| Discussion | |
From: mda
Keywords: Commander
Remote Name: 24.147.232.222
Remote User:
Date: 06/10/01
Time: 18:37
As you know, I came up with a different approach to the BTCrit optimization, based on supporting a GOTO mechanism for jumping directly to a given sequence #. The advantage of this over your suggestion of adding an explicit sequence number range is that the GOTO mechanism is simpler and more general, it required no interface changes, and it keeps the optimized state-saving logic nicely encapsulated in a single svm_save method. Since the BTCrit optimization was the only justification for contemplating a way of specifying a range of sequence numbers, I'll hold off on that idea.
|
|
| Contents | Discussion | |