I decided to just let the thing go as long as it can run and wound up
getting a stack overflow. Once that happened, I traced crash to a
routine that draws the trajectory on a world map. (this is not yet
implemented in the new sim) Once that draw routine was commented out,
things worked.
I'm still puzzled as to why the problem manifested itself by speeding
up when RB3DSpace windows were opened. I would have expected a massive
slowdown if the stack was getting whacked too much.
Anyway, thank God for small favors!
Back to your regularly scheduled program...
On Aug 12, 2005, at 9:00 AM, Joseph Nastasi wrote:
Okay, this one is really kicking me!
In Wings of Mercury, I can have multiple windows that contain
RB3DSpace controls open at once. I recently did a big overhaul of the
main loop of the program, eliminating all threads (though in 2005, we
can now control the priorities! my timing blows) and have everything
contained in a short interval timer. I use the milliseconds technique
that produces a delta time that can be used to modify simulated
movement, etc. Almost exactly like the Open Plane by the Strout Game
Think Tank. :-)
It was working very well up until one revision ago both on OS X and
Window XP. All of a sudden, after making some minor changes (yeah, I
know) the speed of the simulation starts racing like it's in fast
forward mode. It only happens on XP, not OS X. Okay, so I go back to
the older version. Still does it. Go back another version. Still does
it. and so on...
I keep copious development logs and there is no reference to seeing
anything like this before. Besides, this version has only been in
development for a couple of months, so I'd probably remember anyway.
I've done all sorts of tracing and can't figure out how it could be
screwed up.
But here's what I just discovered: If no RB3DSpace windows are
visible, the timing looks fine. As I start to open Rb3DSpace windows,
the simulations starts to speed up. Once all three are open, I am
running at almost 15x normal speed.
I'm using 5.5.5, BTW and have the latest version of Quesa, though
neither of those have changed in the last couple of months. It does
this when running in the IDE and as a compile application.
Now here's another piece of the puzzle. The new simulator, which is
more of "MS Flight Simulator for spacecraft" and therefore allows
creation of vehicles from the ground up, uses the same
timer/millisecond technique and does not exhibit this problem at all.
The ONLY difference is that there is one main window that has two
RB3DSpace controls on it. (one for views and one for the control
panel) It works fine on both platforms. (actually, it works _better_
on XP)
Obviously, the must be _something_ I am doing wrong. My RB version
hasn't changed (well, there were sometimes I used 2005, but the
debugging issues stopped that) and neither has Quesa, OpenGL or even
my Windows version.
So, other than crying in my beer (which I am, at this point,
considering replacing with a vat of rum), I'm asking for ideas on why
opening windows with RB3Dspace controls would speed up program
execution?
Thanks for any ideas,
Joe
--
Joseph Nastasi
Pyramid Design - a software development firm
http://www.pyramiddesign.us
Voice 609 601-0814 Fax 609 601-0815
Products:
A-OK! Spacecraft Simulation System - http://aok.pyramiddesign.us
A-OK! The Wings of Mercury
FTP Suite for REALbasic - http://ftpsuite.pyramiddesign.us
Proud Sponsors of REALbasic Olympics 2005
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
--
Joseph Nastasi
Pyramid Design - a software development firm
http://www.pyramiddesign.us
Voice 609 601-0814 Fax 609 601-0815
Products:
A-OK! Spacecraft Simulation System - http://aok.pyramiddesign.us
A-OK! The Wings of Mercury
FTP Suite for REALbasic - http://ftpsuite.pyramiddesign.us
Proud Sponsors of REALbasic Olympics 2005
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
|