Serious bug identified in SmartDashboard/NetworkTables -- robot hangs

[JamesTerm]

Originally Posted by virtuald

NetworkTables is a really useful idea, and so I think it would be worth it for the maintainer of the code (eg, FIRST/WPI) to make the code better -- there's a lot of ways it could be improved. However, if they're going to make improvements in the future, then my recommendation is to rewrite it -- and create some unit tests/etc. If done properly, it will make it a lot easier to fix any such bugs in the future.

I agree with that... it seems now that all efforts are currently focused on SmartDashboard 2.0. That's a whole another subject I don't want to get into now.

It should be noted that I've been using a win32 version of network tables port quite a bit... I've found it even useful in NewTek development to watch variables that dynamically change. That said... it has never once locked up or crashed... so this is partly why I could never find the issue... I can't reproduce it in win32 environment. I believe the last issue remaining deals with the time it takes to connect to the time it takes to make the initial first write. I'm thinking of putting a sleep in there as well as taking a closer look at ConnectionMonitorThread::run()... I suspect this thread may not be sleeping in some cases... but I could be wrong.