Hi, thanks for posting!
NCrunch is fairly religious about catching exceptions and reporting them in its own trace log. Because of its size and constant rate of change, the trace log is only stored in memory. Any crash that brings down the whole IDE is doubly evil - not just because its a fault in the product, but also because it's appeared in a place the error handling hasn't managed to trap it.
I think it's unlikely the crash is being caused by the execution of a specific test, at least, in a functional sense. Tests are always run in isolation outside of the IDE, which means that when they bug out, you should receive the crash report from a different process.
It's been a long time since I've seen a functional issue from NCrunch crash out the IDE without the error trapping kicking in. Possibly this is a complex issue involving one or more of the following:
- Memory issues (i.e. heavy memory consumption causing the process to run out of memory)
- Disk space issues
- Instability in the IDE (can be caused by unstable VS packages, warped VS installation, etc)
When the IDE crashes, is there any chance you can attach a debugger and grab out a stack trace? I hope the dying instance gives you enough time to do this. Otherwise we may need to try and solve this problem deductively (i.e. by progressively uninstalling things until its gone).
Cheers,
Remco
