On Fri, 2007-11-30 at 11:23 +0100, David Olofson
wrote:
On Friday 30 November 2007, Krzysztof Foltman
wrote:
[...several points that I totally agree with...]
If you use integers, perhaps the timestamps
should be stored as
delta values.
That would seem to add complexity with little gain, though I
haven't really thought hard about that...
It does have the significant advantage of eliminating the hard upper
bound on the range of time that can be present in a buffer (and with
'null' events, eliminates any such limit entirely, ala SMF). More
annoying to work with though..
Yeah; someone has to add the "null" events, and the delta nature of
it, obviously. You could hide that in event handling
calls/macros/inlines of course, but still...