them from your account if you want
to maintain them yourself).

> > I'm not 100% sure about removing object-evict from core, because of the
> > problem described by #1112; this might require core support somehow to
> > make the GC know which objects have been evicted.  Do you have an idea
> > on how to make this work when eviction is moved into an egg?
> 
> Well, "object-become!" might be useful here, as proposed in the
> ticket. Another option (for users) would be to evict into a static
> buffer that can not be possibly "stolen" by a heap-enlargment.
> 
> But instead of making GC even more complicated as it is, I would
> recommend to just punt on this issue. The functionality of
> "object-evict"/"object-release" is very application-specific and a bit
> obscure and should be considered sufficient to those who want to use
> it.

Sure, that's fine.  Perhaps if we ever decide to fix it properly, we
can provide a hook for the egg to use.

Cheers,
Peter
-- 
http://www.more-magic.net



reply via email to

[Prev in Thread] Current Thread [Next in Thread]