When Camino crashes (which it does do from time to time, even when it's apparently doing nothing), theres a crashlog left behind that I can poke through. But is there any way to tell what it (or the system) is dioing when the "Spinning Ball of Death" comes up?

Don't tell me this is Cocoa's GC rearing it's ugly head.. Why is there not a variant GC algorithm for comptuers with gobs and gobs of memory that will maintain a "emergncy" heap where things could be allocated while the GC is going on. Then the only atomic operations would be copying blocks from one part of the heap to another, and updating the handle table (I'm thinking of the Schore-Deutsch-Waite (sp?) algorithm that sweeps and then copies/compacts the heap.) With a third "emergency heap", the GC could (within the limits of the atomicity mentioned above) operate preemptively, as long as *new* accesses happened in the emergency heap. Then after the main heap was compacted, the emergency heap's objects could be filtered back into the main heap one at a time. (Of course, this may be terribly naive - our Data Structures prof seems to think we need to know about GC, but only introduces us to Mark/Sweep and the SDW algorithm.. Nor does he recommend Jones and Lins "Garbage Collection" as a supplementary text <AARGGH!>

Jim Witte
[EMAIL PROTECTED]
Indiana University CS

_______________________________________________
Camino mailing list
[EMAIL PROTECTED]
http://mozdev.org/mailman/listinfo/camino

Reply via email to