Keeping a list of instances and garbage-collection

[Bill Baxter]

Part of the reason I wrote it and made it available was to serve as a
focal point for such critiques.   If you think it doesn't work and can
fix it, please do so!

--bb

On Mon, Mar 30, 2009 at 7:00 AM, grauzone <none at example.net> wrote:
> Jarrett Billingsley wrote:
>>
>> On Sun, Mar 29, 2009 at 4:42 PM, Leandro Lucarella <llucax at gmail.com>
>> wrote:
>>>
>>> This was discussed several times in the past. For example:
>>>
>>> http://www.digitalmars.com/d/archives/digitalmars/D/learn/weak_references_13301.html
>>>
>>> http://www.digitalmars.com/d/archives/digitalmars/D/learn/Soft_weak_references_8264.html
>>>
>>> http://www.digitalmars.com/d/archives/digitalmars/D/announce/ANN_WeakObjectReference_-_class_to_hold_weak_references_9103.html
>>> etc.
>>>
>>> I hope it helps.
>>
>> The one provided by Bill:
>>
>> http://www.dsource.org/projects/scrapple/browser/trunk/weakref
>>
>> seems to work fine, and has the advantage of working in both Phobos and
>> Tango.
>
> First, I doubt this actually works. The WeakRef stores the pointer as
> size_t, but the GC is conservative and will still recognize the size_t as a
> pointer. The unittest in the existing code only works, because he uses an
> explicit delete on the referenced object.
>
> To actually hide the pointer from the GC, you could XOR the size_t value
> with a constant. Note that you need to be very careful with the
> WeakRef.ptr() function: what happens, if the GC invalidates the object, and
> then the user calls ptr() in parallel, before the GC calls
> rt_detachDisposeEvent()? The user will get an invalid pointer. As far as I
> remember, rt_detachDisposeEvent() is supposed to be called when all threads
> have been resumed (after a collect() run). This is to avoid deadlocks if the
> dispose handler locks something.
>
> Secondly, this should be extended by a ReferenceQueue (like in Java). As
> soon as the object referenced by the WeakRef is collected, it is added to
> the ReferenceQueue associated with the WeakRef. (In Java,
> ReferenceQueue.remove() also provides a roundabout way to notify the program
> when a reference has been collected.)
>
> And finally: why is this thing not in Tango?
>
> Maybe a Tango dev could comment on this and the correctness issues mentioned
> above?
>