[Greg McKaskle]

Sorry I didn't get to this earlier. Work stuff.

The crash log is very helpful.

I'm pretty sure that the "Too much error data!" messages occur when the robot buffer for errors won't fit into a status packet. The status packet will contain some of the messages, some will be left out, and some text about messages being left out will be appended to the message display and log file. I'm not sure that this has anything to do with the error, but since that code is exercised pretty rarely, it could give a hint.

Can you give more context on what is going on when this occurs? THThreadDestroy() is a wrapper for the posix threads or whatever thread package that LV is using on the OS. As far as I know, LV doesn't often destroy threads during execution. This seems like LV is either exiting or is reconfiguring itself for a different execution and a data structure in the thread or owning the thread has been corrupted.

Do you think this would be pretty repeatable? Could you send your code to us and tell us how to run it in order to cause the crash?

In the meantime, it may disappear as you change your code. Please keep you spidey sense alert for what you think it was related to. In particular, these sorts of corruptions are generally due to a DLL or .out that hands LV a bad structure or writes past one of LV's structures. The LV generated code is obviously capable of the corruption too, but those are less common.

Feel free to contact NI support or my PM if you need more info on how to get this to us.

Greg McKaskle