>It is simple to ensure a particular cursor in that case. Just change
>the cursor to something else and then back to the cursor type you
>want. I'm not saying that cursor handling in Director is flawless,
>but we've had well documented discussions about how changing the
>cursor thousands of times is not advisable. Therefore, I minimize my
>cursor-switching code whenever possible, rather than, say setting it
>repeatedly in exitFrame, prepareFrame, or even mouseEnter handlers.
Couple of things here, you say just change the cursor to something
else and back again, but then that leaves you with the problem of
working out when the cursor has been changed by something outside of
your control, and you would probably get it wrong like Director does
(incidentally, even if a problem only happens in authoring, it's
still an annoyance. Imagine I said don't wrong about that poor
performance, it'll be ok in a full screen projector, you would still
complain that the performance was bad in authoring).
The other thing is that I wasn't talking about changing the cursor, I
was talking about setting it, even if it's already set to the one you
want. I can prove that that isn't a performance drain at all. Put
this script into a frame, and stretch it over a few hundred frames (I
did 300). Set the tempo to 999 fps:
global thecursor
on exitFrame me
repeat with a = 1 to 1000
thecursor = 279 - thecursor
cursor thecursor
end repeat
end
on beginsprite
thecursor = -1
starttimer
end
on endsprite
put the timer
end
I put a small break in the script channel at the end, so that the
endsprite would work.
The results of this test were that it took 4063 ticks to change the
cursor 300,000 times, with the overhead of a repeat loop and going
through frames of the movie. Even including the overhead as if it was
setting the cursor that caused it, that would mean the cursor
changing costs about 1/4 of a millisecond, not very long.
But that's the very worst case, and not what is happening with the
cursor technique I'm using. Next I commented out the first line in
the repeat, both lines in the repeat, and then just the second line
in the repeat. The times were then 45, 19, and 37 ticks respectively.
Let's take the 45 ticks, which has the overhead of a repeat loop, and
the navigation, and the cursor calculation. Even so, we're only
talking about 2.5 microseconds (2 and 1/2 millionths of a second).
Still not down to what I'm talking about. The 45 and 19 ticks times
show that the 300,000 cursor thecursor's that I'm doing take 26
ticks, or about 1.4 microseconds each.
So, I'm sticking to my current approach for controlling the cursor,
because I really believe that a microsecond and a half per frame is a
reasonable expense for the peace of mind of knowing that the cursor
looks how I want it to.
[To remove yourself from this list, or to change to digest mode, go to
http://www.penworks.com/LUJ/lingo-l.cgi To post messages to the list,
email [EMAIL PROTECTED] (Problems, email [EMAIL PROTECTED])
Lingo-L is for learning and helping with programming Lingo. Thanks!]
