Quote:
Did you by chance happen to have Windriver's target debugger connection open when this occurred in your pits? If so, did it report any abnormal terminations or errors?
If not, you might consider doing so as this can provide a wealth of information when things go awry. You don't have to be "debugging" the code at the time -- you can be running an "deployed" program. The nice part about it is that it will report any task failures/terminations to you as well as stack information if available.
|
No we didn't. Good idea for us to try if we can replicate this when we get our robot back.
Quote:
Are you able to go into a bit more detail on what your dashboard task is doing exactly? Particularly in relation to CANJaguar objects, as well as frequency of iteration.
We send dashboard data during disable as well, but do it as part of the normal disable processing routine rather than as a separate task.
|
I'm not familiar with the "normal disable processing routine". Is that a Java or Labview construct ? We use C++ and spawn a separate simple polling loop that polls our 5 jaguars for encoder values, limit switch states, current draw, etc. There are probably 15 or so Jaguar calls in a polling loop with a 50ms sleep (~20 Hz period). This thread is launched immediately at the end of our robot constructor.
Thanks,
john