When Intern timeouts on exit, throws a TypeError

[kitsonk]

When a NodeJS environment does not try to shutdown on a timely basis after Intern has finished executing, Intern logs that it thinks NodeJS has hung and then exits, but does not provide an exit code and then throws a type error:

Node.js hang detected; make sure to close all sockets, timers, and servers you opened during testing!
TypeError: Insufficient number of arguments.
at TypeError  <native>
at null._onTimeout  <node_modules/intern/lib/exitHandler.js:37:13>
at Timer.unrefdHandle  <timers.js:312:14>

[csnover]

The exit codes are normally provided by the process.once('exit') event listeners. This is some new API change/regression in Node.js if it can’t handle the zero-argument call to process.exit. What Node.js version does this start happening on?

[jason0x43]

It doesn't seem to be due to an intentional change in Node. At least, the current docs still list the exit code as optional, a number of Node's self tests call process.exit(), and simply doing what the exit handler does (calling process.exit() in an unrefed timer) doesn't produce that error, at least not in any of my trials. (I tested with Node 0.10.35 ,0.11.0, 0.12.0, and 5.5.0.)

[sholladay]

There have been some problems with unref'd timers recently, I wonder if they could be provoking this somehow.

nodejs/node#1264

Even though that ticket was.closed in late 2015, they are still calling it out in the release notes under Known Issues.

[jason0x43]

Is this still a recurring issue, or has it been overcome by events? I was getting a NodeJS hang message consistently when using one of the Dig Dug tunnels (it wasn't shutting down properly), but never saw a TypeError.

[jason0x43]

Closing this due to inactivity