Arlo Belshee on 2 Jul 2003 20:40:02 -0000 |
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
[ALACPP] More C++ GC stuff |
As I look more closely at the implementation for the GC that I mentioned, I wonder how much of it will remain in a final solution. The interface is very nice; I would certainly keep that. However, the threading model is, um, not present, and the exception handling is "almost" correct. Um. Yeah. Also, there are the aforementioned performance issues. My current thinking is that this would make an excellent reference implementation. There are a number of tricky issues that he has worked out. Some of them are rather subtle (such as how he avoids collection of objects during their construction). So I might use it as the basis for a real system, or I might write my own, based on learning from it. In any case, it would be a worthwhile project. A gc_ptr class with real performance and correctness guarantees could be very useful. GC with nice syntax that lets me use the stack too! Now that would be awesome. Arlo _______________________________________________ alacpp mailing list alacpp@xxxxxxxxxxx http://lists.ellipsis.cx/mailman/listinfo/alacpp