gback at cs.utah.edu
Wed Jan 27 12:50:58 PST 1999
I think kaffe's doubly-linked lists are more historical than anything else.
They're not strictly needed for the current collector, but they would indeed
be needed for a treadmill.
> I'm planning a copying generational collector
> which should do away with (among other things) a lot of the
> block-management overhead, but even so I don't think I can get away
> with less than a word per object of structuring information.
For kaffe? That would be great.
Note: Jason Baker is also working on new collectors for kaffe.
To help facilitate this, I attempted to separate the collector from
the rest of kaffe as cleanly as possible by creating the Collector
interface. This interface is not perfect yet, and it's certainly
not cast in stone yet. For instance, this interface does not
include the necessary elements for any type of incremental collection.
In any event, I'd appreciate feedback on whether the defined
interface is sufficient and how to improve it.
More information about the kaffe