Garbage Collection Idea

[Craig Black]

"Sean Kelly" <sean at f4.ca> wrote in message 
news:e5prin$1jcc$2 at digitaldaemon.com...
> Craig Black wrote:
>>
>>> Another potential problem is this:
>>>
>>> # class Foo { ... }
>>> # Foo bar = new Foo;
>>> # void* quxx = cast(void*)bar;
>>>
>>> What happens now?  We can no longer tell what type of memory is pointed
>>> to by quxx, but we *still* need to scan it.  The only way I can think of
>>> to solve this would be to mark each segment of memory with the kind of
>>> allocation (class, struct or otherwise), and what the type is.  That, or
>>> you can just revert to blind scanning, but then it's no better than the
>>> default GC (and more complex code-wise).
>>
>> I didn't initially consider this case, but it wouldn't be to hard to 
>> accomodate it.  Each allocation unit, whether it is a class instance or 
>> not, would have to get an id that describes its identity.
>
>     char** c = cast(char**) malloc( HEIGHT );
>     memset( c, 0, HEIGHT );
>     gc_addRange( c, c + HEIGHT );

Good point.  With my approach, the GC API would have to be altered such that 
you would not be able to add a range to it without specifying an identity. 
However, you could have an "unknown" identity, in which case the GC would 
work on the allocation unit in a traditional fashion.

> Where would this identity data go?  Another example:
>
>     class C {
>         int val;
>     }
>     int* p;
>     {
>         C c = new C;
>         p = &c.val;
>     }
>
> The instance of C must be kept alive because there is a reference to it... 
> but to a member element, not the root of the object.

The GC has the range information, so this is not a problem.  If the pointer 
lies within the range of the allocation unit, then it knows where the 
identity information would be kept, (either at the beginning or end of the 
allocation unit).

-Craig